 for one question for Andrew and another question for François. I loved Andrew's presentation, but the message of caution, I know in your conclusion you were quite pragmatic, but the message that is coming is that redistribution might be a problem. And this is particularly concerning because we think that if inequality is bad for growth, then the way or bad for growth and for society, the way we have to do it is to redistribute, and if redistribution itself is a problem, what are we left with? And I'll go to your analysis is that when you came to the analysis of growth spell, you said you had excluded some observations, and this is cross-country observations that you have excluded, and we know that most of the countries with less data are also those countries that could be highly unequal and developing. The question is whether you have systematically excluded what will have given you a contrary view of your analysis. I don't know if I have time to ask a question to François. Okay, let me ask a brief question, François. Your presentation is quite mind-provoking. I want to ask about the future of fight against inequality. I'm from Oxfam, I do research for Oxfam, and we've been talking a lot about fight against inequality in the Davos, and you've presented that many more people are coming up with fight against inequality projects, including even Obama in the future. But you've also said that it will be more difficult to do that kind of, to fight against inequality. So my question is, are these programs premature? Should we wait? Is it that we don't know enough in terms of instruments to fight against inequality? What kind of world are we going to see when everybody gets to fighting against inequality without knowing what kind of instruments to use to fight against inequality? Thank you. I think there was one right behind here. Mine is very quick, and it's to the IMF presentation. I know that you are using your cross-section data, and we know that these kinds of studies are so sensitive. Although you're adding that thing, you've checked it through the specification. How if you change the sample size, maybe rope some countries? How if you change the period? How if you change even the definition of inequality? Do you think still you remain with the same finding, so stick to the same implications? That's number one. Number two, which I think is much more interesting, you're sort of adding that as a causality. But again, I know this is in simple panel cross-section data, where you do a lot of averaging. There are a lot of things going in which you don't know. Unless you have a time series, you focus on only one country or one variable. Few variables, so you can establish causality. How strong your argument that you are trying to establish causality in this sort of regression? Thanks. Take one over there. Yes, this is a question for Francois. I might have missed it, and we didn't see the table of contents for the 2A and 2B handbooks up there very long. But I was wondering if they include anything on, efforts to estimate the world distribution of income, wealth, the level of poverty in the world as a whole, that you and others have been involved in working on, or should we wait for 2C? Okay, Mr. Minister. Just find very interesting presentations. Just one quick question for Andrew. You are quite interested in these use of markets and after taxes and transfers, redistribution. But why don't you also use in your regression? Because when you are doing redistribution, you are measuring there the effective, but like in the case of Brazil, you have a huge tax burden, you make a huge effort, and we actually don't change that much the genie, so maybe to throw in the regression something on the effort, like the size of the state, the size of transfers, because this may be not reflected in the actual redistribution observed. Okay, Andy, you want to start? Let me start with the last one first, just because it's right in my mind. I mean, that's an excellent suggestion. I can't remember, usually the size of government comes in negatively in these regressions, and I suspect that would be the case that if we threw in size of government, it would still be negative, so there's a nice, and we could also, I think it would be useful to look more carefully at, one could imagine constructing an efficiency and we've done some graphs relating the amount of genie change to say the size of government or the share of social spending, but we know that so much of government is not even meant to be, is not meant to be redistributed, you know, military spending or something like that, but even within social spending, it's well known that lots of education, most education spending in a sense in poor countries is not redistributed and so on, but we could look at that. On the first question, I should have said the views expressed in this presentation by me are not those of the IMF or IMF policy necessarily, and we had, you know, it was not, but I am, you know, at the IMF and we had, after we wrote the paper, when we wrote our blog entry, we spent many days internally in the institution discussing the wording of the blog entry to get the message just right, so it is still kind of controversial within the IMF in some ways. But what we're trying to say, what Jonathan Ostry and I are trying to say is that the aggregate data do not suggest cause for inaction on redistribution, even if you're worried about growth. Of course you have to be worried about the details. We have another IMF paper, not we, but there is another IMF paper on how to redistribute, how to do it more efficiently. It's clearly critically important as we heard this morning and so on, but it's just a caution. It's just you have to, you know, apply it cautiously. We didn't select samples for results. We did, for example, one of our restricted samples is the one that Frederick Solt suggests to use. He has his own restriction of the sample where he thinks you'd get decent redistribution data and that's the restricted sample I showed there is his restrictions. It's true that when you restrict the sample you tend to lose more developing countries and we have limited ability to check and see if our results are true by continent and things like that because it's just you run out of data. It's not just the OECD or the non-OECD that's driving the results, but beyond that it's hard to say. We do use time series and cross-country variation. It's a panel for the growth regression and you need both to get the results and certainly to do anything about causality we need the time series measure and I will actually respond very briefly in a way, at least it's not really a comment on me but something Francois said about theory. Obviously our paper is guilty of too little use of theory and perhaps to many other sins but that's one of them. It's hard to do both right. We have a project that's financed partly through some cooperation with the UK's DFID where we're trying to take macro models on a specific country to put in enough richness and heterogeneity to confront distributional data and talk about the effects of real exchange rates on distribution and vice versa and that's taken a couple of years and we don't really have a paper yet so it's hard. As well? Yes, thank you. Maybe I would like to start with a few remarks on what Marcelo said and what Andy said about the point by Marcelo that when you look at the size of the government and you want to look at and you say that there is maybe not very much impact on the change in the growth inequality between pre-re-distribution and post-re-distribution income. This is true but the problem is that when you look at post-re-distribution you are missing a very important part of the government expenditures. You are missing all the public spending so if you want to do that you should get into what is happening with education expenditures and things like that and maybe the picture would be completely different. So I think this is a very, very important point. Again, this is an issue of consistency of the concept that you are using. On the point by Andy I didn't mean to say that because you were presenting this paper presenting the IMF but I remember that maybe for the first time in the World Economic Outlook 2007 there was a chapter on inequality in the world and this was the kind of signal I was emphasizing. Okay and on the two little theory this is something that we hear we see very, very often in the one of the examples that I could not cover the example of technical change many people would tell you look because there is a skill biotechnical change then the wage the wage-skill ratio has to increase because of that. This is not true. It has to shift abroad but at some stage it should stop to increase even though the technical progress is continuing. This is a supply response and this is exactly the kind of two little theory that I am trying to to stress. Now on the on the first question about the kind of pessimistic view that I expressed the point to that today I mean again we have to make a big distinction between developing countries where the redistribution system is quite developed for some people even other developed and developing countries where there is still a huge margin to do redistribution. The point we are facing today in developed countries is the fact that maybe this is technical change maybe this is something else but there are forces of more inequality in those economic systems and we would like in some cases to be able to go counter to that increase in inequality and this is where the instruments are limited simply because we cannot anymore increase the income tax rate as we want simply because we see that there is more inequality. This is something that cannot be done because immediately we know that capital will flow to the neighboring country or people will leave the country. In France we have one of our movie stars who decided Gerard Partieu decided one day to go to live to Russia because of that because he said ok I am paying too much tax this is enough and we also know that there are some suburbs in Brussels where people are essentially French people living there because they are trying to avoid taxes so this is something which is developing and the only way we can go against it is to find some kind of international cooperation in fighting tax havens etc etc there is something of this type going on I mean the attack of the United States government on UBS and in general on Switzerland is something of this type and because of that Germany and France and the UK are following suit so this may be the direction but this is the kind of constraint that we are facing in the same way if we want to improve the distribution then the only thing which is left is education but when you have when you are in countries where the level of education when the average level of education is already very high it is very difficult to make more to create more equality in that way it is in that sense that I am saying that instruments to fight additional inequality in developed countries are becoming much scarcer finally on the world distribution of income yes of course there is a chapter on the global distribution of income which is written by Soudi Rano and Paul Siegel which is a survey of that literature all right thank you very much and thank you for very nice presentations and a good discussion I would have hoped we could have continued but there will be plenty of time to speak to both Dandy and Frazade during the day so with that join me in a big hand for the presenters