 Thank you very much. It's a pleasure to be here and I thank the organizers for inviting me to this very nice event and to present some reflections in front of a very expert audience. I must say that this presentation is a little like an advertisement for a book which will be out in a few months before the end of the year as a matter of fact, which is the handbook of income distribution volume number two. And in order to understand the title of this presentation and the way in which these thoughts are being put together, I think it is interesting to go back to the genesis of this handbook of income distribution volume number two. Fifteen years ago with Tony Atkinson, we edited this handbook of income distribution at Sevier. In those days we cannot say that income distribution was within the mainstream. We all felt like on the side of the main economic areas and we made a huge effort to put together many very good contributors and to make sure that we were able to give a rather comprehensive picture of what income distribution analysis was about and what we knew and what we didn't know in those days. Because of that we were very surprised when the book came out to see that the publisher had added to the title handbook of income distribution volume number one. We may at least say that there will be no volume number two. We've said we put together all what we could. There is nothing left. And that was it. Then three years ago we received that collection, Canaro and Michael Intrilligator, asking us whether we would consider doing volume number two. They said, no look, there is no way we can make volume number two. Over a few years before, another Oxford handbook on economic inequality had been published. So we thought that there was nothing to add to all this. And one day we were having a dinner together in a nice restaurant in Paris and we said, okay, if we had to do that, what would be the chapters of this new handbook? And we started brainstorming and very quickly on the napkin of the restaurant we were able to have 20 chapters which we thought were really quite interesting and we had not realized that things had gone so quickly and that so much knowledge had been accumulated during those 10 years in those days, now 15 years. So this is the reason why finally we decided to move on and to have another handbook on top of the Oxford one, on top of the first one, and not only to have another handbook but to have a handbook with two volumes that will be number two A and number two B. And this is the table of contents of two A. I will not go through all the chapters. You see that it is very, very comprehensive. We have four parts, concepts and approaches, evidence, theory in two B, and finally policy. And this presentation basically relies on the introductory chapter that we are just finishing with Stoney and in that introductory chapter we are trying to give a flavor of the issues which are being discussed. Throughout the handbook, but this is a very long one and this is a short chapter so it is really a very, very impressionistic view of this and we are trying to add some personal reflections on the state of this field. And this is what I will be presenting. I will take up the various parts of the handbook in some cases very, very quickly because these are issues which are discussed later on in this conference so there would be some redundancy in dealing with these issues extensively. Okay, so let me start with this first part which is on the different facets of inequality and to tell you more or less what we think and what the handbook tries to achieve. The first aspect of this diversity of inequality and again and I believe this is a real progress that we made over the last 15 years. The fact that today when we look at inequality we are immediately convinced that there are various ways we can do that. We can consider different definitions of income, we can look at family income, we can look at individual learnings, we can look at wealth and we know that what we will be concluding from descriptive analysis might not be the same in the various cases. What you have on this picture is various inequality measures corresponding to different concepts of inequality in the case of the U.S. And in the case of the U.S. you have some parallel evolutions basically because we are in a very favorable case where inequality has increased so much in the U.S. that all the definitions are more or less consistent in showing this inequality. But this view that we have to look at those various aspects not only the type of income or the kind of income concept that we use but also to look at the unit. Whether we are looking at individuals within families as with intra household inequality as when we look at gender inequality this kind of thing. Whether we are looking at groups of people we know that the conclusion that we will get from descriptive analysis will not be the same. There will be differences depending on the kind of source of data source that we are using. Already this morning we heard about the difference between tax data and surveys. Indeed in some cases we find different evolutions depending on the source that we are looking for, that we are looking at. And the main conclusion from that is that the progress that we have made is that we are now convinced that we have to take the whole picture rather than to look at only one of these curves. I would say that 15 years ago too many people would have been happy looking at one of those curves. And if you look at the case of the US and find that as I said before there is consistency between the genie of gross income, family income between the top 1% in coming from the tax data from the wealth distribution. But if you look at all those curves you will see that there is a curve which has a completely different pattern which is increasing right since 1950 which is the inequality of individual learnings. And this is a curve which is more at the end it goes to the middle of the chart and here we can see that there is, even in the case of the US, depending on what we are looking at there are differences. And what we like to do is we need to look to see what is the kind of relationship that we have between those various concepts. So this is the one dimension of the progress made over the last 15 years. Another progress has to do with the fact that we moved toward another definition of inequality which is what we called in this chapter with Tony, beyond income inequality. And we are using beyond income in the same analogy as beyond GDP as was mentioned in the morning by Marcelo, the work by Sen, Stiglitz and Fitousi. And the idea behind that goes back to almost the beginning of the current episode of inequality measurement when Rawls and Sen basically had a question about equality of what? I mean this was a Sen's question, a very well known Sen's question, but 50 years ago almost Sen has a question already got the intuition that looking only at monetary inequality and income was not enough. And since then there has been a lot of activity going on and it's quite interesting to see that it is only in the last 15 years more or less that a huge emphasis has been put on this aspect of inequality. And in the handbook we have not less than four chapters dealing with these issues. And I think it is interesting to show you the kind of thought evolution behind this and we try to put together a kind of simple framework which shows the different concepts which are used by people and the way in which they are trying to quantify beyond income inequality. And I will simply look at this very simple list of concepts. At the beginning we can use the Sen's concept of functioning which is basically the way in which people live. And in this vector of functioning, A sub i for individual i, we can make a distinction between income which is y and the non-income dimensions. So y would be income or consumption expenditures and the x would be health, would be the relationship with your neighbors or with your family. It would be the kind of job status in the society that you have etc. And we can work with that. It's okay, what is the inequality of function? And what we call multidimensional inequality measurement basically is about this. And of course there is a chapter and the author of the chapter is in the room. One of the authors is in the room about this. But you can say okay, the way in which we can do that, the way we go from multidimensional concept to inequality, we have to go back to one dimension. So the way in which this is done that people are using an aggregator function which put together the y and the x into a single scalar and then inequality is measured on this single scalar. And there are many ways we can do that and of course for every aggregator function we'll have a different way of measuring inequality and a lot of things have been written on this. Now this aggregator function is very arbitrary and presumably you would like to be able to use the fact that people are different in their preferences among all possible functionings. And this is a second concept, individual preferences among all the functionings, which depends on the individual. And the issue is is it possible to take into account these individual preferences? And one approach is the one which has been proposed by Marc Florbet and François Maniquet and Conde Canc, which is to use income equivalent approach using these individual utility functions and estimating individual utility functions on satisfaction data. The third concept is again a very, very timely concept because it goes from preferences or ordinal utility to cardinal utility. Then S is for satisfaction and satisfaction comes from the utility that people get from the set of functions that they have plus individual characteristics which explain why people with the same functionings may enjoy a different level of satisfaction. And some people instead of looking at income instead of looking at health or the other non-income dimensions of welfare or functionings would be looking at satisfaction. And many people have tried to look at the way in which the inequality in satisfaction is changing over time and across countries which raises some technical issues because satisfaction is not continuous, it is a categorical variable but I don't insist on this. And then if we move up in the degree of generality of the concept then we have the inequality of capabilities. And what is a capability? A capability is a set in which people can choose the functioning. And this set Q depends on some individual trade, individual context parameter. And the issue would be to measure the inequality of those sets. This is something that we don't know very well how to do. This is technically extremely difficult. So many people instead of doing that simply look at the inequality in the individual parameters Z. And when you look at the human development index for example, this is what this corresponds to. And then some people have tried to generalize that using individual data and coming from surveys. And in the same line we have people who are trying to measure inequality of opportunities. And the way to do that is to define types of people using the Zi. And then to look at the way in which for various types of people the distributions of the functionings where are changing or not the differences in distributions of the functionings across different types. This shows you the conceptual elaboration of this beyond income inequality school. And again there is a huge activity in this area. And for the moment I don't think that we have found simple ways. I mean there are some partial results which are interesting. In particular all those results on multidimensional poverty with deprivation counting. But there are still efforts to be made. And this is definitely apparently a very important area of research in this whole field. I go over these because this is the detail of what I just said. On the data I will not say more very much. Simply to say that there are three areas of progress, revenues of progress. One is that over the last 15 years we have this rise in experimental economics. And in the field of income distribution we are using actual statistics on income. But in experimental economics economists are generating their data to some extent. And they are trying to see how people feel, how people behave in the front of some specific situations. And this is really providing wealth of new data. And this is certainly something that will be going on now for quite some time. Administrative data and in particular the top income source coming from tax data is something which is more and more widely used. And it is certainly something which has to be recommended when it is possible to do that. And finally a very important progress has been historical data. This is for those picketies 2001 book on tax data in France. Now this has been applied to many countries. And this means that in many cases you are able to look at the evolution of wealth inequality. In some cases income inequality going back one century for income. Going back two centuries at least for wealth inequality. In the handbook there is a chapter which is focusing on this. And it is really quite impressive to see the kind of data and the kind of information that researchers have been able to dig out. And obviously the big progress is the increasing availability of income and consumption surveys. But there are still comparability problems across countries, across time periods. And because this is something that will be discussed in the afternoon with this Journal of Economic Inequality panel. I am not insisting on this. Let me say a few words. I would like to say a few words on economic theory. But let me simply say what I wanted to deliver. Economic theory is there is a lot of activity going on in income distribution analysis through theoretical models. At the same time there is a lot of research being done on empirical data. But the link between those two research streams is not completely clear. Very often empirical data analysts are using theory in a very, very simple, almost naive way. And in many cases theorists are satisfied with very general results which may not be that relevant when looking at data. And to some extent there is in some case too little theory and in some other case too much theory and not enough empirical analysis. And what we are trying to do in this interview chapter is to give examples of that. So one example is really about the role that this is something which has been discussed again and again. What is the role of skilled bias technical progress in the increase in earnings inequality that we have observed in many countries in the world. And we are trying to show that most empirical analysis are very, very short on the theoretical level. And because of that the kind of conclusion that they get is not really completely satisfactory. I would like to spend some minutes on the second example which is automation. Because I believe that this is giving a very nice view about what we can do. And then I will not say very much about the disregard of transitional past. The theorists tend to look at steady states and the problem is that we know that in the real world to go from one steady state to another steady state may take in some cases 40, 50 years. If you look at the PKT book and this distribution of wealth and this idea that when R minus G is increasing there will be more inequality. Yes, at the steady state but because this is an intergenerational model to get to the steady state of an intergenerational model you may need one century. So in one century many other things may change and the kind of relevance that we get from this kind of empirical analysis in order to do policy analysis is not completely clear. But because of time I don't think I will even talk about the second topic. I prepared something for a much longer presentation I know, two minutes. So maybe you'll be missing the most dramatic part which was this one. We are basically imagining what the world will be in 50 years from now, maybe 40 years from now. And we are providing some possible hypothesis about what is going on and the reason why inequality is increasing. And in particular the reason why the capital share in GDP is increasing. But let me conclude looking at the role of policies. In the field of policies again over the last 15 years there has been a big change in the policy context. In particular when 15 years ago income distribution, income inequality was not within the forefront in policy debate. Today there is a kind of official adoption of distributional objectives. The MDGs is probably the best example but we just heard somebody from the IMF talking about inequality and redistribution. This is something which would have been somewhat odd 15 years ago. Maybe Andy will not agree with that but have been outside the IMF but close to the IMF on the other side of the street and this is really the feeling that we had. Davos is talking about inequality, amazing. Obama we know is about maybe to launch a war on inequality. So this is something completely new. At the same time and this is quite interesting there is a kind of pessimism in developed countries about the potential to continue with the kind of redistribution system that we have and some pessimism about the possibility to go against the increased inequality that we observe using the instruments that we have. And basically this is because globalization is making many of these instruments inefficient or ineffective. In particular the tax system is becoming ineffective because of the mobility of capital and of course mobility. In some cases the mobility of people. And at the same time we have what we heard this morning with Marcelo Neri's presentation where the decline in inequality and that we will be hearing more today, the decline in inequality in Latin America. And so this is a very interesting change in the landscape of income distribution policies in comparison with the situation 15 years ago. Now when we look at the future, I think that the future is quite right for developing countries. We have heard something of this type for this morning. Economic analysis is definitely confirming the fact that there has been a huge impact of policies in the Latin American case. And we know that because, and this was also something that Andy Berg said, because the redistribution system is very little developed in developing countries or in emerging countries, because of that there is a huge scope for intervention in many of these countries. So from that point of view the pessimism is not really on the side of developing countries and this is exactly the opposite. But in developed countries the fact that the conventional instruments which will not disappear, I mean social security will stay alive and redistribution benefits, tax and benefit systems will remain alive. This is not the issue. The point is that more inequality will be very difficult to fight with the instruments that we have in our hands. And because of that it may be time to be thinking outside the box and maybe to get back to in some cases old ideas about income distribution and income distribution policies which may have been forgotten. But ideas like basic income or something that Atkinson at some stage called participation income, things like this old idea by Thomas Paynes to create a kind of universal inheritance. Everybody would be narrating at 25 or 20 some amount of capital. Things of this type are probably were considered as utopias at some stage. If indeed we are in a system where inequality will keep increasing and there are some reasons to believe that this will be the case then it is time to look maybe more carefully to this kind of interventions. So thinking outside the box is really quite important. Some people are trying to do it in the handbook but not enough. This means that new ideas are most wanted and I certainly hope that 15 years from now there will be another handbook. I will be dead by that time. But a new handbook with truly innovative ideas and outside of the box ideas. Thank you very much. Thank you very much.