 I was joking before I got up here that I was told I have to wait for 10 seconds before I start saying something. So I think those 10 seconds have gone by now. So Excellencies, distinguished colleagues, friends of the UN, ladies and gentlemen, it's a really great pleasure for me and also a great privilege on behalf of you and you wider to welcome you warmly to the 2017 wider annual lecture. And a special welcome to today's lecturer, Professor Sabina Alkira. Each year, UNU-Wider hosts the wider annual lecture delivered by an imminent scholar or policymaker who has made a significant contribution to the field of development and social sciences. We are most grateful that Professor Alkira accepted our invitations to be the 2017 annual lecture. Sabina is the director of the Oxford Poverty and Human Development Initiative, OPA HI, and in her lecture today, she will discuss the implications of using the Global Multidimensional Poverty Index and other poverty measures for achieving the United Nations 2030 Agenda for Sustainable Development. Focus will be on Sustainable Development Goal number one, ending poverty in all its forms everywhere. In other words, Professor Alkira will be lecturing on a top international development priority and on how we go about measuring progress towards its achievement. The New York State Poverty and Human Development Initiative, which Professor Alkira heads is an internationally recognized research center within the Department of International Development at the University of Oxford. Here Professor Alkira has been working consistently at the forefront of the theory and measurement of poverty and well-being in a way that goes beyond the traditional focus on income and growth only. Another colleague, James Foster, who is both a OPA HI Research Associate and Professor of Economics at George Washington University, Sabina devised the Global Multidimensional Poverty Index as a new and inspiring method for measuring multidimensional poverty. And the approach has, to give just three illustrations, been adopted by the Mexican government, the Bhutanese government in their Gross National Happiness Index, and by the UNDP. Over the years, Professor Alkira has been called upon to provide input and advice to many initiatives seeking to take a broad approach to well-being. Examples include the Commission on Measurement of Economic Performance on Social Progress, instigated by President Sarkozy, the UN Human Development Report Office, the European Commission, and the UK's Department for International Development. Sabina was also involved in 2016 in drafting the Stockholm Statement on the Set of Principles for Economic Policymaking in the Contemporary World. You will find this statement at the wider website translated into about 25 languages. Professor Alkira's research interests include multidimensional poverty measurement and analysis, welfare economics, a Marche sense capability approach, and the measurement of freedoms and human development. So, Professor Alkira is a researcher that does not shy away from the hard and difficult questions and issues that we are facing, both in our profession and more generally. With these introductory words about today's speaker and the theme, I wish to conclude by welcoming, my welcoming remarks by noting that I'm delighted that the wider annual lecture 21 will be delivered today on 24 October as part of the UN Day celebrations organized jointly by UNEWider and the UN Association of Finland. Seven UN entities and related organizations, the UN Association of Finland, UNDP, IOM, Finland Committee for UN Women, the Finnish National Committee for UNICEF, Pakulai Appu, and Finland Partnership have all joined the celebrations by exhibiting their work outside in the foyer. I think this is one of the best ways of going about celebrating UN Day. Now to today's speaker, Sabina, a warm welcome to you. The floor is yours. We look forward to your lecture. Excellencies, distinguished guests, friends. It is very much an honor to be here and I'm grateful to Wider for this opportunity to fin for his kind words. And Wider also plays a very special role in my mind when the time came to found OFI, our research center. This was the place I came for advice to try to understand in a sense how research centers can continue to have a broad intellectual horizon, can continue to do research to high standard. And yet also engage with the policy world and with those most directly engaged in the issues of development. And I've been very grateful to learn, although not perhaps very well to emulate some of these features of Wider I admire so much. There's a lovely line in a ditty by Edgar Leslie that goes, Taint no sin to take off your skin and dance about in your bones. Now, obviously this is stuff and nonsense. We can hardly throw off our skin and muscle and fat and hair and dance around as grinning skeletons, as lovely and invigorating as it might be to do so. But just the thought of doing it, of being able to dance as naked bones, uninhibited and unencumbered, feels as if it's just a little bit of a joke to dance as naked bones, uninhibited and unencumbered, feels intriguing, liberating and somehow clarifying. I mentioned that because as unlikely as it may see, I hope we can take some of that freewheeling, creative and curious sensation into the next hour when we will talk about a rather serious subject, which is how people are acutely poor. Ending acute poverty is of ethical importance and human urgency. It is a collective priority within and across nations and most of all for the protagonists of poverty. And it is so because despite fantastic gains, far too many continue to go about their lives in abject, dangerous and harsh conditions, unable to shape lives with meaning and dignity. But in the next hour, I'd like to step back from the many complexities and isolate one aspect only, which is poverty measurement. The question is whether we are able to use multi-dimensional poverty measures to expose a relevant, if limited, skeletal structure of poverty in such a way that makes the task of addressing poverty somewhat clearer for all concerned. My own thought on this is not fixed. Our commitment, all of us, must be on the desired state of no poverty and not on any particular skeletal arrangement. But for the next hour, I'll present a structure of poverty measurement that may, if well executed, be of some use. And I do so knowing that this work is not my own. The methodology was developed with James Foster, as Finn said, a leading, precise, and powerful mind across the field of poverty and inequality measurement. And the applications were all undertaken with students and coworkers in our research center, OFI, whose brilliance and care and steely determination are wonderful to work with. But let us begin, as is natural for us here at Wider, with Amartya Sen's capability approach. And I'd actually like to start with a paragraph from the 1990 book edited with Martha Nussbaum, in which Sen had an article called Capability and Well-Being. He wrote, turning to poverty analysis, identifying a minimum combination of basic capabilities can be a good way of setting up the problem of diagnosing and measuring poverty. It can lead to results quite different from those obtained by concentrating on inadequacy of income as the criterion of identifying the poor. The conversion of income into basic capabilities can go with varying levels of minimally adequate income. The income center view on poverty, he argues, may be misleading in some respects. That paragraph now seems prescient. For its claim as to the persistent relevance of paying attention to non-monetary dimensions of poverty, stands empirically validated by extensive micro-dated analysis, which was not available in 1990, by participatory insights and voices of the poor and other such studies, longitudinal and ethnographic studies, social movements, and so on. This conceptual articulation by Amartya Sen recognized implicitly the importance of numbers and measures of poverty, even crude measures, such as human development index, as a guide to action. And more recently, of course, the statistics have gained public prominence in the Stensiglitz-Veutussi Commission and its successors in the data revolution and in many, many different fora on statistical capacity building and strengthening statistics. Miguel's Tsecoli's words well encapsulate the motivation for improving statistics and measurement because of their link to action. A number, he writes in Spanish, can awaken consciences. It can mobilize the reluctant. It can ignite action, generate debate. It can even, in the best of circumstances, end a pressing problem. And of course, that is our topic today. But until recently, as we know, I think, Sen's insights and other work on multidimensional poverty was, in a sense, marginal. With the advent of the Sustainable Development Goals, multidimensional poverty appears to be graduating from the margins to take a more central place. But it's not yet certain how that space will evolve. But clearly, the development now is framed multidimensionally with the SDGs being an integrated and indivisible balance of three dimensions, economic, social, and environmental. And turning to poverty, the second sentence of the pivotal document of the SDGs transforming our world, the 2030 Agenda for Sustainable Development, reads, we recognize that eradicating poverty in all its forms and dimensions is the greatest global challenge and an indispensable requirement for sustainable development. Here, poverty is definitively recognized to have multiple forms and dimensions. And those include extreme poverty, the measure of $1.90 a day, as one particularly high-profile component of a wider concept of poverty. And the shift of emphasis is sustained throughout that document with the phrase, many forms of poverty recurring seven times. And the goal one has a number of targets of which just the first target focuses on income poverty. Complimenting it, target 1.2 is to reduce at least by half the proportion of men, women, and children of all ages living in poverty in all its dimensions according to national definitions. So the discussions leading up to the SDGs also highlighted the importance of multidimensional measures, multidimensional understandings. So for example, in the UN Secretary General's report in December 2014, trying to set a positive and ambitious tone for the SDGs, he argued that poverty measures should reflect the multidimensional nature of poverty. And in the same December 2014, the UN General Assembly spoke of the need to develop complementary measurements that better reflect multidimensionality of poverty and of well-being. And turning to financing Addis Ababa Accord in 2015 also called on the United Nations and international financial institutions to develop measures of progress that recognize this multidimensionality. So the space is there. The concept from Amartya Sena is there, the recognition that numbers guide action. But what in a sense and how should that space for better measures be filled? So we might turn naturally to the indicators of the Sustainable Development Goals, and these were approved in July of this year, the list of 232 distinctive indicators that were proposed by the UN Statistics Commission. And their proposal clearly echoed what has become a pivotal phrase within the SDGs of the need to leave no one behind, to have measures that can be disaggregated, to break down the silos and have policies that are integrated, multi-sectoral and indivisible, and to recognize the priority of eradicating poverty in all its forms and dimensions. But the list of indicators of the SDGs leaves the precise definition of multidimensional poverty measures open. And we'll see that later. So at this point I just wanted to pause and say that we are here celebrating UN Day in Finland, but there are also other agendas that intersect and like the SDG agenda have created a space for wider measures of poverty. But these agendas might be regional or they might be national. And so if we think of the 2063 Africa agenda, we could also find space. And so there seems to be a turning of the tide and an interest in better measures and structures of poverty that also pay attention to the non-monetary aspects. So what would multidimensional measures do? Clearly they would look across different dimensions, perhaps including perhaps supplementing monetary measures. And they would also somehow reflect the priority that we see of poverty reduction. So choosing different SDG indicators that reflect poverty and so merit the distinctive priority given to its alleviation. There's a big focus also on recognizing the interlinkages across indicators and a multidimensional measure which draws on different indicators for the same household at the same time organically does this by its very structure. And clearly a measure would need to be disaggregated in order to assess over time if groups were being left behind. But I think beyond all of these more measurement-focused desiderata, there's also a very strong need for a measure to be able to engage very practically with policies and will come a little bit later to those, policies of resource allocation, targeting, policy design, coordination, monitoring, evaluation. So if we think of what measures should fill this space, there is a very obvious document to turn to. And that is the report monitoring global poverty led by the commission established by the World Bank and chaired by our late and beloved Sir Tony Atkinson. The opening lines of his report recognize the controversiality of monitoring global poverty, however it is defined. He recognizes that some people will consider the exercise to be futile. It is simply too demanding to find measures of poverty that are relevant globally or that can be measured with adequate precision for the task. However states that this is not a view he shares and not one that underlies the report that was released. And so the aim of the report was to say how can global poverty be measured? And clearly the focus, therefore, is not on precisely identifying the number of people, the level of poverty, but of accurately measuring its trend over time. And the second gloss is that that trend must be measured not just as a precise point estimate, but with standard errors, with a recognition not only of sampling errors, but of also the many other kinds of non-sampling errors that affect the data used to make these estimations. And so it's important to be able to both measure and convey publicly the margins of error. So the commission was concerned only with monitoring the extent of global poverty and its reduction, and it had three parts, monitoring extreme poverty, which was the monetary sections, and then beyond one goal 1.1, complementary indicators and multi-dimensionality, and making it happen. And so the resource document, if you don't know it, that I would warmly commend to you is the second part of this report, which is looking at non-monetary indicators that might use to monitor global poverty and what characteristics they should have by what principles they should be designed and precisely what dimensions and methods should be used to develop them. So, Sir Tony Atkinson, of course, was one of the progenitors of discussions about social indicators in Europe and recognized the need for a broad engagement in terms of the process of indicator design and the transparency by which the indicators that finally are launched were justified. And so in developing complementary measures to a $1.90 a day global poverty measure, he advised a number of different principles. One was that a non-monetary measure or set of measures should be truly global, covering not only developing countries, but also high-income countries or vice versa. He specified later how that could be done, perhaps in ways that accord with the aspirations of different country blocks by using the same indicator or dimension, but with different cutoffs. Also, the indicator itself had to be transparent. It had to identify the essence of a problem so we could understand if increases in the indicator were good or bad. And the definition had to be generally valid and have a clear normative interpretation about the changes over time and also robust and statistically validated so that there is a literature about its definition and construction. And also the global indicators needed to somehow be in accord with national data and national measures. And so there needs to be, and we'll say more about this, a conversation between the global and the national measures, which is perhaps more active than has been in the past. Furthermore, if a dashboard of different indicators is to be used, the dashboard should somehow be balanced so that the component indicators, roughly equal in importance, would span the issues of interest. And finally, in order not to tax the data producers too much, he recommended, insofar as possible, to design complementary measures that use existing data or where they must be extended to do so using existing instruments or administrative data. So drawing on this process, there were two recommendations that I think must inform any efforts to think about the structure of multi-dimensional poverty measures. The first was recommendation 18 that the bank should establish its own requirements about measuring a small set, a parsimonious set of complementary indicators that would include an overlapping measure will come to next and make sure that these were fully represented in the activities of the international statistic system. So where the sustainable development goals have a great breadth with 232 indicators, this report also advocated parsimony to have some high visibility indicators that would get special attention. Indeed, he gave a draft, the report gives a draft proposal of what the dimensions should be of complementary indicators, nutrition, health status, education, housing conditions, access to work, and personal security, for example, from violence. So that is already a step forward, a step which draws on a huge amount of literature ranging from studies of the dimensions of poverty from voices of the poor to studies of the dimensions of life of the sans-diglitz and Fatouci report to many other data exercises across Europe about the appropriate social indicators. But the report continued to then think about what a dashboard with these six component indicators would be missing. And what it was missing was the joint distribution of deprivations, that is, being able to look across and see who is deprived in several dimensions at the same time. Do those with low levels of education also suffer from poor health? Do those who are at risk of violence also have a lack of access to work? And so this need for an understanding of the interrelationships, the interlinkages across dimensions motivated the recommendation to have a measure of overlapping deprivations. And for those in Europe, this will be very familiar, that if you think of the Venn diagram of three different deprivations, then you can see who is deprived only in one, the outer clear circle, who is deprived in two of the three indicators at the same time, and who is deprived in all three. And that reflects a counting approach where you're counting the number of deprivations that a person has. It's very simple, very intuitive, but it conveys information that a dashboard of six isolated indicators cannot convey. So the final recommendation of the report was that a multi-dimension poverty indicator based on the counting approach should also be developed and reported. The report specified it was not proposed that the indicator include monetary poverty dimension. There are a number of national multi-dimensional poverty measures that are official statistics will come to mind. There are a number of national multi-dimensional poverty measures that include the income dimension. And these are the model in a sense of this recommendation, and not, for example, Mexico, which does include income at 50% of the weight. Furthermore, in terms of methodology, the recommendation admisaged the counting approach being implemented in terms of the adjusted head count ratio, which is that extends the foster gear index into multi-dimensional space that we have had the privilege to develop and try to work on its use a little bit. And it has two components, the percentage of people who are poor, or the incidence of poverty, and the average share of deprivations, the average percentage of possible deprivations that poor people experience, which is the intensity of poverty. So it's a kind of a poverty gap measure, for those of you who know in multi-dimensional space, and the product of those two is the adjusted head count ratio. So now what I would like to do is spend about 10 minutes and give you a briefing on a implementation of the adjusted head count ratio which is inadequate, incomplete, does not have all of the dimensions that were recommended in this report, but shows what is possible using a consistent set of indicators for over 100 developing countries. Very briefly, the Global MPI was first developed, co-developed with the UNDP Human Development Report Office for a launch in 2010 at the 10-year anniversary of the Human Development Reports, and it therefore sought to measure directly acute multi-dimensional poverty for over 100 developing countries. And in doing so it was clearly data constrained. But there were 10 indicators for which sufficient data were available for its construction. First of all, you are deprived if any member of your household is malnourished. You're deprived if a child has died. If nobody has completed five years of schooling in the household. If a child is not attending school up to the age at which it would complete grade eight. If you do not have clean cooking fuel, improved sanitation, safe drinking water, electricity, if your floor is dirt, sand, or natural you're deprived. And if you do not own more than one of a small set of assets, radio, television, telephone, bicycle, and motorcycle you're deprived in assets. So from these indicators, remembering the Venn Diagram what one does is construct the deprivation profile of each person or each household to see what indicators they are deprived in at the same time. These indicators naturally were designed in the time of the Millennium Development Goals but they match or they very partially reflect SDGs one, two, three, four, six, seven, and eleven and we'll talk at the end of some changes that might be possible to better align them. They also are not perfect. In some countries not all of the assets are present. In some countries we only have nutritional data for children and not for women or not for men. In some cases we have only whether a woman said a child had died in a household and not whether the man also said so. So there are some as in any global measure there are some very strong differences between countries that are thoroughly documented but that do exist. So how is this information go into a national measure? I was told not to put a lot of equations on the slides so I'll try to do it very simply. Consider three-year-old Nahato from Uganda and the deprivations that she experiences. Her house has a dirt floor. Her house is made of poles and of mud but there is a solar lamp and so she has access to electricity and indeed to clean and sustainable electricity. She is one of ten children of her mother Namubi who is 38 years old and there are problems with school attendance because some of her older siblings have not been able to attend school though it's $2.75 US for four months of schooling. The global MPI leaves out work and of course the working conditions are vital but they were not available and the working hours are long but there is also food insecurity and undernourishment in the family. Any poverty measure also overlooks many riches and cultural or relational aspects of life so for example Nahato and our family will be considered MPI poor but there are also characteristics like being outgoing and self-confident that they enjoy and at night they dance together to the music from a radio shared between neighbors so clearly any poverty measure particularly a global one will overlook a number of aspects it will not have data for example on work or on violence or on access to health care but it does measure a subset of indicators that may be relevant to their lives. So having constructed in a sense the deprivation profile for Nahato and her family the last step is to identify who is poor and this is done by fixing a poverty cutoff in which a person by this measure is poor if they're deprived in one third or more of the indicators weighted and they're weighted equally across dimensions across indicators within dimensions Nahato is deprived in 50% so she is poor so the adjusted head count ratio then looks at the percentage of people who are poor because they are deprived in one third or more of the indicators weighted and it multiplies this by the average intensity so Nahato intensity was 50% and on average the intensity in Uganda is some number and multiplying two together gives the adjusted head count ratio you're adjusting by intensity it also gives a measure which has strong characteristics like the ability to be able to break down by indicator which is what it makes it which is a vital use for policy as we will see. So I've given this a very light touch but for those with any interest there is extensive documentation and in particular reflecting on the Atkinson Commission's requirements there are standard errors and confidence intervals statistical inference about the levels and the trends of poverty published. There are robustness tests both to the poverty cutoff to the weights to the deprivation cutoffs and there's clarification of the axiomatic properties that the measure respects and so what does this look like when it's implemented so in the 2017 Global MPI we used demographic and health surveys for 55 countries multiple indicator cluster for 38 and PAPFAM for 3 and national surveys and these are quite familiar so I'll just look at a couple results but really I'd like to focus again on the structure of multidimensional poverty measures and not on these findings but in 2017 we updated 25 countries each year we update the countries for which new data are available and the country coverage is such that we cover 5.4 billion people quite good coverage in sub-Saharan Africa with 96% of the population covered East Asian Pacific with 95, South Asia with 94 and low income countries it's 99% coverage also in lower middle income countries and 92% across all middle income countries population coverage so it's very much a measure focused on developing countries it's not a global measure does not include high income countries but within its population there's perhaps some interest so what does this kind of measure give you at in one period of time clearly you can see how many people are poor 1.45 billion across 103 countries you can see where they wake up a billion of them wake up in middle income countries 72% 48% in South Asia 36% in sub-Saharan Africa again keeping in mind the sustainable development goal focused on leaving no one behind it's necessary that the measure be disaggregated and for a direct measure like an MPI that's straight forward so for example within Afghanistan we can see that poverty ranges from 25% in Kabul to 95% in Urozgan or 94% in Nuristan so quite a range but again the policy traction comes not just at looking at the level of poverty but the composition of Myanmar we see in the Rakhine district where the tragic Rohingya genocide is underway we see that it is the poorest district of Myanmar and we also though can see the composition of poverty for the Rakhine and how it is higher but actually in this case not dissimilar to that of Myanmar nationally except for higher levels of malnutrition in particular the MPI you might think is useful when there's a large disparity of levels of poverty but it's interesting then to look in Chad where on average 87% of people are poor but there are two regions Lakh and Wadi Fira where 98, 99% of people are poor so what's the value added of having a multi-dimensional measure if you look at the composition and that's the beige and green stripes you see in the beige which is Lakh rates of child mortality and nutrition are much higher and you see in Wadi Fira that deprivations in water are signally difficult and so actually the policies needed to confront multi-dimensional poverty even in these very, very high poverty regions can be somewhat distinct and so trying to pull that out can add value so the global MPI is broken down for 988 sub-national regions but of course you want to look at other groups so one of them is disability and DHS now have incorporated disability status questions of the Washington group into their future surveys and so you can see in Uganda that 76% of people who live with a person in the disability with disability in their household are poor and 69% of those without and finally on the global MPI we disaggregate by children and we find that 48% of that 1.45 billion people who are multi-dimensionally poor are children and furthermore that children are over-represented even given their presence demographically in the population clearly most of them at 84% live in South Asia and Sub-Saharan Africa in India, Pakistan Nigeria, Ethiopia more than half of the poor children live and again we can see that compared with adults children are more deprived not just in school attendance or malnutrition but in absolutely every indicator of the global MPI so age disaggregation even when it is not precise and we might want an individual child poverty measure which we are doing in some places but even this kind of disaggregation can bring useful information out so that's a brief overview of the global MPI the question is is tracking such a structure imperfect as it is over time useful for the sustainable development goals target of cutting by half or ending poverty in all its dimensions so let me say a little bit about how the data are treated to look over time first of all they are very rigorously harmonized in every detail so this is Cote d'Ivoire and in 2005 the survey didn't have nutrition so we drop it in order to have a comparable portfolio of indicators in Sierra Leone they didn't have male malnutrition in 2008 the first period so we drop it to have rigorous computations in Central African Republic they didn't have a mobile phone in their asset basket back in 2000 so it is dropped so there's a very rigorous harmonization across time which is in the public tables we are in the process of a big study we will launch in about six months which will be then tracking trends in global MPI over time I will show you a few results and I know there's an interest here in wider in Sub-Saharan Africa so I will share with you the coverage of 35 Sub-Saharan African countries in 234 sub-national regions this is one of the papers that we've published it was published this year and across these countries we can see the speed of reduction with Ronda being the fastest followed by Ghana, Liberia and Comoros we can see the relative reduction even of countries that have low poverty with South Africa reducing its MPI the fastest relative to its starting level in 2008 we can also identify sub-nationally regions and this is a map that shows the regions that reduced MPI even faster than Ronda which was our fastest country their runaways, their really positive outliers and understanding what went right in those regions is interesting we can compare trends nationally with trends in $1.90 a day the $1.90 a day trend is green and the red is reduction in head count ratio it should have error bars and so we can see that the trends are distinctive for the two indicators and so there's a value in measuring the trends of each because they don't necessarily mirror each other and we can also look a little bit more within countries so for example Côte d'Ovoire there's a population growth in 18 of the 30 countries in Africa that significantly reduced poverty and that wiped out the increase in the number of poor people increased although the level of poverty decreased but the interesting thing for policy again is how that happened and the number of poor people increased but the interesting thing for policy again is how that happened and taking apart the indicators in Côte d'Ovoire we see that there were significant reductions in the number of children out of school in child mortality inadequate sanitation unsafe drinking water and a lack of assets and it was changes in those five indicators that drove the national results and we can also of course break it down sub-nationally to see for each component region so this is how the headline looks at trends and in a sense can track progress but then can also be a tool to see how that progress happened and perhaps to incentivize change in the next period in terms of leaving no one behind when there are disaggregated statistics then you can look at the poorest and see if it reduced poverty the fastest so this is a happy story in Côte d'Ovoire in the poorest region so on the horizontal axis the poorest are on the right hand side and it's a race and so it's reducing poverty the fastest and so it's catching up it's a positive story in terms of leaving no one behind and in eight of the countries we saw that happy story with the highest reduction in MPI occurring in the poorest sub-national region there's also a question of whether the trends nationally match the trends in national or in Côte d'Ovoire 90 a day poverty so in India for example we see a positive pattern with monetary poverty reduction these are the regions again with the poorest on the right hand side and with the same axes and the lowest dots reducing poverty the fastest and we see that on the left hand plot which is monetary poverty that the poorest regions reduced monetary poverty the fastest but on the right hand you see that they reduced multidimensional poverty the slowest so the biggest gains were in Tamil Nadu or Andhra Pradesh, Kerala other states of India that were not the poorest to begin with so this divergence of sub-national patterns is quite important and again calls for these indicators to be monitored separately and this finding was taken up and republished in the global monitoring report 2015 a little bit of the stories that could come out of changes over time with a incomplete multidimensional poverty measure but one that covers a good proportion of low income and lower middle income countries and of course we can also see patterns by indicators statistical significant reductions in each MPI indicator for example for nine countries or reduction in every sub-national region in terms of the SDG aim to cut by half poverty in all its dimensions there were two countries that had a 12 year period between surveys which is not advisable but in their case they both more than have the MPI incidents during that period and other countries have also done so across several periods so certainly it is possible within 15 years to have global MPI from very different starting points so that's the end of really presenting to you a question which is whether this skeletal structure of poverty measurement would be useful in tracking trends over time useful both to see how we are doing but much more importantly to give information to policy actors so that they might be able to catalyze and accelerate the reduction of poverty now for those of you who need a little bit of a tea break I'll do three minutes that will be of most interest to young people and students which is where you can get all of this on the web and the rest of you could step back if you need so all of what I've presented all of the tables are freely accessible as Excel files online not only do we report one poverty line but we report three poverty lines not only do we report poverty but we also have a destitution measure and alas 706 million people are destitute meaning that they have severe malnutrition or open defecation or not even one year of schooling we also have all of the indicator level detail and the desegregations that I've shown in the tables and things like standard errors or confidence intervals or retained sample if you're really a geek and the strictly harmonized data that we have is online and as I said a lot more of such data will be forthcoming soon if you're interested in a particular country you can download the country briefing and look at graphics that visually display the information and finally if you want to make your own infographics you can use the interactive data bank choose the information that you would like to put into your PowerPoint or your presentation and clip it out so if you want to look at Cote d'Ivoire sub-nationally and Cote d'Ivoire in the region of its neighbors and see how its poverty is related to Ghana, Burkina Faso, Guinea, Mali etc you can do that so that's a question and a tool and I'll come back to that question later but what I'd like to do first is ask whether information in the global MPI is already present in other indicators whether they're $1.90 a day whether it's an index like human development index or social progress index global peace index is there any value added to an index like the global MPI not considering its structure but also its empirical results so first we turn to the $1.90 a day and I very much see it as a necessary complement to the global MPI the global MPI uses the health surveys and multiple indicator cluster surveys that do not have consumption or income poverty so it is impossible to include them in it and there would also be questions as to whether the volatility of household level consumption data meant that their aggregate would be an accurate representation of their monetary poverty over the last year so I do see them as essentially complementary and of the 103 countries for which we have global MPI in this 2016 update we have $1.90 a day for 86 of those countries in 10 those measures come from the same year in 24 the $1.90 are more recent and in 52 the MPI data are more recent there are MPIs for some countries without $1.90 measures like Afghanistan or Algeria or Egypt or Myanmar, South Sudan or Yemen but there are also estimations for countries where we lack global MPIs including Chile, Costa Rica, Iran, Malaysia and Venezuela so in a sense the measures complement each other and there are also some differences in terms of their country coverage though the number of countries is by and large similar for both of the measures if we look at the message or the head count ratios and that's what we can compare this is a scatter chart with the total of $1.90 a day poverty on the bottom axis and MPI on the vertical axis another way of presenting it is this chart in which the height of the bar is the head count ratio of multidimensional poverty and the black dot is the head count ratio of $1.90 a day poverty for the same measures so what you see is that they agree on the low income countries low poverty countries but there's quite a bit of disagreement between the more poor countries and because we are talking a little bit about the need for standard errors we put confidence intervals for the MPI head count ratio so you could see that even when we consider at 5% the confidence intervals it doesn't really change the overall relationship between the two measures very much now let's look at some of the other composite indices and then we will briefly talk about the structure of composite indices vis-a-vis the structure of a poverty measure so one of the best documented and most rigorous indices is the global peace index which has 23 indicators of violence or fear of violence and it's clearly of interest because the global MPI lacks personal security which was one of the dimensions of great interest and it has qualitative indicators which are banded, weighted and aggregated and there are robustness tests and extensive documentation so if you're looking for how to write a document on composite indicator structures it's a very good one to look at and the 23 components do not overlap with the components of the global MPI and so as we might expect the measures diverge empirically and they clearly complement each other so the total population is very low for the 95 countries for which we have data on both measures another index which is put forward or has put itself forward in the context of the sustainable development goals is the social progress index which includes three dimensions human needs, foundations of well-being and opportunity and four dimensions and each component then itself includes a number of indicators of different kinds for example, for nutrition there's undernourishment, depth of food deficit, maternal mortality rate child mortality rate and death from infectious disease so these are put together it's sort of a massive undertaking in terms of the number of indicators but it covers many of the same similar domains as well as additional ones on opportunity and environmental quality and personal safety the relationship to the global MPI in terms of the country levels is higher for the 73 countries for which both indicators are available and we could go on but the relationships are not as strong for example with the Lugatum Prosperity Index or the Ease of Doing Business Index because we might wonder whether countries with a strong business sector have a better environment in which also to redress poverty the Fragile State Index is important because half of the poor children live in fragile states but interestingly again there's not a strong association the strongest of all the indicators is with the Human Development Index primarily through its education component but also somewhat through its life expectancy so a question we talked about the conceptual need for multi-dimensional measures that complement monetary poverty measure drawing on Amartya Sen's work we then looked at Miguel Zecheli like Mahbub Al Haq who recognized the importance of numbers to incite policy action and looked at the space in the Sustainable Development Goals for multi-dimensional measures of poverty and of well-being so what would be the value added of having a multi-dimensional poverty measure alongside these other composite indicators I would look a little bit at some of the features of the measures first of all remember the diagram of Sir Tony Atkinson of the overlapping deprivations to do a counting based measure requires aggregation first across dimensions for the same person or household and then across households and that's a different order of aggregation from every composite measure so composite measure like the Human Development Index Social Progress Index or the others first aggregate across all people this means they can draw from different survey instruments from different years they can combine data for different base populations but it means that they do not reflect the joint distribution of deprivations so if you have a society like this in which deprivations are spread across four people these being the people or a society like this in which one person has all four deprivations if the marginal numbers are the same the Human Development Index or any composite index would have the same value so they're not going to reflect the joint jointness of deprivations in contrast counting based measures will be able immediately to capture this but they will be limited by requiring all the data to be from the same data source and also by requiring a common unit of analysis whether it's the individual or the household a second observation is something that I very much learned from James Foster which is the need to have very clear understandings of definitions between different kinds of indicators so one might think of an indicator of well-being which existentially essentially measures the size of a distribution the income per capita for example and that would be different from an inequality measure which essentially measures the spread the ratio of two income standards a poverty measure following CEN 1976 that is an absolute poverty measure would first identify everybody who is poor and then focusing only on them aggregate their information across the population to have a poverty measure so a poverty measure reflects the base of the population and it's quite useful to mentally distinguish them because what you find is that many composite indicators combine different kinds of measures well-being inequality and poverty and they also will include other statistics perhaps that do not relate to people but relate to small arms or to the environment or that relate to mortality statistics like numbers of homicides if you look at the sustainable development goals and your goal indicators and you just think of the definition of poverty not the content of the indicator but the structure of the indicator then over 60 of the 232 SDG indicators are poverty in structure they identify a condition they identify the people who experience that condition and then they aggregate across the population to reflect to create the SDG indicator so it might be interesting to really look at those SDG indicators and think how do we summarize at least a subset of them into an overall measure that is consistent and perhaps how do we do so with accounting methodology a couple other distinctions between a global MPI or accounting based measure and the composite indicators is that because the composite measures are from different surveys they often come from different years and the frequency of updating those different surveys will be different and so it may be that there will be a very long lag in updating some component indicators furthermore the surveys used for composite indicators may not be disaggregable by the same populations and so it's generally more difficult to disaggregate the broader indicators and there may be less possibility again because of the combined data sources of having standard errors of having harmonization in a strict sense over time and also we could go into this more the weights that would be affixed on a normalized 0-1 cardinal index component have a very different character than the weights that are affixed to a 0-1 deprivation of a woman like Nahato and whether or not she's deprived in each indicator in particular as Martin Revalian said the weights across normalized indicators in composite indices imply a marginal rate of substitutability between the indicators at different levels of achievement and across indicators so there may be some value in again thinking of the structure of a poverty measure in also having a structure as was recommended in the Atkinson commission report even if some, perhaps some of the indicators change finally I'll mention in passing that in 2014 UNICEF released a counting based measure for children the multiple overlapping deprivations analysis which covered 40 countries using data from 2008 to 2013 and this is again a very similar measure that also makes an adjusted head count ratio for two populations for children 0-4 and children aged 5-17 it's a tool for advocacy and so rather than a focus on policy that is the need to understand each indicator and how to change it the focus is on really thinking of child rights as indivisible and so they identify a child as deprived for example if they are not immunized in DPT or if they have not had a skilled birth attendance or both and so that's gives a higher number which is very useful in advocacy terms it does lose information that would be relevant for policy if a child is deprived you want to know do they need immunization or did they need a skilled birth attendant so now let's come to the practical part we've seen the landscape and the openness to multi-dimensional measures that in a sense this is moving from the margins to a space of its own and it is invited to do so but where does it fit in the SDG reporting so if you look at the SDG report 2017 on Goal 1 it includes a $1.90 a day measure updates and unemployment updates as well as those on unemployment benefits and natural hazards so those were the four indicators mentioned in this year's SDG report for Global 1 sorry these are out of order I will come back to the SDG reporting but mention that there is a bit of disquiet with the silence around multi-dimensional measures and that's partly coming from many of the countries that use national multi-dimensional poverty measures as official statistics and these countries are designing measures that are not comparable like the global MPI but like their national development plan that are implemented using their national data sources that might include for example employment or internet access or violence or environmental conditions such as are relevant in those countries and they are comparable over time within the country for use and comparisons there so the countries are using these national MPIs quite extensively to complement their monetary poverty measures and all but Mexico have distinct monetary and MPI measures but also for budget allocation for example president Solis in Costa Rica passed a presidential decree by which allocations now must reflect the level of multi-dimensional poverty as well as monetary poverty and population density it's used for targeting marginal regions and marginal groups in a number of countries and it's an associated census instrument it's also used extensively for policy coordination because the MPI gives a headline one number that reflects the work of many sectors and so it's in a sense a common goal politically and so the different ministries can learn how they must work together to shift the dial on that common goal and the minister of health also can realize as he said in Colombia how he needs the work of the minister of transport or of water so that he can do his and again there's a strong emphasis on desegregation so when Mexico launched its MPI for the first time it desegregated by indigenous status Panama launched its national MPI in June and it desegregated and found that in the comarcas of Panama over 90% of people are poor whereas in other regions it's 4% so this wide disparity between the two countries and it's also an interesting set of initiatives around engaging the private sector in doing a business MPI to where they identify who within their employees are poor but according to the national multi-dimensional poverty index and have interventions on their behalf so these countries in a sense are both learning about using the MPIs not necessarily the global MPI to use that directly but using it to prioritize the SDGs in their context that are poverty related and that need the most emphasis and this community includes statisticians and the heads of statistics offices present in different areas it also includes more policy leaders so for example in the high level political forum many countries have mentioned multi-dimensional poverty in their voluntary national reviews and also there were three side events in the most recent General Assembly on multi-dimensional poverty including the heads of states of Honduras, Bhutan, Colombia, Mexico, Chile Vice Presidents of Costa Rica, Panama the Administrator of UNDP and two of the three and a number of high level people trying to think through how can multi-dimensional poverty measures actually be a tool for prioritization and for joined up planning in the SDG environment as well as for monitoring groups at risk of being left behind but when we come to SDG reporting the question is what should be reported and that remains an unanswered question in this table I've just put together some of the national official national MPIs of some of the countries and their corresponding global MPI what you can see is that they're very different in Armenia 29% of people are poor by their national MPI and less than 1% by their global MPI and the national MPIs tend to be higher except in the case of Mozambique and Pakistan but if you go to the SDG indicator platform the space is filled for 1.1.1 the dollar 90 a day for 1.2.1 the national income poverty measures but it is blank for multi-dimensional poverty and it's quite interesting it's because countries are custodian agencies of the 232 indicators only for two indicators 1.2.1 and 1.2.2 the multi-dimensional poverty measure and there is at this point not a way for countries to enter their data and so it's a pause because this will be corrected but in that pause there's a question of what should be reported should it be a comparable MPI like a dollar 90 a day there would be some arguments for saying that that would be better for example the target is to reduce by half the proportion of men women and children living in poverty in all its dimensions and so a comparable measure would make sense like a dollar 90 a day because you are identifying an unacceptable level of poverty and wanting to bring it down but then we have the clause by national definitions which is so important in the SDG environment that indicators be nationally owned nationally constructed and as we have seen when countries come to design their national MPIs like their national income poverty measures they are different often they are more ambitious than the global MPI so there may be space for two indicators as in the case of monetary poverty a comparable one and a national one but in any case as the Atkinson commission report articulated there is need for a very strong conversation between national measures and comparable measures a greater understanding of how they relate to each other and of how the relative priorities that each embody can synergize with each other so in closing I just wanted to say that I have shared one potential structure skeletal structure if you will of multi-dimensional poverty but I've also very much pointed out its weaknesses and its weaknesses are largely because of data constraints it does not have empowerment or violence or employment or many of the other aspects that it might reasonably contain at the same time one wants to be very grateful for the data that exists in the demographic and health survey and multiple indicator cluster survey all of this disaggregation is impossible without those data that are free and publicly available but if we look at the more up-to-date data for 83 countries and 5 billion people we can ask what might be possible in the near future in terms of improving the global MPI and we looked across 31 indicators in quite some depth and to our interest and perhaps disappointment a little bit there's not a huge margin for change at present in the data that exists if we want changes to involve at least 70 countries and 3 billion people but we can move to stunting or combine stunting with under nutrition we can have age specific body mass index we can add roof and wall to the flooring indicator and perhaps replace electricity with overcrowding as it becomes less relevant so there's some changes in the margin but there's also a very strong need to continue to explore better ways of charging geospatial data on violence or thinking about short work modules for inclusion in these surveys so I think my time is coming to a close and so I hope that we have shared a little bit of what are the possibilities in terms of developing a better global multi-dimensional poverty measure what value added it might have what kind of rigor it might offer that would in a sense add value to the landscape of existing measures each of which do important work of their own so we began by basking in the lovely image of Edward Leslie that ain't no sin to take off your skin and dance about in your bones and we've tried to take that freewheeling curiosity into the sober matter of poverty measurement we explored whether multi-dimensional poverty measures could describe the skeletal structure of poverty in a way that might better evoke understanding and ignite action and this exercise was very incomplete because it mainly left out policy analysis participatory and community engagement private sector interventions and many of the real work of reducing poverty but the question that I put before you and the one that we must ask is whether even this exercise as limited as it is might clarify the task and the priorities of reducing a small set of deprivations that are interlinked catalytically and so I just want to close unapologetically with the words of the Atkinson commission that did put before us the value of thinking about measurement alone even as we know we must then move on to the next step of policy and analysis the estimation of the extent of global poverty is an exercise in description as the Marta Sen has written description as an intellectual activity is typically not regarded as very challenging however as he goes on to say description isn't just observing and reporting it involves the exercise possibly difficult of selection description can be characterized as choosing a subset of possibly true statements on a subset a subset of them on grounds of their relevance understanding the choices underline the monitoring indicators and their full implications is challenging there will be differences of view but it is hoped that the ensuing debate will bring together all those concerned and provide a basis for action to tackle one of the gravest problems thank you so much for taking us through such a variety of themes that are sitting not just for this goal for this SDG but in effect relates to all of the rest this means that there's a lot to discuss so I'm happy to say that we have allocated quite substantial time for the questions and answers section so we will now basically enter into that part of the conversation I really hope that we can open up ask questions to Professor Akira and get a lively debate I'm going to sort of take about three questions at a time and then see whether we cannot get the dialogue going because in the usual wider way dialogue informs everyone so who would like to start okay I don't see so many hands okay the start down there in the back hi, my name is Rachel and I'm a research fellow here at UNU wider I was wondering if you could tell us a bit more about some of the policy lessons that are emerging from the analysis of the global MPI in particular for instance if you think about changes the countries that have changed most rapidly in the global MPI measure over time is there anything that jumps out in terms of the sort of policy steps they've taken or comparing for instance the global MPI with income based poverty measures where there's the greatest gap is there anything that comes out in terms of policy there thank you okay yes here my name is Pascal Do I am the author of a PhD dissertation the role of higher education in poverty reduction at the University of Tampere and within this research I used two contradictory titles one was why poor people remain poor by G and another one was why the rich countries became rich by Reinhard published in 2007 I was going to find out if the MPIs are cotominous to mitigation of poverty and if it is I would like to know how you situate the role of higher education because when I wrote the thesis not only education and higher education but science when I wrote this thesis I argued that science higher education was very central even in addressing the 8M digits and if you side with me then I would like to ask another question would I have the impression that when I see the example of Africa that the MPIs undermine the power side of the game political poverty what do you think thank you okay we will go here in the front hello I am Carlos Grading I am a research fellow here at Unowhider and I would like you to stand a little bit on the added value of using an aggregate multidimensional index complimenting how we can establish a consumption poverty measure used by the World Bank I have no doubt that we are adding something with this new index but I still wonder what is exactly what we are adding because if these dimensions were highly correlated with consumption we would get basically the same measure maybe with a different level of the poverty line higher or lower or smaller we always should be look at different dimensions separately but the fact of aggregating all dimensions in one index so it is the lack of correlation between consumption and this aggregate indicator where we find the added value and I wonder what is or what do you think it's the main source so I could think in terms of measurement error it's not a good measure of well-being because it's more difficult to report than years of schooling or having a dead child etc which I think they are easier to report in the serving maybe it's because in some areas income doesn't buy better education or better health because you need some infrastructure you need schools, you need health care centers or maybe it's because households could achieve their consumption above the poverty line making sacrifices in their health or in their education because for example using child labor or working longer hours so what's your understanding of what type of households are we adding to the global poverty picture that we would miss using the consumption poverty measure only thank you quite challenging issues policy relevance education, Africa and the value added of an aggregate so so I'll take them in a reverse order two different things one is the relationship between multi-dimensional poverty and monetary poverty so at the household level the mismatches are quite large even if the levels of poverty are similar so I'm sorry Sabina or you cannot no, people are shaking their heads so we need some technical assistance here now Sabina that's very different so the mismatch between income and multi-dimensional poverty at the household level is quite large for example in Bhutan 12% of people were monetary poor, consumption poor 12.6 were multi-dimensionally poor 3.2% were both overall 21% of people were poor but only 3.2 had both types of poverty and if you look in China, if you look in Chile if you look in different countries where the poverty measures come from the same survey you can look at that mismatch and it's consistently high why? we don't know that requires qualitative work as you said some of it will be the volatility of the consumption poverty measure some of it may be household spending patterns but we also need to know more because these are households maybe where the service provision is similar another cause of mismatch would be service provision being quite different in one area than another and so you do see places where there's plenty of service provision but perhaps not the livelihoods and so the MPI levels would be low or vice versa in the region Gaza since around Bhutan or Papua Nogini in Indonesia it's the same the monetary poverty is very low but multi-dimensional poverty is very high because there are no services but there is for various reasons in the two regions income so I think they're capturing different kinds of deprivations now what does an aggregate add to a dashboard and I didn't show the slides on that but if you were to look across the 10 indicators of the global MPI then you would find that 75% of the population were deprived in one of them 3.9 billion and so it's quite capacious it would identify a lot of people if you said if you had any of these deprivations you were poor so a feature of multi-dimensional poverty aggregate is that it in a sense it focuses in on people who have multiple deprivations simultaneously and so your resources your fiscal resources which are limited are going to be contained a little bit to the people deprived in a third or half or whatever the poverty cutoff is at the same time and so there's plenty of empirical work showing what the difference between just having a dashboard and where any deprivation reflects is it equal in importance in a sense and having an aggregate multi-dimensional measure I don't know if that helps and to Dr. Pasquale in higher education so you asked if reducing MPI was co-tournamentist with ending poverty and clearly it's only partial so like the target said poverty in all its forms and dimensions and all its dimensions would not be possible to measure so an MPI only tracks the indicators that it contains so for example if it doesn't contain livelihoods or if it doesn't contain violence and those are important deprivations then zero poverty could be coincident with high deprivations in those areas I think your question about higher education was a different one and it was really saying in an environment in which multi-dimensional poverty is going down perhaps for people without higher education what is the role of higher education and an educated population in driving that change and that's a much more interesting question but a complex one so it's not one that we've looked at in terms of looking at the average level of higher education prevalence and seeing if that coincided with a faster reduction in MPI for the other portions of the population but I think it would be interesting to explore and to Rachel clearly that's a very important question and a very ambitious question we don't attempt to have cracked it we've done a few regression analysis we've looked at a few determinants and what's very clear is that however you measure growth elasticity the elasticity of growth and multi-dimensional poverty are very low so we don't have the same relationship you get with income poverty it's not a surprise back from Bourguignon and others at the World Bank showing the correlations are different so the drivers would tend to be more associated with proactive social policies and public expenditures as well as activities by other stakeholders it's also interestingly not necessarily good governance at least not in low income countries so we did multi-level analysis my colleagues did it didn't find it and also if you think of Nepal which actually reduced its MPI from 26 to 2011 the fastest it didn't have a government for part of that time it's it but somehow they were able to sustain the social expenditures and also they had high remittances and high back flow of educated labor into the country so it's a very complex story I think it's country by country it's not enough to be able to generalize yet but again our data are online and so I'm hoping that others might take them and do the findings we haven't been able to okay thanks for this we take the next round yes here in this middle a little bit in the front hi this is Jonas Uotinen from the University of Turku I'm a PhD student at the economics department I wonder is there a way for the MPIs or a modification of it with a possible addition to it to also inform the rich countries on their aims as it seems there are possibly a bit lost if we kind of try to move beyond the gross domestic product then it's like okay what shall we aim then for and this relates to to also the what is still missing from the capability theory there was for example in in his article who pointed out that it seems that deducing from the like or he stated that the increased freedoms in United States as by the increase of purchasing power over the 20th century has come has led to greater anxieties and other mental disorders instead of it being the contrary so it could there be some kind of addition that could lead us towards understanding of of these issues some of the things that come to mind is perhaps some Aristotelian of realisation of the human potentials and what those be as well as the developmental trajectory of the human according to the western and eastern developmental psychologists and thinkers so these are some of the things that come to my mind that possibly could be added or something like this I would like to hear your opinions on this but I think there's a hand in the back thank you I'm Jim Syngian researcher University of Helsinki you admit that poverty is a very fluid concepts and my question is simple what's your definition of poverty and who is poor I'm a research fellow at Wider I have a two part question and I would appreciate your thoughts first in your researcher boots and then in the policy maker boots if you can imagine to Trump the first one is regarding the SDGs and implementation of the monitoring an OECD report last year tried to put to work as many of these indicators as possible the most developed and let's say data rich OECD countries would not have most or at least I think more than 80% of these indicators ready lying around my question to you is for the developing countries where these gaps are even larger what would be the indicators you would kind of focus on in data collection where do you see the biggest gaps the second part of my question relates to limited resources in the sense of where do we get more data from there's also a big push also championed by the World Bank for identification for development now that has great potential for example linking and composite indices of course the different data sources but it's also a very big exercise usually it is argued from the side of we can more easily distribute benefits and we can more easily monitor now for your exercise where do you see the value added or would you say actually you would prefer that more modules would be hooked on the already existing surveys that you work with okay another three relevant to rich countries definition of poverty major data gaps and where do you see that very good so Johannes in terms of the rich countries certainly there's nothing that stops them also from having a multi-dimensional poverty measure at least the EU's silk is comparable across countries and we've been involved in that silk too and are involved now in that silk three exploring those options but your question was wider than that and it might be relevant to mention the case of the gross national happiness index which has health education living standards like a poverty measure governance time use in the environment and then community culture and psychological well-being and in terms of developmental states it was interesting when there was an international community interacting with Bhutan's gross national happiness index one of the lessons that they got quite interested in was that they realized that happiness is a skill and you can learn to do it and so in a sense also coding where people are on that skill of happiness which maybe people think is Buddhist but you can also from mindfulness or from psychology you can enter into that path an interesting thing is when Bhutan updated its gross national happiness index in 2015 from 2010 there were statistically significant increases in material you know in income in jobs in healthy days in the services of water and sanitation and roads and electricity in education so all of those went up significantly and overall there was growth in gross national happiness index but in the psychological well-being category every indicator significantly declined so that was satisfaction with quality of life positive affect negative affect and spirituality sense of belonging declined a feeling of etiquette and courtesy declined and so what was interesting is because you had that panoply of indicators you could see that the simultaneous movement in different directions and I think that that's quite useful for in a sense catalyzing a conversation about where societies are going and how whether that matters whether they like that change or might want to redress it and to Jimson we work with most of my work is not on the national MPI on the global MPI most is with national governments who are doing their national MPIs when I'm not doing research and clearly the priority is that the voice the protagonists of poverty the poor people would have a very strong input so for example in El Salvador there was a two-year participatory exercise with different communities after the government already had a draft measure and the draft measure had health education living standards and employment in it and after the participatory exercise they added lived environment and violence because those came out so strongly from the communities and it seems very important that a measure of poverty should reflect the experiences articulated by those in poverty at the same time you need very much this not to be only a statistical exercise because if the minister of planning the minister of finance if they are not on board and don't see its relevance they won't use it and so there's also been an interesting engagement in countries setting up different kinds of committees that in a sense have the user of the statistic so that by the time it's launched like in Ecuador President Cordrea launched it the head of statistics explained it the minister of planning Nancy and the minister who is doing the targeted social programming both spoke of how they were going to use it and that means that a statistic has an audience anticipating its reduction and thinking towards action so it seems that both involving the poor communities and in a sense the users of the statistics is key alongside of course the statistical community and to Pia it's interesting in terms of data gaps what is what the we are secretary to a network of 53 participating countries who are designing or using multi-dimensional poverty measures and what a number of them are saying now is that really it's up to them to prioritize and they may not prioritize the missing data they may prioritize actually making some steps based on the information of data that they already have and I think localizing or whatever term is going to be used but setting priorities is very much the topic of the day because it's impossible to do them all and for me there's a big danger in multi-dimensional poverty measures if there is a explosion of them of simply statistical overload there's too many numbers and what we observe is that when there is a multi-dimensional poverty measure then in a sense it brings together 10, Costa Rica has 20 indicators which is too many but it's a lot but it brings them under one umbrella and so in a sense they get a kind of serious attention and the danger with 232 participation as we all know and so trying to find ways of making it easy for again statistical users to engage with the data that do exist in an action way I think is more of a priority and so that's that does I'm just thinking of the countries they are all doing new surveys they're all adding little modules but it's the ones that really seem to be more important in their context and regarding the identification there's also a downside for example in the Aadhaar program in India there's people who will be left out people for whom it doesn't work the fingerprint or there's a risk of it being used in a way that it may not be completely constructive and might be exclusionary or harmful rather than just providing benefits and so I think there are a lot of ethical questions it would be ideal if we could match different data sources but the distance from here to there is quite big and so what we observe is that in many of the conversations on big data or the geospatial or the merging of data sources it's very exciting but if we want poverty data now at the moment that definitely includes poor people it really has to come from household surveys and I see that one of the dangers is that there are champions of big data there are champions of geospatial this or administrative vital statistics that but there's really no champion of basic household survey data that's extended in the way Sir Tony Atkinson recommended to look at missing populations and you know some of the well-known problems of household survey data but I do see that as an area that desperately needs ongoing investment and is being overlooked in the conversations thank you yes here in the front and then we will let's take the front first hello thank you I'm the chairperson of the Finnish Association here in Finland my name is and I'm also the president of the European Union of the Deaf my question is more of a comment to the MPIs and about the dimensions of poverty one of those is education and I'm thinking or my question is what would be the other other aspect would be also language and the use of language and also the barriers in using language because education you can get to education but if the education is given in other language than what is your natural language then it's going to be a bigger barrier to you than getting into the education getting the information if there's information and you don't get it in your own language it's not usable to you so that's also a barrier in your life if you don't get the information that you need and you cannot do the conversation with different people and you also face this even greater barrier in your education and in acquiring knowledge what is your thinking about in the MPIs and poverty connecting to this access to information and to language so can you measure this and also can you eradicate poverty by this and what is the relationship with language and I'm not a statistical person myself and I'm thinking about what is your opinion of these I'm more of a linguist and an educationist so this is where my question is coming from thank you there was in the middle there was a hand here in the middle hello, good afternoon my question is here before I ask the question I would like to say something here I think being a poor child today is most likely to be a poor parent in the future being said that if you think the individual can be trapped in a poverty if it is how does the MPI will elaborate to explain it and how does it help to reduce the poverty to cast SDG goal and my name is Rajiv Casey from Nepal, thank you thank you and then here in the centre next to the one who is taking all the video from today hi I'm Surabhi from India so I just wanted to ask how do you find solutions how would you implement these goals in places that are so diverse even if you consider nations each a country like India is so diverse so the struggles of people would differ geographically and I'm a little nervous but if you see poverty it's the struggle to I think it is when people are struggling to get access to health and education and these are constructs and I see poverty there when people are trying to get access to this because it's imposed on them in some way and if and as you say the surveys are done at the national level and even there the data would be skewed because they wouldn't probably be accessing every region I'm interested to know how the data set is collected related to a question but again how do you connect these goals to places that are culturally very diverse sure so how does this relate to the use of language access to information I guess the second question was about intergenerational transmission of poverty and how that relates to the multi-dimensional index and then how do you really do this implementation across such diversity very good in terms of language the example I can give is from Columbia and it is when they realize that their national measure was not relevant for the indigenous community and so they created an indigenous MPI with the community and their number one request was for education in their language and so it was a very articulate demand for changing the specification of education to include the language of instruction and so far that's the only country that has put language into their national MPI but because that has visibility in other countries that have an indigenous population at least that has been discussed so that incomplete but it's what we have to Razif's question in terms of intergenerational so the global MPI uses repeated cross-section data but there are studies with panel data and with panel data you can see in a sense the chronicity of multi-dimensional poverty you can also see what deprivation changes so people drop into poverty and four if you isolate the people who are chronically poor across time periods you can see what is the caustic combination what are the deprivations they always have now clearly that will be partly a function of how you've designed the measure but it's interesting also to see what combination tends to be associated with chronic poverty so at this stage we're simply doing that kind of a descriptive analysis but a colleague is thinking then of extending the poverty trap literature and doing more analysis on that but right now we're just going through different sets with panel data to try to uncover those combinations that tend to regularly be associated with chronic multi-dimensional poverty there's also different papers for example Luis Felipe López Calva of the World Bank found in one paper that multi-dimensional poverty was associated with chronic income poverty another paper found the opposite so it's still a literature in a time of flux but interesting questions to Sabida I think no in terms of your question no it's a very good question and I must say I'll give a little advertisement for sense and solidarity I don't know if you've read it Jola Wala Economics for Everyone by Jean Dres which came out last month it's a fantastic book and it's very much a book which has relevance to these conversations and to your question because he articulates how research on poverty cannot go without a very real interface with poor communities and their activities and that certainly Jean Dres embodies that in his own work and in our observations in India the MPI work has gone ahead at the state level not nationally so Andhra Pradesh for example has a state level MPI Assam had one it's a state level human development report so those have been the areas of interest but I think in the I gave earlier the example of El Salvador but some of the interesting work is when NGOs understand an MPI it's very simple it's like a counting based measure and so the NGO sector can do it easily in their own communities you don't need fancy software so if you think of the program of targeting the destitute families or whatever it's a counting based measure if you have three of these nine deprivations you're identified as eligible so it's an intuitive work approach that can be used locally and not just nationally and I think that makes it more relevant as you said to the different deprivations of different cultural groups so that those are some suggestions but in India that's yeah we're not doing anything nationally your question about data so we are waiting for the NFHS4 data which should come out in December January in India and it'll be representative of the district level nationally so there's a huge survey but it will have some representativeness but again it may not pick up on specific contexts of different regions okay we will take the next round so we will up here almost in the front good evening my name is Obin Gottfried I'm really interested in social and public policy and I know what you are coming up with is going to be some of the basis for most policies formulations within national and international levels you talk about education in relation to poverty eradication and I'm really interested when you talk about having access at least to the grid 8 and I really want to know how getting up to grid 8 for example can't really effect poverty reduction or eradication eradicate poverty looking at most especially if you look at developing countries or curriculum structure and mode of transmitting knowledge do you think or is there any kind of empirical evidence which shows that once somebody reaches grid 8 it can really affect promoting or eradicating poverty that's my question thanks okay and then there was one here who is next to you Charlie good afternoon my name is Jovin enlightening presentation thank you very much if I made two questions in two weeks the United Nations conference of parties is happening what would be one question you would like these decision makers to have in mind in order to move the conversation forward that's one and second question is as you churn all this data with MPIs and so on I'm curious to know whether you see companies, startups, SMEs trying to plug into these data using big data technologies etc in order to move things forward in terms of applications in terms of making use of these great sets of information that you produce thank you okay here in the front hello I'm Jukka Birthela from University of Tampere and also affiliated with UAU Wider I'd like to hear your views on moving beyond poverty to inequality to what extent would it make sense to use the same underlying indicators to measure multi-dimensional inequality because now it seems that poverty is measured both in monetary terms and then in a multi-dimensional whereas inequality continues to be measured only in terms of either income or consumption okay on the eradication of poverty on the use application of the generation of data that has been taking place and then on can this be replicated for inequality very good in terms of the eight years of education that's basic education and so we just used it because it's one of the levels if you look at the national MPIs that are being developed most countries employ the number of years of compulsory schooling which is in effect at the time and they vary a great deal what we do in the global MPIs we had actually a range of different cutoffs for that variable and we consulted we implemented six years eight years ten years eight years is also not compulsory in all countries because it's the basic education standard and it's beyond primary school it seemed one that we would that we went with but I don't you can cite certain papers but when it's a global measure it's going to be messy and so the eight years is going to be relevant in some places and not relevant in others but similarly a measure of acute poverty will not be relevant in all places and that's a consistent population of deprivations during in terms of there were two different questions in terms of the business MPI that is taken off in Costa Rica Costa Rica has a national MPI and actually its development was co-funded by the business community and then the largest bank was curious whether any of its employees lived in MPI poor households and they wanted to do a survey using just the questions they needed to construct the national MPI and they did that and to the interest but also consternation of the leaders they found that there were there was a non-insignificant number of employees who were living in MPI poor families and furthermore they weren't only in the lowest paid jobs there were some in the middle paid jobs and these tended to be because of high dependency ratios disability at home, high unemployment among youth so they looked at the composition, the profiles of deprivations and implemented some vocational training programs in order to recruit some of the unemployed people and give them proper skills and some other programs that's then taken off and now I think 80 institutions 80 in Costa Rica are trying to replicate it so that's a point to watch and some of them are international firms so it may move outside Costa Rica soon so I hope that's enough in terms of inequality theoretically it's completely possible so what you have is you have a matrix of people and dimensions and if you take the 01 whether or not they're deprived in something you could do that or you could take their overall achievement deprived in each indicator you could take the reverse of that which is whether or not they've attained a certain cutoff in that indicator and so you have a score of weighted attainments which is currently meaningful for the population and then you can take that vector and you can make any inequality measure you want tile one two Atkinson, Genie whatever 90 to 10 ratio I think empirically the problem is not on the measurement side so for example I mentioned that if you take a union measure of the global MPI then 75% of the world are poor and that is largely driven by cooking fuel so many people cook with wood but they may have good chimneys so it's actually not a health risk but the survey doesn't have the right question so it's what I would call a spurious deprivation it looks like a deprivation but it's not for them in the case of the global MPI we censor it, we clean it if the people are not deprived in one third of deprivations 60% of people in Bosnia cook with wood but they're not all poor then we censor it from the data set so what you would need to do for an inequality measure is limit it to those variables where the deprivations are accurate in a sense each attainment or each deprivation is accurate and you don't have some indicators so some of the global MPI indicators are very rarely censored, they're quite accurate and some particularly cooking fuel is quite inaccurate so that would be the difficulty, James and I are doing a paper now both looking at inequality among the poor and then also looking at this attainment matrix and Atkinson-Bergenione inequality measures and what you can do multi-dimensionally with very similar techniques as a counting based poverty measure okay further questions okay yes, here there are so many that second rounds will be difficult hello, I'm Soumya I'm also a visiting scholar at Wider I was wondering if multi-dimensional poverty would be much lesser in those countries where gender relations are better because you're looking at these basic services water, sanitation and particularly at children so children are much better if women are also much better off in the region so what are your views on that thank you okay, one in the background Hula Wlinstel from NORAD I'm just curious on the question of aid allocation coming from an aid agency how do you see the MPI and the impacts if you compare with income poverty and the 190 if you try to think about allocation by countries or by sectors or by themes that's one and the second one is the use of MPI in terms of guidance or policy reform you talked a little bit about it this obviously is a complex question but I would be interested if you have any examples that there is dialogue with governments and if you see anything in terms of expenditure patterns and thinking of expenditure patterns okay, thanks and then here in the front and that I think Tony Addison from WIDA the very first WIDA annual lecture was given 20 years ago in March 1997 by Nobel laureate Douglas North and the world has made quite a lot of progress in 20 years on poverty partly due to the MDGs and the SDGs so if we could look 20 years into the future what do you think the world of poverty will look like how much progress will we have made and which countries might be ahead and which countries behind so if we go forward 20 years what will our world look like okay three small questions one on the balance, the reference to gender the whole question of aid allocation and policy reform and how the index might help inform that and then we'll see here in 20 years very good so on gender I must say one of our disappointments is that the data that are available do not permit us to make a gendered MPI we could make a women's MPI but we can't have one where we can disaggregate meaningfully by gender we've disaggregated the global MPI but we haven't published the results because it's not correct it just reflects the demographic structures so I think really we have called for in fact our network I mentioned we're the secretariat of our network the network designed a gendered survey and proposed it for the SDGs so that we could have within the household multiple respondents we looked into a feasible and inexpensive way of sample design for doing so and we designed the survey it didn't go anywhere so I think really there I would join my voice with many others and call for better gendered data going forward because we can't really use this data and answer your question because we simply have no information on intra household sharing that being said we have done individual child poverty measures and we're doing more of them we're doing a lot of them where we use the global or the national MPI for a number of indicators and then for example in our cognitive development and nutrition or health we look across the age cohorts of children and have individualized indicators and those we do show up gender differences so that's I think a step forward and there are data for doing the child poverty measures in the mixed surveys but it is a disappointment and it's baffling that we don't have better data in terms of aid allocation it's not an area in which I'm an expert but we had a little bit of a look this year because of the first of all the disparity between the low income country category and where the MPI poor people live with 72% of them living in middle income countries which is similar to the percentage Andy Sumner found with monetary poverty when we looked at the aid distribution we used different definitions and they're published in our policy briefing of this year we found that the distribution across low and lower middle income countries was actually surprisingly more balanced than we would have anticipated but what we found was that the allocation per poor person according to the MPI was very very noisy with India and China having less than a dollar per poor person per year and other countries having a huge amount now clearly that's influenced by PPP clearly it's influenced by national public expenditure patterns that aid flows tend to complement but there's I think a lot to be unpacked and so all we can do is we can say per MPI poor person what are the aid flows to these countries in nominal terms and then the conversation must go beyond that with others and in terms of policy change I could give the example which President Juan Manuel Santos of Colombia would give in that case which is he launched the MPI in 2011 and he set a target for its reduction and then he worked with Mackenzie and he worked with public expenditure people to try to figure out the best patterns of fiscal expenditure to be used he had a committee that met twice a year the ministers could not send deputies they had to meet with him annually they updated their national MPI and in between they updated it using administrative data to try to look at the trends and then when an indicator was not moving to plan they had responsive policies and so they were able to greatly accelerate the reduction of multi-dimensional poverty they'd analyzed it back since 1996 using the same survey so they could expand, accelerate reduction but it was using the same fiscal envelope and so the same budgetary envelope just spending it better so that's the kind of story as I mentioned Costa Rica is another one that it's invested a lot in allocation because they found there was zero allocation to some of their MPI indicators and that the allocation to the poorest regions was not comparable to the extent of poverty in those regions so the rebalancing of existing resources seems to be a very common first step to have an immediate effect Looking into the future? I'm not very good at that clearly my hope would be that the acute poverty could be eradicated but one doesn't know going ahead and so I'm not sure how sensibly I can address that I think that what's clear is that as countries come to very low levels of multi-dimensional poverty like Mexico then they redefine their ambitions and they use a similar structure but now Mexico is 1.2% according to the global MPI of poverty but 43% according to their national MPI because their aspirations and frames of reference have changed and so in terms of poverty and development hopefully there would be a re-articulation of the appropriate sets of goals particularly if climate change permitting and all of that we are able to reduce the kinds of acute poverty that we now face and I think also that the psychological domains and that the other domains of well-being may come in more strongly as we are able both to measure them and to think about sensible and pluralistic policy responses to them there seems to be an appetite so in a number of countries not just Bhutan we are seen not only the poverty measure but now them wanting to do a linked well-being measure that includes the dimensions of poverty but then goes beyond it to some of these soft things as an experimental measure not yet to have the kind of seriousness of a poverty measure but to try to keep these within the field of vision and that I think will be in a measurement terms quite difficult but interesting Sabine can I sort of just add one observation and we've been spoken about data and so on and I just hope you will bear with me just making one reference to wider work on Vietnam we have actually put out a whole series of YouTube videos on what the data revolution actually means in practice so I mean those are sort of want to try to get a sense of what some of these things actually mean in concrete practice out there and what you can do once you start over a period of 10-15 years having built up a panel data set sort of what does it take to actually get to the point that some of these questions can be addressed in greater depth and so on I'd like to sort of just make that point because I genuinely agree with you on this issue that we do need to make sure that household surveys do not get sort of disappears in this so big data is now all over there and it can answer all our questions big data cannot answer all our questions we do need to get down to the household an individual level and collect that data in order to really come to grips with it and so I'd like to make that point if I may just as one final observation or question I participate in the formulation of the SGGs and I mean I was in numerous meetings and of course there's the number of indicators and measures and so it grew bigger and bigger being after all trained as an economist I started getting a bit worried because I was thinking okay how do we operationalize all this I mean I certainly agreed with the political intentions behind and so on but I did get a bit worried about it and then every time I would sort of ask about do we actually need this measure or this indicator and so on I was constantly confronted with the following statement rights are indivisible you cannot sort of start trading one dimension off against another one you have to have them both I'm kind of pondering what you think about that I hope my question is clear because it's something where I believe that at least the economics profession but also other professions have a big issue that we are not always kind of managing to to get to click among us so I was wondering whether you had some reflections on that well I think two or three things one is that not all SDGs might be framed as rights so that's why yes there is the language of integrated and indivisible there because they're interlinked but perhaps there's also a language of prioritization and then if you think of the incremental realization of social and cultural rights there's also a recognition that with limited resources it's essential you know it's impossible to advance on all of the fronts together and so prioritization and setting medium term goals is essential because it is impossible to do otherwise and I think that that actually where countries are engaging the SDGs and not all are that's the exciting point is when they feel empowered then to look at that basket of goods and select the first step of actions to be undertaken and that we'll actually put them towards the goal of realizing the SDGs but do so in a way that also coheres with Agenda 2016 or with their national development plans or with some other priorities which are also salient because it's not only the SDGs that have a voice in so many contexts sure time has run out I hope that first of all you will join me in saying thank you very much to Professor Akira this was great as an afterthought there is a reception there is a glass of wine and there's a bit of food also outside can I invite everybody to join us in the reception and let the conversation continue and thank you very much