 At the outset, I just want to take a couple of minutes to first of all welcome you, but secondly, to express my gratitude for the excellence of this whole conference. In my long career, I may have attended at least a hundred conferences, probably more than a hundred conferences, and certainly this one ranks among the very best, if not the best that I have ever attended. So I think the fact that it was an overwhelming success is very much related to the enormous efforts that the present director, deputy director, staff have put into preparing these conferences. But of course, the contributions of previous directors and previous staff over three decades have been instrumental in creating what I consider, and this is my opinion, what I consider to be the leading development center in the world. So again, my deepest congratulations. Now, I think if I had to find three words to capture the essence of this conference, it would be vision, venue, and variety. There has been great vision in the choice of topics. Many topics are highly contemporary, but they were linked to historical precedents in really, I think, a very insightful way. The venue, again, I've been all over the world, I don't know of any comparable venue in terms of efficiency, in terms of facility to move from one place to another, even size. And then variety, notice three Vs. Variety in terms of international distribution of participants who have people from all over the world, but also variety in an intergenerational sense. You have some very old people like myself, but then fortunately, many, many young members of the profession who I feel confidence have learned a great deal by participating in this conference. So in summary, and again, on behalf of the participants, and I probably am the oldest member of this very distinguished gathering, I'm not necessarily distinguished myself, but I'm talking about the gathering. But on behalf of all of the participants, I would really again want to thank Fintop, Tony Addison and the whole staff for the wonderful job they have done. So now let me turn to some logistic issues related to the present session. Each of the presenters will have 20 minutes and unfortunately, I will have to be very strict about it. Fintop has asked for 10 minutes at the end of the session. So we will give him 10 minutes, which means that if everything goes well, we should have up to 15, 20 minutes for comments and questions from the floor. So now let me very briefly introduce the panelists, and we're going to follow the following order. Martin will start. Martin's presentation is fairly general, as I understand it. Peter is somewhat more specific, and Santiago Levy, I think, his orientation tends to be somewhat more policy-oriented. So Martin Reverellian has been the Edmund Villani Professor of Economics at Georgetown University, and he was previously the Director of the Research Department at the World Bank. He has researched extensively on poverty, and in fact, I can't think of anybody who has done more research on poverty than Martin, and on policies fighting poverty. In 1990, he proposed what has become to be known as the $1 a day poverty line, which over time has been somewhat inflated to $125 and even $2 a day. And he and his colleagues at the bank have monitored progress in terms of poverty alleviation, essentially using this poverty line. He is the father, and also I would say the main architect of Povkalnet, which is the data system which practically all empirical poverty analysts use today. He has advised numerous governments and international agencies. He's written a number of books, over 200 papers, and he's a senior fellow of a number of institutions. And again, for the sake of time, I will not give more details. Now, just one second. Peter Lagnar is Professor of Economics at the University of Amsterdam. He was until very recently the Research Manager of the Poverty and Inequality Group in the Development Economics Research Group of the World Bank. He first joined the bank in 1992 after completing his PhD in Economics from the London School of Economics. He has focused mainly on poverty measurement methods and rural-urban economic transformation. A long-standing area of his work has explored so-called small-area estimation methodologies which allow economists to develop poverty maps. Again, he's one of the main architects of poverty maps in developing countries. He has also taught at the University of California Berkeley, the University of Namu in Belgium and in Japan. And then, finally, but of course not least, Santiago Levy has been Vice President for Sector and Knowledge at the Inter-American Development Bank, I think he was named in 2010. He's also a senior fellow non-resident at the Brookings Institution. Previously, he was General Director of the Mexican Social Security Institute from 2000 to 2005. And under his tenure, the Social Security Institute instituted a number of important reforms. From 1995 to 2000, he served as Deputy Minister at the Ministry of Finance in Mexico and he is the main architect of the extremely well-known social programs Proglesa and Opportunidades that benefits the poor. He holds a PhD in economics from Boston University and he was a post-doctoral fellow at Cambridge. So after this set of introductions, let me immediately turn to Martin and his first presentation as you can see is on Towards Better Global Poverty Measures. Okay, thank you very much, Eric, and thank you to the organizers for inviting me. I think I came to the very first wider conference in 1985. I think I was probably the youngest person in the room then and no longer, but I remember it very well. A much smaller event and a different venue but also in Helsinki. Okay, today I'm gonna reflect a bit on the relevance and continuing use of the absolute poverty measures that have become standard in the developing world. And I gotta talk about some of the, some limitations we've realized for some time but talk about some possible solutions too. These are really suggestions for things to think about going forward. I don't actually think the main things we do need replacing but I do think they need augmentation. The guiding principle here, I guess there are two. One is that poverty measurement has to be socially relevant. It can't be out of step with prevailing thought and social policy and that's key. It's relevance is crucial. And the second guiding principle which I'm gonna talk about here is that it needs to be welfare consistent. I'm gonna argue that existing measures have some inadequacies from that point of view and in particular we need to augment existing current absolute poverty lines in two important ways. This is my only picture and it's a bunch of cat leaves. Cat is a mild narcotic, I'm not even sure we'd call it a narcotic, that's widely used in Yemen and Djibouti. I know it well from Yemen. You chew it for hours on end. I tried it once, it didn't have any obvious effect. But the point about cat is it's a social inclusion need. In Yemeni society you have to participate in cat sessions. There are kind of male cat sessions and female cat sessions but the non-participation in such sessions is a problem for you socially and economically. It's part of, it can be considered a basic social need. My first point today is really about what I call the elephant in the room in absolute poverty measurement, the very existence of these social effects on welfare and what do they mean for the way we measure poverty. I'm guided here by, in many, this and many other things by Amartya Sen and a comedy made in 1983 in a paper, an absolute approach to poverty measurement in the space of capabilities translates into a relative approach in the space of commodities. And it's that translation, how you go from some agreed conceptualization of welfare in an absolute sense to a commodity bundle or an income level that provides the money metric of that level of welfare. How do you actually do that in practice? You've all seen this kind of relationship that we see across countries between the national poverty lines and average income on the horizontal axis, the log of private consumption per capita on the vertical axis national poverty lines, and we see that gradient. It's not perfect, obviously, there's measurement area, there's variance, and but strikingly, we see as countries become richer, their poverty lines rise. The question is why do we see that? Why do we see that relationship? There are two explanations, and we kind of don't face up to this in my view. One is the existence of social norms, higher standards of welfare used to define what poverty means in richer countries. If that's the reason, then I think we have to stick to absolute poverty lines because we want to judge welfare by the same standard in different countries. We don't want to say that somebody, if we agree that two people at the same level of welfare, then they both should either be poor or not poor. But there's another reason why we might see that relationship. That is the existence of these social effects on welfare. In other words, that you need a higher level of income, a higher level of real income in a agreed sense, a higher level of real income to achieve the same level of welfare in a richer country. Now if we acknowledge the existence of these social effects on welfare, and I think now the evidence is overwhelming, then what does it mean for poverty measurement? The problem is a deep identification problem. The problem is we don't know from the data we have which of those explanations is right. And actually I don't believe we'll ever know. The re-implication of that is we're gonna have to think, given that uncertainty, we're actually gonna think about bounds on the true welfare consistent poverty measure. And I've shown in a paper that the true measure must lie between two bounds. The lower bound is this standard absolute poverty measure. There is the dollar a day measures if you like or dollar X a day. The upper bound is what I call a weekly relative poverty measure. The true welfare consistent measure allowing for social effects on welfare lies somewhere between these two bounds. Doesn't that lie outside this interval? But where within that bound we'll never know unless we can pin down how much of that gradient we see in poverty lines is due to social effects on welfare versus differences in social norms. To implement the bounds, I've been using this very simple characterization of the relationship between national poverty lines and average income. It just shows the upper bound is the bold line there. It has a slope of one half over about $2 a day and the lower bound is $1.25 a day. This is all in 2005 purchasing power parity. That's what it looks like for the developing world. So this blue line is what you've probably seen before. This is the lower bound. So that's the incidence of poverty, the proportion of the population living below that poverty line over time using the lower bound versus the upper bound. The welfare consistent measure is somewhere between those two bounds. Now, it still looks good from the point of view of overall progress in reducing the incidence of poverty using the upper bound. So that's good news, poverty rates are still falling. But in terms of numbers of poor, they're not falling enough at the lower bound to prevent rising numbers of relatively poor. So in this picture, I give you numbers of poor, absolutely poor, the lower one where we see a decline and the blue line is the numbers of relatively poor. In a sense, it's not too surprising that that's happening. Our success against absolute poverty, our success relative to the lower bound is coming with increasing numbers of relatively poor people. The second challenge we face is possibly more subtle, but something that occurred to me a long time ago, but I've only recently had time to figure out what to do about it. When we talk about poverty frequently or when people talk about progress against poverty, they often use this expression, no one left behind. I saw it recently in the 2013 UN report on these SDGs. And where they said the indicators that track them should be disaggregated to ensure that no one is left behind. And we see that repeatedly. When we look at what's happening to average consumption in the developing world, again, you probably know this, in around 2000, this is the mean consumption per person per day, in around 2000 we saw this sharp change in trajectory. Fantastic, the rate of growth in the developing world is roughly doubled before and after the beginning of the present millennium. But the question is, were the poor left behind in that process? This picture, neither this picture nor the absolute poverty measure I showed you before answered that question. How can we go about answering it? That's quite a challenge. But I'll first motivate it by pointing out that a lot of people are asking the question. Economists are not answering it, but a lot of other people are asking the question. And we see these claims repeatedly. I like this first one from Bunky Moon. The poorest of the world are being left behind. We need to reach out and lift them into our lifeboat. When I, in the past, when I've seen statements like that, I've been mildly horrified because it doesn't seem to be true. I kind of think, well, haven't they read my papers? I mean, don't they know that's not right? And I think about it some more and I start to wonder, well, maybe it's the difference between how they're thinking about poverty and how I'm thinking about poverty. In fact, we hear the contrasting stories and most economists, like myself, will talk about how growth is reducing poverty or at least it's associated with poverty reduction. And we see pictures like this. This is another version of the same picture again. This time it's for various poverty lines just showing you that progress in reducing poverty relative to the lower bound, that progress is robust to the poverty line and given you lines from 50 cents a day to two dollars a day. So all that's kind of understood. That's taken for granted. What I want to say now is how can we understand these conflicting views, one side saying people are being left behind, another side saying we're making a lot of progress against poverty. And the answer in my view is to recognize that there are really two conceptually two very different approaches to thinking about poverty. What we've normally thought about is what I call a counting approach where we think about either counts of poor people and possibly each person counted separately or weighted counts like a squared poverty gap or the Watts Index and so on. These are all within what a framework I call a counting approach. There's another approach which I think motivates the way a lot of people look at the world and what I call a Rawlsian approach. They look not at counts, but they look at the lower bound of the distribution, the lower bound of the support on the distributions we've talked about. They look at the lowest level of living and I ask is that increasing? I'm not gonna argue that's how you should think about poverty, but I wanna point out that's what a lot of people are thinking about. They're looking at that lower bound. I have a paper that gives you much more motivation, but we can draw that motivation from many sources. We can draw it from moral philosophy. We can draw it from social policy. We can draw it from reading the newspapers. I now see it all the time and there are many examples. We also see it in social policy. The largest anti-poverty programs in the world today are motivated by raising the floor. They're not motivated by, they're rhetoric at least. They're not saying reducing numbers of people. They're saying raising the floor. I'm thinking particularly the Debar program in China. Debar means minimum livelihoods, a program that the largest cash transfer program in the world by far in terms of numbers of people, which deliberately tries to raise the floor. The National Rural Employment Guarantee Scheme in India. Same idea, this time using employment to try to assure a minimum level of wage income. Again, raising the floor is the objective. Whether it attains that objective is another question, but the rhetoric in social policy is very much about the Rawlsian approach. So how can we implement it? What do we do in practice to make this possible? Conceptually, what we're trying to do and what I'm trying to do in this paper is make a distinction between two ways in which a community of distribution function can shift. I've given you two sets of CDFs here. On the left, in both cases, we have first order dominance. So further by the counting approach, poverty is unabiguously fallen. But looking at the picture on the left, the floor has stayed put. The lower bound of the support is identical. Look at the picture on the right, the lower bound of the support has risen. I just wanna make that clear in how we measure poverty, make it explicit. How can we go about doing that? I've derived an approach which is operational. It has to be operational with existing data, but it has to be theoretically sound. And a little bit of maths, you can actually figure it out. You can derive a formula for the expected value of the lowest level of permanent consumption, recognizing that that's unobserved. When we measure a consumption or income in a survey, the lowest observed level of consumption or income in that survey is definitely not. We cannot be certain that that is the lowest level of permanent consumption. There's a couple kinds of transient effects in the survey. Maybe you interviewed some household that was sick that week and they didn't eat much. There are many reasons. Transient effects in surveys are a commonplace thing and measurement errors, we understand that. So the observed lower bound in a survey is not a reliable indicator. We need to think about a probabilistic approach which recognizes that the observed lowest level of income is not with certainty the lower bound of the support. Once we recognize the need for that probabilistic approach where the probabilities are not normative, they're positive, the weights we attach to observed levels of living are an effort to try to get at the expected value of the lowest level of permanent consumption. We can back out this really cute formula. The expected value of the minimum level of consumption in any distribution is the upper bound, think about a set of people with an upper bound to that income above which we are certain you are not the poorest person. So there's a set of people within which we think the poorest person is found but that set has an upper bound in terms of consumption or income and that's this y-star. It's not a poverty line, it's the income level above which we think there's zero probability of being the poorest person. So I have to make a judgment about that. And secondly, we have the squared poverty gap and the poverty gap index for that y-star. So again, the formulae point us to ratios of phosphatoglyphalic measures in a somewhat different application. We're not using them to actually implement the counting approach, we're using them to implement the Rawlsian approach. That's kind of cute because that means I can estimate the expected value of this lower bound of the support. I can estimate that expected value from existing poverty measures. I can back it out and this is what I get. The lower bound is 67 cents a day. About half of the dollar 25 a day poverty measure. That's an average, averaged over 30 years from 1981 to 2011. But the variance over time is minimal. Here I give you a gain, that picture with the average consumption, average consumption level in the developing world and that blue line at the bottom, that's my estimate of the floor. One of the amazing things about statistics is that believe it or not, that blue line has a statistically significant positive slope. It's got a T ratio of about five. But you can see for yourself, there's been very little progress in raising the floor by this measure. All of our progress has been in reducing the numbers of people living near the floor. The lower bound has not risen. Another way of looking at that, go back to the CDF. So clearly what I've established is it's the pattern in the very top panel here. First order dominance again, but with no increase in the floor. And if we just calculate the absolute gain across percentiles across the 30 years in the developing world, the absolute gains go from basically zero all the way up to some very large number at the top. Almost final point. Let's take a longer term perspective on this. I've got very interested in the history of poverty reduction in rich countries. And one of the striking things I find going back combining data sets, the Boogie on Morrison data set with some new data, what I find is that the way the rich world today escaped poverty over 100 years between 1850 and 1950 was strikingly different to the way the developing world is escaping poverty today. It's different in the following sense. The rate of progress by the counting approach in the developing world today, annualized rate of progress is greater than in the rich world over that 100 years, annualized, but the rich world did much better in raising the floor. The rate of progress in increasing the floor in today's rich world at the time it was as poor as today's poor world, that rate of progress in raising the floor in the rich world was about double the rate of progress we're seeing in the developing world. Lots of things to think about why that might be true and my paper discusses some possible explanations since I'm short of time I'm gonna just conclude with two slides. I think we can talk seriously about these bounds. I think we're probably still gonna focus on the lower bound but I think we should keep the upper bound in tow. In other words, we should be judging poverty. You should be deemed poor if you're poor by a common international standard or you're poor by a standard that's typical of the country you live in. We need both those criteria. The upper bound is the latter one, obviously the lower bound the former one. And I argue that is welfare consistent. I'm not trying to depart from Sen's original Maxim. And finally, we can also measure success in leaving no one behind. It is tractable. We can measure or make operational. I'm sure it can be refined and improved but at least there's something on the table. Thank you very much. Thank you very much, Martin. So Peter is next. Good morning everyone. Let me start by thanking the organizers very much for this privilege of being allowed to come and present to you this morning in this panel on poverty, in this conference on the big themes of economic development and the future of development economics. And I chose to spend this morning talking about the experience of economic development in one small village in India. I thought it would be possibly an interesting application of looking at some of these themes that we're interested in in this conference of the big themes of economic development. And it also allows us to sort of think about certain aspects of income distribution and poverty that perhaps we can't touch on as easily using sort of large scale household surveys which are typically the bread and butter of poverty analysis. So I'm gonna talk about economic development in Palampur. Basically I would like to convey essentially two sets of ideas. The first is that I would like to show or illustrate to you that the economic development that's been occurring in one small village in India over a period of roughly six to seven decades has been one that actually mirrors or reflects many of the big themes of economic development that have been coming up at this conference, issues of structural transformation and so on. And these can be found and seen at the sort of very granular level in the context of this one little village in North India. And then secondly, I would like to sort of think a little bit about some of the distributional issues and some of the observations that we take away from that experience in this village of Palampur and just sort of think about what are some of the important issues that arise that we may wanna try to do more about capturing in the kind of analysis that we do typically with the normal data that are available to us. So in some sense it's a plug I'm hoping to make for doing this type of research with these village studies that take a very long term perspective on economic development in the context of a very focused small community. And I'll come back to that in some of the discussing comments. So first of all, Palampur is a village in North India and the Gangetic Plain of North India and the state of Uttar Pradesh in the district of Muradabad district. It's a small village that's been studied for approximately, we're into our seventh decade now. We've most recently collected data in 2015 and we've been collecting data in this village over seven decades now starting in the 50s. The kind of issues that have been examined and looked at include some of these big questions in economic development like the impact of the Green Revolution and the impact of diversification in the rural economy. We've looked at institutions in the village such as sharecropping and tenancy and we've been looking at how conventional economic theories allow us to understand what's going on in the economic development of this one village. The village was chosen not randomly but it was chosen with the intention of capturing some of these big processes that we're interested in. It's not too close to Delhi. It's about 200 plus kilometers away from Delhi so it's not been contaminated in a sense by this very large city but it's located in a fairly densely populated part of the country and so there is proximity to urban areas with daily commuting and so on going on and that's something I'll come back to. It's a village that has been cultivating wheat and so it's been very much part of the Green Revolution experience that occurred during the 1960s and 70s. You can't really talk about a typical village but we can certainly say that the village is not particularly unusual in any kind of way and that's something that has actually continued over the years. It's not as though somehow during the process of our investigation of this village somehow suddenly a large aluminum smelter was located right next to it and thereby changing entirely the trajectory of the village. So nothing like that has happened. It's not particularly unusual but of course we can't say anything more about India as a whole than what we can say about this particular village. This is where it's located. It's in Uttar Pradesh in the northern part of Uttar Pradesh. It's about 30 kilometers to the south of Muralabad which is a fairly large city and about 18 kilometers to the north of Chandosi which is a much smaller little town but both of those cities have exerted quite an important influence on the village. The village has been studied since the late 1950s. First it was studied in a very detailed way by the Agricultural Economics Research Center by investigators from the Agricultural Economics Research Center in Delhi, part of Delhi University in 1957 and then again in 1962. Very important aspect of this study is that there's a census of the village. So every household in the village was interviewed and information was collected on every household that was in the village and that tradition has been carried through all the way to the present. So unlike many other village studies that are perhaps familiar to you where there's been a sample of households selected that have been investigated or studied closely, this is a case where we've actually been studying the entire population of the village and that's gonna be important for some of the observations I'll be making about income inequality for example. It was studied in 1974 or 75 by Christopher Bliss and Nick Stern who spent nine months living in the village and published a book describing their experience where they were focusing very much on agricultural practices in the village and trying to understand the tenancy market and so on. It was then again studied in 1983, 84 by Jean Dres and Nares Sharma who spent 15 months living in the village and conducting very intensive field work. Very detailed information was collected not just on the conventional type of economic variables of interest like income and so on but also very detailed qualitative information that was collected and assembled in a huge mountain of diaries that were collected and so on. So there's very detailed information of both a quantitative and a qualitative type. It was then subsequently resurveyed in 1993 with a kind of a quick survey which did not attempt to collect income data in great detail but did collect information on demographic structures and occupational profiles and so on. We then had a resurvey in 2008 and 2009 that was conducted by Himanshu who's a colleague very much involved in this project and who's also attending this conference here today. He spent two years living in the village with the collaborators and collecting probably the most detailed information that we've collected yet, including for the first time also consumption data which we've yet to analyze in great detail but that's been a bit of a departure from previous work but also very detailed information on aspects that have not received the same kind of attention in the past, gender relations and so on. This is much more detailed information that we have now. And then most recently in 2015 we used these computer handheld tablets to do another quickish survey to get an update on occupations and the demographic structures and so on. So we have data that extends all the way through to the present day from the 1950s and this is quite a unique setting in which to study economic development in one village. Just some quick figures about the village. It's a village that in 1993 had just over 1,000 people living in there and that's increased to 1,270 by 2008, 2009. It's a smallish village, the average household size of just over four. A key point is that the village has a quite a established and well-defined caste structure which is a big theme in understanding what's been happening in the village over time. There's three main castes that are really key players in the sort of economic development story of the village. They comprise the takours which are ranked at the top of the caste hierarchy followed by morows which are a traditional cultivating caste and then there's the jattabs who are a sort of disadvantaged caste at the lower, at the bottom of the caste hierarchy. These are the three big population groups. There's a whole series of other castes in there but those are the three main castes. The main economic activities in the village today are agriculture, it remains an agricultural village but there's also very important diversification out of agriculture into non-agricultural activities. That's been a big part of the story of economic development in Palampur. The kind of crops grown are wheat as I already mentioned but also rice, sugarcane, a cash crop called menta which is a main ingredient into I believe the chewing gum industry and that's exported around the world and the number of other crops and so on. Some of the main public amenities include a school. There's a railway station, not far away, temples, ponds and so on. Some of the big drivers of change over this time period since the late 1950s has been first of all population growth. In 1957 there was just over 500 people living in this village and that's expanded to 1255 by 2008, 2009 and expanding further with the most recent data. A very big development factor for us as economists of course is that the rise in per capita incomes, incomes have risen from 189 rupees in 1960, 61 prices to 411 by 2008, 2009 with a big jump occurring between 62 and 74 as part of the Green Revolution and then again a big increase in the subsequent period between 74 and 2008. Wheat yields have increased considerably and this is part of, this is the Green Revolution story and this big changes here occurred to a large extent between the 50s and 60s in the early years and 1974, 75. This was when the Green Revolution was mainly introduced, the new seed varieties, fertilizers, but a very big part of that story has been the expansion of irrigation to the village became fully irrigated by the time by the 1970s whereas that was far from the case in the early two rounds. By 2008, 2009 the yields have expanded even further. This is the consequence of further intensified irrigation allowing now for the first time through triple cropping during the course of the year and also the sort of intensification of cultivation through the use of technologies like tractors and so on which only sort of made their appearance by roughly 1983, 84. Wages, daily wages, agricultural wages have risen dramatically in the village and this is something we can see in the daily product wages where you're roughly receiving two and a half kilos of wheat per day for working in the field as an agricultural laborer in the 1950s and 60s and that's up to nine kilos of wheat per day if you work in the fields today. So that's a very, very big part of what's been going on to welfare in this village over time and I'll come back to that. In terms of the population groups, the population shares of these three main groups, the Tacors, Morales and Jatabs has increased over time with some of these smaller casts that I haven't emphasized so much here kind of diminishing in size and in number. There's migration into and out of the village is not really a very big phenomenon but some of these smaller casts have been moving out of the village. There was one cast called the Passis that moved into the village during the 1940s and 50s and that have been making their way out of the village again more recently. So what migration has occurred has been mainly of that kind but the big groups in the village, the Morales, Tacors and Jatabs have tended to stay in the village and their representation in the village population has increased over time. We've been using as our main indicator of wellbeing or of analysis a measure of income which we've attempted to construct very carefully and in a way that's uniform across all the survey years so that we are able to make comparisons over time. The interesting thing, the luxury that we have with a very intensive field study like this is that we can really subject this income measure to all kinds of checks and cross-checks and so on so we're fairly confident that this is a very accurate measure of income. Measuring income is of course really difficult to do and in a rural setting where you're doing cultivation and so on these things are very tricky but at least we have been in the position to do extensive checking and verification of the income data that we've collected, particularly in the more recent rounds of the data when more of us have been directly involved in the field work. What's been happening to the structure of the economy of the village? Well, as I said, it's an agricultural village largely and historically it was overwhelmingly agricultural and it still remains a very important part of the economy. Agricultural still remains sort of the backbone of the village economy but it's no longer the case that it's the only story of economic development. A very big part of what's been going on is an expansion or diversification out of agriculture. The big part of that, a big feature of that is the phenomenon of commuting on a daily basis out of the village to some of these nearby towns. So we have a kind of a process that's not unlike what Arthur Lewis has described of people moving out of agriculture and moving into the towns. It's taken the form not of permanent migration to the city but of commuting from the village to the nearby towns where there's all kinds of casual daily wage work that's underway. So that's a growing part of the village economy as well. We can see that here reflected in terms of the share of income coming from non-agricultural activities. Roughly 40% of households receive more than half of their income from non-farm sources by 2009 and that was well below 20% in the 50s and 60s. Non-farm income has just really expanded and become dramatically more important in the village over time. We can see that also reflected in the terms of the kind of occupations that households and the number of villagers that are involved in non-agricultural activities. It's important to emphasize that the non-farm employment that occurs in the village is not sort of regular salaried employment largely. It's very much of it is a kind of a casual, not terribly productive, it's not terribly attractive. It's often dangerous in the sense that it exposes the employee to all kinds of health risks and hazards and so on. So it's not necessarily always particularly appealing but it's considered appealing to the farmers, to the villagers. Agricultural daily wage work is very tough and very difficult. This climate makes it often very, very difficult physically and it's very much sought after by villagers to find some kind of source of employment outside of agriculture and that's what we do see. Self-employment has been expanding over time as well. So it's a combination of daily casual wage employment and self-employment that's been really the big growth over the past decades. We can see that the share of income from non-farm sources has been increasing from 34% to 52% for the village as a whole and a very important development and that's something to take note of given our interest in distributional issues is the rise in incomes from non-farm sources that are going to the jatabs, this lowest, this cast that's located at the very bottom of the social structure. This is something that's relatively recent. It was not happening up to 1983 but then between 1983 and 2008, 2009 has been something of a revolution in the extent to which the jatabs have become involved in this non-agricultural sector and that's been a big part of our understanding of what's going on both in terms of inequality as well as poverty reduction in the village. So what have been the big distributional consequences or outcomes in the village over time? Well, I just would like to sort of focus my remarks here and sort of three things. There's been a very significant poverty reduction and this is something that resonates also with what we've been observing at the all India level. It's not inconsistent with what we've been observing at the all India level. The progress has been most striking in the more recent decades. There is also an increase, a very sizable and discernible increase in income inequality within the village. This is something that doesn't resonate as well immediately at the all India level but I'll come back to that with some remarks about that aspect. And then one thing that we've been able to look at and we've been quite interested in looking at recently is this whole issue of mobility, income mobility and its connection also to these issues of opportunity that we've been discussing at this conference. And what we have seen and what we do see in Palampur is this very dramatic rise of this lowest cast in recent decades. And so that's evidence of really important social mobility that's taking place. At the same time and touching on some remarks made by Marcos Yanti at yesterday's conference, we do also see some evidence of a Gatsby curve. So if we look at intergenerational mobility, there's actually some evidence that as the villages become more unequal, there's actually a decline in intergenerational mobility in the village and we're able to provide some evidence of that and I'll come back to that. So just on the poverty front, what we have is conventional estimates of poverty irrespective of what kind of measure you use, we'll find a clear decline in the headcount rate of poverty or some other conventional measure of poverty. I do wanna go back and try out Martin's lower floor story. We have done all the stochastic dominance analysis but we never really focused on that minimum income and here we don't have a survey, we actually have a population census so it'll actually be interesting to see what's been happening to the minimum incomes in the village over time. So I do wanna go back to that. But sort of conventional poverty measurement suggests that there's been very significant poverty decline. One thing that we've been able to do in Palampur which is one of the nice features of a rich study like a Palampur studies that we've been able to complement our analysis of poverty based on income with at least some investigation or some acknowledgement of other dimensions of welfare. Then the way that we've done that in the Palampur study is to introduce an idea, a concept of what we call observed means which essentially takes the form of the investigators who spend a very long time in the village and really get to become very familiar and very on almost intimate terms with all of the villagers. They're able to observe these villagers in a way that extends well beyond the calculation of a simple income measure. They're able to look at these villagers in terms of their lifestyle, in terms of intra household distributional issues and so on. And they're able to come and obtain an assessment of how well that household is doing that's kind of separate and not necessarily entirely lined up with an income measure. And so we've constructed over the years in 1983 and in the subsequent rounds a sort of measure of observed means. Now this is a very much a relative measure. It's a way of ranking households rather than the cardinal measure that actually provides a precise number. But it allows us to rank households in terms of a very broader concept of welfare that implicitly takes into account all kinds of other dimensions that may be of importance in the village. And then we can compare how households are ranked when it comes to ranking in terms of incomes or when it comes to rankings in terms of observed means. And what we see is that there is at least in this particular exercise, there's a fairly nice correspondence in terms of the kind of rankings we get using our very carefully constructed income measure and the rankings we get using this observed means ranking. And just to show one aspect of that here we have, we're looking at the three main caste groups and we're putting them into five categories or roughly 20% of the village population in each of these categories extending from being considered very poor to being considered rich. And we see that the jet tubs in 1983-84 on the basis of income were very clearly concentrated in the very poor and poor categories. Whereas the morows who had been benefiting largely from this green revolution technologies, they were the traditional cultivators, were really taking advantage of these opportunities and were really making the most of these new technologies and were doing well in agriculture and therefore seeing relatively high incomes and were therefore ranked at the sort of higher categories of this distribution. In 1988-2009 with a diversification of the economy away from agriculture and basically pressure of population on the land it's become much more difficult to sort of make ends meet relying on agriculture. And the morows whose strategy had been to focus largely on cultivation have found it difficult to sort of maintain those same living standards that they were able to achieve in the earlier rounds based on their cultivation practices. So they've dropped down in the welfare distribution and jet tubs who we've seen were taking advantage of these new non-farm opportunities have been able to take advantage of that and actually rise at least for some of them in the income distribution and are now sort of located along the full spectrum of the income distribution. If we ask whether a similar pattern would be observed if we had done this whole exercise in terms of this observed means criteria we see that the picture is very similar. Again, with the jet tubs at the bottom of the distribution and the morows at the top of the distribution in 1983 and then that changing as we move to the more recent rounds. So again, this is reflected in the observed means classification as well. All of this simply to say that this village study and this detailed information allows us to look at welfare and welfare rankings in terms of the sort of standard criteria that we often use as economists but also to complement that with other indicators that are of importance and other dimensions that are of relevance. Turning now to inequality, I've already mentioned that there has been a noticeable increase in income inequality in the village and that's just documented here in terms of the Gini coefficient as well as any other measures of inequality that one might be interested in and it's really quite striking how inequality has risen from say 0.27 in 1974, 75 in terms of the Gini coefficient to 0.37 and that's a very big increase over time and we don't typically see those kind of big increases very frequently. So there's been a very noticeable increase in inequality and the question arises, well what's been behind that? One of the exercises we carried out was a sort of standard decomposition of inequality of income inequality by sources of income and there we can see that whereas in the earlier rounds of the data most of the inequality could be attributed to cultivation income by the 2008, 2009 round of the data more than 50% of overall income inequality could be attributed to the incomes coming from non-agricultural sources. So this diversification of the economy away from agriculture into non-agricultural activities has had as a consequence a very big increase in inequality and we can see that kind of contribution coming from these non-farm income. So inequality has risen dramatically as a result of this expansion of non-farm income opportunities. We can ask as to, we've already shown you quickly that the jatams have been rising in the income distribution and the murals have been dropping down a little bit and what we do see confirmation we see confirmation of those kind of processes when we do a decomposition of inequality in terms of population groups with caste differences accounting for less of the overall income inequality that we observe in the village than what we had observed in earlier rounds. 13% of the inequality is now attributable to between caste differences. Most of the inequality is a within caste phenomenon. One development that we had been observing and that we had become quite alarmed about up to the more recent rounds of data had been that the jatams as a group had been falling steadily behind and further behind the rest of the village. So it had been the case that up until 19, up until the recent rounds of the data, the jatams as a group had been falling behind the rest of the village population. That seems to have, again, probably as a result of this entry into the non-farm sector, that seems to have reversed itself with the jatams now no longer being as falling behind and in some sense catching up to the rest of the village. And we can see that in this partitioning index figure of 20%, relative to 36% in 1983, 84%. So just a couple of remarks about what we've been seeing in Palampur and thinking about how that relates to what's going on in India as a whole. We've been noticing quite a significant increase in inequality occurring in Palampur and this has been occurring as a result of a process of non-farm diversification that if you look at the all India level, if you look at the NSS data, you do find evidence of a similar process underway in India as a whole as well, this diversification of the rural economies occurring and the NSS data, for example, find plenty of evidence to support that. Yet at the all India level, we don't see a significant increase in inequality. And in Palampur, we do see this very significant increase in inequality. So can we think a little bit about how to reconcile those findings? Well, one of the possible issues is that we've been measuring inequality in terms of income, whereas the NSS survey typically measures inequality on the basis of consumption. So it could be the fact that these different indicators are telling a different story. That's one possibility. Another possibility that has come up already at this conference is this whole question of people thinking about inequality in terms of absolute terms, as opposed to us measuring inequality using relative measures of inequality. It could be that there's a lot of debate about inequality in India and a lot of excitement about inequality because people have in mind very clear increases in absolute inequality. And that's just not reflected in the relative inequality measures that we're using. But I would submit that there's a third possibility, which I think is an important one to pay attention to, which is that it's perfectly possible for all India inequality measured in a relative measure to not change over time, while at the same time, village inequality in all of India's half a million villages is increasing. That's not at all inconsistent. And it might actually be occurring. It could well be occurring that we're seeing village level inequality increasing, but that all India level inequality, things are not changing much at all. And that would be possible only if the differences between villages were in some sense narrowing over time. And maybe such a process is underway. It's well worth exploring further. Finally, just a few remarks about mobility. I think I've already documented the rise of the jet tubs and the decline of the morows. We find evidence, if we look at transition matrices across adjacent years, we find evidence of considerable mobility, both of upward as well as downward mobility. And we also find evidence of declining immobility. So if you look between 57 and 83 in terms of income, 22% of households were ranked in the same quintiles. If we look at between 83 and 2008, 2009, that had declined to 18%. So there seems to be some evidence of increasing mobility when we just looked at the standard intra-generational mobility. And that's consistent with what we've been seeing, understanding in terms of the relative movements of the jet tubs versus the morows and so on. However, it's interesting to think about inter-generational mobility as well. It's a distinct type of mobility. It's the mobility where you ask how a father's income is able to allow you to predict or anticipate what your income is going to be. And it turns out that there has been a lot of discussion following Alan Kruger's well-known discussion of the Gatsby curve that as economies rise, as inequality in countries increases, inter-generational mobility might actually be declining. And then we see this relationship this is taken from Alan Kruger's paper. We see that there is this positive relationship between immobility across generations and inequality in countries. Well, we decided because of the data and the richness of the data that we have that we could actually investigate this to some extent in the Palampur data as well. And essentially the story we find is very similar to what we've been observing at this national level in the Gatsby curve graph. What we're seeing is declining immobility and increase in the inter-generational elasticity of incomes between fathers and sons. And we first do the comparison between 1957 and 1983. And then we do the comparison between 1983 and 2008, 2009. And as inequality has risen in Palampur, we're seeing evidence of inter-generational mobility declining, and this is obviously a source of considerable concern. So let me just end very quickly just to say that I think there is a real value to looking at some of these issues of income distribution and poverty dynamics in the context of these village studies. And in India, there's a very rich tradition of doing this type of work. There are lots of village studies underway and they are periodically assembled and summarized. And I think this type of analysis can be very helpful to understand what's going on. Many important components of the story of income distribution are often statistically invisible in our household surveys, whether it's a village or whether it's looking at a small town. It's something that we should pay attention to. And there's a very interesting study that this is not just to suggest that this is not just a phenomenon of a developing country. Very interesting study by Bill Easterly and colleagues recently of a single block in New York City over four centuries has documented also this very interesting process of up and down dynamics of how the economy of one block in New York City can be changing dramatically over time. So let me stop at that and thank you very much for your attention. Next we have Santiago Levy. Thank you, Eric. And good morning to everybody. I won't be using slides, so I'll be speaking here from the table. Let me begin first by thanking Weider. It's a real pleasure to be here. Thank you so much for the invitation. And as Eric mentioned at the beginning, I'll try to speak a little bit more from the policy perspective and needless to say, very much from a Latin American view, perhaps some of my remarks will be useful for the regions of the world, but I really don't claim any knowledge of other areas. I'll speak mostly about Latin America. I'll begin by saying that in matters of poverty, there's actually good news coming from Latin America. If you look at the period from 2000 to 2013, 2014, years for which we have the latest income expenditure surveys, the news is basically good for the region as a whole. If you take the $4 a day poverty line, poverty count fell from 45% to 25%. If you take a stricter poverty line, say a two and a half dollar a day, poverty fell from 28% to 14% of the population, about half. So those are very, very good news. Secondly, poverty fell in almost every country in the region. It's a really economy, it's sort of a continent-wide phenomenon. Thirdly, poverty did not increase in 2009, 2008 as a result of the world financial crisis. And fourthly, even though the region has not been doing so well after the world financial crisis and in fact there's been a major growth slown in the region since 2011 to 2015, poverty has not increased. So we compare these facts with what used to happen in Latin America during previous crisis or other periods. This is really different. It's notable on the whole, it's very good news. And I think it's sort of a substantive achievement of the region that needs to be sort of recognized. What explains these notable achievement of the region over the last 15 years or so? I think broadly the story has two large components. One is growth. Certainly, particularly the first decade of the century was very good for Latin America. On the back of spectacular conditions in international capital markets and on the back of a commodity boom, the region increased the growth rate to about five, five and a half percent, which is almost double the traditional growth rate of the region in the 90s, which was on the order of two and a half to three percent. So growth is an important part of the explanation, but it is not the whole part of the explanation because as I mentioned before, from about 2011, the region has been slowing down. Sadly, every year since 2011, the growth rate to the region has been less than the previous year. And in fact, the growth rate for 2015 will probably be zero. So the fact that poverty has continued to fall despite the fact that there's been a growth slowdown says that growth is not the complete part of the story. I think the second part of the story has to do with the fact that poverty policy is broadly defined in the region improved noticeably over the last 15 years. There's been a real change in the attention of governments to designing and putting into place poverty focus programs that have really made a difference in terms of the living standards of many people. So what I wanna do now is speak a little bit about these programs, my assessment of where they are, and then come back and pick up on the growth issue and then say, what does that relate to poverty in the region? As many of you know, Latin America was pioneer in putting together these targeted transfers programs to the poor that were made conditional on some kind of socially desirable behavior, behaviors associated with sending children to school or having members of the family go to health clinics or some other behavior that was thought to be important to be stimulated. And many countries design programs that would target transfers conditional on these behaviors. Depending on which country the region you look at, these countries are now targeting anywhere between half a percent and 1% of GDP directly on these families. This is different from what it was in the 1990s and targeting anywhere between half a percent to 1% of GDP on the lowest 15, 20% of population really has meant a difference in terms of income poverty, of reduction of income poverty, and has also made a difference in terms of reducing inequality. There's of course a lot of heterogeneity across countries in terms of the size of these programs and the coverage, but by and large, almost every country in the region has a program of this sorts, sometimes focused on the early associated with pensions, but usually focused on the whole members of the household associated also with behaviors, as I mentioned before, having to do with schooling and having to do with health and nutrition. This has been a substantive achievement, I think, in terms of policy, but I think that there are four important issues that looking forward we have to pay attention to thinking about these problems. First, there's still an issue of coverage and targeting. I think some of the numbers that Martin was showing us before probably suggest, even though the data's not all that was there, at the very, very, very bottom, perhaps people aren't being able to help as much as they have. So there's still an issue of targeting and there's still an issue of coverage and these issues will become important in the future years because the fiscal situation of many countries in the region is not as good as it was before and trying to really reach some people haven't been treated so far is an important challenge as to be there. Second, yes, these programs have been very useful in changing the allocation of household time, a lot less allocation of time of children to work, a lot more allocation of time of children to go into school, some allocation of time of the household to doing things associated with their health and their nutrition, but the deeper indicators of human capital accumulation are not as good as we'd like them to be. So food consumption has gone up, the diets are much more diversified than in the past, but if you look at deeper indicators of nutrition like anemia, the advance has not been as deep as we'd like it to be. Some indicators of unstunting have improved but have not improved as much as we have and if you look at other indicators, yes, children are going much more to school than they were doing in the past, even youngsters are going much more to school. There's been an important change in the ratio of girls to boys so that in many countries now what you see is that attendance by girls is actually higher than attendance by boys, a real break from what the situation used 20 years ago, but if you ask questions about how much learning is actually taking place in school, the news is not so good. So first point, coverage and targeting, second point, impact on deeper indicators of human capital not so much there and this brings me to the third point, I think that has been an unbalance in many of these programs, paying attention to the demand side of stimulating the demand for education, stimulating the demand for health, stimulating the demand for attention to nutrition and similar things have to do with human capital but insufficient attention on the supply side associated with the quality of services. So maybe you have the kids going to school but the teacher is actually not there, you have the household showing up at the health clinics but the medicines are not always there. The quality of services is still a challenge and the political economy behind the quality of services is still an unsolved issue in the region. It has proven a lot easier for governments in the region to, if I say so bluntly, redistribute money and it's been a lot more difficult for governments in the region to actually organize services from the point of view health and education in particular and ensure that in remote villages and in peripheral urban areas, the quality of the services really implies that there's a translation into better human capital. And the fourth flaw with this program as I see it and this is probably a design flaw of some of us who were thinking about these programs 20 years ago and now with hindsight, we realize that the important design flaws have to do with the fact that I think there was a substantial underestimation of the problems associated with early child development. There was a thought that if children had better nutrition and if children went to school, that would be enough. And I think issues that now we understand better than what we did 20 years ago associated with other indicators of child development, social emotional development, language development were not paid the attention that they should have been paid to. And what you have now is evidence that shows that children from poor families are going to school but by the time that they're reaching school, they're one and a half, two years behind children of similar ages because their language development and other metrics have not stopped. So there's an important agenda to incorporate early child development as part of the human capital programs that are being done there and that's not there. So there is a substantive agenda from the policy point of view strictly thinking about poverty programs in the region and that substantive agenda in the years ahead has to focus in my view on the quality of targeting and the coverage of the programs. On the political economy of service delivery, we did understand the incentives on the household side to consume more human capital. We did not understand the incentives on the government side to supply more health and education and more things to do with human capital. And thirdly, to design cost effective ECD interventions, early child development interventions that are cost effective and that can be brought up to scale. We have a lot of evidence of very small programs on early child development that yes, we can measure very carefully what the impacts are. We have very little evidence of what can be scaled up programs that can be done massively at a cost effective way and incorporated into the human capital issue of poverty alleviation. That said, I think the balance from what we were say 20 years ago is on my view on the positive. The key question, and this brings me to sort of the second part of my intervention, the key question then from the point of view of poverty policy looking forward is, is the accumulation of human capital in poor households translating into higher earned income? Is this more human capital that they're acquiring as a result of all these programs implying that they can earn more income on their own? And there the evidence that we have is very, very scanty. They're not sufficient studies that really can do a panel or to a follow up of households and find out whether in fact these households are actually having more income on their own as a participant in the labor market. But the little evidence that is beginning to come in and now in some countries, these programs have been there for 20 years or so, the little evidence that we're beginning to see is that the answer is most likely not. There is not evidence that there's been translating of this into higher income for the poor. This is again has to be taken with a grain of salt because the numbers as I mentioned before are not really systematic for the region as a whole, not deep enough. But this issue of higher earn income for poor households has to be placed in the context of are these economies growing enough to generate this? And here the difficulty, the analytical difficulty is separating a transitory growth that the region observed particularly between 2002 and 2009 associated as I mentioned at the beginning with particularly good conditions in international capital markets, particularly good conditions associated with the growth in China and world commodity prices that did bring a kind of a growth spurt in the region into a perhaps a more long-term trend in which growth in the region is really very, very slow and it's mostly associated with the, perhaps it's the biggest possible in Latin America which is why is productivity growth in the region stagnant. So if you take away the contribution of population growth and all that and if you look at growth indicators to the medium term in the region, what you find is that productivity growth, total factor productivity has pretty much stagnated in the region for the region as a whole. Variations across country but by and large it's a very, very dismal picture. Now I think the challenge is to begin to make a bridge between the people who think about growth and productivity and the people who think about poverty. This has been by and large two sort of separate areas but I think we now have to think about total factor productivity growth, not as a growth issue per se, we have to think about a poverty issue. What is it that needs to occur for productivity, particularly labor productivity of the poor has to do? If we're going to break the intergenerational transmission of poverty and if it's going to be the case that the new youngsters, poor youngsters that are entering the labor market are gonna have higher earned income than their parents, it must be that they have higher productivity jobs than their parents. And this higher productivity jobs for the new cohorts of young people that are entering to the labor market are not being there. This is in my own personal view, the key analytical challenge in terms of poverty policy for the region. We do understand a lot better money of the issues having to do with human capital accumulation and we do understand the challenges that are specific to what needs to occur to improve on that dimension. I think there's a lot less consensus on a diagnosis first of why is it that total factor productivity in the region as a whole has stagnating and secondly, what would be the sort of policy combination that would bring to back breaking that. We don't have a canonical model, we don't have in my view a systemic understanding of that issue. What I think we have today is a paparita recipes that really are just thrown out and so one day we focus on micro credit as a way maybe it's gonna help and raise their incomes and then the next day is gonna be the pushing small and medium sized enterprises and granting credit to small and medium sized enterprises and then the next day we think about training programs that this is gonna be the real key that it's gonna bring about the change and then now we think that the real issue is about that the quality of education has not been up to standards. Each of these component probably is right but we don't really have in my view, I might be wrong, a systemic understanding as to are we really just shooting with a shotgun all over the place or do we have a clear understanding of exactly where is it in each country that is a real substantive constraint for real total factor productivity growth. So policy in my view a little bit, I'm being a bit sarcastic here, is kind of a flavor of the month kind of policy between SMEs or training programs or the quality of education or micro credit and the research agenda looking forward in terms of poverty alleviation policy has to be focused on this and in particular has to be focused on the centrality of the labor market as the central mechanism by which the question that I asked before is greater accumulation of human capital for the poor going to be translated into higher earned income for new cohorts of workers that are entering the labor market. Without focusing on the centrality of the labor market I think we're not gonna make progress. This is very conflated with issues of formal and informal participation in labor market. There's a plethora of views out there. There's yet a canonical model that will bring sort of a good sense of what we have to do. So this is a research group. The invitation then is to focus our attention in poverty policy to of course continue working on these issues having to do with human capital accumulation and some of the challenges that I mentioned before. There's still an agenda and there's still things that can be done there and these things will improve the welfare of the poor. But we have to broaden our view when we think about poverty policy we've got to place this in the context of growth and total factor productivity and we have to think about the deeper questions about the labor market, productivity growth and labor productivity in the labor market. There is where I think the onical challenges are. So let me stop there. I think there are many achievements, many things to feel proud of and there are many many more things to be worried about. Thank you. Thank you very much. We have just heard some I think very rich presentations and covering really the continuum all the way from the global level to the regional level to the village level. What Martin did was to focus on the need for accurate and relevant measurements of poverty, particularly for the most deprived among the poor. Santiago looked at regional trends within Latin America showed the impact of a number of social protection projects on poverty raised a number of issues in terms of the interaction, interrelationship among growth and poverty. And finally, the Peter Lionel concentrated on the economy of the village looking at poverty issues within the village context and again emphasizing the impact of technological change and policy changes within the village. And to use an analogy which is probably quite far fetched as some of you know, I have a small estate in Northern California which is a small Redwood forest. And the analogy that came to my mind is that what Martin was doing was to focus on how to nurture the giant Redwood forest, really global view. What Santiago did was to concentrate on a forest which combined different kinds of trees but the idea again was how to nurture these trees so that they would grow. And finally Peter looked at it from the standpoint of a bonsai tree. How does one nurture a bonsai tree so that the bonsai tree actually grows? So it was a very rich and of course a complete, relatively complete picture of this continuum. So at this stage I opened the floor for questions and comments and as usual what we're going to do is to take maybe two, three questions, rounds of two, three questions, give the panel a chance to answer them and then move to the next one. Rolf? Thank you very much, Rolf from the Hoover Institute of Social Studies. Like Eric, I'd like to congratulate all the presenters also presenting a presentation with clear presentation with our PowerPoint. I enjoyed that also very much. I actually want to link up the first and the last presentation. I liked very much Martin's proposition to look more at the bottom and I liked his conclusion was very descriptive that in the sliced countries the bottom was lifted faster than in developing countries. And Santiago Levy mentioned the absence of attention to the labor market. And I just want to ask Martin because he said in his paper he has more explanations which he mentioned. Is there a link between the setting up of the labor market institutions in the 19th century, minimum wages, social welfare in developed countries which actually was a big contributor to lifting up that bottom and then coming back to Santiago, if that is the case, what are the challenges then for developing countries to achieve that same process and what that implies also for wider research in the future. Thank you. Thank you very much. Yes, sir. Gentlemen. Good morning. This is Anand from University of Bradford. Thank you very much for all the presenters. I'm going to ask a question on the bonsai plant that is to Peter. I think as someone who is interested in case studies I have drawn a lot of inspiration from Palanpur studies and I think these have brought a lot of rich, consistent kind of understanding of village level institutions and so on. Just I have two questions. One is whenever I try to propose these kind of studies university research ethics committee really demands quite a lot of attention. So by the time these studies started in the 1950s perhaps research ethics and concept of informed consent was quite different from how it is today. So I just want your advice on how you reconcile the question about extractive nature of this kind of research and how we could answer research ethics committee that this kind of research is justified, number one. And number two, in longitudinal studies how do you avoid Heisenberg effect that the very fact that you are doing these studies information gets out and maybe some people who moved to the village or because they want to be included in the studies not knowing what these studies are about. So how do you overcome Heisenberg effect? Thank you. Thank you. One more question. Yes, please. Thank you very much to all the presenters. Guillermo Cruz is from University of La Plata in Argentina. I have a question for Santiago but I welcome all the presenters to the convene. Thank you for the thoughts on this. But I wanted to know to what extent do you think welfare reform, let me call it in those grand terms or integrating more these programs, these now not so new programs into the welfare system, some of them old in the region and this perhaps goes beyond the region and other developing regions or integrating these programs into the old or the new systems. How much do you think that is a first order effect on productivity or more of a second order or something good that we want to achieve but that might not affect productivity and the labor market as much? I want to know your thoughts about that, please. Thank you. So I'll give a chance to the panelists to respond to the first set of questions and I already noted a few future questions. Martin, you want to? Thank you for the questions. I'll just address Rolf's question. This is something I'm still trying to figure out and ongoing research, I just sort of figured out how to do the measure. I figured out how to do the measurement in a reasonably satisfactory way and now I'm trying to explain the differences that I see in the evolution of the floor in different countries and the cross-sectional differences as well. My working hypothesis at this stage which is consistent with everything I know is that a huge role has been played by social protection policy and social safety net policy broadly defined including, of course, minimum wage legislation which began in New Zealand in 1906 but there are a number of things happening in the late 19th century, early 20th century in that period up to including just after the First World War in the United States across Western Europe as well which I think were key in raising the floor. We see an effort now in the developing world to do just that. We're seeing a rapid expansion in social safety nets in developing countries. 20 years ago you could count on one hand the number of countries that had anything you could reasonably call a social safety net now it's getting quite common. It doesn't mean that they're very effective, however and I've been looking into that and bottom line is the coverage of the poorest is abysmally low, it's just not reaching them. The reason isn't so much poor targeting in the conventional sense, it's really poor coverage of these programs. So yeah, that's my hypothesis at this stage but again it's something for further work. Thanks very much for the questions. I guess the Palampur study as you noted goes back a long time and I'm not sure, oh sorry, and I'm not sure to what extent there was extensive inquiries and there's a lot of investigation of these whole issues of consent at the time of the AERC going into the village in the 1950s and so on. Over the years we have had, it's been an ongoing process, I don't remember of us having any particular questions raised about this issue with respect to the recent rounds of data that we've been collecting. I think possibly one factor has been that we've not been investigating or trying to look at a particular intervention in any kind of way, it's just been more like an anthropological study in the sense that we just go to the village and we sit and we observe and we ask questions about what people are doing but we're not trying to follow up on any particular interventions of any kind and that may well be one of the reasons that there has been less querying around the Palampur study along these lines. I should leave it, my colleague Himanshu is here and who is in charge of the data collection in the more recent rounds, whether there was any specific questions raised say by JNU at the time of us doing the more recent data collection. I'm not sure of that but perhaps we could follow up with him after the session. On the Heisenberg effect, I mean this is something that we do obviously, this is a question that's been raised periodically with respect to the study, it's clear that when you go into a village and it's not a big village and you spend a year in the village pestering the villagers with all kinds of questions about why they're doing this and what they're doing and are they sure that it was really 50 kilos that they harvested as opposed to 47 or something like that, that you might affect behavior, you might affect the way the respondents might give answers to questions and so on and I don't think that we can claim that it did not happen, it's very difficult for us to make that kind of a claim. Now on the other hand, we do want to note that between these rounds of survey data collection there were often years, sometimes decades of time so that it wasn't as though this was a continuous process all throughout the seven decades of field work and it also became the case, it also seems to be the case that as you're collecting these data, at first the villagers notice you and are somewhat curious and ask you questions about why you're there but very quickly you become part of the landscape and they just get on with their lives, there's really a very deliberate effort made largely to stay uninvolved in any of the villagers' activities and so on to sort of remain outside and as villagers became aware of that they also just sort of at certain point they said, okay well we'll just go on with our life and we didn't see, I don't think we have any examples of really egregious changes in the behavior of the villagers as a result of the field work being undertaken but obviously it's certainly not impossible that there were aspects of the villagers' behavior that were affected and I don't know how you could get around that doing this type of study. So thank you for the questions. On the question by Rolf. So I do think that the institutions here are really critical and are really really important the way you think about policy. Broadly in my view I think that attempts to redistribution through interventions in the labor market in Latin America are most likely gonna fail than work. This varies very much from country to country of course but a key empirical fact that needs to be paid attention to is that for the region as a whole more than 50% of the labor force is in the informal sector. For some countries, you know Bolivia is 80%, for Peru it's 61%, even for a country like Mexico 60% of labor force is in the informal sector. It's in fact few countries mostly in the Southern Cone, Argentina, Chile, Uruguay, perhaps a little bit Brazil that have less informality. When you think about that you've gotta think about the fact that this is endogenous to the incentive structure. So if you then change the incentive structure in the labor market, there will be a response by firms and workers. So most likely because there's imperfect enforcement of these regulations, you're not gonna really reach the sort of people at the very, very bottom that you thought you were gonna reach. So from the point of view of poverty if you really, really want to reach the very, very bottom in most countries in Latin America you don't think about the minimum wage. You have to think about a minimum income and some other mechanism to redistribute that avoids the labor market, that's kind of rough. On Guillermo's question, what I think has happened is something at the same time good but that needs to be sort of changed. I think that the countries of the region have constructed what I sometimes call the tropical welfare state, which is a patchwork of contributory and non-contributory programs that have very conflicting set of incentives for workers and firms. This has been good from the point of view of many more people now have an income transfer at old age than what they did before. Many more people now have access to health insurance than what they did before or to disability insurance or to protection against risk. And this is definitely welcome. On the other side, we'd have to think that there's firms and workers react to all these incentives and that the choice of self-employment and being employee, the choice of firm size, the choice of contractual structure of the firm is endogenous to all these incentives and this hurts productivity and it goes sort of back to my original point. In my own work, and I've only done numbers for Mexico, what I find is that the productivity cost of all these interventions can be high. Controlling for firm size in formal firms are 35 to 50% less productive than formal firms. These are large numbers. And the size distribution of firms is very skewed. So I'm not arguing that only the labor market interventions associated with quote unquote the tropical welfare state is behind this. There are many other issues having to do with taxation. There are other issues having to do with credit. But clearly among the incentive structure, the firms and workers are looking at are the incentives that are being from the social insurance mechanism and from the labor regulations, severance pay and all that, minimum wages in some cases. And this, in my view, needs deep reform if it's going to be productivity conducing instead of being an obstacle to productivity growth. Thank you very much. Just for the sake of maximizing time, could I ask the individuals asking questions to be as brief as possible and also to the extent possible, the replies to be brief. I already have two people on my list, Florence and Alain and this one. So one more maybe, Rob. Okay, so Francis, would you please? Oh, thank you. I thought there were three absolutely fascinating and enlightening presentations. But for each of them, I wonder if they'd considered also using a non-monetary measure of poverty, multidimensional poverty, something which captures, in a way, the real aspects of poverty. Because if you think of each of them, if you think of Martin's presentation, it's capability poverty that Amartya Sen is talking about and Martin is trying to capture. If you think of Mexico, there's this sort of irony or paradox that poverty seems to have declined and yet real human capital has not improved or not so much, so a multidimensional poverty measure would automatically capture that, changes in nutrition and so on and so forth. And similarly for Palanpour, the well-being measure is one thing, but we would like to know what had happened to nutrition, what had happened to real deprivations. So I just suggest that along with your monetary measures, you should add this. Thank you, Alain. Thank you. The process of what's happening in Palanpour is very interesting in the sense that land appears to be a curse and the mobility is more associated with non-farm opportunities. And is that because there is fundamentally a land market failure, land fails to concentrate, land fails to create opportunities for those who could develop better incomes via access to the land? Is it due to lack of migration and to then land not being available or are there ways that are restricting land transactions that create this land curse in a sense? Then in San Diego, there are industries in Mexico that have been spectacular gains in total factor productivity, especially in the car industry around Monterrey. And yet real wages have fallen, real wages have fallen. The work of Harley-Shacon, for example. And so is that the labor market failure or is that the drag of the labor market failure overall, which in a sense is preventing total factor productivity gains from being reflected in rising wages? And if there is this drag, then is there a way of delinking, in a sense for workers in those particular sectors, the total factor productivity of linking the total factor productivity gains to the wage via labor, via representation, via bargaining, things which are to a significant extent missing today. This is Supermanian. This is for you, Martin. I found your presentation very interesting and perhaps this is a little self-serving because it has commonalities with some of the work which I have been doing recently. It's interesting to note that the head-count ratio comprehends the count conception of poverty but ignores the floor, while the measure which you propose comprehends the floor conception but tends to ignore the count part of the picture. And there might be something to be said for combining both of these in a single measure and one possibility which I would like you to consider and which I have advanced in some of the recent work that I've done and I'm sure this is not the only way of doing it. It's the following one. It's useful to recall that your measure which I read with great interest is actually the average income below the upper bound that you stipulate, less an expression for inequality in the distribution of incomes below that upper limit. Now in exactly the same spirit, one could think of fixing effectively the head-count ratio at say some, the poorest income, poorest X percent of the population. And to consider, let us say, an indicator which is simply the average income of the poorest 20% adjusted for inequality. And one possible merit of this, although you might see it as a demerit, is that one dispenses altogether with a need for a poverty line, which itself as you know, has been a subject of some controversy to put it mildly in the discussion. Thank you. Thank you, Rob. Post FAO, I liked all the presentations, but particularly interested in Santiago's presentation on the effectiveness of the social transfers. I think what the bottom line, what you're trying to say is that the social protection measure, the cash transfer programs, only become truly effective if they're well-embedded with other policies. And I think not for nothing in Mexico, they have now changed the program's name again from Progresa to Prospera to indicate the linkages with productive interventions. I think what would be interesting to look in the house that works in the Latin American context, but at least at FAO, the research we've done in African context in how social protection is helping poor farmers and rural livelihoods, that there may be three key mechanisms that previously may have been underestimated in the research. The first are local economy effects. If you give poor people a bit more income, they spend it and typically spend it locally, so you get a multiplier effect, and we find it's between, for each dollar spent on the cash transfer, you get a dollar and a half to two and a half dollars back in local economy return. But it works better if other things fall into place, if you have infrastructure, so there's market connections and so on. The second is the impact on assets. Are people able to buy agricultural tools or other things, or is the social protection enough to prevent them from selling off a cow or a plow which otherwise would have prevented their next cycle? And the final, maybe what we find is the most important point, particularly for those that we don't want to leave behind at the bottom end, is that the social protection seems to be giving them back the dignity and the entrance into social networks, the community participation and so on. And that has a very strong productivity effect. We're interested to see if Latin America contracts a bit different, but whether that could also be explored further. Thank you very much. We have 10 minutes and four questions on the floor which have to be replied to. And if we want to have another round, let's try to again be extremely brief. I'd like to have one more round before Finn essentially closes the session. Okay. Very briefly, both in response to Francis Stewart and to Professor Supermanian, I'm all in favor of a kind of, I like the expression dashboard approach. I'd really not in favor of combining indicators. I typically don't find this very useful. So in this case, in this case, sorry. It can't move anymore. Move, move, move, move, move, move, move. Technical problem. And a minor problem with the existence of this table. But okay, I'm not in tall in favor of, okay, put it this way. I think poverty is multi-dimensional, but I don't think you can combine it into one dimension. I don't think you can, it makes any sense at all to be combining these things in some arbitrary ad hoc way. I don't find that useful. I don't find it useful, policy. I don't find it useful, monitoring. So in response to both, I don't think that's the right way to go. I think we need, well, what I'm saying is that existing approaches in an important way do not capture everything we want to know and we need supplementary measures. And finally, I think in response to Francis, one thing is to say you're measuring poverty in terms of incomes and another is to use a money metric of welfare. And that's what I'm arguing. There's no problem with measuring poverty in terms of income as long as you're using money metrics of welfare. Welfare is a function of many things. And as I've argued, it's also a function of social effects, relative deprivation, shame, social exclusion. These things are aspects of people's welfare. What one is struggling for is a money metric of those things. Be careful about the critiques of the income-based poverty measurement because sometimes they miss that point. Quick answer to, the issue of land and agriculture is very interesting. I mean, what I showed you very quickly in one of the graphs was this is the very important influence that population growth has had on the village so that per capita land holdings in the village have declined by a factor of five since 1950s to the present. There isn't space for enlarging the village land. There's just a growing population and so per capita land holdings have really declined precipitously. Wheat yields have continued to rise. So it's not as though it's just a backwater area. There's nothing happening. It's quite dynamic. There is a lot of agriculture going on and it's quite intensive and there's three crops per year that are taking place. There's all kinds of new seeds and new crops being experimented with and so on. So it's quite dynamic from that point of view. The tendency market exists as an institution and has existed throughout the whole survey period as a method or mechanism whereby land can go from those who have an abundance of land to those who would like to operate more land than they have. And it has continued to function. It's also been evolving in very interesting ways and we do have separate research that's describing how that tendency market has been evolving. And then finally, I would point that one thing that has been noticeable, one indirect effect from this non-farm sector has been the very, very dramatic rise in agricultural wages which have made cultivation much more costly and exercise. And so farmers complain bitterly and it might be partly also the Enrega might have also had an influence here, but farmers complain bitterly about the lack of agricultural laborers at times when they really need labor assistance. So that might be also be a factor behind some of the struggles that the farmers are having at the moment. With respect to the non-income dimensions, one of the, I was trying to emphasize an ability to track changes over a long period of time and so I was focusing on our income measure which we have attempted to try to produce in a way that is comparable or as comparable as possible over the various survey years. We have of course been collecting all kinds of other data on nutrition, on educational outcomes, on intra-household distribution, on gender relations, on political activism or organization in the village, and these are all being described in the research that we're doing. They don't all necessarily lend themselves as well to these comparisons over time as our income-based measure does and so that's what I've been focusing on in my remarks here today. Quickly on Francis' question. I'm sort of on Martin's side of this. I think there should be various metrics that we look at at the same time and try to pull together a picture from the various metrics, but I think trying to aggregate all these metrics into a single indicator has a lot of conceptual problems so I think it's kind of this dashboard approach. I feel comfortable with that. On Alain's question, in a way Alain, the bet 20 years ago was that in fact, all these reforms were gonna be sufficient such that the large firms and the formal firms would be able to over time take a bigger share of employment, a bigger share of value added and a bigger share of the economy, so to speak. The bet didn't work. 20 years later, I don't know what it'll work in the future but certainly for the last two decades, that bet didn't work. What has occurred is that you have a subset of firms that are fairly small that show very high productivity growth but the large number of firms, 70, 80% of all firms have stagnant productivity growth and the weighted average of that translates into basically stagnating productivity growth. So in fact, I think we have to rethink and think whether in fact this notion that all these small informal firms will fade by themselves automatically as a result of growth is actually not the opposite unless you do something about these firms and what is incentivizing their existence and their continuous repetition, growth will not be able to pick up and that's perhaps the way to think about it. Lastly on Rob's question, I think some of these effects that you're speaking about about the multiplier effect in local regional economies there are present in the Latin American case except that as opposed to Africa, you're talking about a much more urbanized region. So they dissipate a lot more because a lot of the poverty is in the urban poverty. So that you really need to think more systematically about how to participate, not in a small region labor market but sort of in more larger regional or in some cases, depending on the country, look at national labor markets and what is their pattern of participation there? Thank you very much. I can now accept two more final questions and they have to be very, very brief, yes please. Thank you very much, Marty Chen. It's a panel dear to my heart because I work on the poor. I grew up near Palampur and I believe in labor market interventions but Santiago, I want to take up something we've discussed in the past which is I think we have policy schizophrenia. This is Robby's term regarding the informal economy. So there may be policies that incentivize firms to be informal or small but there are also policies that make it very difficult for them to become productive and that penalize them, criminalize them. So I think we have to look at the whole spectrum of the policies and that's my answer to your question about systemic understanding. Thank you. Chico will have the last word. Thank you. Chico Ferreira from the World Bank. Thank you for three excellent presentations. A very quick, narrow question for Martin and a quick, broad question for Santiago. For Martin, I really like this kind of Rawlsian floor and then an upper and a lower bound for the counting approach. Is there any way you can help us get rid of what I've begun to think of as the tyranny of the PPPs? The purchasing power parity exchange rates which I used in this lower bound that you would still I guess under your proposal use. You know, as you know better than me but I've begun to learn recently are every time they get re-estimated they change dramatically and they use ring countries one time and then they use a core set of prices the other one year they survey rural areas the other they don't and so it's a mess. Yes, I'll be very quick. So, you know, is there a way we can get rid of the tyranny of the PPPs in some way? And for Santiago, in this conference, you know, we've had occasion I feel there's a bit of cognitive dissonance between across the regions. So we hear Justin Lin and others talking about flying geese and industrial policies and other sort of things and I always find myself sitting near a Latin American and kind of shake our heads in disbelief at this. Are we the last converts of the Washington Consensus? You know, why are we so skeptical of these things? What do you think of the role of the state to increase productivity? Thanks. Great, thanks, Chico. I don't see any way of, unless you go to strongly relative poverty measures, which I definitely don't think we should do, the kind of measures that are used in Western Europe, half of the mean or half of the median, unless you go that route, you will need to make price level comparisons across countries so there's no way of avoiding it. But there are way better, one of the fundamental problems with the PPPs is that they've never been designed for the purpose of poverty measurement. They were designed for national accounting purposes and I've been complaining about this for 30 years and it still hasn't changed. The big thing that has to change, in my view, is that the ICP, National Comparison Program, has to make its price data public. It has to be the micro price data, not just the basic heading price data. Once those data are in the public domain, we'll start to see a flourishing and I think initially an instability but a final stability on something that's more appropriate for poverty measurement. So, very quickly. Yeah, absolutely, Marty. I mean, you have to look at both sides and I always think about, there's a very large tax on formality and then there's kind of a subsidy in formality and surely, I fully agree with you, many of these firms are taxed or extortion. There are many, many barriers for these firms to go from the informal to, and that's part of the incentive structure that they're facing and that's part of the outcome. So, both sides, I don't really see much paranoia. I think they both add in the same direction so in agreement with you on that. Chico's question is extremely, extremely difficult, extremely, as I was listening again yesterday, not yesterday, before yesterday, to Justin Lin. So, let me make a broad statement. If you think about historically the nature of the Latin American states and the nature of the Asian states, this is a major generalization but I think the key difference is perhaps the Asian states had a central growth mandate and the Latin American states had a central redistributive mandate and in a way, it generates very different structures of the state, very different mechanisms of intervention and very different views as to what is priority and what is not priority. What I think broadly Asia has been able to do is to really focus the instruments of state intervention on growth and in Latin America, relatively, we have focused a lot the instruments of state intervention on redistribution and we have yet to find the mechanism to make that compatible with growth. So, I'm not either for market state, I'm sort of agnostic, pragmatic on all that, it's more like how to... Well, we've essentially completed, I think, an extremely full session and I now want to thank the panelists for, I think, extremely interesting presentations and I also would like to thank the participants, the audience and the patients and it probably is the last big conference that I will ever attend, so I will keep very good memories of this event for future references. So, let me now turn the mic to Fentarp, who I understand has some final words for us. We should still applause. Dear friends and dear colleagues, we have once again gotten off to a good start to the day. Please excuse me for taking the floor now, but this is actually the last plenary session of the wider 30th anniversary, so this is the only possibility I have to address you as conference participants. Martin, Peter and Santiago, you exemplify in the finest way possible the nexus between research and policy that we wanted to bring out in this conference. Moreover, I would like to take the opportunity to thank you, the chair, dear friend and mentor, Erik Thorbeder. You have been an inspiration for many, many years. Erik, you also took a lead in thanking the wider staff. Now, many of them were not here when you expressed your thanks. I therefore asked whether they could join us now because I would like to convey to you that it's not possible to bring 600 plus people from 75 countries together for three days in Helsinki without a very dedicated staff to plan, organize and implement such an event. And the wider staff, I don't know, could I ask you to stand up? The last 12 months have been tough. You've been fantastic. Now, let me also thank you, the wider global network, for coming to Helsinki on this occasion. Chairs, presenters, respondents, each and every one of you have contributed to a wonderful celebration of development economics and wider's first 30 years. A special thanks to those of you who took the time to interact with our younger colleagues. We must not, not even when celebrating three decades of accomplishment, lose sight of the fact that the development job is very far from done and our younger colleagues are the future of the development economics profession. They deserve all of the attention of those of us who have lost a bit of hair. I do as well on this special occasion wish to express a particular word of thanks to Professor Ernest Aiiti, chair of the wider board and vice chancellor of the University of Ghana. Ernest, you've been a constant source of inspiration to the wider senior management, always providing admirable judgment and kind, quiet advice. For this, thank you so much. My sincere thanks also goes to Ernest's excellent colleagues on the wider board, Professor Ravi Kanbua, Dr. Kristina Kuvaya and Dr. Su Ling, who are here with us today and to Professors Elizabeth Satterley and Anne Case who could not join us in this occasion due to their teaching obligations working with the young and new generation. I'm very grateful that you have all of that experience and that you have been willing to share that with Tony and I. Now, to be sure, there's much more to come today, not least the wider annual lecture this afternoon, but we'll be joined by yet more guests, including Pristin Marte Artisari. Now, he was one of the initiators of wider three decades ago, together with this year's wider annual lecture, Professor Amartya Sen. Now, I'm sure you would want to ask questions this afternoon. Could I ask you to prepare yourselves? Turning to the wider events next year, I can already now say that wider will hold two 2016 development conferences. The first will be organized by a team of younger wider researchers, whom you will see on stage this afternoon. Put more on wider's conferences next year on the new wider website and I began this conference by requesting that you do take a look at the soft version and give us feedback, so we make sure that the new wider website will be a really interactive and great tool. Now, on the wider website, you will also find more on the 2014 wider work program on the triple challenges of transformation, inclusion, and sustainability, addressing three core concerns, Africa, gender, and development finance. Now, I was sitting in this session, we've just been in, thinking about the fact that while we have been sitting here, there are actually 16 countries studies of the labor markets in Latin America that have been going up on the wider website in one of wider's projects. And on the day before the conference, we had a researcher meeting on six countries studies in Africa focused on the interlinkages of the labor market and poverty. And I made the point in my opening remarks that we in the process of launching a social assistance politics and institutional database called SAPI. I was interviewed for my present job at wider a few years ago and one of the people who interviewed me asked me, so is it now going to be all about foreign aid? And yes, we have indeed seen sessions on foreign aid at this conference, but I hope that this conference has demonstrated that at wider, we do keep the wider perspective, the many different perspectives on development in mind and that we do try to bring people together, to bring the network together to address all of these many challenges. And you will on the wider website find all of the information on no less than 15 ongoing projects on a variety of challenges under the headings I've just referred to. Finally, I wish each and every one of you safe travel home, after the dinner reception tonight, of course, and we should tonight be ready to celebrate wider at 30. Thank you, enjoy your coffee.