 My presentation is the result of the ambition of a silly man who after having started the inequality in Latin America and found some result Decided to look in three weeks to what happens to inequality in Africa So when I saw this conference, I told Finn or perhaps I should come and present some of the results for Latin America to see whether they make sense for Africa and then he said no You just write the paper. So this is the last three weeks. What was I done? Now, what will we discuss? Well, I think that the general tendency I mean Let's say during the last 20 years ago, I remember There was a huge debate on global inequality a la Milano, which I mean among world cities now now there is a much more attention to within country inequality and That is also due to the fact that we do see that in some countries the majority of the countries of Latin America and Quite a number of countries in Southeast Asia is inequality has improved massively now the situation of Africa sub-Saharan Africa remains broadly unexplored and then we see it here We can we keep hearing very very different opinion So what I try to do and you will I mean I have to apologize if the results are still very preliminary We're trying to look at what happens to in a quality in Africa using Latin America as a benchmark Not that Latin America is particularly similar to that but in Latin America. We have a very clear story And so we are trying to see whether some of the factors which have been Important that in America are relevant in Africa and many are not and what are the factors which would not be? Relevant in Latin America are relevant to Africa They I can't display the conclusions that I would say then 60 70 percent of the cases Africa is his own drivers inequality drivers Now one reason why I also looked into inequality in Africa is that I run by mistake Into this paper by Salah Martini Pinkowski Who argue that in Africa inequality and poverty felt much faster than what people believe Now so and you know this is a paper which has received quite a lot of attention But I mean I think that the bit the paper is based on bad statistics and so They reconstruct data for some 48 countries while in my case I am able to reconstruct it for 28 countries then they make all sorts of assumption about the shape of the distribution and then they come up with this With this shape and then of course the story is that the since this happened in from the 1990 onwards It means that liberalization globalization is good. This which is also questionable from a policy perspective Now this is what happens to Latin America and in Latin America. These are well measured data in Latin America We have what is called said luck? Which is a big project financed by the World Bank and I think you and the P in which the University of La Plata Standardized all the micro surveys going back for 20 years using the same way of computing the data Missing income imputations grossing up and so on and so forth. So these are fairly reliable data They're not perfect data and they're not perfect. What what Andy says is basically As the efforts by Atkinson size and Piccetti and the other show Basically what when we measure the income distribution, we measure the 90 percent percent as or 1995 and we don't really know what happens to top incomes and that is also true in Latin America. Now, there are two tree of this Assessments of inequality based on tax returns, but basically we have one for not over time We have one for Argentina in 2002 then we have one for Brazil and so on so we have no time effect The key point is that is unlikely I mean, I will not tell you why is that the even if I assume that The the rich have become richer and then brings like the financial sector Accumulated the greatest share of national income in Latin America is unlikely that we'll have fundamentally altered this one now Alvarado has shown with a very nice pay The correct a genie can go one way and the normal genie can go that but for sure We know that there has been a redistribution between the 1995 bottom part of the distribution now Now what can I reproduce this chart for Africa? No and first of all because you wouldn't come out like that, but basically because first of all there is no said luck So I mean, there's not a standardized database So what we do we started to go back to all the sources of within country in equality And these are with which I'm proud to have created in wider now I'm also very happy to hear from Finn that this is a new Data set would be released anytime with two thousand more genie points And then of course we went to the World Bank and we use pop call and then there is Milanovic Wyd Which has more or less the same data plus something else and now the World Bank has also a set of I don't know many 200 consumption data Consumption service which they are standardizing like what it is like But standardization in this case it requires so many assumptions that I'm not sure that they will solve the problem The example they gave me is that in one serving Burkina was a consumption survey So they asked the families What do you consume of millet goat? Vegetable oil and Medicines for instance only for consumption item were measured in the subsequent survey 300 So I mean if you want to standardize those would be very difficult. So this is the situation We have a weak informational base But so what we did is basically we compile the data according to all these different Compilations and then we see if there are sort of a common trends They could all be wrong perhaps but we certainly are comforted by the fact that In several cases that are Common trends emerging then there are some outliers and then some series including from the World Bank or with in which Time series, which is not a complete time series. You have a six or seven data The gene goes up by 20 points or goes down by 20 points, which is impossible So we eliminated one and then in the end we retain 28 countries out of 90 But fortunately these are the most important African countries So so we these captures 95% to the population on GDP Now we do come up with the different trends Which are like that now, I think that I mean I mean when we did that basically I didn't do any particular statistical test We just look at the scutter and then we will do it Now so you have some countries where there is more or less mean this is the average of this country So there is variation and so so you have some kind of about eight in which there seems to be a decline Then others where there is the climb but then followed by an increase then there are other countries Which are increasing all the time and then others in which perhaps the increase is being Reversed during the last few years Now the question is it Now let me repeat it again Now we will go back to this date and look at them even more carefully We will try to prop into a lot the World Bank will give us the the new data Actually, they already send them to me, but I was the send them to me one day before I left So we will track gains all possible sources and then we have to say we have to try to make a story out of it Now when you have weak data basically you have to rely more on economic theory and on that we do know a little bit more about Africa So now the method is the same one which I mean the way I'm proceeding is the same one Which we're using for the Latin American study So we do distinguish between what we call immediate or statistical causes normally There are many case studies in which like the Niconio did a very nice study for five countries In which you take two points in time and you do a sort of a statistical decomposition Which tells you that ink the genie went up or down because of changes in factors in their Different types of income shares and in their own concentration coefficients So there are standard the composition techniques learn minutes sake Milano which and others Now that tells you For instance that labor income went up is this is the source of the decline in genie But it doesn't tell you why it doesn't tell you the the causes the phenomenon So you have to look at what we call the underlying factor now you when you look at the underlying factor You cannot use the same decomposition technique because it applies to 40 50 countries and You have variables like macro variables, which are not the composable in the same way So here what I'm trying to do is to summon out is on the basis of the literature what happens to The immediate the term so the statistic now in Latin America the key point in the analysis that declining skill premium What what drove the decline in inequality during the last 10 years more or less and that is because it has been a massive increase in the supply of skill labor Which reflected bigger for some educational area did we find that? Same is it equally important in Africa? No, if you look at that we'll see some data on education and then is it the skill premium important perhaps in the urban sector like Cogniforms that is what he finds also for some of his paper But not not generally then so these are all the reasons so there's been a stable demand for skilled workers rising demand for unskilled now In Latin America as the fall in the urban rural income gap being a central point The key factor in explaining the climate one not except in two or three central American countries, which are much more agricultural So so here we already start seeing that the paradigm for explaining it is that in one region is basically the supply the amount relative supply and demand to skilled labor and in the other one is the the size and Trend in the rural urban income gap now the question is on what does it depend this urban rural income gap? Well, first of all, I mean luck I mean there's increases in world agricultural prices and not only for copper and oil and so on and so forth if you look at them The price of maize this is double, you know So now of course most of the maize produced in Africa is eaten in Africa But there is a modest transmission of all prices on domestic prices, which may produce an incentive second, I mean the exchange rate policy and I think that I think Andy and other people have mentioned before Basically now there are more reasonable macroeconomic policies. I mean there's not that is Overly deval overly value overly appreciated exchange rate now thirdly in In part of Africa and these I think is one of the good part of the story There is a little bit of a maize green revolution and perhaps there are others which have not been able to capture But one of my students has just done the thesis on the green maize revolution in eastern Africa Which shows and I will show some chart that there's a and then liberalization of agricultural markets So basically these factors they are central to the analysis of the decline of the urban rural income inequality urban rural income gap in Now in America there has been Here we have Armando Armando Marietta's who wrote the chapter in our book and basically I think that There is a long debate, but I mean it's quite certain that The bull's collar and all these type of transfers mean they They are they've been important and Pais de Barros who is a Brazilian economist basically argues that 2.5 genie points were explained by the bull's collar and various types of transfer now these transfers are basically Budget finance, I mean they are paid by the government now in In Africa except for the southern African pension systems basically The transfer system it does not operate like that Remittances in Latin America remittances turned out to be surprisingly Equalizing against the theory now in Africa. I don't know I mean I haven't been able to find evidence in this regard yet Now what are the underlying the terminus then I mean you say you can start saying well both First column is Latin America and the other one is well In terms of trade, I mean I think that data shown then even in regression analysis We do find that that the terms of trade gains for the oil and metal sector basically are highly equalizing The production function yesterday I was talking about production function I mean the production function in that sector is very capital and skill labor intensive So if the oil prices go up there will be some Spanish engineer will make more money But not not not many unskilled local workers The foreign direct investment in which in Africa they go mostly to this and in Latin America as well The most they go mostly go to the primary sector. They're equally an equalizing the market remittances We don't know and then for an aid for an aid is irrelevant in in Latin America and highly relevant in and then here there is a chart Where is it? I lost it Okay, there is a table. I mean basically in Latin America in Sub-Saharan Africa, then the low income comes to receive a shock of about six point of GDP per year And that if it is properly used, this is a very very important to correct here on this now Exogenous changes may dependency rates mean in Africa. There are no changes in dependency rate because birth rate have not fallen Self-wound revolution is we don't know how to measure it as yeah, so perhaps is relevant And this is what Radley says now. There is a decline in it HBI hates I mean there is no number I don't see any more papers but I wrote a book on the economic impact of HIV AIDS in Africa and basically Here what we say this is the prevalence rate in southern Africa and you see that there is a decline But very very tiny one. So it is unlikely that the improvements on the front they have Reduced the rate of degree of impoverishment of poor families. So that is unlikely and Conflicts conflicts in Latin America are irrelevant here. They are they do are relevant and then we measure them with them Now then then you have the growth pattern and I think that in Latin America It is slightly equalizing the top one And now this is all countries Confounded and without control so they cannot really say much But this is the second one the bottom one is the one for Africa. So basically it's totally neutral So growth here appears not to reduce income inequality so And what on the other side? We we see is that this type of growth is a growth where the resource countries when there is a massive Primarization of the economy so the countries which are yes, thank you the campus which are With a high share of natural resources on GDP between 92,000 and 2010 You know on average now resources account for 42% of GDP and these are These are resources which they are oil and metals and Muslim oil and metals. So we see that for about 18 countries The natural resources have increased quite a lot and that tends to be an equalizer Now policy measures now both country I think that they followed similar policy have more reasonable macroeconomic policy with one difference is that in Latin America in Africa They the the reduction of the debt has basically largely been driven by the big initiative So we were very I was very skeptical myself, but actually in the end you do see that this is delivered now the other Trade policy I mean in in Latin America turns out that trade policies and equalizing during the first period and then neutral in the second So an equalizing 80 90 in None and neutral in the second in Africa. I don't know yet And as I think that that is a major point now if I listen to the African debate many people complain about China displacing everything taxation now taxation actually Amazingly Africa has increased taxation So the whole idea that the state capacity is that bus can be improved and if you look at the central bar I mean the central column, I mean in seven or six or seven countries the main source of taxation are direct access now Educational policy, this is a major point and you see this is the ratio skilled workers With second and tertiary to the other one So and here is the proportion of rural population and it's a little bit like what Eric in another way show before So basically there is no increase in second and tertiary education in countries with very high rural Rurality and the question is it is the government was doing a terrible job Or is it the fact that families in a backward agricultural society do not want to send children to school Now this is the eels maize eels Hectograms practice and actually you do see these are Malawi and Zambia and actually you do see that there is some No, the political fact has been in Latin America in our Analysis, we do also pay attention to both number one whether the country is democratic or not And second it turns out that the proportion of left-wing regime Tens I mean left wing means Brazil or why social democratic or more radical matters Now here is now the scale corrects it, but I mean it's quite striking that the period in which inequality fell in Latin America. I'm almost done in basically shows that the very sharp rise on the Left center center left regime and very sharp decline in the right-wing regime Now in Africa we don't have a coding of the country by the political regimes, but do we have sort of this? Trends about democracy and democracy and hypocrisy and you see that there is some some improvements But not really much of that now. I think this is that there are Now so some of this data we try to I mean this is very preliminary Okay, but then first of all we find that GDP per capita growth tends to be equalizing if you multiply if you interact it with the agricultural production index Which means if growth takes place in agriculture, then then you will have lower poverty and lower inequality Now the share of work with tertiary turns out to be no significant But if you take the share of work is with the without education tends to be unequalized Then the external debt for some reason doesn't turn out to be significant The taxation turns out to be highly significant Turn turns for trade if there are interactive variable mineral rich turns out to be highly unequalizing remittances tend to be slightly Equalizing aid on GDP doesn't turn out to be nothing for the moment and FTI foreign direct investments is as in Latin America alien equalizing because it goes to the resource sector and then that conflicts isn't equalizing and the quality of public administration is Now I conclude I mean I think that on Based on this highly preliminary work. I mean what can we say in terms of reducing inequality in Africa? well, I think that It is a massive need to reduce the rural urban income gap and then I start that Eric show that I mean there is decline in the share of agricultural work without increasing GDP That that is a sort of worry and then what you want to see is that a non decline of the share But with an increase in income in rural areas, so we don't want to use the Lewis model, but the Rani's fame model You know, so why how does it happen? I mean these are the oldest fact of the revolution liberalization sector and so on and so forth Then the second thing is which I didn't mention so far is that now there are something like five percent of the land in Africa has been taken over by land grabs and Now in part and and there is a nice chart which shows that this happens not Countries like I don't know Central Africa the public but in countries with fairly fairly low landman ratios So the land grabs do not occur as one would expect in land abundant countries But they occur everywhere and perhaps even more in these countries Now I think that the other lesson is that there must be efforts up for intensifying a vacation. Yes Let me just finish this one. No, then effort in tax collection and some transfer now the final point that was also mentioned in the prior section is democratization and And if I listen to most of the African debate is that avoid reprimarization I think in Ghana in the crowd in December there is this initiative on Africa inequality and I think it's very much based on avoiding being Brought back to the colonial division of labor. Thank you