 Then I'll go a bit quickly but what I'm doing is a quite simplistic approach to see if we can find an index that can help us not only say if a growth was pro-poor, but by how much. And first off I base my approach on a relative approach to pro-poor growth because here we've been discussing inclusiveness and I think the speakers before were talking about a relative approach to inclusiveness whereby the poor benefit is proportionally more than the non-poor and I think this is important to mention. So the increasingly recognition of inequality within development discourse I will not go through it because I'm pretty sure all of you are well aware. For me it's important to see that inequality is not anymore something that left is that red people would only defend but it's something that now it's mainstream because the negative effect it can have on growth, spells, length and poverty reduction rates and so on. When it comes to the terminology pro-poor growth this is evolving I wrote this a couple of years ago I was doing other things in between and in between pro-poor growth has stopped being fashionable now it's inclusive growth. So what I'll try to focus is on a very narrow definition which is mainly on the pro-poor growth concept there have been two main approaches one is the absolute one that Ravalli for instance used to defend I don't know what he thinks now but whereby any growth that reduces poverty is pro-poor while what I'm claiming together with much more famous people than me is that it should be called pro-poor even only if the poor grow faster than the non-poor and that would be the relative approach and I think this is a no brain actually because if a growing with the non-poor benefited more than the poor then you could also call it a pro non-poor growth would be the same definition that pro-poor growth which okay it's absurd so I will focus on the relative approach and that means that what I'm looking at is at a growth in which the poor benefits more disproportionately from growth and then using a sharply decomposition we can look at any change in poverty we can look at the growth component or the distribution or redistributional component that means any change in poverty you can attribute it to whether everybody the minimum went up or whether the distribution of income changed in favor of the poor so this is generalized by short rocks and from the shappity composition and then I look at I first look at what there was available in terms of propurness and there's several methodologies many of them are based graphic methodologies like the one we've seen on the growth incidence curve I think that's very useful but that doesn't give you a measure of by how much that pro-poor is or you may have an incidence curve that is slightly progressive or slightly regressive or that it does not allow you or that there is not a clear dominance on the previous situation which you wouldn't be able actually many cases to say that was pro-poor it wasn't so when it comes to indices that you can actually measure so far in the literature that I know at least the Quanian perniers and son have an index that is the I call it the Kaquani index but I did a consultancy for Honduras I tried to play this index to find out and then I realized that in fact well they have two definitions the first one when the per capita growth is positive second one when per capita growth is negative that means first that you have values that may not be comparable because these are two different formulas then you have that when changes in poverty due to growth aren't close to zero which sometimes happens especially if you have a population growth and economic growth that are the same then the mean income may not have increased actually in that case that shoots up to infinity in the case of per capita of negative per capita growth you would have that if you have that these two are equal but if the difference sign again you have infinity so how do you compare an index with itself when you have actually asymptotes in between and any values that are in principle quite that they cannot cure quite often so what I looked at is instead of getting a formula in which you have a division yeah I tried many different formulas yeah and at the end I thought that the the easiest one was okay if we are looking at how poverty changed because of distributional issues then we could actually just take the distributional component of changes in poverty and that what that tells you is how much poverty changed due to changes in the distribution in favor of the poor and I think it's a pretty simplistic definition and it actually works what I wanted to know is well okay then does this apply to any poverty measure so can I take the composition in any poverty measure and that will tell me always would measure in a good degree if the poor benefited more than the poor the answer is no I took the headcount index and and and actually you can think of a change and increase a general increase in the income in which no poor passed over the poverty line headcount is unaffected yeah and then you wouldn't this wouldn't show yeah and I see that for this case of the headcount index you could have actual distributional changes that would clearly affect poverty but it wouldn't show in the headcount index so what I did is I applied it to the poverty gap and in fact I was quite lucky to find that I wanted to see if if this distributional changes in the in the in the poverty gap would actually be related to the growth rate of the poor and that of the non-poor although you they have a problem because in the in the in time one you have a group of poor in time to you have a different group of poor so which one do you take and that's that's why I avoid taking the group at the beginning of the group at the end by doing the decomposition and for the case in which the number of people in society and the number of poor don't change and then then I can prove that this distributional company of changes in the poverty gap is proportional to the difference in per capita income of the whole society and that of the poor alone so what I'm saying with this is actually the definition that this index actually is really directly related to the growth difference between the poor and the non-poor but without being strictly that because then you have an identification you have problems of anonymity if you look at the poor at the beginning and the poor at the end they may not be the same no so then then everything that follows is straightforward yeah depending on whether the if the index which is the changes in the poverty gap because of distribution is negative that means that poverty gap decreases because of changes in the distribution and that's a proper growth or a pro non-poor growth because not everybody know non-poor is rich so very often we say pro rich but and then in the case of recession what you could have is that distributional changes have favored the poor or the non-poor even if it was a recession or it's harmed the poor less or the non-poor less and then I would call it an anti-poor anti-non-poor and from here one could actually establish a proper growth rates which could be the growth rates per capita and plus a factor of inequality aversion and so on but that's always a bit funny to introduce I think I'd rather think of a stand-alone index that can tell you how much pro how pro-poor growth was in that period and from here one could one could define some some indices to evaluate for instance you could think of given the growth over the last period how much pro-poorness or how much distributional changes would we have needed to achieve a certain poverty or poverty gap reduction goal like having it or eliminating the poverty gap or so on meaning I agree that growth is a necessary condition but not sufficient to reduce poverty and then this at least I could tell you how much effort you should be doing in the distributional site to contribute to poverty reduction and the same you could also and do the same a bit like as a goal for the future and say by 2030 we want to reduce poverty to so much counting on certain growth then we wouldn't need so much changes in the poverty gap because of distribution I'm not dealing on policies at all yeah so don't ask me on policies because if I knew that okay so here I applied the data to Honduras and just so that you show to show what one could get and get in the index at least what I managed is to classify each of the spells that I look at and then see if in this spell distributional changes were in favor of the poor or not and in the case of Honduras that so for the moderate and extreme poverty lines so you have that it was anti-poor and anti-poor contractions but they have pro non-poor pro non-poor in fact this is the only one that has progressive changes and it's actually it's after the hurricane meet hit Honduras so it's quite interesting the hurricane hit and in fact the poor were less affected than the non-poor which I don't know if that's very consistent but that's that's what it could be that the the rich lost more assets imagine or at least more quantifiable monetary assets and and this is just to have a look at the at the possible tools for evaluation and we could see that this the changes in distribution here were like 1.2 percent of what you would have needed to eliminate or to have the poverty gap and here that overall over this period 2004 2007 the changes in distribution that the poverty gap changes because of distribution were in the other way that they should have gone and they were of a magnitude of 30 percent of what you both needed to achieve the 2015 goal of halving the poverty gap so why am why am I claiming that this index is useful I'm claiming that it's easy to interpret this is distributional changes in the poverty gap any politician can understand that hopefully and it does measure growth it does measure it does measure how proper it is and it's really related to to a disproportional benefit for the poor so at least I can claim that for this definition this seems to work so what I'm what I'm doing now is calculate this for the whole world income service and hopefully in a few weeks or months I'll have this for all world countries and spells if possible that's a yeah that's it thanks thank you