 I thought that was a really very rich session which reflects the richness of the book but also has some new stuff, so that was great. So let's open it up for questions. Ma'am, let's collect three, okay? So I'll do one, two, three this time. So it's about the back. So I'm slightly surprised that there was no mention of government failure in all of the discussion because in many developing countries, who are the focus of this work? The government is probably the biggest investor in education and all of that. So if any government policy fails, that will certainly influence the mobility in education. For example, in Ghana, we used to have three years of education for senior high school and then the next president came and said, I don't like that, I want to make it four years. So that literally truncates the educational path of a generation in that period. There could be free education in other places. The next government come into office and say, I don't want to sponsor free education. So that will feed into that. Anything in developed countries like the US, for example, income mobility can actually be affected by the government as well because if you are a single household person to qualify for social welfare, you have to end less than 15 end results. So if you are a risk adverse person, there's that tendency to try and work at the minimum so you can keep getting that benefit. So there's no incentive to work and get 2000 or 2005. So there's that cap on how much you can make and that limits your mobility in income. So why are we not looking at how government can influence hardware or downward mobility in income and education? Was that aimed at a particular speaker? I think Kuma Singh can start and then. Let's just collect two more. I just wanted to be thinking about the other lady or something. Just a set of presentations. My first question goes to Okuno. Yes. So in your description of how you measure mobility, you said you created a dummy. So I'm just wondering what happens when there's actually zero. So if it's positive that there's upward mobility and if it's negative that is downward mobility from my understanding. So how do you deal with the zeroes? I'm just wondering if you consider the situation where you're looking at maybe three mutually exclusive categories where you have upward and downward and there's no situation where there's no mobility. I'm just wondering your thoughts on that. And my last question goes to Patricia. So I don't know whether I missed it or maybe it is because of lack of time. So from your presentation, was there an empirical analysis? And if you did do this, what were the results showing? Because from your discussion, you were showing three main things that it wasn't very clear to me whether there was some empirical analysis that was done and what the results should be. Thank you very much. Thank you. I have one specific question for you. I knew that there were problems. That's a general question I made them at that time. A specific question on the co-residency bias. Can you say something more on the magnitude and the direction of this on the data? Thank you. So let's start. Who wants to start? Well, let me start. I mean, obviously, I totally agree. Government failures or lack of investment, public investment, education, so on is very important. The chapter that Jerry Bearman did on human capital showed, so Jerry did a great review of all the literature on human capital and mobility. And I should just say that he then developed that chapter to a book, which is a Cambridge University Press book that we will publish very soon in our new series with CUP, which is going to be the CUP Elements in the Internet of Economics. The book is pretty much just going to happen. I'm going to come out very soon. Open access, as always. So in the book, he develops it much more. The argument is that he makes it the early childhood education is the most critical mobility. It really happens, and I think Shiko kind of traditionally talked about it in your first slides. The first five years of what happens in a child's life is really important. I mean, obviously, what happens at school, what happens at the university, all that, but it's just the first five years. That's what the evidence is showing. And there's quite a good literature now, especially perhaps Latin America, more than Africa and Asia, on how early childhood investment has this huge effect on mobility of the child. And I think that's where what has to, the government has got gas resources on human capital investment. Maybe that's where it should spend looking at the evidence on this. So I'll just ask you to read Jerry Redman's chapter on this, on this, and then the book is going to be out coming out very soon, because that's really, I think, very important for us, because we have, in economics, neglected that a little bit. We have neglected that a little bit because we always felt that primary education, secondary education are quite important. We didn't ever think that the first five years is where cognitive ability gets developed. And cognitive ability is really fundamental in mobility. But if you get impaired, and we know that differs a lot, so cognitive ability differs a lot, as I think you showed, should go through this morning, by income status. So you are a poor kid. You can't go to an early childhood education program because that's only for rich children. Right away, you're disadvantaged, right? And that's where the government can step in. That's where I think there's a huge role for the government. I think we should think more about those sorts of programs than often what we think about, which is schooling and so on. The other thing I should say, the one problem I have with the education mobility literature is only talk about quantity, not quality. Quality education differs remarkably across the developing world. But there is yet to see, I have yet to see a paper that looks at quality of education and looks at mobility because we don't have, again, measures of, say, the measures we use tend to use on the piece of measures on ability over generations. And I think that's something we need to think about it. Why are we not yet there on looking at education mobility where we're in quality and not just focusing on quantity? Because quantity, in my view, is a very imperfect measure of education these days. Thanks. Who's good? It was, okay. Your question was what was the extent of bias? I can tell you that certain type of co-residency where you see the father and mother and only a subset of children, the bias for the relative persistence measure was simple, the slope estimate was something to the point of, like, 30%. So that was pretty big. And relative to that, the standardized measure had a lot less bias, something below 10%. So that's actually our Journal of Human Resources paper, if you wanna look at it. For the intercept, I don't recall the bias numbers, but biases, I mean, the attenuation bias of 30% is big enough that you need to worry about it. And I also wanted to come on the government action. I just want to add a little bit. I think while we talk about social norms, institutions, and all this, government action is already, this is everywhere in the mind of everything, okay? So it's not just government policy related to education, it's government policy. For example, you can think of trade liberalization, having roads constructed, things like that. All of these things, actions have implications for mobility or immobility. And a lot of the time, what the intention of the government is, as we say, that you want to have all children do better, in on average, a lot of the actions do, but a lot of the actions, the unintended consequence is that it increases persistence too, sometimes. So that's where the frontier literature is right now, okay? With that. Yeah, so let me, so you had a question about the mobility measure, right? So that's a question we also have been thinking and wondering, so what I didn't have time to talk about is other walks that we have been doing is using other measures as a robustness, similar to Chettis, like P25 measures, Samash, Paul Novosad's, bottom 50% measures. So we are actually on the works, but we didn't want to show it now in the essence of time. So yeah, that's a very valid point. Thank you. Kunal, do you want to add anything? I mean, I think we, again, this question of how to measure mobility. So the zero one is essentially upward is one, zero is no movement or downward. So that's the way we defined it. And this is the Allison, so the paper, Allison econometric has a similar thing, similar method. You could argue that we're not allowing therefore the continuum of mobility downward, upward and not changing. And I think the continuous measure is probably better in that sense. It is the upward, the decodermus measure is easier to think about because it's easier in a regression framework, but I agree, we should, the continuous measure is probably more informative. Exactly as you said, yeah. So yeah, so I think you asked about whether, what I was talking about was part of a specific, I guess, empirical exercise. So my chapter was more of a, set the stage type of chapter for the book where my job, as I interpreted it was, was to ask ourselves the question, why is mobility lower in developing countries? Then the way I decided to answer the question was, okay, let's look at the general development economics literatures and all the questions that we are answering in that literature and see how it can help us in the mobility literature to see where those potential mechanisms are. So the things that were mentioning, we're not part of a new empirical analysis, we're taking from results, empirical and published results from the literature to say, okay, we see that if we give money to grandma, then the prime age individual in the household, they migrate, okay. That sounds like something that might affect mobility. We see that if we reduce the information friction, lower income students are more likely to go to college. Okay, that speaks about mobility. So that was what I did in the chapter. So all the findings that, if I use that word, I apologize, but all the results that I was describing were results from the literature where they were not part, this chapter did not produce any new empirical results. Great, I'll just add one tiny point to Konau's point about quality of education, which is that in the inequality of opportunity literature, because we obviously don't have to have the same outcome and we don't have to have quality of the education for the parents. There are a couple of papers looking at the effect of inherited circumstances on test scores, but that's only in the IOP literature and you're absolutely right, not in the IGM literature. Okay, a second round of questions, I think there was a gentleman there, and if you're behind, if you can't see me from the screen here, please jump, but we have Murray. So let's start with you, sir. So for example, you can have, so it has to do partly with the networks thing that you were saying, so it has to do with getting a job in the public sector, right, but it can also have to do with allocation of public sector infrastructure, right? So communities that are more privileged, like might be able to engage better in some clientelist like relations with politicians so that they end up having like more. And so I feel that this could potentially be a big driver for higher mobility. So what do you think about that and... This is for Kunal, Terence Gomez from UNO-Stimulia. I'd like to, you didn't have time to build on the point, but you raised it that from the book, it became clear that an interdisciplinary approach is much better to get insights into the problems that we are trying to address. I'd like to push you on that point. What exactly is it about interdisciplinary research that you found or the book, the people who contributed to the book found in terms of how this can help solve problems? I just want to make one more point. In 2008, after we had the global financial crisis, there was a big debate about economics. And the University of Manchester, there were students protested against the way in which economics was being taught. And the debate then arose also about the same point that you're making, that there needs to be an interdisciplinary approach. A political economy approach was just mentioned. So I want to put this point to you to see how when you were doing this book, how did these ideas come true and what is this new approach, this new interdisciplinary method that you're talking about? Murray, maybe you can speak really loud, so I don't have to... To pick up on the point about education and quality of education, I'm just wanting to push the panel a bit, I guess, on, okay, but what are we to make of this? Chico just referred back to the fact that this does link to qualities of opportunity. Do we want to adjust the way we measure education? Or are we going to continue to measure education the way we are and then work out in specific contexts like South Africa, for example, where there's been huge progress in years of schooling, but it hasn't translated into equality of what education gives people in terms of mobility, and especially, you know, there's a gender dimension to that where they bump into a labor market that's skewed, but there's just a general dimension of that where years of schooling mean different things now. And so that's just an issue. Do we want to adjust what we think of as education as like an equality of opportunity or the way it drives mobility? Or are we happy to use the measure that we have and then treat and look into the context as to why it's working differently? The thing that worries me about the letter is that we produce these correlation coefficients, right, of educational correlations, but what we're saying implicitly is that in different contexts, they're very different because of the quality relation. Thanks. Thank you. Maybe let's start with Farhad and work this way, or... Okay. Well, let's start with Kunal. Nobody wants the mic. So I'll answer Miguel's question. I'll leave the other questions to you. Political economy, clientelism. So I didn't talk about that in the chapter, but that's a very good suggestion. I'll make just one more... So the short answer is yes, it's something that would impact the estimate we do, which is the intergenerational correlation. Just in terms of what you were saying, it would add one more channel of transmission, well-connected parents, partially in South Africa could be ANC or whatever connections you have. So I guess the short answer would be yes, that's a channel to study. But then it also made me think about another point, which is I don't know if we want mobility to come from public sector jobs. So I guess that's an empirical question, I suppose, or maybe ideological question. Is that what we imagine mobility to become is that everybody from a lower income background gets a job in the public sector. So it could be one of the ways we create mobility if the public sector allows better screening of good workers from certain set of mental health, affirmative action, whatever it is. But I guess I would have to think harder on whether promising public jobs is the way to increase mobility, which is not what you're saying, by the way. It's just that it made me think about that. So yes, on the first question and then thanks for actually making me think of a second question. So let me answer Terence's question and I'm going to use my response to Terence's question to answer Marie also, actually. So the point, so what did I learn or what did we learn as economists from the papers that we commissioned from the non-economists? So first is, economists are absolutely obsessed with income mobility, okay, and then educational mobility. But you see hardly any papers on occupational mobility. The answer is, why should we worry about occupation? Ultimately, occupations are getting higher incomes, but we should worry about occupation. Why? First, because we don't have the data on income that we need. There's no way we're going to get the data on income of the kind that we have for the US, the Nordic countries, and so on, tax data along with the fact that tax data does not include informal sector incomes. So we're not going to get that. Let's be honest. Just accept that we will not get the income data to look at mobility, the way that scholars have been looking at mobility in Nordic countries and in the US and so on. So income is not going to get us very far, okay? Secondly, occupation itself is important. If you think about agricultural worker in India, what does the person want? The person wants to get out of being an agricultural laborer, wants to be a teacher, wants to be a government clerk. Occupation makes a big difference in a person's life. That person's not income is important, but the person is much more worried about getting out of a bad occupation. And this is why sociologists have been very important to us because they've been saying, you guys should think about occupation, which is also a proxy for social class, right? So we can learn as economists a lot from those sociologists, especially because sociologists are using a lot of different methods as we do in a different way. They don't look at causal work. They look at a lot of descriptors, mobility matrices, but we can learn a lot from that kind of method. The second thing we can learn is that I don't think I'm very much for causal evidence in this literature. We don't have enough causal papers, as Patricia said. So we need more papers which are causal for what shifts mobility. But we know that ultimately, as I was saying, mobility is not even one factor that shifts somebody's lives. Things are happening at the same time in their person's life, the child's life. And this is where non-economists can help us using qualitative methods to tell us about what other things matter while we are still looking for our natural experiments and our policy experiments. So I do think bringing methods together can help economists to do better work in mobility. So I'm looking at an instrumental kind of approach that we can do better work in mobility as economists by actually bringing in other disciplinary insights. So that's the second thing I would say, that first, get out of income, education is important, but education, my response, education is a mechanism, not an end in itself. Nobody wants to be super-educated to have a poor quality job. And the point is that what we see, as Arasthup showed, you can get educated, but you don't often get good jobs. Why? Discrimination, inequalities in the provision of infrastructure, education, so on, all make a difference. Or other things, networks, as Patrish said, make a difference. So education mobility is necessary, but not sufficient for getting us either of income mobility or more importantly, occupational mobility. Exactly as I think Mario is saying that if you spend too much time worrying about education mobility, we might miss the fact that that's just a mechanism. That's something we want to see to get us what we really want, which is occupational income mobility. And that's why we need to be careful about education mobility. Sorry, she's looking at you, but I think we just sort of say that it's interesting, but it's not really enough in the literature mobility, right? So let's stop there. And maybe I will disagree a little bit with Kunal too. Yes, exactly. So the answer you're going to get from me is going to be an economist answer. I completely agree that educational quality is the one variable that we want to capture. And if you really think about it, the quality of education should be reflected in workers' productivity or in other words, in income. And this is precisely why the developing developed countries literature focuses almost exclusively on income to measure economic mobility. We are focusing on education simply because we don't have income data. It's the same problem. We don't have education quality data. If we had quality data, I am sure that we would be able to come up with some kind of index that would tell you what the productivity effect of this education. Now I am going to disagree a little bit on the occupation mobility because this is something that actually I started my life with. I mean, very fast paper that we did was looking at occupational mobility. The problem with the occupational mobility is that the mobility measure is going to depend on at what detail you defined your occupation. Whether you know the overall farm versus non-farm, or whether you go in much more detail ICIC codes for this. That would be number one. Number two is that just like education quality, content of a job or occupation changes over time. Right? And that is not something that just looking at the occupational category you will be able to tell. And that's another reason I think why we should try and get data on income in developing countries. Okay? And I do think that for a whole lot of developing countries there is enough data now that in maybe five, ten years we will be able to talk about income mobility too. So that's my response. No, no, I don't want to add anything in this. Yeah. That was a very interesting exchange and I think there are things to both of that. It's very good. I like that. Any other questions? One here. Marcus. Miss anybody? Let's start with you, sir. What role does what role the equality of outcomes and place in the developing countries do we have sufficient data to look at the great catch-pick only for the developing countries for example? And then maybe related to this is that the, I mean one difference that it's large between developed and developing economies is the extent of redistribution which is way larger in the rich world. So I mean this probably also has quite a bit influence on the, on at least some of the measures of inequality of opportunities. So I'm Marcus Yantian, Stockholm University. I'll direct this. I have a comment to Patricio which you may want to reflect on and then some more generic question. The comment to Patricio is that we know a lot about persistence even though a lot about the causal effects on persistence of different things. We know a lot less about mobility in the sense of the expected deviation of people's outcomes from their expected value given their parental income. And we know even less about the causal impact how different programs would impact on that, what you might think of residual variation. So persistence and mobility are interchangeable if errors are homoscedastic but they're not. So do you have any kind of thoughts on that if we should learn a lot? This is independent actually of whether or not you're looking at poorer or richer countries. It's true of the both. But the second question is somewhat more generic and sounds silly. But I actually think it's not which is, could you elaborate a little bit more on why we should focus on mobility. So for instance, what are the welfare economic reasons for looking at intergenerational mobility? I claim that they're actually much less obvious than the most people think and in fact this kind of tips things in favor of inequality of opportunity rather than looking at persistence. But I'll leave it at that with that easy little question. Thanks Rahul. Okay this time let me start on this side. Okay, I don't think I remember all the questions okay. The one that Marcus asked last for why we should care about mobility. That is the question I can probably give some answer. And to give that answer I have to go back to one of the example that Chico already always said. This I had used many times so you can tax me on this which is that inequality is something like cholesterol. There is the LDL and there is HDL and one of them is good, one of them is not good. And inequality is similar. Some inequality is good because it motivates people to work hard for the reward of it. So a lot of the innovations happens because of that motivation and that is the type of inequality that we should be tolerating certainly and even encouraging to some extent. So that is good for welfare no doubt about it. Some of the inequality is of the LDL quality which is not good and there is actually some empirical evidence that this is not just good from the point of view of what society thinks what is good. It is also not good for human capital accumulation for long-term economic growth and so on. So I give you an example. If you think that distribution of talent is more or less random in that case if a poor family cannot send a kid to school that certainly has the negative impact adverse impact on human capital accumulation and through it it also has negative impact for both economic growth and welfare. And that is the reason I think we should be looking at mobility and that is precisely why I think that intergenerational persistence is something that gives us some sense of you know what that friction is and that is going to be my answer to Marcus. So time flies when you are having fun so let me ask you each to answer those questions and make your final remarks in about one to two minutes. I'll just give a very quick response to Yuka. So Yuka one of the key stylized facts we know again we have this in the book and we use Victor's great dataset on the opportunity is that the great Gatsby curve. The more unequal society is in outcomes now that is a correlation is not causal and you can easily imagine the causal relationship can go in both directions more unequal societies lower income households have less to invest in the two kids education clearly it will affect mobility. On the other hand one can also assume the higher mobility can lead to more poor income households in a relative sense increasing incomes over time catching up with the richer households and that will lead to greater equality. So the causal relationship works on both directions I don't necessarily think that it works in one direction and I think it would be really good if we did more work on this because I think it would be really nice to think of how we can causally estimate which way the relationship is stronger if we do believe it goes both directions right. I just want to say very quickly this book that we did was actually starting point for more research and social mobility I hope that you know the book can help us to do some more answers methods and causal work and also mixed methods because definitely and I think there's a sense also what gets from the questions there's a lot of work to be done really for developing countries. We are really at the beginning of where we should be and hopefully you know many of you here are thinking of more work on this because there's really a need for much better quality work on causes, drivers, methods, measurement all of that so hopefully that's the something that we want to say this book is not an end point it's a starting of a conversation on mobility that's actually a very good way to end but now I have to ruin it but I'll just say answer Marcus points about yeah your predicted values sort of versus deviation from that actually let me stay on that point that's a good agenda for research right so it reminds me of the paper I have with Miles on occupation precisely at the top right you could think of drivers of policies interventions that help improve or decrease persistence in specific parts of the distribution like why should I get that this job why maybe the government can intervene that is that fair right so I'll just say that that I agree with you on the non-linear idea yeah and with doing more work on that okay well on that note then let's please thank our presenters thank you very much