 Okay, thank you very much, and we'd have to turn to everybody. I must say that this presentation, and as a matter of fact this session, has a kind of dual role. On the one hand, we will be going over evidence at the global level, and in several countries, about inequality and trends and changes in inequality. But at the same time, this session is also a kind of first step in launching a project, which we have called for the moment, Inequality in the Giants, and now you may understand the choice of the four countries that have been selected for this session. And the idea being, as the title said, to look at what is going on in those countries, and the way in which the evolution of inequality within those countries may affect global inequality. Initially, I plan to go over a series of issues on global inequality and various estimates. What I will do, I will try to focus on several points which justify the project I just mentioned, Inequality in the Giants. And basically the argument will be the following. When we look at global inequality, we find, certainly over the recent period, over the last 10 years, let's say, that looking at the data as they come from household surveys in the various countries in the world, that inequality has gone down in the world. And this is mostly due to the fact that between country inequality has gone down. And I would say that this is a rather robust finding. And several people who were disagreeing before about the way in which inequality in the world was changing now agree on this. But at the same time, we know that we don't observe inequality very well in many countries. And one of the big changes which took place in the way in which we think about inequality these days is the emphasis recently put on top incomes by several people, including Tony Atkinson, Thomas Piketty, and many other people. So it is quite possible that the conclusion about declining global inequality be very much affected by this underestimation that we have of inequality in some countries. And what I would like to insist upon is this ambiguity and the fact that our project will be trying to get to resolve part of that ambiguity. So let me move on if I'm able to... Yes. Basically, I will be looking at several things. The basic question is, is global inequality rising, declining, or is it constant? Does the answer to that question depend on the concept of inequalities that are being used, the data sources being used, and possibly the normalization taking place in those data sources? Is the recent history of a few specific countries essentially responsible for the most observed changes? And here, of course, we have in mind what happened in China influencing very much what is happening at the global level. How to handle missing data at the top? And the decomposition between within and between can to inequality. What does matter most? And I would like to mention here a very important fact. There is a huge literature on this, recent literature on this. I'm listing the various contributors to that literature, and you can see that there are many recent papers in this list. Now, let me move quickly to the essence of what I want to say. And let me simply illustrate those various points with some charts which will show what is happening with global inequality when we modify or when we take into account some of the points which I've already mentioned. Now, I will base the estimates that I will give of the evolution of global inequality on the following methodological framework. The data sources are essentially two. The OECD income distribution database for developed countries, which is now some database which is well known and rather precise and in agreement with most people who are working with those data. And the other database is the World Bank Povkal database, which is for most developing countries. And both databases are relying on micro data sources and handled in a very rigorous way. I'll come back a little later on the issue of the comparability of sources across countries. The sample is a constant sample of 108 countries which cover 95, 96, maybe 97%, not 95% of the global population. And those countries have been selected so that we have at least three observations during the 20-year period, 1990 to 2010. So we're able to follow the evolution of inequality. The income concept will be in general what is coming in the survey, which may be the mean income, I mean income in some cases or consumption in some other cases. I don't understand why the word mean is there when talking about distribution. Sorry about this. But many people, including myself in previous work, are normalizing the survey data by GDP per capita or by household consumption expenditure per capita. I'll say a few words about why we do that, but it is important to know whether we have changes in the result depending on this. Calculations will be made alternatively with two sets of purchasing power parities, the one corresponding to 2005, and the recent one based on price data gathered in 2011 and which in some cases leads to different estimation. And finally, I want to say something about the impact of China in the evolution of global inequality. Here is a key chart. At the top, you have the evolution of global inequality as measured by the Gini coefficient among all citizens in the world, so at least 95% of them, and the blue curve is what comes out directly from the household surveys. And you see that this curve is declining very slowly at the beginning during the 90s and much more rapidly in the 2000s. The change is quite striking. The change is almost 5 percentage points of the Gini. This corresponds to the increase in global inequality more or less during practically all, let's say at least half of the 20th century. So this drop in 10 years is quite striking, a very important fact. You can also see that the Gini coefficient is very high. In 1990, it is 0.74 and for all of you who know little about the Gini coefficient in countries, it is very rare. And I don't think I ever saw a country with a Gini coefficient for income or consumption equal to 0.74. But maybe I'm wrong on this. The two other curves are what happens when we normalize. Why should we normalize? Basically because we know that some countries have not the same definition of income. Some are using income in the surveys, some others are using consumption, and even consumption is not defined in the same way everywhere. So because of that, it might be helpful to normalize all the survey data so that the mean welfare coming from the surveys is comparable across countries. And to do that, we use national accounts. This is very much debated in the sense that there are many reasons why national accounts differ from surveys in terms of definition. And it may also be the case that, and it is the case that in several countries, national accounts are maybe themselves wrong. So this is worth what it is, but it is interesting to see here that it doesn't matter whether you normalize or not. The absolute level of inequality is changing, of course, but the evolution is parallel. We have a small decrease in the 90s and a big one in the 2000s. So this is, to some extent, the starting point of the starting evidence that we have on global inequality. This is a side remark about the coverage of household surveys in comparison with national accounts. The green curve here shows you the ratio of the mean income or consumption per capita in the surveys versus the household consumption expenditure per capita in national accounts. And you see something quite interesting here, the fact that the coverage is quite clearly changing over time with the coverage of surveys which goes down. The meaning of that isn't clear. Is it a statistical problem? How the surveys become weaker today than they were 20 years ago? Or is there something else? There are national accounts are defined in a different way, but I think it is quite important to notice that. And the brown curve at the bottom going down is the same thing when you look at GDP per capita. And of course, because GDP is higher than household consumption expenditure, the curve is below. But we have the same shape of the declining coverage, which is an important phenomenon. And the red curve about GDP per capita at PPP 2005 simply to give in that picture an idea about what are the absolute values involved in this. This chart is about what is the impact of changing the PPPs and moving from 2005 to 2011. And you see that the blue curve is with 2005, the brown curve is with 2011. You see that there is less inequality with the new set of purchasing power parities in particular because some big countries in particular China is richer with the PPP 2011. But again, we see that evolution is completely parallel. And the bottom curves are about poverty. And this is poverty with the two sets of PPP figures. And this is a message that I'm completely cheating showing those curves because the blue curve, which is said to be purple here. The blue curve should be much below the green curve. And it is a big problem because shifting from 2005 to 2011 PPPs, poverty in the world went down enormously because of the change in the purchasing power parity of a few countries. Now this is a big problem. There is actually a commission working on what to do with this issue of the PPP correction in the future. Here what I'm doing simply say, assume that the poverty would be the same with the two sets of PPPs in 1990. What would be the evolution? And the evolution is the same. So again, we have problems of levels, but the evolution in all these data is quite convergent toward this drop in poverty and inequality. And this is the impact of China in all these. So you have the initial curve, the blue curve for the world inequality, then the brown curve is a world without China. And you see that instead of a decline in inequality in the 90s, you have an increase in inequality, but then the drop that we observed for 2000. What is interesting is the way in which we should try to neutralize to some extent the impact of China on the global distribution. And there are two assumptions here. One is take China away and the other one is keep China, but assume that the growth rate of China over time has been exactly the same as in the rest of the world. And you see this is a green curve and you see that the impact of the Chinese growth is much lower in that case. Finally, this is the decomposition of the change in equality into between country inequality and within country inequality. Between country inequality does not take into account inequality within countries. And you see that here the overall change is essentially due to the drop in the between country inequality. The within country inequality, which is a lower curve, is slightly increasing, but very little. Most of the action is in the between country inequality. Now let me now look at what is the thorny issue in all this. Many people would say those figures are wrong because you are simply missing the most important part of the action, which is what is going on at the top. And what we know, we know that because of what I call the top income revolution because this is something rather new in the field of income inequality. Measuring top income shares from tax data has added a new dimension to inequality measurement. And in particular household surveys are found to be most frequently underestimating inequality because top incomes are missing. And because of that, people say, look, if in your data it is okay that the top income shares is increasing, then what you say about the global inequality is wrong. Maybe it is increasing because within countries inequality, inequality is increasing and you don't take that into account. And this is what we like to look at. Now the problem is that top incomes are observed only in a few countries today. And second, combining those tax data with household survey data, which are not defined in the same way. The units are not the same. You have individuals, tax units, households, depending on what you are looking at. This is something difficult. And this is what I'm calling the daunting task of correcting for missing top incomes in global inequality estimates. There are very courageous people in the world and in the profession. And they tried to make that correction, in particular in a paper which will be out in a few months, I guess. Lakhner and Milanovic tried to do something of this type. And in a paper that was published in the recent, the second volume of the Handbook of Income Distribution, Anna and Segal also have a correction for that. And what they find, they find that inequality, of course, is higher when you make the correction. This is hardly surprising. But they find, and this is the case for Lakhner and Milanovic because Anna and Segal are not covering the whole of the 2000s. They stop in 2004 and there are problems even there. So this relies only on Lakhner and Milanovic. They find that inequality might not have gone down at all in the recent period if you take into account that. Now, what is the kind of methodology that they are using? Lakhner and Milanovic make a very simple assumption and this is linked to a chart that I've shown before. They say all the discrepancy between household surveys in terms of the mean income and national accounts is due to top incomes. So all what is missing when you compare the total income collected or in surveys with the total income in national accounts, all that goes to the top quintile of the distribution. And they make, this is a kind of change that they make. And not surprisingly, and given the fact that you have seen the coverage of surveys is declining with respect to national accounts, they find that inequality has not gone down but has increased slightly. At least the downward trend disappears. In Anna and Segal, they do something more sophisticated to some extent. They regress for countries where the share of the top 1% is available in tax data. They regress it on the share of the top quintile in household surveys and on the mean income of the country. And they find a relationship and they apply that relationship to all countries in their sample. I must say that I'm a bit reluctant on this, especially given the fact that having mean income as a variable here means that as countries grow, necessarily there is more and more inequality in the world, which is quite surprising. So maybe it works for a small period of time, but I don't think that we can rely really on this. And when they do that, they find that inequality would have increased much more, would have increased in the 90s and substantially so. Now, I don't want to criticize, I mean, these were nice attempts, but I don't think that at this stage we are equipped to do this kind of correction. At the same time, I would like to insist on the fact that this hypothesis that inequality in top incomes may completely change the way in which we look at the original global inequality is certainly not something that we want to discuss. And there is one big evidence on this, which is the fact that practically, I mean, in the majority of countries in the world, it is okay that the non-labor income share of GDP has increased quite substantially over the last 20 years. And there is this paper by Karababounis and Neyman, which show that and there is more evidence lately on this. So this is a sign, I'll stop in two minutes, this is a sign that there may be something going on there. So I will not go over that, I'm simply here listing what is robust and what is less robust in what we know about global inequality. And this is a summary of what we have said. I want to insist only about the last following point. When taking into account missing property income in household surveys, it might be the case that the evolution of global inequality would be reversed. This is a possibility, but if this is the case, then I believe that we are still in front of a kind of big historical change. For two centuries more or less, inequality has increased in the world because inequality between countries was increasing. And this is the 19th, the Industrial Revolution, et cetera, et cetera. And inequality within countries was either constant or declining. And if this were true, then it would be exactly the opposite of what observed during two centuries, with inequality between countries going down and inequality within countries going up. Which I think is a very interesting question. And of course, a question that begs the next one, which is, is it okay that there is some sustainability between and within inequality in the world? And let me finish on this. The inequality in the giant project is an attempt at answering this specific question by doing what? Simply focusing on big countries in the world where the data are in general better than in other countries, which of course represent a large part of the world population. And if we know more about what is going on at the top in those countries and what has been actually the evolution of inequality, then we should be able to make progress in the measurement of the evolution of global inequality. Thank you very much. Thank you.