"CONTENTdm"@en .
"Digital Himalaya"@en .
"Journal of Bhutan Studies"@en .
"Santos, Maria Emma"@en .
"Karma Ura, 1928-"@en .
"2017-12-14"@en .
"between 2008-06 and 2008-08"@en .
"A scholarly bi-annual publication on the social, cultural and economic aspects of Bhutan."@en .
"https://open.library.ubc.ca/collections/dhimjournal/items/1.0365186/source.json"@en .
"page 1-50"@en .
"application/pdf"@en .
" Multidimensional Poverty in Bhutan: Estimates and Policy\nImplications\nMaria Emma Santos* and Karma Ura\"\nAbstract\nThis paper estimates multidimensional poverty in Bhutan\napplying a recently developed methodology by Alkire and\nFoster (2007) using the 2007 Bhutan Living Standard Survey\ndata. Five dimensions are considered for estimations in both\nrural and urban areas (income, education, room availability,\naccess to electricity and access to drinking water) and two\nadditional dimensions are considered for estimates in rural\nareas only (access to roads and land ownership). Also, two\nalternative weighting systems are used: a baseline using equal\nweights for every dimension and another one using weights\nderived from the Gross National Happiness Survey. Estimates\nare decomposed into rural and urban areas, by dimension and\nbetween districts. It was found that multidimensional poverty\nis mainly a rural phenomenon, although urban areas present\nnon-depreciable levels of deprivation in room availability and\neducation. Within rural areas, it was found that poverty in\neducation, electricity, room availability, income and access to\nroads, contribute in similar shares to overall multidimensional\npoverty, while poverty in land ownership and water have a\nrelatively smaller contributions. The districts of Samtse,\nMongar, Chukha, Trashigang and Samdrup Jongkhar are\nidentified as giving the highest contribution to overall\nmultidimensional poverty. The methodology is suggested as a\npotential formula for national poverty measurement and for\nbudget allocation among the districts and sectors.\n' Oxford Poverty and Human Development Initiative (OPHI), Oxford\nUniversity and Consejo Nacional de Investigaciones Cientificas y\nTecnicas (CONICET)-Universidad Nacional del Sur, Argentina.\n** President, The Centre for Bhutan Studies, Thimphu.\n Journal of Bhutan Studies\n1. Introduction\nFostered by Sen's (1985, 1990, 1999) pioneering 'capabtiity\napproach', there is now an increasing consensus that poverty\nis an intrinsicaUy multidimensional phenomenon. This has\nled scholars to propose different multidimensional poverty\nmeasures. However, some of the proposed measures seem to\nhave incorporated a multi-dimensional perspective at the cost\nof giving up the simplicity and intuition that characterise the\nunidimensional measures. Departing from this, Alkire and\nFoster (2007) propose a new family of multidimensional\npoverty measures which is a variant of the extensively used\nFoster, Greer and Thorbecke's (1984) class of one-dimension\npoverty measures (FGT from now on). The dimension adjusted\nFGT measures keep the simple structure of the one-\ndimension case and satisfy a set of convenient properties,\namong which decomposabitity across population subgroups\nand the possibtiity to break it down by dimension are useful\nfor policy purposes.\nIn this paper, the mentioned new class of measures is applied\nto estimate multidimensional poverty in Bhutan. Bhutan\nconstitutes an extremely interesting example of how a country\ncan define development goals, tailor its policies to these goals,\nand see them materialized. Since 1961, the country\nimplemented coordinated efforts towards development\nthrough consecutive five-years-plans. In particular, the\ncountry has made significant progress in extending the\naccess to safe drinking water and sanitation, protecting and\nmanaging the country's natural resources, providing basic\nhealth care and increasing the access to primary education.\nHowever, more can sttil be done in some of the mentioned\nareas as weU as in others. Within this development agenda,\nthe MiUennium Development Goals play a key role since\nBhutan is seriously committed to contribute to the realisation\nofthe Millennium Declaration.\nIn this context, this paper intends not only to present\nestimates of multidimensional poverty in Bhutan, which\nwould complement the income poverty estimates performed\n Multidimensional Poverty in Bhutan\nby the National Statistics Bureau, but also to suggest the\napplied methodology as a potential formula for budget\naUocation among the twenty districts, and within each\ndistrict, among the different gewogs, the lowest\nadministration units.\nThe data used in this paper correspond to the 2007 Bhutan\nLiving Standard Survey. It constitutes a unique data source of\nthis country, representative both at the national and district\nlevels. Estimations are performed for rural and urban areas\nconsidering five dimensions and also for rural areas\nexclusively, with two additional dimensions. Each measure is\nalso estimated at the district level, and in aU cases, using two\nalternative weighting structures: a baseline of equal weights\nand another one with weights derived from the ranking of\n'sources of happiness' identified through the Gross National\nHappiness Survey.\nResults confirm that, indeed, income deprivation should not\nbe the only considered dimension. Deprivation in other\ndimensions such as education, access to electricity and room\navailabUity in the house, are significant both in rural and\nurban areas, and not necessarily related to deprivation in\nincome. AdditionaUy, deprivation in access to roads is a\nsignificant component of multidimensional poverty in the\nrural areas. Land ownership in the rural areas and access to\ndrinking water in both rural and urban areas, seem to be\nrelatively less important. It was also found that\nmultidimensional poverty is mainly a rural problem, which is\nparticularly important given that 74% of the population in\nBhutan live in rural areas. When analysing at the district\nlevel, it is found that Samtse, Mongar, Chukha, Trashigang\nand Samdrup Jongkhar are the five districts with the highest\ncontributions to aggregate multidimensional poverty.\nHowever, even in the other districts with lower contributions,\nimprovements in the mentioned dimensions are sttil\nimportant.\n Journal of Bhutan Studies\nThe rest of the paper is organised as follows. Section 2 briefly\nrevises the literature on multidimensional poverty measures.\nSection 3 presents the methodology used in the paper\n(measures estimated, data-set used, selected dimensions,\ndeprivation cutoff values and weighting structures). Section 4\npresents the estimation results. Finally, Section 5 contains\nthe concluding remarks.\n2. Literature review\nSince Sen (1976), the measurement of poverty has been\nconceptualised as foUowing two main steps: identification and\naggregation. In the unidimensional space, the identification\nstep is relatively an easy one. Even when it is recognised that\nthe concept of a poverty line-as a threshold that dichotomises\nthe population into the poor and the non-poor- is somehow\nartificial, it is agreed to be necessary. Greater consideration is\ngiven to the properties that should be satisfied by the poverty\nindex that will aggregate individuals' data into an overall\nindicator. However, in the multidimensional context, the\nidentification step is more complex. Given a set of\ndimensions, each of which has an associated deprivation\ncutoff or poverty line, it is possible to identify for each person\nwhether he/she is deprived or not in each dimension.\nHowever, the difficult task is to decide who is to be considered\nmultidimensionaUy poor.\nOne proposed approach has been to aggregate achievements\nin each dimension into a single cardinal index of weU-being\nand set a deprivation cutoff value for the weU-being measure\nrather than for each specific dimension to identify the\nmultidimensionaUy poor. This approach has some practical\ndrawbacks, in particular, in that it is based on a number of\nrestrictive assumptions, such as the existence of prices for aU\ndimensions. Moreover, it does not agree with the conceptual\nframework of the capability approach which considers each\ndimension to be intrinsically important. Then, each\ndimension with its corresponding deprivation cutoff value\nneeds to be considered at the identification step of the\nmultidimensionaUy poor.\n Multidimensional Poverty in Bhutan\nIn this perspective, two extreme approaches have been\ntraditionaUy used. On the one hand, there is the intersection\napproach, which requires the person to be poor in every\ndimension under consideration so as to be identified as\nmultidimensionaUy poor. Clearly, this is a demanding\nidentification criterion, by which the set of the poor is\nreduced as the number of dimensions considered increases,\nand may exclude people that are indeed deprived in several\nimportant dimensions. On the other hand there is the union\napproach, which requires the person to be poor in at least one\nof the considered dimensions. Clearly, with this criterion, the\nset of poor increases as the number of dimensions does, and\nit may include people that many would not considered to be\nmultidimensionaUy poor (Alkire and Foster, 2007, pp.8). The\nunion approach has received important support both in the\ntheoretical and empirical literature. In particular, Tsui (2002)\nand Bourguignon and Chakravarty (2003) adopt it for the\nmeasures they propose.\nTsui (2002) develops an axiomatic framework for\nmultidimensional poverty measurement (which includes\nsubgroup consistency) and derives two relative\nmultidimensional poverty measures, one of which is a\ngeneralization of Chakravarty's (1983) one-dimensional class\nof poverty indices, and the other is a generalization of Watt's\n(1968) poverty index. He also derives two absolute\nmultidimensional poverty measures, i\nBourguignon and Chakravarty (2003) distinguish two groups\nof multidimensional poverty indices, depending on whether\nthey consider dimensions to be independent or to have some\nsubstitutability or complementarity. Those that consider\ni The distinction between relative and absolute poverty indices is\ndue to Blackorby and Donaldson (1980). Relative poverty indices are\ninvariant to changes in scale, such as a doubling of the poverty line\nand all incomes, while absolute indices are invariant to translations\nor additions of the same absolute amount to each income and to the\npoverty line (Foster and Shorrocks, 1991). In practice, relative\npoverty indices are the ones that have been most frequently used.\n Journal of Bhutan Studies\nattributes to be independent satisfy what they call the One\nDimensional Transfer Principle, by which poverty decreases\nwhenever there is a Pigou-Dalton progressive transfer of the\nachievement in some dimension between two poor people. The\nprogressive nature of the transfer is judged by the\nachievements of the two poor people in that specific\ndimension, independently of the achievements in the other\ndimensions. These indices are additively decomposable. The\nsecond group of indices are non-additive -ie. non\ndecomposable- and by choosing appropriate values of the\nparameters they can reflect either a substitutabitity or a\ncomplementarity relationship between the dimensions. For\nboth groups of indices, extensions of the FGT class are\nproposed.\nOn a more practice-based perspective, the Unsatisfied Basic\nNeeds Approach, widely used in Latin America, also uses a\nunion criterion, identifying as households with unsatisfied\nbasic needs those that are deprived in one or more of the\nselected indicators.\nIn view of the two prevailing extreme criteria to identify the\nmultidrmensionally poor, Alkire and Foster (2007) propose a\nnew identification methodology which, whtie containing the\ntwo extremes, also aUows for intermediate options. Assume\nthat there are k = l, , d considered dimensions, and that\nci represents the number of dimensions in which individual\n/' = 1, ,nis deprived, then an individual is considered to be\nmultidimensionaUy poor if c. > k . When k = 1, the approach\ncoincides with the union approach, whereas when k = d, it is\nthe intersection approach. For 1 < k < d , the identification\ncriterion lies somewhere in the middle between the two\nextremes. Then, for the aggregation step, they use the weU-\nknown FGT class of poverty indices. The resulting farmly of\nmeasures satisfies a set of convenient properties including\ndecomposabitity by population subgroups and the possibtiity\nof being broken down by dimensions. These last properties\n Multidimensional Poverty in Bhutan\nmake it particularly suitable for policy targeting. Additionally,\nthe class includes measures that can be used with ordinal\ndata, which is very common in a multidimensional context. A\ndetailed description of this class of measures is presented in\nSection 3.2.\nA final note must acknowledge the probably most popular\nmultidimensional poverty measure, which is the Human\nPoverty Index (HPI), developed by Anand and Sen (1997),\ncompanion index of the Human Development Index (HDI).\nBoth indices are periodically estimated by the United Nations\nDevelopment Programme for all countries to monitor the level\nof deprivation and development correspondingly with a\nbroader perspective than income. The components of the HPI\nare survival deprivation (measured by the probability at birth\nof not surviving to age 40), deprivation of education and\nknowledge (measured by the adult literacy rate) and economic\ndeprivation (measured by the average of the percentage of\npopulation without access to an improved water source and\nchtidren under weight for age). In developed countries the\nindicators for each of the components are specified according\nto the higher living standards.2 An important advantage of the\nHPI is that it only requires macro-data, which can be\nespeciaUy important for countries in which micro-data\ncoUection is sttil at its beginnings and its quality is not\nassured. However, it has some disadvantages. Clearly, the\nthree selected dimensions can be argued to be arbitrary as\nwell as the weighting system used to calculate the measure.\nWhen micro-data sets are available more informative\nmeasures can be calculated, with a higher number of\ndimensions and alternative weighting systems.\n2 In particular, the survival deprivation is estimated as the\nprobability at birth of not surviving to age 60, the deprivation of\neducation and knowledge is defined as adults lacking functional\nskills, the economic deprivation is defined as the percentage of\npopulation below 50% of he median adjusted disposable income, and\na social exclusion component is also added, defined as the rate of\nlong-term unemployment (lasting 12 months or more).\n7\n Journal of Bhutan Studies\n3. Methodology\n3.1 Data\nThe dataset used is the 2007 Bhutan Living Standard Survey\n(BLSS) conducted by the National Statistics Bureau (NSB).\nThere are 9798 households in the sample and 49165 people.\nThis is the second BLSS performed; the previous one was\ndone in 2003. Both surveys have foUowed the Living Standard\nMeasurement Study methodology developed by the World\nBank. However, the 2007 survey has more than doubled the\n2003 sample size and it has also extended the coverage, so\nthat the sample is representative both nationaUy and at each\nof the 20 Bhutanese districts (Dzongkhags), in rural and\nurban areas.\nThe unit of analysis to identify the poor is the household.\nHowever, households are weighted by their size (as weU as by\ntheir sample weights), so that results are presented in\npopulation terms. Table A. 1 in the Appendix presents the\ncomposition ofthe sample.\n3.2 Multidimensional poverty measures\nThe poverty measure applied in this paper corresponds to\nAlkire and Foster's (2007) family of multidimensional poverty\nmeasures. Before introducing it, it is convenient to clarify\nnotation in the first place.\nLet M\"' denote the set of all nxd matrices, and interpret a\ntypical element y e M\"' as the matrix of achievements of n\npeople in d different dimensions. For every i = l,2,...,n and\nJ' = 1,2,...,d, the typical entry y.. of y is individual i's\nachievement in dimension j. The row vector\nyt =(yi\,yi2,----,yidS) contains individual i's achievements in\nthe different dimensions; the column vector\ny \u00E2\u0096\u00A0 = (y1 -,y2 \u00E2\u0096\u00A0,....,y\u00E2\u0080\u009E,-)' gives the distribution of achievements\n8\n Multidimensional Poverty in Bhutan\nin dimension /' across individuals. Let z, > 0 be the\nJ j\ndeprivation cutoff value (or poverty line) in dimension j.\nFoUowing Alkire and Foster (2007)'s notation, the sum of\nentries in any given vector or matrix v is denoted by \v\,\nwhtie /u(v) is used to represent the mean of v (or | v\ divided\nby the number of entries in v).\nFor any matrix y, it is possible to define a matrix of\ndeprivations g = [g. ], whose typical element g.. is defined\nby gtj = 1 when ytj < z \u00E2\u0096\u00A0, and gtj = 0 when ytj > z \u00E2\u0096\u00A0. That is,\nthe ij entry of the matrix is 1 when person i is deprived in\ndimension j, and 0 when he/she is not. From this matrix,\ndefine a column vector of deprivation counts, whose ith entry\nc. =| gt | represents the number of deprivations suffered by\nperson i. If the variables in y are cardinal, then a matrix of\nnormalised gaps g = [g. ] can be defined, where the typical\nelement g]}.= (z}.- yi}.) / z}. when yij 0,\nwhose typical element g\" is the normalised poverty gap\nraised to the a-power.\nThe methodology to identify the multidimensionaUy poor\nproposed by Alkire and Foster (2007) compares the number of\ndeprivations with a cutoff level k When each selected\ndimension has the same weight, the possible values of k go in\nthe range of k\u00E2\u0080\u0094 \ ,d. However, the methodology also\naUows other weighting systems, which will be explained at the\nend of the section. In general, for any weighting system, let\npk be the identification method such that pk (yi, z) = 1 when\n Journal of Bhutan Studies\nc. > k , and pk (yt, z) = 0 when ct < k . That means that an\nindividual is identified as multidimensionaUy poor if he/she is\ndeprived in at least k dimensions. This methodology is said to\nbe a dual cutoff method, because it uses the within dimension\ncutoffs z \u00E2\u0096\u00A0 to determine whether an individual is deprived or\nnot in each dimension, and the across dimensions cutoff k to\ndetermine who is to be considered multidimensionaUy poor. It\nis also presented as a counting approach, since it identifies\nthe poor based on the number of dimensions in which they\nare deprived. When equal weights are used, when k \u00E2\u0080\u0094 \ , the\nidentification criterion corresponds to the union approach,\nwhereas when k \u00E2\u0080\u0094 d, the identification criterion corresponds\nto the intersection approach. This identification criterion\ndefines the set of the multidrmensionally poor people as\nZk \u00E2\u0080\u0094 {/ : pk {yi; z) = 1} . Once identification is applied, a\ncensored matrix g (k) can be obtained from g by replacing\nthe ith row with a vector of zeros whenever pk {yi, z) = 0.\nMatrix ga (k) can be defined analogously for a > 0, with its\ntypical entry g\" (k) = g\" if i is such that c. > k, whtie\ng^ (k) = 0 if z is such that c. < k .\nA first natural measure to consider is the percentage of people\nthat are multidimensionaUy poor: the multidimensional\nHeadcount Ratio H = H(y,z) defined byH = ql n , where q is\nthe number of people in set Zk. This measure is the\nanalogous to the unidimensional Headcount Ratio, and it has\nthe advantages that it is easy to compute and understand,\nand that it can be calculated with ordinal data. However, it\nsuffers from the disadvantages first pointed by Watts (1969)\nand Sen (1976) for the one-dimensional case, namely, being\ninsensitive to the depth and distribution of poverty, violating\nmonotonicity and the transfer axiom. Moreover, in the\nmultidimensional context, it also violates what Alkire and\n10\n Multidimensional Poverty in Bhutan\nFoster (2007) call dimensional monotonicity: if a poor person\nbecomes deprived in an additional dimension (in which\nhe/she was not previously deprived), Hdoes not change.\nConsidering this, Alkire and Foster (2007) propose the\ndimension adjusted FGT measures, given by\nMa (y; z) = ju(ga (k)) for a > 0 . When a = 0, the measure is\nthe Adjusted Headcount Ratio, given by\nM0 = ju(g (k)) = HA, which is the total number of\ndeprivations experienced by the poor (| c(^) |=| "Periodicals"@en .
"Bhutan"@en .
"dhim_jbs_18_01_01"@en .
"10.14288/1.0365186"@en .
"English"@en .
"Vancouver : University of British Columbia Library"@en .
"Thimphu : Centre for Bhutan Studies"@en .
"Images provided for research and reference use only. Permission to publish, copy or otherwise use these images must be obtained from the Digitization Centre: http://digitize.library.ubc.ca/"@en .
"Bhutan--Periodicals"@en .
"Bhutan--History"@en .
"Bhutan--Civilization"@en .
"Multidimentional Poverty in Bhutan: Estimates and Policy Implications"@en .
"Text"@en .