 I think one of the things that's almost certain when you ask to present is that you'll prepare probably double the number of slides that you have time for. And then if you interact that with a draconian chair who sends an email the night before telling you he's going to manage your time, you know that you're never going to finish your presentation. But I'll give it a go. So this is a paper that tries to do three basic things. One is to describe trends and even in a comparative static way trying to understand inequality in Sub-Saharan Africa. Secondly, to give you a little bit of a story about structural change. If one thinks of this morning's presentation from Justin Lin in the context of Sub-Saharan Africa and then finally sort of a whistle-stop tour of possibly some of the drivers that may be influencing and shaping inequality. I should say at the outset that in almost every single case we are not working with micro level data. If you're within the World Bank or the UNDP as the case may be, country level agreements will give you access to the micro data. And as wonderful and as nice as Kathleen and Chico and the World Bank staff are, they're legally bound by those agreements and cannot give you access to the micro data. So for individual researchers working with such data, it actually becomes quite difficult. So most of our estimates are based on the POV-CalNet data. So we all know the background, or at least if you work in Africa, the background to this sort of, is this the final frontier as a continent where six of the 10 fastest growing economies in the world have been from Sub-Saharan Africa? And that's really been the story if one thinks of the public debate and the discussion around Africa. But of course lurking beneath that lies the slightly bigger concern around whether such growth rates or whether this growth spurt has generated distributional outcomes. The literature as you know is wide and deep. I've summarized it in the paper which you can have access to as sort of three big storylines. The first is essentially that you will get that growth accompanied by, if I can find mine. So growth accompanied by a rise in income inequality will reduce the growth poverty elasticity. And that's just a simple way of saying that inequality is often the thief of the growth poverty relationship. That if you have large increases in income inequality, it will reduce the elasticity. There is the path dependency issue which I'll come back to later that a high initial level of income inequality will make it much, much harder for poverty to fall in the face of high growth rates. And related to that is the notion that you may, you often find a fairly wide range of growth poverty elasticity but not so much with inequality and growth. But that's a literature that's well known to this audience. So what is our headline result? Well, we took the POV CalNet data and looked at the average genie coefficient for Africa and all the detailed data issues are here and in the paper. And we came up with an average genie of about 0.43. When we looked at other developing countries, the average genie was 0.39. Statistically significant, right? The difference. We took different measures. We took the ratio of the top 20% to the bottom 20%. And that ratio was 10 to 1, right? In the case of Sub-Saharan Africa. And it was just under 9 for the rest of the developing world. If you split the sample from Sub-Saharan Africa into income groups, low income, lower middle income, and upper middle income countries, you get a similar result. What you tend to find is even by income group categories, income inequality levels in Sub-Saharan Africa are higher relative to a sample of developing countries. And that was, I think, a fairly new result, as simple as it is, that we didn't have a sense of what African inequality looked like. So we scratched a little bit more. And the Komongorov-Smunov test for the significant difference shows that these can be rejected. So these guys are actually significantly different from each other. And so even in a distributional sense, you've got higher levels of income inequality in Sub-Saharan Africa, the blue line relative to the rest of the developing world. It turns out, though, and if I had to pick a sort of one result that surprised us is the second bullet here, that the data shows that there are these seven outlier African countries and we list them there, the sort of Angola car Central African Republic, and then a whole host of, interestingly, Southern and Eastern African economies, which drive this result. When you remove those seven outlier African economies, there's no significant difference in that previous table between the mean African genie and the developing countries. Here's a huge economic history project. And if Finn was here, it's a big wider project. Is there something going on? Some would argue that there's the economic history of colonialism and conquest in Southern Africa that's very different to West Africa. Andrea mentions that in his paper as well, that may explain this difference. But certainly without these seven outlier African economies, inequality levels between developing country regions and Southern Africa disappears. The movements in the genie over time, as Kathleen has sort of referred to as well, but for us, it tends to get messy because now we're working with absolute estimates. We're trying to average across surveys, and we're not sure about the quality of the surveys. We take them all from POV, CalNet, and so on. But essentially, you get a result, and I will be skipping a lot of slides as a forewarning, that looks like so. If you look at the three right bars, right, that essentially over the period, 94 to 2013, you've seen an overall decline. That's the red bar in inequality in Southern Africa, driven principally by the lower inequality economies. Now, if you think of that second slide about relationships between inequality and growth, makes sense, right? You've got a part dependency in high-income inequality economies that make them much harder to realize growth, which is inequality reducing. But we note that this is, you know, the whole sort of Africa is not a country dialogue. Just remember that this sort of trend can be generalized to the continent, but certainly not at the country level. So if you look at the country level where we could get decent sort of series of data, you see it's fairly messy. So you've got economies where you've seen constant, some declining, and some increasing in income inequality. So we tried to look at the GDP story. So we clearly pick up as Kathleen in their paper. They also refer to it as that there isn't a Kuznetz relationship. Gary Fields has written a textbook looking at global evidence on the Kuznetz relationship. It isn't evident in the African context. We have a very interesting result for us, though, where you certainly see where the green line is the strongest and most positive relationship. And the green line is the fitted values for high inequality African economies. What this suggests is a certain part dependency in the growth dynamics of these highly unequal economies that they're more likely as they grow to continue to replicate the patterns of income inequality. And I'll skip this. So one of the things Chico mentioned again in his introduction is, of course, the relationship, if you like. So we know about the poverty numbers in Sub-Saharan Africa, but what is interesting, and this is the visual of Chico's introduction, which is based on colleagues of his at the World Bank and their macroeconometric model. If you look at the growth poverty elasticity for the rest of the developing world, that's the light blue bar with controls, relative to Sub-Saharan Africa, what you have is a lower growth poverty elasticity in Sub-Saharan Africa. Sub-Saharan Africa realizes with the same growth rate a lower poverty reduction relative to the rest of the developing world. And so that for us is really part of the story about inequality dynamics in Sub-Saharan Africa. So if I switch now to thinking about structural change, thinking about inequality dynamics and how they may be replicating themselves or not, one of the storylines that we develop in the paper is this notion of structural change that you need or you possibly need, if you think of a Roderick and Justin Lin view of the world, large scale wage employment possibly or most likely in manufacturing to generate the kinds of rises in income that would reduce inequality. So we look then in the paper at changes in to monster table, but changes in sectoral employment, sorry, changes in shares of GDP by sector. Again, this is an intro into the kind and the quality of the data you have in Sub-Saharan Africa. You cannot get more detailed sectoral breakdowns than this. And in mining, mining sits an industry out of interest and we've managed to isolate manufacturing. The red rectangles refer to the regional changes in manufacturing share of GDP. And every single estimate there is negative except for East Africa in the post-2000 period, which is almost zero. And what you've seen is essentially what you would have to interpret as deindustrialization in Sub-Saharan Africa. Nowhere do we actually observe and in fact some of the work we're doing on another project at a country level replicates this except maybe for Ethiopia, but if we take Ghana, Nigeria, Uganda, South Africa, what you've seen is either flatlining of manufacturing as a share of GDP or in fact a decline of manufacturing as a share of GDP. And so the story for us is really that if you want to generate large numbers of low wage employment, you need manufacturing to be front and center of that growth dynamic. If you don't do that and you depend, as we show up later, on resource based growth dynamics, you will more than likely, because of the nature of a resource based growth path, you will more than likely replicate patterns of inequality and possibly also increase inequality levels. And so, sorry, I should have shown you if you look at some of the data here, most of the growth has been elsewhere, particularly in services on the employment side. But what you've effectively got is a growth dynamic of the six of the 10 fastest economies that we've been hearing about has been built on the natural resources set pretty much. So you've got high commodity prices that have generated double digit growth rates in some of these African economies that are essentially capital intensive in nature, do not generate large numbers of jobs. And the consequence has been that individuals have moved from, in some economies, from agriculture directly into urban informal employment, variously defined, which simply replicates the patterns of unequal growth. Another way to think about it is that you want to be in the second quadrant. You want to be as part of a growth dynamic, you want to be reducing the share of industry when things of mining, right? And you want to be growing your share of manufacturing as a share of GDP and very few African economies are doing that. In fact, what's happening is a growth path and a growth dynamic that's based on, as we saw in the data in the table, declining shares of manufacturing, right, in this quadrant here, and an increase in the share of mining as a proportion of GDP. And that first bullet gives you the exact estimate that 35 out of 40 African economies in the sample for which we had data have seen a rise in the share of GDP of mining and utilities. And in fact, you could pick an African country, particularly a fast-growing African economy that's not a fragile state, and look at the sectoral data and you will see that. You'll see a rise in the share of mining and a decline in the share of manufacturing as a percentage of GDP. And that for us is symptomatic of this lack of structural change and then to go to complete the story that then replicates a pattern of inequality, a pattern of growth that is highly unequal. So with that in mind, we turn to, as I said to you, this is a laundry list of what could be some of the drivers of inequality in sub-Saharan Africa. It's not exhaustive and it certainly deserves close attention and you could add to the list. The first is sort of an Asimoglu and Robinson-type story which is about the history, the economic history of post-independence sub-Saharan Africa where you had small European populations that retained wealth, highly extractive and arguably poor quality administrations that then focused on law and order rather than economic development. The purpose or the the result of this sort of institutional setup is to replicate the patterns of inequality unlike in other forms of other developing countries that went through a post-independence period. You then also have very strong evidence of ethnic fractionalization. There's some serious econometric work which suggests that ethnic fractionalization will increase either behaviorally or structurally income inequality levels in a society. And that for us needs to be taken into account when one thinks of patterns of inequality in sub-Saharan Africa. Now there's sort of almost an assumption that resource dependence on its own will increase income inequality levels and I'll go through in the next slide the supposed reasons for that. Of course there's an endogeneity problem at the outset and without the kind of data that we need you can't actually resolve that. There is some work that tries to do that but on the basis of cross-country evidence and so on. But all we do here is some some sort of correlation type work. Firstly if one looks at resource dependent versus non-resource dependent African economies what you do find that the KS test comes back showing that the quality of the distributions cannot be rejected. You don't actually have significantly just different distributions genies for resource dependent African economies versus non-resource dependent economies. So the story in and of itself is not empirically true although you'll see there's a little bit of a teaser here the tail end of the distribution. It seems like you can get very high levels of income inequality popping out of resource dependent economies. We then ask the question do you find that resource dependent economies in sub-Saharan Africa are more poorly governed and this is from Revenue Watch Institute that gives you a resource governance index. So you've got Norway very well governed right at the bottom end you've got Myanmar and seven of the 15 economies that are the most poorly governed as resource dependent economies in sub-Saharan Africa. And the channels are clear I described them there but these are license-based industries. If you have a license-based industry it means that the the president or the minister is able to hand out licenses to preferred groups or interest groups. It creates a setup for corruption, for rent seeking behavior and therefore also more unequal outcomes. Extractive industries generally because of the economies of scale are more likely to involve monopolies or large multinationals which again influences distributional outcomes. And then finally what you will also find is that most of your natural resource sectors are capital intensive. They're not going to be generating large numbers of wage employment or large numbers of jobs for every dollar of capital investment that then further induces unequal outcomes. Let me switch to I've got a story about labor markets but let me quickly switch to human capital. As a driver of income inequality we all know the story about human capital accumulation and so on but in the sub-Saharan context there's two things that are going on. One is much lower enrollment levels at the secondary and at the tertiary level. There's also a story about ECD early childhood development so you'll see that there's a drop-off at the beginning. ECD there's almost no early childhood development happening. So one thinks of the health economists and public health officials telling about the first hundred days or the first five years of a person's life. There's very little enrollment in early childhood development but then there's a complete collapse post-primary in enrollment rates throughout sub-Saharan Africa and so you've got an enrollment problem at the post-secondary level and you've got massive problems in terms of quality. This is data from the Brookings Institute which suggests that a third of all African children are operating below the minimum threshold level for either reading or mathematics. We can show other data using the TIMS globally standardized tests which again the blue bars are bad performers and you'll see the majority of performers in South Africa, Botswana and Ghana which are the African countries for the TIMS test are performing below par. So you've got very low quality outcomes in the schooling system. For every 10 hundred children that enroll in the primary schooling system in sub-Saharan Africa only four make it to a tertiary sector, to the tertiary institute. So what you have is an educational system that whose job, if one thinks of R is greater than G, the famous equation for the year, I think, or the decade. Human capital accumulation is supposed to do the job. It's supposed to reduce income inequality. What you've got is a schooling system that's unable to push people from primary into secondary schooling and then secondly when people are in the secondary or tertiary parts of the schooling system the quality levels are incredibly low and that then just reinforces the patterns of inequality. I've got a little bit about gender where again you've seen very little progress and the evidence does show that if you close gender gaps you will close overall inequality gaps in a variety of indicators. So very quickly my conclusions because I think I've got about a minute left. So on average what did we find? Both at the mean and at the median you've got inequality levels in sub-Saharan Africa that are higher than other developing countries but notice the seven outlier story. So if we remove those outlier African economies there's no difference. So you've got inequality that's been driven at least at the aggregate level by these seven economies. We have some evidence for the 94 to 2010 period that overall inequality levels have declined but again the country data shows that there's a lot of heterogeneity there. Our drivers of inequality we sort of settle on the four key factors I'm sure the others. One is this natural resource dependence which we think through the channel of rent seeking poor quality institutions and so on feeds inequality. Secondly and for me crucially is the lack of a dynamic large-scale manufacturing sector and that's the medium through which you create large numbers of low wage jobs which will narrow the income distribution. I didn't have a chance to go through the labor market section but that's essentially what we show. And if you look at the as a consequence of that low contribution from manufacturing you've got a large share of the labor force in sub-Saharan Africa that are either in low wage jobs in the urban informal sector or not even wage jobs but low paying jobs or in agriculture. So you've got these almost low earners at the bottom end of the distribution and then you've got a growth path that's dependent on natural resources and that just serves to stretch the distribution. And then finally I've given you the story about human capital but it's certainly true that secondary and tertiary enrollment rates and improvement in quality are essential for reducing African inequality. Thanks very much.