 Okay, thank you very much. So this work has been done in collaboration with Andrew Dabal and Vascom Molini from the World Bank and Francesco Zasketino from Italy from the Second University of Naples. So the focus of this work is on Nigeria and it basically builds upon recent narrative according to which the poverty reducing benefits from stable and sustained growth pričo je bilo odpočenite na zelo, da je bilo vsega srečenost, zato je bilo vsega innekovalitva. Na zelo, da je vsega innekovalitva vsega innekovalitva, ki je bilo vsega innekovalitva na Zelo, pričo je bilo v 2004 in 2010, kdaj je bilo vsega vsega površnje, kako se je vzelo, da ne zelo, in inkružilo inkružilo, zelo z delovim inkružim. Vzelo po vzelo, da jih zelo, nekaj tudi je vzelo, da je vzelo v zelo vzelo. Zelo, da je vzelo, da je vzelo, da je v Nijegiriji, v raznih svojih, je tudi v zelo vzelo, prejzivno povoljizacije. Vzelo, da bi se vzelo vzelo, in ljup, početih vseč je vseč je vseč vseč je vseč je vseč je vseč, je vseč ta vseč bezpočetne, ki se pristimem vseč. Ako je predsačne, in je vseč na koncentratih vsičnev vseči, nekaj vseč, nekaj vseč. The second view of polarization, which is a more general view, regard polarization as clustering of groups of populations around the local poles of the distribution — let's say, local modes or means of the distribution independently on where they are located along the income or consumption scale, but notwithstanding these differences we sreč da bi vidi that both the views of polarization share the common idea that the more polarized the societies is, the more social conflict is likely to arise. Opravdu, že Nigerija prihodi skupanje, na poške načinovacji, na nozzle v državci, in tudi je, da se poradovače. Pod del vreselosti izgledaj, srednosti na radosti na nesrednosti včasno. Na vse, da je povedal, je je občasne občasno, in izgovoril stvar pa izgavov z svojih svih konflikov. Sveče tako povedljače na vsega izgovoril izgovoril na jeziru. Na vsega problema, ko ješel, je, da se vsega vsega zelo vsega sezera v �gavov z naprejženja, ki je vsega vsega kaj pa razjela. Zdaj, najbolj problem, ko je vznik, kaj je zelo vznik, vznik, ki je dobročen, kaj je dobročen, kaj je dobročen, kaj, da, z različenem počke, zato je zelo vznik, kaj je dobročen, vznik, kaj je dobročen. Nezaj, da smo počeli do vsem izgledali prizavljenja, nezavljenja, nezavljenje, nezavljenje, nezavljenje, nezavljenje, nezavljenje, nezavljenje, nezavljenje. Ko se počeli? There are several household survey on consumption expenditure in most African economies. The minimum welfare variable, which is used for this kind of analysis, is consumption. And we also refer to consumption. There is, for example, the national living standard survey and the follow-up, which is a harmonized national living standard survey, which are the most frequently used by, for example, the National Bureau of Statistics to explore, to monitor progress in poverty reduction over the resilience. But actually the National Bureau of Statistics in Nigeria also conduct another important consumption hazard survey, which is called the General Household Survey, which is conducted in collaboration with the World Bank. What we do here for constructing the data, which should be, we hope, comparable over time, is to take the most updated and reliable data on household consumption expenditure in Nigeria. That is the GHS data, the General Household Survey consumption data, and then use them to reconstruct backward consumption expenditures in time. This approach is called survey-to-survey technique, which is basically an exercise of this kind. We estimate the model of local consumption expenditure using the most updated and reliable data, and we use the estimated coefficients on this data to reconstruct the consumption expenditure backward in time with reference to the National Living Standard Survey, which is the oldest one we have at our disposal. This is because essentially the two household surveys are not comparable between them, because there has been a revision of the data collection methodology in the meanwhile, and also they are not comparable because we did some preliminary checks, some experiments. For example, we calculated some preliminary figures on inequality and poverty levels, and the figures differ substantially, so we were forced, we can see, to use this kind of methodology to reconstruct at least two distributions at different points in time, which are the most distant between them when possible. In this case, we are able to cover about a decade. Why this kind of exercise? Because these are diagnostic plots, but I will prefer to say you about the methodology we use for export polarization. Why this kind of exercise to reconstruct comparable data series on consumption expenditures in time over a relatively long time? Well, sincerely and basically, because if there are no sudden shocks, phenomena like inequality, poverty, polarization, change very slowly in time. So, significant changes can be detected only over a relatively long period of time. So, we need before data series, which cover at least a decade, and we succeeded in reconstructing at least data series distribution, which cover a decade. And secondly, because the methodology we use to explore polarization, and this kind of methodology is quite different from the other methodology, which has been used by previous work on polarization in Nigeria, like, for example, simply computing, let's say, standard measures of polarization, and comparing them between time is a nonparametric methodology, the one we use, and it's called the Relative Distribution. It is based essentially on the comparison of two distributions in time, and hopefully the distribution, which are very distant in time, and lack, for example, in our case, because we have a distribution for the most recent data in 2013, and they reconstruct the simulated one, we can say, for 2003. So, we were able to cover our data at least a decade. And this kind of methodology we use is based, as I said, on comparison to distribution. Let's say, for example, the one at time t-1 in 2003 in our case, and the one at time t in 2013 in our case. Simply the relative distribution, it is simply the ratio between the two densities. And in particular, there are some advantages which are linked to this kind of methodology. In particular, the possibility to explore the sources of changes you can see when proceed to compare distribution. For example, this distribution enables us to perform the composition into location-shaped changes. We can see only composition due to some kind of median or mean shift, or the changes do only redistribution. And so, changes in distributions compared over time only due to shape changes, modifications in the shape. And furthermore, and finally about the methodology, this kind of methodology also give us the possibility of quantifying what we can say at the graphical level, because it brings a set of indices which are called relative polarization indices. And in particular, the first one, the median relative polarization index, gives you the possibility to see if there is movement from the middle of the distribution toward the tails. And moreover, it is decomposable into other two indices, the lower relative polarization index and the upper one, which give us also the direction other than the magnitude of the change. For example, we can say in this case, if we can see, sorry, in this case, if there is more polarization in the lower tail, which is called downgrading of distribution or more polarization in the upper tail. But let's see the results. In this table we have some preliminary summary statistics which concerns the data. The second column, this one, is the one referring to the most recent dataset, the one coming from the general household survey and referred to the years 2012 and 2015, simply because it's made in the second half on the first year and in the first half of the second year. The statistics shown in the first column on the table instead refer to the simulated distribution, the one obtained by means of the survey to survey technique and they refer to the data, the summary statistics shown there, which come from the national longitudinal living survey. This is the oldest data we have at our disposal and the one we have simulated. Apart from the changes, which are shown by the increase in the mean and median expenditure, also the changes in dispersion of the distribution, the most notable features are that there is a deterioration, we can say, between the two points in times of the consumption share of bottom percentiles of the distribution and an increase of the consumption share of the top percentiles of the distribution. We also calculated inequality measures, as, for example, the genie entire indexes and both show an increase of inequality. And obviously we have also calculated polarization measures, standard polarization measures, like, for example, the first Wolfson and the one given by the closest Banerre and both, even so, most likely, compared to inequality phenomenon, show an increase of polarization in this case. But let's see how it works in this case on the ground, the relative distribution methodology. In this figure you have in the upper left corner simply the nonparametric, estimatively kernel densities of the two distributions, which are at the beginning of our time period, in 2003 and the end of our period, 2013. You can see these lines give you the median of the distribution. You can see that there is an increase, a location shift from one period of time to the other, like we saw in the summary statistics in the specific tables. And there is also an alteration, a change in the shape of the distribution that was moved from 2003 to 2013. The subsequent graph, which is called relative distribution, is simply the ratio at any design, if you want, but also if you like, at any percentile of the distribution of the 2003, which is the reference one, to which we compare, the most recent distribution, and you have this kind of pattern. What you see, there is an increase in the share of households in the desires of the reference distribution, the 2003 distribution above the median, of the median of the reference distribution, which is the 2003 one, and a decreasing share of households below the median, but the same. So there is a general change, we can say, in the distribution. But what is the main source of this change? Like we see, but it's also visible here in the table, there is an increase of the median. So we have probably a location shift, and the location shift is shown here. This is the location effect, which operated in this kind of general change, overall change that we can observe. And the location effect is quite significant, but as most of the shape, but this is not the same in terms of redistribution. But the shape of this curve is quite similar to this one, but it's not really the same. There are other changes, which probably are due to the fact that other than increasing location, location shift of the distribution, there are also changes in shape redistribution. Net of this median effect, net of this location shift. And these changes are given in the last plot. Here you have, as you can see, a U-shaped pattern. And this kind of U-shaped pattern is simply polarization, because you have an increase of the share for households in the lower design of the distribution, an increase of the share of households in the upper design of the distribution, and a decrease of the share of households in the middle of the distribution. This is the changes that are only due to changes in the shape of the distribution that is in redistribution. Net of the location effect. Both the effect have operated to produce this kind of overall changes. When we move to quantify our polarization phenomena, because as I told before, there are also the possibility to use polarization indices. We can say that both indices, both, sorry, all the indices are positive, and also they are significantly positive. This first value tells us that there is a kind of process of polarization in the consumption distribution of Nigerian households, and this polarization is both a polarization in the lower tail of the distribution and in the upper one. This kind of polarization that we can say at the national level, it's obviously because Nigeria is exactly heterogeneous, it's a big economy, it's a populous country. It's not entirely homogeneous within the country, but for example, it varies from zone to zone. There is a south of the country, which is more polarized of the north one. I stop here, and I skip also the conclusions, because the problem is a kind of summary of the results. I just do the last remark, which is the fact that this kind of analysis will prepare the ground for further research on the link, which is supposed to exist in the literature between the fact that more polarized society are more conflict prone, and in Nigeria, basically, we have probably both the features of this kind of further research. Thank you very much. Thank you. In a few minutes, and I'd like to then collect a few questions from the audience and then give our presenters a few minutes to respond. So, let's start here in the front of the room. We have microphones that will be coming to you because we're recording.