 and that he's here, so let's hear. Okay, all right, let's hope that stays. Okay, so following in the ideas yesterday, I want to do two things. One thing is to give you my view of what is a general setup to think about the problem of inequality. I mean that sort of as a physicist. I mean that I want to create a setup that's general, that could go all the way from physics to biology to society and that speaks to fundamental quantities. But then at the same time, these quantities become instantiated in particular contexts like in a society with under certain circumstances. So I want to show you how I think we can do that. And what that forces you to do is to think of inequality in terms of a component of a process. And that process is the process effectively of change and of growth in the society. Growth is basically we think about often as economic growth as we were discussing yesterday, but I want to also put a lens that's a little wider. You'll see why that came up yesterday of human development. What does it mean for society in some sense to live better? And that will resonate with some of the ideas yesterday. So that will allow us to go from structural inequality, which we measure all the time, to something that's dynamical, that's changing and therefore opens up possibilities for the future. And then I'll bury you in data. I'll show you lots of data that's spatial that speaks to what this session's about, which has to do with how this inequality is expressed and how they connect it to some of these quantities. At the end of the talk, you'll see that we have a lot of work to do to bring these pictures together. But nevertheless, I think that there's a context that's interesting. All right, so I'll skip these things. I study much of what I do today. These days has to do with studying urban context around the world, but you'll see that much of the data that we use is really from cities around the world, big and small, in many different contexts. So it's really exploring the generality of how humans live together that we could actually get to some of these points. This slide I didn't show yesterday, that's sort of my closer collaborators, but there are many more people, and I'll highlight them as we go. You'll see some of them will be reflected in some of the work I'll show you later. But happy group, right? And as I said, I'm interested in the problem of development. So let me jump through this. I'm interested in the fact that under certain conditions, I like to speak sometimes of situations where a lot is going on, like in a city or in your life, a lot is going on, but nothing happens. You don't have change that's systemic. But if you look for the places that do create change, a lot of these places tend to have to do with cities and urbanization. So again, I showed you this Shanghai 25 years, completely different society. It's easy to see in the built environment, but the most important thing is that people are different and live differently, right? They do different things in their jobs. They eat differently. They have different clothes, everything. Okay, so and just sort of in this, just to situate us with a couple of ideas, I wanted to show you also this, because in many ways inequality has sort of a pretty twin, right? Which is diversity. So it's the fact that in complex systems in general, not just in human societies, every agent, every organism, every person is actually different, right? This is not true in physics. In physics, every electron is the same, strictly the same. Every, you know, every particle is the same. So the differences in complex systems in agents are become intrinsic. They have to do with how many resources, what information, what's your history, all these things are intrinsic to the agent. And this is fundamentally different. This is both a source of many problems when you look at it from that side, but also the beginning of everything, right? So Darwin, what was interested in, and I think that this is a nice quote, right? There wouldn't be such thing as beauty if there were no diversity, but also means that there wouldn't be such thing as ugliness all at the same time, right? So, but the point, you know the problem he was trying to solve, right? He was trying to solve the problem of biodiversity. Where does that come from? Is there a bug or a feature? It's a bug, it's a feature, sorry. But you know, and so diversity and inequality have something to do with each other and that's tricky. So we need to have diverse societies, diverse complex systems, but at the same time that creates inevitably sources of difference that can be expressed then in inequality. So there's something we need to specify in order to not throw away the baby of the bathwater when we talk about these things. Okay, so development as a process I think is clear. The word itself implies that there's motion, that there's change. I want to talk really in terms of inequality as a process. How is it generated? Both enlarged and diminished. And in order to do that we need to talk about, I used physicsy words, statistical dynamics, but really what I want to do is to think about these processes of growth and development. I'll be clear about what I mean by that in a society, but also in some sense implicit in biology. And the program here is quite ambitious because we started yesterday with economists that have been interested in the problem of economic growth. And part of the solution to the problem of economic growth, which is sort of a tentative solution, it's more qualitative than it is quantitative today, has to do with the fact that in human societies we use new information to use the same amount of resources in different ways and therefore create value. Often this goes under the words of knowledge or technological change, but that's seen as the engine of what allows societies to change. So we need to get to that and talk about what that means for you and me as individual agents and how it actually gets aggregated in collectives and into societies. And that program is still to be done and I'll show you it's part of what I mean by creating a general picture. This also has time scales. So if you think about yourself as an individual and any biological organism again, you have a life course. So in fact, that's a curious question. I'm not gonna go there, but why did biology, why did evolution program us to die is a very interesting question. But then of course, there's sort of national development, ideas that we often have in mind when we think about and Jim had those ideas in mind, Robinson, about how nations as collectives change. So the different scales. I'm just throwing some words at you so that you see that we have to do several things at the same time. Okay, so I want to run a thought experiment with you. I kind of mentioned to some of you I was talking yesterday. I do this to my students just to disorient them a little bit. So bear with me. So here's a device we just invented. This is in the spirit of Einstein who's in the back, it's a thought experiment. And so here's sort of a device that measures inequality in the society. I put there an injustice gauge. I'll get to why injustice, but it should be clear already. And I want you to tell me, okay, you're the planner or you can imagine a society that is different from what we have today. Where should I put it? What's your choice? What's the level? You know, gene equal efficiency, right? They go between zero and one or zero and a hundred. Just to situate you, the United States is something like 0.47. It's seen as a fairly unequal society. So where should we put it? What's your choice? Okay, so let me just show you a few examples. So sorry, it sounds like a game. But this is New York City, it's 0.51. Same as in Babwe, quite unequal. This is income, sorry. Or maybe Japan, quite equal 0.33. More or less, these numbers you know, they change a little bit every year. So where would you set the gene meter? 20, okay. I don't think we ever had any society, large society with 20. Yeah, yeah. So why 20? Yeah. What would happen, oh, how do people feel? Well, 34 sounds pretty good, it's a nice society that exists, right? You know, my point about showing you this is that it helps a little bit, right? We have some reference points. But it's not enough. We have to know what's going on, right? What? 43. 43, why 43? Okay, okay, so it's kind of hard to know, right? So my point is that this is not the right question, right? We can measure it, right? We've been measuring it this way and other ways. But it's not the right question. We've got to know what that means and whether people are living well or not and who's living well or not. Okay, so the first jumping point from just measuring has been this one. This is from a BBC article, but it reflects a lot of literature, particularly a different standard literature that comes more from moral philosophy and development studies and so on. Which as you say, as you can see here, I hope speaks about basically the idea that it's not just about measuring in genie, but it's trying to have a sense. And people do have a sense. Anthropologists think of fairness and unfairness, justice and injustice as sort of a human universal. We're very attuned to whether things are fair or not. It seems that in every society we do that. But of course it's not easy to know. But the point about this is to go from structural inequality, measuring a cross-sectional difference in a society where we don't know really what's going on and how people are living that we can suspect. But we bring in effectively those assumptions to what we measure to more functional inequality. Think about it's unfair or fair. But then you have to say unfair or fair for what, right? So the question continues. Okay, so what I want to show you now is sort of how I think we can think about this a little bit as physicists. This is from an article I published a little while ago and this is part of a large program. It's also sort of in my book. But it's the idea that if you have an agent in complex systems, they have certain general characteristics. So forgive me, I'm gonna start very general. And the difference again to a physical particle is that an agent in complex systems has to have behavior in the world. So dissipates energy by the basically two first laws of thermodynamics and needs to get energy from the environment in order to survive and get its next lunch and keep going and it has to have behavior. So that means that the energy that it gets from the environment allows it to exert forces, exert behavior, so literally consume that energy in terms of behavior, but it needs to be able to do that and it needs to know what to do with that energy. The moment you have energy or something in your bank account, you have to know what to do with it, right? And so in order to know what to do with it, you need to have information. You have to have a decision about if the environment does this, I'll do that, et cetera. That is information, literally. And so an agent in complex systems, all agents, this goes back to Schrodinger, many other people at the beginning of life, need to do two things. They need to internalize energy or resources and they need to internalize information. And they need to express these two things together in terms of agency, in order at least to get their next meal. But of course for people, we do many other things with it. So these two things basically call attention to two quantities, energy, which we know a lot in physics, all physics is built on energy and it's conservation. And then information, which is somewhat different quantity that we deal less in physics, but it's very interesting because it produces effectively attractive forces between people. It's often good to collaborate with other people. There are no things that we don't know when together we can do and get resources or do something in the world, write a paper together or whatnot. So the other element of this is that this all works on a cycle of finite duration. So usually there's a time that we can take to get all the stuff done. So for humans, more or less, the time is the life course. How long do we live? For other systems, you may be less important, like in the city or nation, it's not very clear that there's an expiry date. But it's made out of course of agents that have that preoccupation. So this allows you to actually write all the stuff in math and so on. And a lot of this program recently, this is what Jordan's been working on, several illustrations of this. So I just want to advertise this work. Where is he? I don't see him. So there it is. But part of the idea is that to get information you need to learn, to get resources you need to apply the information well and you can close that loop and start modeling this in general terms. So what does that look like? So this is kind of the simplest thing that you can do. You write this essentially as an acuasion for, this is the energy conservation acuasion, right? If you're a physicist. I'm calling it resources, yeah. It's an open system, energy can come in. But here I'm just saying that resources, which is energy or money can change in terms I'm gonna use more economic seed language. So there's an income and there are costs. So there's input of energy and output or dissipation. So that's a Y minus C. This actually leads to a spatial equilibrium in cities but in economics becomes a budget constraints. Just basically the monies can serve in the short term or the energies can serve. So this is what physicists know. But then the new part is that this usually is written in finance, in biology as a growth rate times resources. So the growth rate as you can see is the difference effectively between incomes and costs. So if you can accumulate a little bit you can keep growing. But the idea is what is that quantity and where that comes from, right? This is typically not in a physical system. It's a different kind of dynamics. It's multiplicative. It's naturally exponential but that quantity usually fluctuates. So it's a stochastic system. And so in the simplest cases it's a geometric Brownian motion but in fact in reality those parameters have structure and that's where we need to go. So I just want to show you a couple of things that are consequential for inequality from an equation like this. In general again that parameter, the growth rate can depend on resources, can depend on place, can depend on time. But if it's approximately a constant the solution is well known. It's a log normal. Here it is at the bottom. Does this, no. So there it is at the bottom you can integrate it and what you find is that these three properties it's the property of that solution. The first one is that the actual growth rate is the average growth rate of that parameter over time so that is expected minus the variance of those fluctuations divided by two. This is very familiar to biologists because it means that if you have fluctuations across generations that subtracts from the growth rate. This is characteristic of a multiplicative process. You don't see in additive processes. For example, let me give you an example. Why is that? If I give you, let's suppose we have a hundred euro each and I'm gonna impose a fluctuation on your money. First it goes down 20%, right? How much money do you have? 80, very good. Now it goes up 20% so the fluctuations cancel. How much money do you have? Ah, you lost some money. So this is this. This is actually, so this actually gives you a drag that multiplicative processes are always losing resources. So they always need to be actually creating a little bit of innovation to make up for it. A little bit of growth I should say. I'll get to what innovation is. The second thing is that over the long term the growth, the average growth rate over time which is number two divided by T is actually a constant of motion because the fluctuations vanish. So if you integrate, so these trajectories actually over time become more and more predictable once you take time averages. So they're not identical but self-similar over time. So that means that growth becomes predictable if it's non-zero becomes predictable over time and there's a characteristic time in three for that to appear. And that has to do with the magnitude of the fluctuations in these growth rates divided effectively by the growth rate itself. So it's a square. But what this means is that there's a fog of uncertainty even if you have a growth process there's a fog of uncertainty for a while because of fluctuations. And that depends on how strong these fluctuations are. How much your income and your costs are varying if you will. And only after a while this characteristic time scale can you tell, right? So if the growth rate is very small you don't know there's growth. So for example historically we kind of computed this in various contexts but even like in the West in the Roman period the Roman Empire there was a little bit of growth rate for about 200 years but it's very small. It's about 0.17% of our estimate. So they don't know there's growth. Adam Smith probably didn't know there was growth in modern terms even though he saw that some societies were richer than others. This is something, we became conscious of this only recently, right? So this is kind of interesting. So this means that people also that experienced a lot of volatility in their lives a lot of instability in terms of costs and incomes they don't know that their resources can grow. They don't have their experience. They don't feel it, okay? And you only feel it over a large time scale so about 10 or 20 years. You're told the economic growth works and so on but this is not always present. You only feel it over a long term. This is a bunch of consequences about how it's managed but it becomes a long term problem, okay? So these two parameters that appear in these growth equations have different meanings. So this is more or less where I want to go and then move on. There's an average growth rate as I mentioned which will be related to information how you allocate resources. So this is the knowledge you're gonna have on your context. And there's this volatility that's actually killing you if it's too large and this actually requires control, requires that you eliminate fluctuations as much as you can. The simplest thing that you can do is just it's called consumption smoothing so that you basically don't spend all your money when you get money. You kind of save a little bit and then when times are bad you have a little bit in the bank so you smooth more or less your consumption. That allows you using your reservoir and that allows you to do that but if you're very poor you can't do that so you're much more subject to fluctuations. This goes for nations and so on we just went through periods like this. So one of these mechanisms needs control and efficiency and the other one needs learning. So they are in some contradiction. So we were talking yesterday for example societies they experience a lot of instability. They like control, they want control. So for example, we just went through this political cycle where some people want more control than before and with COVID we had that, right? Whereas you have societies sometimes when everything's going well they become very open, democratic societies and so on and they can grow but they also become unstable from the point of view of fluctuations over time and this also has consequences for agents across the population. So I'm just setting this up. So the last part of this is that an agent is composable in complex systems. You can start with people but then you can have groups of people and cities and nations, right? And then they are working together and what working together means is that they are to some extent pulling their resources and pulling their information, right? They now act together to write that paper, to run that firm, to whatever that is, right? So this is pulling information together and information, I wrote this paper a long time ago but this is well known, is a quantity that actually is not additive, right? It's not equal to the sum of the parts of the information you have and I have because it can be synergistic as it's called. So effectively if two pieces of information are conditionally dependent on a signal then you can actually do better than acting individually. So this creates an attractive force effectively for people to work together and you're familiar with it, you do it all the time. This is why we're social and this is kind of the mystery of humans being social. It's basically grounded on this but this poses another problem, right? Which is, is it worth it to work with other people? Even, so first are there potential for good collaborations but then the mother of all problems is once we make money together how is it distributed back to the individuals, right? So this is the problem of inequality as Stieglitz says it, is that basically it's organizations or capital versus labor and so on that don't redistribute it back fairly, right? So that the agents may feel that it's not worth it to be part of the organization. So this becomes sort of, so this usually means that you need a social contract to be able to work together and when this doesn't work you get people kind of dragging their feeds, other people controlling them, you need defection, you may have sabotage, you have migration, people leave and you may have even revolution. So this is basically all kinds of things that happen when there's unfairness but and then you need public agents usually to create some sort of fairness or distribution which is always imperfect but that creates sort of a dynamical balancing act. So this is kind of the setup that I'm working with. Okay, so last two things I know we're going over time but a little bit with what I want to show you next but what is the theory of unfairness? So I want to just show you two things. So this is an interesting idea. It goes back to Riles, this is very influential. It has its jargon as Jaya was telling you yesterday but it poses that a society that would be fair is a society, this is a structural argument in which people would not mind swapping roles at random. So I go to Matteo and we can imagine that if the world is fair we change places and I have your life and you have mine and we're just as happy, right? Or any one of you, so imagine that, right? So most people are poor so it ended up maybe better off. The Bill Gates would end up living in the slum in India and so on but we will be okay with that society if it was a fair society, what society would you build? So this is, I like to call it preference symmetric invariance in language of physics but it means that you don't mind that you basically are pre-muted with somebody else. So this is his idea but so this is very powerful and it's still used a lot in ethics but then Amartya Sen came in as you heard yesterday from Jaya and the idea is that this should be procedural. It's not just about a structure in which people don't mind changing positions but it's the ability to do things, to have agency, to use your resources and information to hope to have a better life. So this is where essentially this idea of justice is now, conceptually, and it's been operationalized in terms of measuring human development so I just want to take you there but this is basically where it's part of a process, a process of becoming who you want to become if you will in your context. So the question about unfairness or fairness, equality and inequality to many people is boiled down to this problem. Okay, so I want to show you then now a lot of data but first I want to just emphasize a couple of points that you need to say something about function, inequality. I'm gonna show you scale dependent when you measure it and apply to many different quantities so you need a logic, I'll use the logic I gave you and then you need a standard. Inequality is always a relative quantity so you need a reference frame, say, am I compared to a flat distribution or what and so on. So just to summarize for all complex systems this needs to be specified in context obviously but inequality scale dependent applies to many quantities but there's an organization of these quantities in terms of resources which can be wealth or energy in terms of information, there can be strong inequalities of information which is consequential about who you can become and what you can do and time. And this is also of course, you may be not have resources or information because you don't have access to other people because you're segregated and so on. Okay, so these are three quantities, right? Not two, not two. And they work together, they're not a dichotomy. The argument here is that in the most general terms you need these three things together. Let me show you why. So this became actually a measure of human development completely independently inspired by Sen's work but recently what we did is to localize measures of human development which are based on quantities like this all the way down to the neighborhood scale this is where the spatial part of the talk starts. So this is from a recent paper that just came out and this was inspired by the work of Mahmoud Olhak in UNDP, United Nations Development Program. He was a mayor of Amartya Sen's, they had gone to college together at Cambridge and apparently they had conversations and I think at some point he saw Sen said, you know that stuff you're writing is super qualitative. I mean what you're talking about even and so he decided to actually operationalize it. And what did this become? There's a history of this index but it became essentially the product so these three things are necessary of educational, life expectancy and real income. So you see how this becomes energy in what you can buy locally, knowledge which is proxied by education, you could argue with that but it's a proxy and life expectancy which is a measure of health but it's also a measure of your planning window and how well you can execute it. And so this became an index that's been measured now for over 20 years, maybe 30 years and ranks countries that became sort of a race to the top of who's better and you see there at the top a couple of years ago and the United States is number 17 now in the world. I think you just dropped another set of places ranking after this one. And it's number 17 because of inequality because health and education in the U.S. are very unequal, very local as you know and that produces better results compared to nations that do better and the nations that do better are mostly European and Asians, social democracies, advanced countries. So these mongrel that's neither capitalistic nor socialistic and so on are the countries that are doing better, okay. So across scales so trying to do justice to the title of what we're trying to do here you have a distribution and what you can do here on the bottom is to go from a national picture which UNDP has been producing for a long time to smaller scales. So basically because these characteristics actually of individuals and households your education, your lifetime, your real income then you can basically measure them locally their capabilities of individuals not nations, right. So you can bring them all the way down to as long as far down as the data permits and when you open these windows what you find is extreme inequality at the local scale, right. So there's a map there of Chicago. Chicago is always the proxy child for inequality, I'm sorry. But what you're seeing here, I don't know how, sorry, I thought this at a point. Okay, so I'll show you in more detail in the next slide but on top you see that logistic is actually do a little bit better this is the usual scaling but actually there's a lot of randomness partly because we made, we took into account costs there's also a lot of inequality there on panel E, okay. And so for Chicago for example just to situate you this is what it looks like if you don't need Chicago if you don't know Chicago you don't need to know but there's something that we describe as the south side and the west side which are poor and the north side which is rich we can see there in the circle where we live which is Hyde Park, the University of Chicago. So there's a college town effect the college town is actually other paradigm for the places that have the highest human development because they have educated people who also make good health decisions they tend to live long and they have decent real incomes. So, but you have places in the city that basically have worse human development than China or Mexico for example and other places that are better than the best country. So there are extreme inequalities in this way and these are very local so across the neighborhood like you see here for Hyde Park you go down the block this is the hospital in the middle and then you have about 15 years decrease in life expectancy. How did that happen? Obviously that should be something that people should understand and they're trying to but they do this very mechanically without this picture of development in mind. And anyway, so you can do this for cities and here it is and there are many of these college towns that do well you can spend time with the day if you'd like there's an interactive map that show you the places that create and destroy human development the worst places in the U.S. tend to be jails or asylums and so on so you have data all the way down to those scales. And what I wanted to show you is just this which is that if you now look at human development versus almost any other quantity particularly in the U.S. at the local scale people have been looking at things that they call neighborhood effects but things that are here on seat have to do with infrastructure no plumbing, children poverty, incarceration crime is also not here but we've done it meanwhile teenage pregnancy which I'm showing you there people in public assistance, et cetera for example measures of health like smoking, disability, poor mental health, obesity teeth loss is actually an excellent measure of people having problems. The last one is Raj Shetty's social mobility index and it's a little strange so I'm not so concerned with it and obesity seems to be a general problem with the U.S. doesn't matter if you have high human development but all of these basically as you see in the just a moment they basically all drop as human development goes up and basically they go to zero around one which is the United Nations standard, yeah. Yeah, yes, exactly, yes, much better. Yes, exactly, that's the point, thank you. Yes, so I call this the Anna Karenina principle of human development, forgive me if you don't like the metaphor but the idea is that when you look at those plots here, sorry, not used to this one, when you look at these plots you see that there's a lot more dispersion at low human development and less dispersion and better values as well at low human development so it's not just that conditions of disadvantage go away but the uncertainty is much lower, people have much more predictably a good life with no problems so that's the spirit in which I'm calling Anna Karenina principle of development, we see this also internationally though the data is not so good is that once people actually have capabilities measured in this way they have these conditions for development they basically have no problem so all happy families are the same but the people that have deficits of some of these quantities they have all kinds of problems and these problems may be different and they may be manifest differently you lose control of your rent then you're still healthy for a while but then you can't keep your home then you live rough and so on and then your health is bad so depending on what happens you have many pathways that create many different problems but there's a lot more uncertainty as well so you see that in the data and you see that on things like financial diaries and so on so this is one aspect of it that has a very strong spatial neighborhood structure we've known this for a while but now we can start to see globally so how well am I doing at the time? not so well eight minutes, that's not so bad so there are three more things I want to show you so bear with me but the second one has to do with the structure in space of quantities and there's data in the US and increasingly around the world that allows us to see distributions all the way down to very small blocks and I want to show you is that some of the objectifications that we do of calling a neighborhood rich or poor, black or white which in the US we do a lot, etc. are simplistic what we actually have is a mixture of people at all scales where the mixture shifts so you need methods that allow you to characterize therefore inequality and distribution across scales and to know what are the effects for example of place on these distributions so this is a little bit like in physics we like to coarse grain you start with something that's very noisy and you create averages and we hope that that's a good characterization of the bulk effects we're all familiar with this in physics but we want now to do the opposite we want to start with a big number on top like GDP growth or an index of poverty or an HDI for a nation and we want to bring it down to each place so this is fine graining and the point is that coarse graining physicists love it but you lose information right, you're kind of averaging stuff out whereas when you go like this you have to put in information so you start with a general theory of a city or whatever but then you need a theory of a neighborhood why does that neighborhood get to be poorer than the average and so on okay so this gives space for people who like theories of particular places like sociology or anthropology often is more local and people that like theories of more average things like economists or physicists but it means that actually they're working at different scales with somewhat different questions okay so the question that we started with here is that this is a map of New York and it's the greater New York and metro area and it's what you're seeing here is average income and you always start looking at this these are block groups so quite small about 1500 people and so this is a complicated pattern right and if you look very carefully there's some reds in the greens and greens in the reds so green is rich and red is poor it's kind of complicated so what is that? you shouldn't average that right people live there so okay so when you look a little bit more in detail so I'm zooming in just to give you examples I actually know the distribution of income for each neighborhood so I'm giving you three examples there so one is this green neighborhood there in the upper east side and you see that most people are very rich so the one in the middle here right most people are sort of in the highest income bracket but there are some people there they're not so rich this is actually I'm going to show you the most extreme neighborhood in the United States it's the richest okay and then you have one that's all mostly poor but there's some people there that middle class and you have a more mixed one which is in Queens okay so this is kind of the logic of it but if you average all this and ask what's the distribution income citywide it's something people see all the time it's approximately log normal so each neighborhood deviates right from that average and deviates because there's you know we can call it spatial selection because you end up with a different concentration of people sometimes by accident sometimes because of structural factors okay so the question you can ask is you know if I know you live do I know how much money you make right that's a question about inequality but it's also a question about predictability it's a question about information right well you can turn around and say well if I know your income can I predict where you live and turns out the amount of predictability is different in different cities in some cities yes you can in some other cities no not really you know there people live everywhere so this tells you how segregated city is in this case by income you could ask the question about race or any other question and what this leads you to is to basically methods informed by information theory there's a long tradition of these methods in econometrics thanks mostly to Henry Thiel and these are basically methods that allow you to apportion inequality or dispersion at different scales because they're they're decomposable across scales so this is what and this just follows from the properties of conditional distributions so I'm gonna go a little bit faster through this but basically you get filters about if you're a rich person where do you live that's the green if you're a poor person where might you live it's the red but then you can basically follows that you can build these information theoretic quantities that tell you how much predictability the neighborhood structure has an income of isa versa and how much do the two are related across the city and that's the mutual information okay so what's the most segregated income group in any city the rich somehow the rich all end up in the same place it's nuts they could live anywhere they could buy an entire block but no they end up in the same freaking place right it's at least in the US we see this in other places but this is cleaner and what's the next most segregated group the poor curious right and then you know the least segregated people in middle class people they more or less can be everywhere you can imagine them living in a rich neighborhood you can imagine them living in a poor neighborhood and so on so part of what's happened in the US this is just for a single year is that we have in the last decade or so since 2007 fewer people in the middle and more people at the extreme so everything's getting more polarized okay so this is basically the richest building and all these people crazy people I mean would you really want to live with the cock brothers and Nushin and so on they all live in the same building and it's stacky look at that anyway all right so there it is 7040 Park Avenue and this basically allows you to say how different each neighborhood is from the overall distribution of the city so how much information do you need to specify about it to explain how the heck everyone ends up living in the same building this is not normal so that's the red but you see that this is also true of neighborhoods that have a lot of poverty concentrated but also a lot of wealth okay and this is basically for the United States it shows you that different cities have like large cities but also cities in the south and new cities like the cities of Texas relatively speaking are cities that have a long a strong alignment of neighborhoods with income versus cities more in the northwest and so on in the purple who have less structures so people more or less live everywhere so this is not the same yeah yes this is American Community Survey data I'm happy to tell you more I'll show you the paper so I just want to finish in my missing few minutes by showing you a bit of a similar picture going internationally so this is now a study of Brazil and South Africa and then I'll finish with our Africa wide study that's just coming up but basically we did the same thing not with income though we did income too but with services access to services this is South Africa it's a study of South Africa and Brazil but has some of the same characteristics these inequalities of who gets service and who doesn't so you see here basically this is going to be water, sanitation decent houses and so on is very local so this is Johannesburg so you see that the purple everyone has services and the blue people don't have services but also if you measure the inequality at different scales from the neighborhood where it's more homogeneous to the city a metro area is very heterogeneous to the nation you basically that see that inequality is scale dependent so you have to be careful when you say genie blah you have to say what scale for what that's it okay this is just warning you okay so what you see I'm kind of going faster is that if you if you measure the access by neighborhood to services and across each city you also measure the standard deviation which is a measure of inequality very simple one you tend to have this diagram that as you have more access there's also more inequality for a while and then when everyone has access obviously inequality has to go down so this is like a Kuznetz curve for services for things that people in the end become universal access services but for a while as you start servicing people this basically happens that you create more inequality so this didn't need to happen you could just basically service each neighborhood the same way that would be the equality line but there's also a maximal inequality line where you basically do it in a way that is concentrated so this is what you always see but this predicts a certain pattern of inequality as something gets distributed in the population so it's kind of interesting this is the end of that story which shows you that in South Africa where I was showing the first plot as services become more and more universal which you see there on the on my left on your right you will see that inequality is coming down this is because everyone getting water right so this is good so this happens a lot there's something coming through at first it goes to place depending how it's distributed it goes to places that are there prepared or richer or have more power but if it's going to be universal service because it's mediated by public investment and so on in the end this is good okay so the last question is if we understand all this sort of scientifically meaning it's predictable it's quantitative we have a model for it we expect how it develops over time can we actually accelerate it and make it better so that you know these inequalities don't stand for so long people expect there will be a period of growing inequality but we're getting there you know all kinds of things so this is a long-term study of slums of neighborhoods that don't have that typically informal don't have infrastructure but but that we've come to understand in terms of what they look like spatially we use a lot of spatial data at high precision as I'll show you and then the idea is that there are ways to deliver those services they're actually better they're more incremental and more respectable people and the place history then has been done in the past with urban scaling with urban planning so so this is kind of a large program and we're in the middle of releasing a bunch of papers that include data for the entire sub-Saharan African continent the first time we have every building for Africa from high precision remote sensing is just a revelation to see how these cities are put together how they're growing how a lot of it is having a pair of urban zones and so on but this relies for if you're a data head like me this is Chicago but it's like what you see here it's Google Earth is a 3D model of a city I live there but it's now this is becoming possible as a mixture of high precision remote sensing AI models and so on that basically are creating complete replicas of the space of cities and this is being done this is an example for South Africa also for informal areas so you can see the tires and so on there right but this is kind of a recreation of each building this is not done at scale yet but it's kind of interesting and the data for each footprint is now created at scale so and this is part of for us it's part of a big survey that we did with non-profits this non-profit called Slamdwellers International which our community is trying to foster their own development by creating data about that deficit so they want to show you this actually started in India in Mumbai but spread and it's very prominent now in many African countries the idea is that by creating local data you can show the authorities that you have lack of certain services and the idea is that you can influence the process by being better than the authorities are creating evidence okay so this is the kind of evidence you get but it's very varied and a lot of it informs you know gender issues for example women tend to be the ones that have to carry water or more insecure when there's no sanitation there are a bunch of other issues like that so when you visit these places what you find is that often when you don't have services you don't have street access so this is a community in Uttvadaam in Accra basically where people made streets so that you have access and along the streets as you kind of see here there are toilets and so on which they don't exist in the houses where people live because they don't have access to the water and sanitation line so this is a place this Fuzini is this guy standing here is the boss and he's a super charismatic guy and he's showing us how they had this problem they had to solve it but basically they created a street network and so what we took this idea to town and basically it's a longer story but here's an example from India here's an example from South Africa again where you have the same thing the city is delivering services but where the streets touch the neighborhood but the neighborhood itself doesn't have services for the houses inside but the community center which is of course created by the community is in the most central place so you see kind of it's also the bar but you see sort of the contradiction here it's not like a city that we're used to living in where each one of our houses you walk to the street and you have services and you have an address and so on so this is spatially that's the nature of the problem and it's a problem of topology I want to spend time on this but it goes back to oil or it's a beautiful story and so on okay so you can kind of come to characterize essentially the spaces of cities and identify parcels within each block surrounded by streets that are kind of internal and have levels of being internal and this allows you to identify from spatial data this kind of problem of access to street so we're using this a lot and this gives you a measure of spatial inequality and then how you could extend the street now working blue in order to provide access so this is excruciatingly spatial right but then you can zoom out and start creating what are the places at the neighborhood scale each block that have services or not how is this distributed you know, fractions of the population is very practical but it also speaks to what we're doing here so this is a city of free town Sierra Leone and you see that most people actually have a little bit of deficit and then there are a few neighborhoods here in yellow which tend to be more rural or in town, hills and so on they have kind of extreme situations but the extreme slums they're very dense and so on actually rare when you do this at scale they're the ones that capture the imagination but a lot of the deficits and inequalities are mild they're not extreme at least for most of the population this is a very famous place that's getting a lot of attention now it's called Makoko it's as they call it the Venice you can see that the boats this is built over the water on stilts it's in Lagos supposed to be the largest city in the world and so you can do this for every city and at scale and so on so you see all this inequality how it's being mitigated through delivery of services where it's bigger and smaller and so on so that's it so let me just finish with this slide which is basically that when you look at the problem of inequality of distribution of difference of diversity scale is important quantity is important but all this tends to be grounded on the ability of people to exert their own agency to be resilient and so on and so a lot of this depends on individuals and households and communities being able to do that it's grounded on people so this idea of processional inequality of human capabilities is very important but that all is mediated by the city and the larger governments at larger scale so this is where basically these scales interact and the mechanism by which things are created and mitigated matter and they have social contracts at various scales so all this matters for this problem and its solution if you have interest in seeing more of this there's a book but I want to just say one last thing which is that a lot of what we can do today with this kind of data is structural we can see these inequalities and we have now this idea that we need to create improvements over time in terms of agency but there another kind of diversity which has to do with the potential for growth cities and ecosystems and other systems need to create ambient diversity as sort of a capital for future growth for future solutions and so on and so if we don't do that it's not only bad for people involved which is often the way we measure injustice and injustice but it's bad for the system because eventually it will not grow produce value, produce its future so there's also the agency of the ensemble versus the agency of the individual and these things are connected and therefore need to be connected in terms of costs and benefits so that problem is back to that collective problem the mother of all problems in terms of inequalities accessing the long-term improvements and reflecting that on individuals that becomes key so thank you Okay, so things changed so the idea now is that we're gonna have like some minutes for discussion now and then have the coffee break and then we have the presentation okay, so any questions to Luish? Any questions? Okay, that's my chart Yeah, I'm not sure we need the microphone maybe we can, well it doesn't matter Well thanks, Luish, for the talk and so I was wondering given that you have looked at so inequalities in different aspects so I lived not in Chicago but very close to Chicago for a few years and I went back a few years I mean, well last year I was there and to my surprise I saw there was like a biking lane which was unthought of when I was living there because it's always full of cars so I was wondering have you looked at inequalities in this sense and adaptations to more sustainable methods of transport in cities? I mean now every major city has like well, I don't know if every major city but many major cities have biking lanes, like different types of and they have reduced some of them like car lanes in just to facilitate mobility of people and I was wondering have you looked at this and how this, I'm sure there are inequalities about this too but I haven't seen anything about it so I'm just wondering There's work that's coming up but this infrastructure as you were describing it's happening many cities so gradually so but it has a lot of the same signatures that you see in the provision of services it happens locally first so there's injustice about or inequality about who gets that bike lane and who doesn't, right? And then its advantage becomes with scale when you can go everywhere from anywhere in a city on a bike which most American cities at least I know if you go to the Netherlands or Copenhagen you can do that but those are systems I've been studying those systems actually that have been there for 20, 30 years that became large systems with all kinds of other things so now part of the problem actually we're studying is that the initial development of these lanes actually are dangerous for the cyclists to some extent so even though it feels like a good thing as you're saying because they're incomplete and because the drivers are not educated and they're often not done to the highest standards it gives the cyclists a sense of security and agency and then the car's still there in some other way so the initial so I wanted to characterize this kind of thing thank you for asking in terms of again some sort of Kuznetzker where at first the provision actually makes things worse possibly but eventually so that should not be a reason to start because the advantages come at scale so I think that that's part of the problem the same, you know, similar thing with a very different example like electric cars electric cars mostly in the US but I think it's true in other places too are expensive and the tax breaks that you get typically go to richer households so, you know they're trickling in as it goes so even though it's a good thing for society there might primarily something that's creating inequality for now hi, thank you so much for your presentation since you are talking neighborhoods and inequality and you also brought up Chicago I was just wondering I'm not a physicist by the way but I was just wondering if your models or you interact with or take into account some of the historical, social, spatial effects of policies like redlining, for example which restricted people of color from accessing home loans and have been shown to kind of have multi-generational effects or in South Africa, for example, the extreme segregation which was imposed by the apartheid state I was just wondering if you've been able to engage with some of these maybe historical policy effects the answer in general is yes but we've had actually a couple super talented postdocs in our group and then I institute that do exactly that I think what we see most of the studies that I see now are the persistence of these historical policies and circumstances into the present and how they may be mitigated or not most of the time there is depending on place and so on but a relatively strong presence of those patterns of deliberation segregation of lower quality housing or whatever in places that received redlining and so on there are exceptions there are places that have changed since then both in South Africa and the US so depending a little bit on the local history and what UN faces on but there are these historical injustices for sure one interesting aspect of neighborhood inequality related to this is that when you think about the time scale for the character of a neighborhood to change it's social economic status it's racial, ethnical composition and so on it's at least decades and when you think about that in the United States more or less every household moves every five years so a neighborhood actually has a lot of churn there's a lot of local movements and so on but somehow it's character persists so they become a place that because of their characteristics tend to attract for a while unless there's something that happens the same type of population sometimes channeling that continued disadvantage so I think people of these sort of ecological effects as sometimes sociologists certainly Chicago talk about it are just starting to be studied but I think we've had challenges studying them longitudinally over time over these timescales but this is coming but definitely there are these persistent injustices I think part of the question has been are they changing or not and what's the right mechanism to address them in terms of compensation but also creating better local conditions I mean in some sense as you can see what needs to be done is to create a decent living standard everywhere both at the individual level and at a place level and that is just not happened even though there are mandates to do that Thanks for the talk so I have just a technical question so going back to this multiplicative growth model so so in this model essentially you have these two parameters so the average growth rate the drift and diffusion yeah but what happens is when you when you introduce redistribution or interaction between individuals then essentially the growth rate is independent of the diffusion of sigma and actually you have a strange effect that the larger is sigma the more equal is that the inequality is inversely proportional to sigma so I don't know how this fits with this so your first point is correct and that was also part of the point I was trying to make is that in general the average and the variance of volatility the independent parameters what needs to be done and has received less attention and sort of what we're doing is what is the theory of these parameters where do they come from so indoigenize them if you want and so it's not just about so most models of redistribution they redistribute resources but in the logic that I try to propose you also may think about redistributing information effectively so that the growth rates equalize or are more or less evenly distributed so our first paper is about that but you have to address these two quantities and their origins independently in general and then to what extent they are also interdependent and that's a lot of the work that in my view remains to be done it's sort of you know it's the theory of the parameters not of the short-term dynamics yeah hi thank you for your presentation thank you for your presentation and I'm not going to explain the work done because I am really many right right so the as you say the um so two things on the one hand there's the human part where you can imagine the idea the idea essentially of metropolitan areas that people in general could I'm addressing your point but it could in principle think about living anywhere it's a shared real estate market and labor market but in practice as you're pointing out and I was pointing out by you before people are constrained in many different ways right then arguably the biggest constrained uh... in modern cities in the US is not so simple but there with me is real estate right that real estate is incredibly varied in terms of price and obviously poor households cannot live in a very expensive real estate in that building actually is a co-op which basically is a club you need to be approved by the board I mean there are many buildings like this in large cities in the US and this is obviously been also a racial thing but also a class thing and so on you need to show you have uh... I think a hundred million in assets at least to be in that building it's a private thing it's like you know still redlining of a sort uh... but there are many many uh... obstacles of this sort uh... the obstacles typically are contravene to some extent but in completely by by public action by services you know sometimes rent controls things that kind of come from the political process more but they're always imperfect right there's always a rats race uh... you know uh... and arms race to to try to keep the city balance and other mechanisms particularly to do with economics uh... and sometimes segregation racism to actually keep the city separate that's the problem I think we're more conscious of them we have better data we can see actually you know in distribution not so much the individual blame distribution by characteristics how neighborhoods are changing but it's hard to intervene in in most cases at that level and know what the right thing to do is but i think that's an excellent question if you had all this data and you see each neighborhood move and so on and the quality of life and the people that are there what would you do uh... and i think to a large extent it goes back to the ideas by rails and others right that you want to make sure the amenities conditions of real estate and so on even for people who are poorer or more excluded are are decent and enabling of their capabilities if that's true then we're ready in a better place and then you can sell expensive things to rich people but if that's not true if that actually is a ghetto or disadvantage or segregated neighborhood then that's a major problem so again i think you have to ask that question along with uh... structural questions but but it's there's work to be done i think we understand better functionally but that in the public realm of policy is very hard still so interestingly enough we are on time so i don't know how i wish for the presentation and then we go for the coffee break and then we are back and then we have more discussions and more questions okay thank you