 Hi everyone, thanks a lot for being here. That was a great presentation. You had a very self-explanatory title, and I don't have the title, it's not so self-explanatory. I will explain what I mean with inequality configurations. It's joint work with Eva Begner and Letizia Barbella. We're all in Marble University in Germany. So the paper is really about multidimensional inequality. And so there's quite a lot of work on multidimensional inequality, and the idea of the literature is basically to take this sense capability approach and to acknowledge that well-being depends not only on income, but on other dimensions like health and other stuff. And so if we focus only on income, we are missing part of well-being, and if we want to measure inequality and we don't focus and incorporate health and all these other things, then we are not measuring inequality of well-being properly. So if it turns out that the people that have high income are also the more healthy ones, it's different than the situation where those that have high income have bad health, if you think of overall well-being inequality. So the idea is to then take into account these different dimensions, and so at the beginning the literature was taking into account like education and health, and then other dimensions have been included. But I come to the idea of multidimensional inequality from a different perspective, and this is a bit what I would like to bring here. That is that I come to it like more from looking at literature and inequality in other disciplines, in political science and in social psychology and so on. And there has been an explosion of literature in these other disciplines to try to understand inequality and to see the relation between income inequality and these other sources of inequality, like in terms of political participation, political influence, self-efficacy, and so on in social psychology. So there have been like lots of interest and what they are like two different types of approaches that have been used. Like one is to focus on the micro approach to see if for example there's a correlation because between where you are in the income distribution and where you are in these other distribution in terms of political power, political participation and so on. That would be the micro approach. But there has also been what I call the macro approach where what they are doing is basically compute inequality on their dimension on political participation and linking to inequality in income at the macro country level perspective. So what we want to do in this paper is to focus on this multi-dimension inequality basically from this perspective to try to understand how different types of inequality relate. So our aim is not to try to measure inequality well-being like more accurately. It's to try to understand how inequalities of different dimensions, like given by different disciplines relate, right? And we do it from a macro perspective, right? And so the aim is really not to measure, it's really to understand the structure of inequality. So if you start thinking in this way, like there are like several questions that emerge, right? Like somebody like do these different dimensions of inequality and equality in political participation in self-efficacy, in health and so on. Are they sort of driven by some single underlying dimension or are they rather like, do they emerge from domain-specific forces, right? So you have in some places a lot of political participation in equality for political reasons, right? Things that have to do with institutions and stuff. Or is there just something that's like there are countries that are just unequally in everything, right? Or are there different types of inequality that relate but others don't, right? And so in the end, some go together but others not. Or some types of inequality are traded off, right? Like there are some countries that maybe have a lot of economic inequality, but lower social inequality or political power inequality. So this is what the aim of the paper is to sort of understand and describe these patterns. And this is what we call these inequality configurations. So the inequality configuration is at a macro level, sort of the amounts of inequality in the different domains that a country has, right? So it's a very simple descriptive paper. So there has been like some effort collecting data on these different types of inequality dimensions, but then the analysis is totally trivial. We are going to do some principal component analysis and some cluster analysis and that's all. So there's nothing fancy. Okay, so the first thing that we needed when we were thinking about these questions is, how do we actually think about these questions? It's not straightforward. How do we think about which type of inequality is related and which don't? And we came up with four type of frameworks or hypotheses. The first one, well, the first two ones are very straightforward and I already mentioned them, right? So one could be that all these inequalities are related. So there's actually just one underlying dimension of inequality that manifests in the different dimensions. The second possibility is that they are domain specific as I mentioned before. But there are two frameworks from the literature that I think that are quite interesting and that we used to guide sort of our thinking about the matter. So one comes from Terborn, he's a sociologist and he has this book on inequality where he basically, it's not very empirically, I mean it has some empirically stuff, but basically it's very much something like theorizing in a sort of simple way, right? And what he basically argues is that there are different dimensions of human life that lead to different dimensions of inequality. So you say that, okay, so as humans we are like organisms, we are persons and we are actors. And these different dimensions of human life translate into a different dimension of inequality. So what he calls being humans of organisms would be vital inequality, inequality in health type of issues. Humans as persons is that we seek respect and status and autonomy, right? And this is what he calls existential inequality. And then humans as actors is that we need resources and we want to act and we are agents. And so this then leads to inequality in income, political participation and power and so on. Okay, so in some sense you could think that this could be a framework that generates different types of relations between inequality, right? So that you could expect that there's a relation between inequality in income, political participation and power and so on that is distinct from inequality in terms of respect and autonomy and so on. Now the other framework comes from one of the many Piketty books that I think that now he's publishing a lot. So that was like Capital and Ideology, a very big one. And he doesn't say this specifically. So this is something that we draw from the discussion in the book, right? So this is empirically based like with lots of historical evidence and so on, right? And he talks basically of different types of inequality regimes in different societies, right? So he talks about that have happened in history. Like they are like slave societies, ternary societies, ownership societies and so on. And from how he talks about them, one can derive expectations about the economic inequality and the social, political inequality involved in those regimes. So the slave and colonial societies would be the most unequal in everything, economic and socially. The ownership societies were those that came up after the French Revolution, particularly in Europe, right? And so these were places that had very high economic inequality, but in terms of social and political equality was supposed to be lower. And this contrast with what he calls the ternary societies, which were those that came before, like the extreme version would be like some medieval type of societies where the economic inequality was not so high, but from a social and political perspective, the inequality was huge. So it was clear that there were classes that had more status and more political power and that was supposed to, okay? And so then there's the social democratic societies that were low inequality and everything. So again, different dimensions that go together, right? The social and political in this case and the economic as a separate thing. Okay, so this is basically the frameworks that we have to think about the problem. And so what is the data? What are the variables that we use? So we measure inequality using genicoefficient and we use income, where we use the wheat precisely. Then we're going to focus on health and in particular on length of life. So we're going to compute inequality on length of life if everyone dies at age 80, there's a lot of equality in length of life. If many people down when they are young or when they are like babies and other people like die very old, then there's a lot of inequality in length of life. So demographers have studied that a lot. Political participation. So this is like the amount of political actions that you undertake. Like protests, signing petitions and so on. Political influence. This is, you will see this comes from surveys and these are surveys about whether you perceive that your MP or your local representative pays attention to you, to people like you, okay? Societ class. So this also comes from surveys and this is self-reported, the social class you belong to whether you are upper class, middle, upper, lower middle, working class, maybe in some surveys and lower class. And self-efficacy. That also comes from surveys and this is about whether you think that it is you or the environment that controls what happens, okay? So of all these things, we are going to compute genicoefficients, okay? So we are going to study why we are going to compute the inequality of all these things, okay? So one thing that is important is that we are not going to relate income or education to these things. We are going to see what is the inequality of this stuff. Okay, so the idea is that in the social class, an equal country is that where more or less everyone says that they are middle class. If there's a country that some people say they are upper class and some people, many people say that they are lower class, this is a highly unequal country from the social point of view. Okay, so the data sources, so as I said, the income comes from the weed. Very useful, having all the countries because we were having lots of constraints for the other measures. Well, for the health, the length of life, not so much. They are the WHO life tables from which you can compute genicoefficients. But the other values were tough. So they come from surveys. They come from the World Value Survey and European Value Survey, but also from other surveys, the International Social Science, now it doesn't come, ESSP and ESS, Social Survey, European Social Survey, International Social Survey and the Afrobarometer and Latino Barometer and so on. And then we have a few imputes. But most come from the World Value Survey and the European Value Survey. Okay, and then we have, so our full data has data from 108 countries. I have the six variables I have discussed. And then for robustness, we use only the World Value Survey to make sure that our clusters of countries don't come from different data sources. And the trade-off here is that the World Value Survey has fewer countries, but it's homogeneous, but it doesn't have social class or political influence. So we do the full, the analysis with the full data and then for robustness with the World Value Survey. Okay, and there's one issue that is important about inequality measurement when you deal with these type of variables. So for income, there's no problem, right? So we use the genicoefficient that come from this with companion. But all the other variables are bounded. And this makes a difference about how we interpret inequality. So I have here some made-up distributions to illustrate the idea. So if I show you this distribution, so I don't know what you might think, but I tell you what I think. I think like, wow, this looks like a fairly equal distribution, right? So there are lots of people in the middle. There are a few poor people, but everyone is more or less similar. Now, if I tell you what actually, this variable is bounded between zero and 10. You cannot have more than 10. And so it actually looks like this, right? And when I look at this, to me, it feels, no, actually there's quite a lot of inequality here. There's a lot of distance between the two, right? And there's a reason why I perceive more inequality here. And the reason is because there's no way with this average level of resources that we have here to have a more equal, unequal distribution than this. So this is as unequal given the resources that there are, the way they are distributed are as unequally as they can. Right, this is not the case here, because you could take some of these people, right? Could do a regressive transfer so that you end up, because there are no bounds, you end up with someone being really, really rich, right? But here, you can't, yeah, thanks. And then in contrast, this one looks more equal, right? Because there's scope for increasing inequality here, if the bounds are the same from zero to 10. Okay, so what this means is that, so if we use the standard genicoefficient actually, so these two distributions are the same, so they give a genicoefficient of 0.1, so fairly equal. Whereas this gives a genicoefficient of 0.2. So these are people that have been thinking about that, right? And so what we do is to adjust the geni following a fairly recent paper that deals with this issue of bounded variables. And basically what this adjustment does is divide the genicoefficient, but the maximum possible genicoefficient that you can have given your average. And this is what we do. Okay, so then this one, the way we are going to calculate it has a genicoefficient of one, because there's no way it can be more unequal. And this has 0.7. Okay, so let me go to the results. So first we do this principal component analysis. So this is just precisely to see, to reduce the dimensionality of these components and to see which dimensions go together, right? And if they're in the end like one dimension that generates all the inequalities or they are separate. So basically you can think if there's a cloud with two dimensions and they are like very related, you can actually reduce the dimensionality by just focusing on the linear combination of the two, right? And you don't lose a lot of information. So this is what this does. And so here the first thing that we show is the proportion of the variance explained by the different components, right? What this shows is that the first component explains almost 50% of the variance. So a lot of these components are very correlated. Like the example I was giving like this. But then there's still 50% of the variance to be explained by the other components. So we use a bit of rule of thumb if you want. Like I think that if then all the components, if no component had more impact than the other, then they would all explain around 16%. So 0.16, right? So basically we take the first and the second component but it's very clear that the first one explains a lot. So basically in some sense this is the main result of the paper. So how are inequalities related? These different inequalities. And the first component, the one that explains a lot of the variance indeed then has as loadings lots of these components. Actually all components load positively. But then a little bit arbitrary to a certain extent but we, thanks. We then consider the components that load the most and this is income inequality, the length of life, health inequality, self-efficacy and social class. So this is the first component. This would be some socioeconomic, psychological inequality. But for the political variables load less and the political variables actually load into the second component. So what this means is that there are two or orthogonal components. One that has to do with socioeconomic and psychological issues and the other one that has to do with political issues. Okay and then we can draw all the countries using as access the two components, right? And so the horizontal axis is the first component, the socioeconomic, health and psychological inequality and then the vertical axis is political inequality. And so you can see the sort of Western European countries and others but a lot of Western European countries have low inequality in both. At the other extreme you have African countries and Latin American countries have high inequality of the first component but they differ for example in the second component. So Latin American countries apparently have more inequality, more political inequality as measured by the surveys than African countries. And then here you have Eastern European countries and Asiatic countries in the middle. We still don't understand exactly why, so we don't have a good story yet as to why this is the case, how political inequality would be so much lower in African groups. Okay, is that the zero? Almost. Okay, so it's robust basically. But the last thing I wanted to show is like how this then affects a life satisfaction inequality. And the first component correlates strongly with life satisfaction inequality whereas the second component doesn't. So it feels that inequality in life satisfaction is linked to socioeconomic, health and psychological inequalities but not to political inequality. And then we have one about the mean of life satisfaction. So is it the case that countries with more of this inequality have lower average life satisfaction and yes, a little bit, but when we control for logic DP, the coefficient goes down. It remains negative, but a political inequality doesn't do anything. Okay, so I just want to say again the results. So I think after being in this conference I think for me what I feel that is interesting and potentially useful of this is that so there's a lot of work on this inequality of opportunity and there are like what's some panels on that. And I think that we should start thinking more about these other types of inequality. Things that have to do with status of relational inequality, things of inequality in power, things that maybe are a bit outside of what we do normally but I think are really important. And I think we don't have measures of that. And I think we should really improve and pay more attention to that. And I think that from the point of view of the normative thinking about inequality, what inequality does, I think is really important and this in the political philosophers have now like this issue of relational inequality has become really important but we are not really taking much attention to that. So we got stuck with the lack of egalitarians talking about inequality of opportunity but there are people talking about other things now. And I think we should join this conversation. And I think that it makes sense in some way that like I was relatively happy in thinking that way about our results because indeed like if you think of like social, psychological inequalities, they do relate or it makes sense that they are related and they have to do with well-being and political inequality is different. It's not so much about well-being, right? But it's about power and it matters as well. So I think that we need to think about this. Thank you.