 10 years ago I stood in this very same room and presented a paper when I was a PhD student. At that time I had joined the PhD internship that Wider hosts every year. So to our fellows PhDs have joined just now, welcome. But also to those who are probably following us live streamed, I would like to encourage you to apply to the PhD internship because it really can boost your career. So the presentation is on the discussion about poverty, inequality and jobs. It is based on these three papers that we just have listened. The first one on eradicating poverty by 2030 and its implications for income inequality discussed by Professor Andrea Cornia. The second one is on how sectorial composition of employment can affect inequality presented both by Professor Yusuf and Professor Andy Sumner. And the third, the role of inequality probably to measure written and co-authored by Professor Al-Qaeda and Professor Puster. My discussion will be based on just one particular point of these three papers, which is essentially the methodology, I would say basically the equation that has been used for the analysis. So I would like to begin with just some highlights of these papers, the aim and the findings because I just have four minutes for this discussion. And then I would like to talk about the multidimensionality of poverty in the development discourse, especially what we understand by a multidimensional dashboard and what is an overlapping poverty measure. The reason for this is that I believe that these three papers that have been presented are certainly very good and all of them talk about poverty or inequality in the monetary space. But we know from the SDGs and this discourse in development that the sustainable development goals are intrinsically interrelated. And these should be measured not only in terms of monetary dimensions, but also non-monitor ones. So that's the reason why I would like to take a detour in my second point. And then I have comments, basically two comments per paper on these formulas that have been used. So the first paper from Professor Al-Qaeda examines whether the 2030 plan eradication of monetary poverty is compatible with expected trends of its immediate determinants. So the determinants in his equation are essentially GDP growth, population growth, income inequality and also food prices. He finds that 20 to 36 percent of the studies of the countries that have been studied will not reach SDG1, which is basically in terms of monetary targets. The second paper by Yousuf Ansamner examines the sectoral composition of employment, thinking about the Kusnet's argument and its effects on income inequality across provinces in Indonesia. So these findings tell us that Kusnet somehow is verified, that inequality increases with the employment share of industry and also with the employment share of some services with high turning points. The third paper by Sven Al-Qaeda and James Foster examines whether inequality can be usually incorporated into Al-Qaeda and Foster 2011 poverty measures. They propose a new axiom, which is a dimensional transfer one, and they find a general impossibility theorem that shows the conflict between the two axioms, but they also propose a way forward for reporting both subgroup decomposition and also inequality. So the multidimensionality in poverty and the development discourse clearly is not new, has been addressed for many years since 2015. Target 1.2 clearly tells us that by 2030 the world would like to reduce at least by half the proportion of men, women and children living in poverty in all its forms according to national definitions. So here I would like to borrow from the report of the Commission on Global Poverty to illustrate this point and what does exactly mean in terms of measurement. The report provides several recommendations, but one of these is recommendation 11, in which it tells us that the bank, the World Bank, should publish alongside the global poverty count a portfolio of complementary indicators, including a multidimensional dashboard of outcome indicators. And of course the recommendation continues. So here I haven't plot the entire recommendation. Recommendation 18 then tells us that the World Bank should establish its own requirements with regard to measurement of non-monetary dimensions and should include a complementary indicators, in particular an overlapping poverty measure. So what does it mean? So here I'm borrowing a table from the report. In this table, which is more or less similar to the examples that Professor Foster has shown, we have a situation where we have five households and then we have poverty measured in terms of four dimensions, each one measured with one particular indicator. And here we have the matrix of achievements and deprivations in each case. So below the threshold we have zero at the privation threshold and the privation stands by one. So the recommendation 11, which we'll talk about the multidimensional dashboard, is essentially looking at each of these dimensions separately, which is clearly interesting and provides a lot of information, but it's only talking about what happens per dimension and not exactly what happens jointly for each household. So in this hypothetical scenario, we have, for instance, in health that there's three households who are deprived in health out of five. There's also three households deprived in education out of five, two in shelter and none in personal security. So multidimensional dashboard will essentially capture this, which is still multidimensional, but it's not looking at the joint situation. This joint analysis, which is captured here by the deprivation score, is essentially telling us that household one is depriving three out of the four dimensions, which is 75% of the privations. Households three, four, and five experience 25% of joint deprivations, whereas household two 50%. If we go forward and the commission goes forward and proposes a counting approach for measurement of multidimensional poverty, we can also set up a cutoff as Professor Foster has shown. So here we would have two people, two households, in this case, living in multidimensional poverty, according to this criteria, and three no. So the sense of deprivation score, which would be the analysis, the basic variable of analysis will provide this information. This side of the table is what we call the overlapping poverty. So how this is related to the three papers that have been presented. So this takes me now to the paper by Professor Cornia. This is the main equation used in his paper. Essentially, he is telling us that the percentage change in monetary poverty is the function of the percentage change in the poverty line, which is still monetary. There's a gross effect composed by the economic part and the population, the demographic one. There's an inequality effect that is also accounting for differences in prices. So my question for this paper is essentially, how can we extend this comparative static poverty accounting model in a situation where, for instance, we think about multidimensional dashboard? What does this mean? For instance, this dependent variable now could be the percentage change in non-monetary headcounts, such as education, health, or personal security, as we just saw in the previous example. But we can also go further and think about multidimensional overlap. So here, the dependent variable could be the percentage change in the multidimensional headcount poverty, according to this overlap measure, or the change in the intensity of multidimensional poverty, using also accounting approach, or even further, think about the deprivation ratios among the poor, which in the counting approach methodology, sometimes they are referred to as sensor deprivation ratios. Clearly, extending this equation is not simple. There are many parameters behind have taken from previous studies would probably require much more rethinking, but also think about what would be the immediate determinants here for explaining these dependent variables, which certainly will differ. Is the genie still a genie measured in the income space? Is it really the same poverty line? I don't think so. Probably here, we would need to consider the deprivation thresholds. And if it's the multidimensional overlap, probably the K cutoff, which is a cutoff of multidimensional poverty. So this is on the paper by Professor Cornia. In the paper by Professor Yusuf and Sumner also uses an interesting equation here. They are looking at the income inequality measure, genie, and many others, where they show robustness to their results. As a function of the share of employment by sector and the square, plus some control variables. So my question here again is how can we extend this analysis of inequality, which is still in the monetary space, to the non-monitor one? Can we, for instance, think about of measures of inequality in education, years of education, or child malnutrition, which is still a cardinal variable, we can still compute a genie? And how can we extend these results and see, for instance, the effect of the shares into these new inequality measures? Of course, the control variables here need to change because here they had education previously, so if we include in education here, we'll cannot be in this side of the equation as well. From the multidimensional overlap, what can we say about this type of decomposition in terms of inequality in the multidimensional space? For instance, the M0 square measure just presented by James Foster. So how can we, in addition, extend the argument of Kusnetz to consider group inequality, which is also something very important in the case of Indonesia, a country that I had myself studied in the past, where we know that group inequality is coming from conflict. So the third paper by Stefanie Kair and James Foster, their analysis and their and the proposition of a new axiom is based on the M0 gamma measures, which is something particularly interesting from this family of measures for those who have been users of this kind of counting approach methodology, is that the M00, which is the head count ratio, or the M01, which is the adjusted head count ratio, given an idea of breadth of depredation, both can be expressed as 01 variables, percentages in this case, or indices in this other case. But when we think about inequality in terms of the M0 square measure, we know that it is a positive number. That's what we know. With larger values, the nothing greater inequality. So my comment is probably a question at the same time, is whether we can provide more interpretability of the M0 square for comparisons over time, where the input variable, which is here the distribution of the sensor deprivation scores, would probably be based on different poverty measurement. In this case, for instance, as you had in a table in the camera, an example, if we have 048 for one particular region and 024 for another particular region, so we know that inequality in one of these regions is double the other one, which is very standard. But when we think about two measures, which are completely different, wouldn't be useful, and this is a question, to normalize the inequality M0 square in measure, into, for instance, something like 01, like the genie, or the tail measures do, which essentially would provide probably a more intuitive interpretation of what inequality is happening over time or with different methodologies. So I'm done. Thank you.