 Olen myself, Fin, jossa meillä on yksinkertaisuus, jossa aloitamme vähän. Tätä katsotaan, että meiltä yksi on VEED, jossa on tärkeää yksinkertaisuus, ja 2 esimerkkejä, joilla on valmiita alueella. Minulla olen saanut sanoa, että VEED on tärkeää historiassa, Ja sitten laitettavasti, mitä olemme tullut tämän vuoksi. Ja sitten, kuten mietin, jossa on tullut finnistoksen, olemme myös nähneet varmasti preliminallisuudet, että mitä voidaan sanoa, että 3-dysipuisiin kuten dataa, jossa meillä on. Joten tämä on planohoitoksen. So I skipped the first part, but the, so I'm, as I said, I'm going to introduce the latest update of the world income inequality database, which we have dubbed the VIT 3.3, and then I'm, then I'm in the latter part, I'm going to just describe to some descriptive analysis and what's currently being done to address inequality, so this is the extent of redistribution across the world. And finally, if time permits, I'm going to show you some preliminary regression results on the determinants of redistribution, so what if you try to explain the redistribution in all countries in the world for which we have data. So let me start about the, about describing the VIT. So this is, as many of you know, this is a collection of information on economic inequality in all countries of the world, and this is freely available from the wider web page. So the VIT was first put together using combining World Bank data and then adding some additional observations, which were basically first collected for a research project, but then it was realized that it makes sense to keep them out as a public good to other researchers as well. So this is, if you like, this can be called like a VIT version one. VIT 2 was then a major revision, which, where much more information was added on the underlying assumptions regarding the observations and also a quality rating to the observations was added. And the latest update of that took place in 2008. And now, and then last year, we introduced VIT 3, where data for seven more years were added, and then we addressed comments by Professor Stephen Jenkins of the LSE, and I come back to those in a minute. So the basic philosophy in the VIT is really the notion that the Gini index can mean very different things, depending on whether it's defined on the basis of income or whether it's defined on the basis of consumption. Then even with an income, obviously we can make a difference between gross income and disposable income, so income before and after government intervention. We can use different equivalent scales, or then equivalent scales are not being used, and the area and population coverage, et cetera, et cetera, can differ. So the point is that the VIT gives all this information to the users. So what it does is, in our opinion, it enables reasonable comparisons of inequality between countries and across years in a given country. A final point I would like to make on this slide is the last bullet, which is that the inequality data in developing countries especially is not collected every year. So an outcome of that is that the VIT actually has many empty observations, but we don't pretend that there actually would be data, so we don't do any imputations for the raw data. So we work on the data, what actually exists in the world. The VIT was evaluated last year, and there's a forthcoming journal, a special issue in the Journal of Economic Inequality. And as I said, Professor Stephen Jenkins of the LSE made a comparison between the VIT and what is called the VIT, the standardized wording of inequality database. And Jenkins' main conclusion was that he rated the VIT as a credible source for work on cross-country inequality, and also superior, by the way, to the VIT. And this is precisely because we give all this information to the user that they can actually know the data if they want to what they are using. Moreover, he had a number of useful additional points, and we did our best to incorporate them all to the revision of VIT last year. And these are being explained in our response to Jenkins in that special issue. Now, the latest VIT is almost ready. It's an annual revision, so we are planning to update the database on a constant basis, and it will be published shortly after the conference. So, obviously, there are some new observations. We have also simplified some of the categorization of some of the background variables, so we felt that it's easier for users to work on, for example, just like a three or at most four equivalent scales than rather than a greater number of those which are currently in the VIT. And most importantly, there's going to be a more user-friendly query and visualization system, which you can already test in the, what is called, network cafe here in the foyer. So, that was my first bit, and now I give the floor to Pin, please. Thank you very much. I don't quite look like Francois Bourguignon. He was supposed to be the speaker on this panel and unfortunately got in golf dinner meeting in Paris and therefore will only be here tomorrow. So, I had to step in at relatively short notice, but I prepared something which I thought I should contribute. For those of you who wish to hear what Francois would have said, there is now a session tomorrow, which will also be covering inequality in China, Mexico, South Africa and India, together with Francois' global overview. Now, in September of 2014, WIDER actually did hold a conference in equality measurement trends, impacts and policies. So, I should not repeat myself. I did make quite a number of remarks which are now available at the WIDER YouTube, but it's not just because I want to show a picture of myself. When you actually go to the WIDER YouTube, you will find a series of presentations on these issues, which I would really strongly recommend. I think this is probably the best set of videos that we have ever made on a conference. And, at the website, you will also find, for example, the information, the presentation, the keynote by the then Minister of Strategic Affairs of Brazil, Marcelo Cortes Neri, who gave the keynote in September of last year. It's also pertinent here to highlight that WIDER has a very strong and very proud traditional work on inequality. Two former directors of WIDER are in the audience, and I wish to stress that what we are doing now is really that we are building all of that hard work that they put in. It's that we are trying to continue to the best we can. In addition to that sort of in-house work, well, there is work beginning with Amatya, but also Andrea and Tony Atkinson and Tony and many others. Now, what is the point that I kind of thought I should try to leave with you? Well, it emerges a little bit out of discussion from Vietnam. I mean, some of you may know that I work quite a bit in Vietnam. I've been going there four times a week, four times a year for 15 years after staying there for three years. And there's a debate about, is Vietnam a success? But then others say it's a failure. And I've sort of been puzzled about, I've been sort of scrapping my head. I mean, why is it that when I sort of try to say, look, Vietnam has been doing very well, then others are really saying it's not going very well. Well, we as economists, we know a formula, which says T times G is equal to 69. This is basically the doubling time times the growth rate is equal to 69. So if you grow by 6.9 percent a year, you double your national income every 10 years. Now, so take Vietnam, take 1986 and assume that at that point in time you had one dollar per person in a family. Well, after 10 years you had two, after another 10 years you had four and after another 10 years you had eight. Assuming, of course, that this rate of growth was equally distributed, all that kind of assumptions. In other words, you moved from one to eight over 30 years. Now, take another individual who had 10 dollars or a family where each individual in that family had 10 dollars per capita. In 1986, 10, in 1996, 20, in 2006, 40 and in 2016, 80. Let's not mention the one with 100. Now, already Plato warned us that we have to be very careful when we are speaking about how we understand the world. The relativity of what we see and how we believe that the world is actually hanging together. So that's one first point of departure. Another point of departure was already mentioned quite extensively by Stiklis in his notes just a few minutes ago. But this is actually a quote from the World Development Report 2006, which I always keep in my mind. It was stated there that the dichotomy between policies for growth and policies specifically aimed at equity is false. So that was the point that Stiklis tried to make. Another citation from the World Development Report is that the distribution of opportunities in the growth process are jointly determined. That for me has always sort of been something that I've kept in mind because it's incredibly important because it highlights why it is that we are focused so much on inequality and trying to understand it and trying to measure it right. Now what this implies is that some policy can involve redistribution of influence and advances or subsidies away from dominant groups. That's what UK is going to speak something about. But I think it's also important here to remember that good redistribution may not always be directed to the poor. This is what Steve O'Connell was having I think in the back of his mind when he last night sort of said something about that well we could just try to give all the poor the amount of money for them not to be poor. But then was that really the right policy? There are trade-offs here. Now I'm in the middle of doing some work with Miguel and Lawrence on using the WID and this is what I'm going to try to report. The questions what are the most recent trends in global inequality? Has global inequality increased or declined? Some of you may be surprised by that question but I actually believe it's a reasonable question to ask. Have these trends been homogenous across regions? Ravi already alluded to part of the answer and now comes I think an important point which is that is the picture of global inequality trends using absolute measures of inequality consistent? With a picture using relatively inequality measures. If you follow the debates surrounding the post 2015 development agenda you might occasionally have been somewhat confused. And I can at least say that there are quite a number of rather confused policy briefs and other things out there at least in the general development discourse. Now the predominant relative inequality measures such as the genia and the tile. Well we know that values remain unchanged when every income in an income distribution is uniformly scaled up. So in other words what happened to inequality in Vietnam between that person who had one dollar a day to two was it eight and then the one who started with ten and went to eighty. Well inequality didn't change. Relative inequality didn't change. That's a number that economists tend to focus quite a lot on and if you look to the literature that has very much been the sort of reference point. And we would often kind of implicitly say well inequality is going up or down based on that relative measure. But what we are sort of trying to work on is that well the less commonly used absolute inequality measures just such as the variance. There values remain unchanged when every income in an income distribution has the same income added to it. Now sometimes you would argue that relative measures have been sort of associated with a more conservative judgment was absolute inequalities more kind of left as if you wish. Now data you already said what is the database we've used. We've used the updated information we have in the WID and what are the results. Well has global inequality increased. Well no it has decreased if you take the green line here the genie has gone down systematically since 1975. To 2010 relatively sort of step by step but you know it has been falling and it is the same picture when you look to the tile but the tile you can decompose. So there you see what is it that has been driving this. Well it's actually the between a between country component that has been driving the fall in global inequality. So said a little bit briefly. It's the rise of China and India which have been which have been growing much faster than what we've seen in development. So it's that sort of it's that difference that has been closing. It's not the within country component that has been explaining this. If you look at the red line that's the between country component and you can see that's the one that's sort of driving. But the one thing to just take from this is that has inequality fallen. Well has it been stable. Has it increased. Well measured by the relative inequality measured. It has fallen. I'm not now saying whether this is the way we should measure it but just us making that observation. Now we can also look at relative regional inequality and this brings home rubbish point very clearly. You see very significant differences between the different regions. This is in this case with the tile index but you can see that the development has certainly not been the same in the different regions of the world. Let's just take Europe as one example. Well some countries have experienced a steep rise in inequality. Denmark my home country Sweden France Bosnia and Herzegovina are countries that have seen a steep rise in inequality since the two thousands. I don't know whether this is why the new government in Denmark now is going to abolish the use of poverty lines. This has now been taken off the public agenda. Resources are no longer going to be used for that. Other countries have observed the decline in inequality Belgium Italy Norway and Ireland. Then there are others with a more flat trend and then again others where you have different sort of trends over this 35 year period. If you look to global absolute inequality estimates this is just using the variance. There are other measures but they pretty much show the same thing. Just boom up there increasing a lot. Think Vietnam think that very simple example I gave in the beginning relative inequality stayed the same. But the one person who had one dollar arrived at 80 sorry at eight and the one person with 10 arrived at 80. Your perception of what happened may be very different depending on which number you're looking at. Now using the way using the data you can also make a series of we call it counterfactual scenarios is really just calculation examples. I'm not saying this is a full sort of counterfactual but you can sort of ask these kinds of question. Well what if India and China incomes per capita and the distribution of income had remained at the 75 levels. Well then if that had happened then global inequality would have instead increased during this period from 0.739 to 0.757. According to the Gini and it would have also increased by the tide index. So what's the conclusion. Well using standard relatively inequality measures global inequality steadily declined over the past three decades. There's a lot of heterogeneity out there that we should work hard on trying to understand as Ravi in my view correctly pinpointed. But we might occasionally be very careful in trying not just to take the North American experience and then impose that on the rest of the world in how we are thinking about what has been happening. When using the absolute measure the variance inequality measure we find the global inequality has increased dramatically. Now I just mentioned that I am Danish and of course for that reason I often refer to Nils Bohr. He was a clever man got the Nobel Prize in physics. Physics is something that you might sort of see as that's really stable. There we really do measure something that sort of kind of we know we're talking about that. Not so variable that's not so susceptible to many things as in economics and social science. Well Nils Bohr did argue in his complementarity theory that with observations where we believe that we see the same thing. We often see something different and therefore will arrive at different insights. And I guess the point is that these insights are not necessarily contradictory or meaningless. They are complementary. I'm not trying to suggest that everything is relative and so on and so forth as Montek was discussing last night. But I am trying to say here that it's very important that we understand that the relative glasses do not give the same kind of understanding like the absolute ones. We need to keep this complementarity in mind. So yes I would strongly echo Atkinson and Brandolini in emphasizing how central the choice of measure is to any discussion of what happened to global inequality. And I may be overemphasizing that some may say that this is rather banal but I can tell you that I have during the last sort of three four years been sitting through one meeting after the other where this point has not been kept in mind. I will basically stop here just with one question. Over the past 35 years relative inequality has fallen and hundreds of millions of people in the developing world have been lifted out of poverty. I think there's reason as has been said before in this conference that this is a major achievement. But what different policies have managed this without the increase in absolute inequality. And how do policy makers minimize this trade off moving forward. I believe that just to something maybe to discuss. Thank you chair. You have something to add. Yes. So I was planning to add something on on some resource on redistribution because while it's the case that the using this conventional measure. Can you hear me now. OK. So Finns pointed out how the using conventional measures world income inequality has actually declined but it's still at the very high level. So then the next question is what should we do about it. And that obviously leads to the question of how much much can actually redistribution affect the living standards of all people in the world. So that's what I wanted to do in a in a working progress project together with Risto Rönke. Who's also at wider. So what do we do. We use the weed to measure the extent of redistribution in all countries in the world for which we have data. We use three possible measures of if you like absolute redistribution. So this is just the difference between the gross income and disposable income genies. So the one is to use the gross income and disposable income genies. The second one is to use the gross income genie and consumption based genie because consumption is going to be afforded by income which is disposable. And then the third measure just to make sure that we capture all possible observations there are is an in a sense a hybrid measure where either disposable income or consumption genies is taken out of the gross income gene. One can obviously also examine relative redistribution. So this is the extent of redistribution divided by the underlying inherent or gross income gene. When we did the calculations regarding redistribution we favor. You remember that in the indeed we have the quality rating so we favor the high quality observations. And if you know the equivalent scale we use the adult equivalence based genies. So the first question is how many observations have we got. So this chart gives you an idea of the number of observations over five year periods in the data sets on the on the extent of redistribution. Now this is the broadest category. So this is where redistribution is measured either based on the consumption or disposable income reduced from the gross income. So as you can see I mean while the coverage or the number of observations increases over time it is still the case that a large proportion of them come from the developed world. So from the industrialized world for example the bottom two parts of the column here refer to the European and the American continents. So another point is that the number of observations goes down even again if we make the restriction that the equivalence scale for the observations for which we have gross and net genes must be the same. So then it drops from something like around 40 to around 30 during the latest decade. If you still then some information about the actual extent of redistribution across the world. So this is this is absolute redistribution. So the vertical axis gives the reductions in genie because of government intervention. And the top lines refer to European countries and then post Soviet economies. Where the absolute redistribution has been around 15 points in the genie index. And then it's already clear from this graph that the developing countries for which we have data you need to remember that for developing countries we have very little coverage. There's not much redistribution going on. And the same can be said if you if you if you take the view of relative redistribution but I skip that in the interrupt of time because it basically gives the same same idea. So given the sort of a relative paucity of observations I wanted to make an aside. To and and and revisit. We revisited an influential study by the IMF from last year. So this is a study by Austria Berg and John Reades which examined how redistribution affects growth. So the idea in their paper so this is motivated from from actually from a more old fashioned idea of Ockens leaky pocket that there would be an innocence this classical classic efficiency equity trader. So that if you want to get a higher growth you need to make some sacrifices in terms of equity. And they they they then use the macro data what's available to explain growth both by closing community quality and redistribution. And the result is that the that there's quite a strong a negative impact on from closing community quality on growth. So income inequality would be harmful for growth. The finding is also that redistribution wouldn't affect growth. So the bottom line of that study is that there wouldn't be no trade off after all in the open sense in the in between growth and equality. So the trouble in the study in my mind is that the is the underlying data. So the data for the inequality and redistribution come from the suite. So the standardized world income inequality data is maintained by a salt. So in the suite if there's no data for a given country in a given year. This is my understanding of what salt does is that both growth and net income inequality values are then imputed based on values from the same country in different years. Or countries in a or from other countries in the same year. So in fact even though we know that the that the for many developing countries we only have inequality data for consumption based measures. And in the time intervals of five to maybe seven years in the suite they are observation observations. And why is that this because these are all imputations. Now I'm not claiming that we have in the week all available inequality data there is but we have quite a bit of coverage. And once we make a comparison of how much information we have in the week versus the suite which is used by the IMF study. We come to the conclusion that actually a very large proportion of the data they use to measure redistribution are based on these imputations or if you like cases. So even if I personally would very much like that conclusion in the study I think I need to say that the I mean in my opinion we really don't yet are in the point where we actually would have the data to support that conclusion. And here's an example on where we compared with Risto the coverage on redistribution in what we can find from the weed versus the coverage in the IMF study. And they run separate regressions on the full sample then what they call a baseline sample and then a restricted sample. And the difference between these samples is the extent of imputation they allow the analysis to use. And as you can see I mean I know that some of the developing country observations are missing from the weed but yet the difference is really striking. Even when we compare the coverage of weed which is the column to the right and then the coverage in the weed in the restricted sample. So if there is still time I don't know. No you have two minutes. Two minutes okay so then I can just within the two minutes I have I can say something about the results on the determinants of redistribution. So what we did was we used five year averages because Gini is not available for all years and we run simple regressions. This is not about causation these are correlations between absolute and relative redistribution and then explain them by some of the economic and institutional variables plus inherent inequality which is the gross income inequality. And in all those models where the results are here from we use Gini based on the common equivalent scale. And we also use the model to predict what is called the redistribution effort. So that's the difference between what we observe in terms of redistribution and what our nice little model predicts. So the difference this earlier work related to this but I think the main difference is that we use a much larger set of countries. I skip the regression results and the predicted redistribution I can come back to those in the end plus these slides will be available from the net after the session. So what are the results? So we find not surprisingly given already the descriptive evidence that GDP is very strongly positively linked with redistribution GDP per capita I should say. And in a log log specification we get an elasticity of around 0.3. So also we find that the ethnic unity is very very strongly positively linked with redistribution and so is the original Merley's idea in the Merley's 1971 optimal income tax model whenever the inherent inequality, gross income inequality goes up. The government should react by increasing redistribution and this is what we observe. Surprisingly the extent of democracy is not positively correlated with redistribution and then we also calculate some of the top five and bottom countries based on the ranking in the predictive level of absolute redistribution. And they include some rather surprising countries. You need to remember that this is relative to our model predictions. Right. So my conclusion regarding the bit on the redistribution analysis is that we need to remember that the data is lacking for many developing countries. I think all of us including we at wider we should do more in making sure that we capture everything there is based on the data we do have at our disposal. It looks like redistribution efforts are very very strongly linked with economic development. And now if we come back to the idea of world income inequality. There are papers around. There are some papers for example by Francois Burkin-Joon who unfortunately cannot be here today but he will be here tomorrow saying that the redistribution in the whole world taking into account things like foreign assistance. ODA, official development aid, only has a very minor impact on world income inequality. It does have an impact on those at the very bottom but then the conclusion I think is that there is clearly a need if you want to address the growing inequalities in many countries. Is the urgent need to build a more comprehensive social protection systems within the nation states as these countries develop. So thank you that was all I had. Thank you.