NEW FRONTIERS IN POVERTY MEASUREMENT

Transcription

1 NEW FRONTIERS IN POVERTY MEASUREMENT James E. Foster George Washington University and OPHI, Oxford World Bank March 14, 2013

2 New Frontiers Demand for new tools Country Demand Policy Driven Today s Topics Ultrapoverty Hybrid Poverty Lines Chronic Poverty Multidimensional Poverty Theme Central Role of Identification of the Poor

3 Traditional Poverty Measurement Variable Single dimensional indicator Identification Poverty line Aggregation FGT (1984) Example Incomes y = (7,1,4,8) Poverty line z = 5 Deprivation vector g 0 = (0,1,1,0) Headcount ratio P 0 (y;z) = m(g 0 ) = 2/4 Normalized gap vector g 1 = (0, 4/5, 1/5, 0) Poverty gap = P 1 (y;z) = m(g 1 ) = 5/20 Squared gap vector g 2 = (0, 16/25, 1/25, 0) FGT Measure = P 2 (y;z) = m(g 2 ) = 17/100

4 Measuring Ultrapoverty There are great differences among the poor General idea behind P 1 and P 2 but not P 0 Greater depth matters Who are the ultrapoor? How to measure ultrapoverty? Several possibilities Deeply Deprived Persistently Deprived Multiply Deprived Spatially Concentrated Deprivation Work in progress: Foster and Smith (2013)

5 Measuring Ultrapoverty Why not simply use P(x;z u ) for a very low z u? Can give misleading picture Although it identifies ultrapoor Aggregation at lower line can reduce measured poverty Separate identification and aggregation Use z u to identify and z to aggregate Resulting P u (x;z u ) satisfies all axioms (including focus)

6 Measuring Ultrapoverty Example Incomes y = (7,1,4,8) Poverty line z = 5 and Ultrapoverty line z u = 3 Deprivation vector g u 0 = (0,1,0,0) Headcount ratio P u0 (y;z) = m(g u0 ) = 1/4 Normalized gap vector g u 1 = (0, 4/5, 0, 0) Poverty gap = P u1 (y;z) = m(g u1 ) = 4/20 Squared gap vector g u 2 = (0, 16/25, 0, 0) FGT Measure = P u2 (y;z) = m(g u2 ) = 16/100 Contribution of ultrapoor to overall poverty? Headcount: 1/4 out of 1/2; Poverty gap: 4/20 out of 5/20 FGT: 16/100 out of 17/100 Suggests FGT already closely tracks ultrapoor

7 Hybrid Poverty Lines Absolute poverty lines z a such as $1.25/day Unchanging over time and space Hence 0 elasticity of poverty line wrt income Useful for comparing countries at similar levels of development for a window of time However make little sense for evaluating poverty across countries or regions at very different levels of development are fundamentally unsustainable over time ie have problems measuring poverty over time and space Ex Is growth good for the poor? Foster and Szekely (2008)

8 Hybrid Poverty Lines Relative poverty line z r such as 50% of mean income Changes with the standard of living Elasticity of poverty line wrt income is 1 However this seems too responsive Wind up measuring inequality, not poverty Empirical evidence: elasticity is between 0 and 1 Citro and Michael (1995) What is available for measuring poverty over space and time? Foster (1998)

9 Hybrid Poverty Lines Hybrid poverty line z = z r ρ z a 1-ρ for 0 < ρ < 1 where z r is a relative poverty line z a is an absolute poverty line ρ is the elasticity of the poverty line with respect to income The elasticity can be estimated Foster and Szekely (2013) for LAC or selected arbitrarily and subjected to robustness Madden (2000) for Ireland or seen as a normative decision To what extent should the poor share in growth?

10 Hybrid Poverty Lines Poverty has two components absolute persons below absolute poverty line relative or hybrid persons below hybrid line above abs. With different policy prescriptions as in the previous discussions of ultrapoverty Other approaches Atkinson and Bourguignon (2001) use a max function Ravallion and Chen (2011) alter A&B to avoid range where elasticity is 1

11 Chronic Poverty Note Previous exercises altered identification The next two alter the variable and identification Chronic Poverty (across many time periods) Multidimensional Poverty (across many dimensions of wellbeing) Closely linked Must have data linked across time or dimensions Must decide how to value different periods income or different dimensions Identification becomes more difficult Foster (2006) and Alkire-Foster (2011)

12 Chronic Poverty First Approach: Components Jalan-Ravallion Identify as chronically poor those whose incomes are on average below a poverty line z Aggregate using FGT applied to distribution of average incomes Assumes Equal weights across periods hence perfect substitutes First aggregate across periods, then see if chronically poor Foster-Santos (2013) Use method of averaging across periods that allows for imperfect substitutability

13 Chronic Poverty Second Approach: Spells Foster (2009) Identify as chronically poor those whose incomes are frequently below the poverty line (eg 2 out of 4 periods) Aggregate using FGT applied to matrices in which the nonpoor spells have been censored out Assumes No substitution of incomes across periods better for consumption? Indeed incomes are not aggregated Instead check how many periods deprived Aggregate spells across periods Each spell has the same value Application: Dercon and Porter (2011)

14 Multidimensional Poverty Everyone agrees poverty is multidimensional Real question is what to do about it. How to measure poverty when there are many variables or dimensions of wellbeing?

15 Multidimensional Poverty Suppose many variables or dimensions Question How to evaluate poverty? Answer 1 If variables can be meaningfully aggregated into some overall resource or achievement variable, traditional methods can be used. Key Aggregation method across dimensions!

16 Multidimensional Poverty Examples Welfare aggregation Construct each person s welfare function Set cutoff and apply traditional poverty index However Many assumptions needed to construct Alkire and Foster (2010) Designing the Inequality-Adjusted Human Development Index Ordinal variables problematic

17 Multidimensional Poverty Examples Price aggregation Construct each person s expenditure level Set cutoff and apply traditional poverty index However Many assumptions and prices needed to construct Ordinal and nonmarket variables problematic Link to welfare tenuous (local and unidirectional) Foster, Majumdar, Mitra (1990) Inequality and Welfare in Market Economies JPubE

18 Multidimensional Poverty Note Even if an aggregate exists, it may not be the right approach Idea Aggregate resource approach signals what could be The budget constraint Does not indicate what is The actual bundle purchased Ex Consumption poverty is falling rapidly in India Yet 45% of kids malnourished Question Is aggregating before identification hiding policy relevant information?

19 Multidimensional Poverty Suppose many variables or dimensions Question How to evaluate poverty? Answer 2 If variables cannot be meaningfully aggregated into some overall resource or achievement variable, or in cases where policy demands an accounting of individual deprivations, new methods must be used

20 Multidimensional Poverty Some go to great lengths to avoid this fact: Blinders approach Let s just limit consideration to a subset of dimensions that can be aggregated, and use traditional methods. Key dimensions ignored Marginal methods Let s only look at a dashboard where we separately apply traditional methods to each dimension. Ignores joint distribution Where did identification go?

21 Alkire-Foster Methodology: Overview Identification of poor Dual cutoffs Deprivation cutoffs - each deprivation counts Poverty cutoff - in terms of aggregate deprivation values Aggregation across the poor Adjusted FGT Reduces to FGT in single variable case Background papers Alkire and Foster (2011) Counting and Multidimensional Poverty Measurement Journal of Public Economics Alkire and Foster (2010) Understandings and Misunderstanding of Multidimensional Poverty Journal of Economic Inequality Alkire, Foster, and Santos (2011) Where Did Identification Go? Journal of Economic Inequality

22 Adjusted Headcount Ratio Concept - Poverty as multiple deprivations Mirrors identification used by NGOs BRAC Depends on joint distribution Ordinal data Dirt floors vs covered floors Qualitative data into quantitative data Transparent Defined by variables, deprivation cutoffs, deprivation values, poverty cutoff Can be replicated and tested for robustness

23 Adjusted Headcount Ratio Can be implemented at many levels Cross country MPI in the HDR s since 2010 Country Mexico, Colombia, El Salvador, etc. State Sao Paolo, Minas Gerais Local village level DR, India, Bhutan Evaluation Impacts on poverty (Smith and Robano 2012) As a coordination tool Ministries in Colombia In constructing other measures Gross national happiness index (Bhutan), Women s Empowerment in Agriculture Index (USAID/IFPRI), Service delivery performance measure (Allwine and Foster, 2011), Corruption (Foster et al, WBER 2012)

24 Intro to: Multidimensional Methods Matrix of achievements for n persons in d equally important domains Domains Persons These entries fall below cutoffs z ( ) Cutoffs

25 Deprivation Matrix Replace entries: 1 if deprived, 0 if not deprived Domains Persons

26 Identification Dual Cutoff Approach Q/ Who is poor? A/ Fix cutoff k, identify as poor if c i > k (Ex: k = 2) Domains c Persons Note Includes both union and intersection Especially useful when number of dimensions is large Union becomes too large, intersection too small Next step - aggregate into an overall measure of poverty

27 Aggregation Censor data of nonpoor Domains c(k) Persons

28 Aggregation Headcount Ratio Domains c(k) Persons Two poor persons out of four: H = ½ incidence Critiques

29 Aggregation Adjusted Headcount Ratio Adjusted Headcount Ratio = M 0 = HA = m(g 0 (k)) = 6/16 =.375 Domains c(k) c(k)/d Persons A = average intensity among poor = 3/4 Note: if person 2 has an additional deprivation, M 0 rises

30 Aggregation Adjusted FGT Family Adjusted FGT is M a = m(g a (t)) for a > 0 Domains Persons However, need cardinally meaningful data for a > 0

31 Aggregation Adjusted Headcount Ratio Observations M 0 uses ordinal data Similar to traditional gap P 1 = HI HI = per capita poverty gap = headcount H times average income gap I among poor HA = per capita deprivation = headcount H times average intensity A among poor Decomposable across dimensions after identification M 0 = j H j /d where H j are censored headcount ratios Extends easily to the case where deprivations have different values

32 UNDP s MPI Published in Human Development Reports since 2010 for over 100 countries

33 33% Identification: Who is poor? A person is multidimensionally poor if they are deprived in 33% of the dimensions at the same time.

34 How do you calculate the MPI? The MPI uses the Alkire Foster method: Formula: MPI = M 0 = H A H is the percent of people who are identified as poor, it shows the incidence of multidimensional poverty. A is the average proportion of weighted deprivations people suffer at the same time. It shows the intensity of people s poverty the joint distribution of their deprivations.

35 Niger Ethiopia Mali Central African Republic Burundi Liberia Burkina Faso Guinea Rwanda Mozambique Sierra Leone Comoros DR Congo Uganda Malawi Benin Timor-Leste Senegal Madagascar Tanzania Nepal Zambia Chad Cote d'ivoire Gambia Bangladesh Haiti Togo Nigeria India Cameroon Yemen Cambodia Pakistan Kenya Lao Swaziland Republic of Congo Gabon Lesotho Sao Tome and Principe Honduras Ghana Djibouti Nicaragua Bhutan Guatemala Indonesia Bolivia Peru Viet Nam Tajikistan Mongolia Iraq Philippines South Africa Paraguay China Morocco Estonia Turkey Egypt Syrian Arab Republic Colombia Sri Lanka Azerbaijan Maldives Kyrgyzstan Dominican Republic Hungary Croatia Mexico Argentina Brazil Jordan Uzbekistan Ecuador Ukraine Macedonia Moldova Uruguay Thailand Latvia Montenegro Albania Russian Federation Armenia Serbia Bosnia and Herzegovina Georgia Kazakhstan Belarus Slovenia MPI vs $1.25/day Headcount Ratios Data Availability 103 of our 109 Countries have income/consumption; only 71 have income poverty data within 3 years of MPI. Income data ranges from ; MPI from

36 Intensity is highest in the poorest countries.

37 Most poor people live in middle-income countries More than twice as many poor people live in middle-income countries (1,189 M) compared to low-income countries (459 M). Total Population by Income Category (2008) 13% 45% 1% 41% 28% 0% 12% MPI Poor Population (2008) High-income Upper middle-income 60% Lower middle-income Low-income

38 Bihar In Kerala India 16% of the population is MPI poor; in Bihar it is 81%. Kerala India MPI The poorest 8 Indian states are home to more MPI poor people than the 26 poorest African countries (421M v 410M).

39 Income vs Multidimensional Why not focus on income/consumption? 1. Gives only part of the story True for all persons that other dimensions matter But especially for children, women Have their own obvious markers for wellbeing and poverty Would be fine to focus on one dimension alone if all deprivations overlapped but they don t

40 Multidimensional Poverty 2. Income/consumption tells only about budget set Not what is actually experienced Declining income poverty and persistent malnutrition in India Behavioral issues?

41 There are different compositions of poverty by dimensions & indicators Three countries: Zambia, Nigeria and Niger. MPIs are similar for Zambia (0.32) and Nigeria (0.37) although much higher for Niger (0.64). Income poverty ($1/day) is similar in the three (64-66%).

42 Niger is most deprived in Education 100% 90% 80% 7% 10% Zambia: relatively 70% more deprived 60% in LS 50% 40% 30% 20% 10% 0% 18% 9% 11% 10% 8% 9% 11% 12% 14% 19% 14% 6% 8% 9% 8% 8% 5% 9% 17% 18% 15% 7% 5% 7% Zambia Nigeria Niger 6% 6% 7% 8% Nigeria: relatively more deprived in Health and Education Years of Schooling Child Enrolment Child Mortality Nutrition Electricity Sanitation Water Floor Cooking Fuel Asset Ownership

43 Nigeria Overall Trends

44 Annualized Absolute Change in the Percentage Who is Poor and Deprived in Inside the Regions of Nigeria 2.0 Assets 1.0 Cooking Fuel 0.0 Flooring North Central North East North West South East South South South West Safe Drinking Water Improved Sanitation Electricity Nutrition Child Mortality Years of Schooling School Attendance

45 Multidimensional Poverty 3. Virtually all income/consumption data are a one period snapshot Cannot differentiate between temporary and deeper forms of poverty Assets are clearly key, but ignored See new paper by Luis Felipe Lopez Calva and colleagues: POVERTY DYNAMICS IN BRAZIL uses non-monetary components and our technology to measure Chronic poverty Country examples: Official poverty measures Mexico and Colombia

46 Multidimensional Poverty in Mexico Methodology & results First released December,

47 Wellbeing Income What are the main features of the new methodology? Current income per capita Population Six Social Rights: Education Territorial Health Social Security Housing Basic Services Deprivations Social Rights Food

48 Poverty Identification Income Economic wellbeing line $1,921.7 U $1,202.8 R $874.6 U $613.8 R EWL MWL Minimum wellbeing line With Deprivations Vulnerable by social deprivations MULTIDIMENSIONALLY POOR Moderate Multidimensional EXTREME Multidimensional Poverty Poverty Without D e p r i v a t i o n s Ideal Situation Vulnerable by income Deprivations Social Rights

49 Wellbeing Income Total Population 2008 Vulnerable by social deprivations 33.0% 35.2 millions 2.0 Deprivation average 18.3% 19.5 millions MODERATE POVERTY EXTREME POVERTY 10.5% 11.2 millions 3.9 Deprivation average 33.7% 36.0 millions 2.3 Deprivation average Vulnerable by income 4.5% 4.8 millions Deprivations Social Rights 0

50 Wellbeing Income Indigenous Population 2008 Vulnerable by social deprivations 20.0 % 1.4 millions 2.8 Deprivation average 3.1% 0.21 millions MODERATE POVERTY EXTREME POVERTY 39.2 % 2.7 millions 4.2 Deprivation average 36.5 % 2.5 millions 3.1 Deprivation average Vulnerable by income 1.2%.1 millions Deprivations Social Rights 0

51 Multidimensional Poverty Incidence Ranks Total of States [20% - 40%) 14 [40% - 60%) 14 [60% - 80%] 4 Source: CONEVAL estimations with information from MCS-ENIGH, 2008

52

53 Social Deprivations Food security Millions of people Income Poverty % 11.7 million % 11.7 million Education -2.3 Housing Basic Services Extreme income poverty Poverty % 48.8 million 0.0 Extreme poverty % 52.0 million Social Security Access to Health Care Fuente: estimaciones del CONEVAL con base en el MCS-ENIGH 2008 y 2010

54 Multidimensional Poverty Index for Colombia and its applications (MPI-Colombia) ROBERTO ANGULO YADIRA DÍAZ RENATA PARDO National Planning Department Division of Social Promotion and Quality of Life September 2011

55 The MPI-Colombia: Proposed by the National Planning Department based on the Alkire & Foster methodology Instrument for design and monitoring public policy Complements the income poverty measure Discussed with the Colombian academy and policy makers

56 Educational Conditions Childhood & Youth Dimensions, Variables and Weights MPI-Colombia Work Health Housing & Public Services Schooling Illiteracy 0.1 School Attendance At the right level Access to infant services Absence of long-term unemployment Formal work Coverage Access to health care given a necessity Improved Water Sanitation Flooring Exterior Walls No Child Labour Overcrowding

57 Results released 2011: Incidence (H) and Adjusted Headcount (M0) for k = 5/15 decreased Fuente: DNP, DDS, SPSCV. 2011

58 Poverty committee Coordinating and monitoring poverty reduction Leaders Counselor for the Presidency National Planning Department Permanent members Ministry of Health Ministry of Labor Ministry of Housing Ministry of Agriculture Ministry of Education Ministry of Finance MANDATORY PRESENCE The President of Colombia 58

59

60 If the Plan is accomplished, if every ministry does its job and effectively spends its committed resources, the MPI- Colombia will decrease to 22% (more than 3 million people lifted out of poverty).

61 Application: Women s Empowerment in Agriculture Index (WEAI) USAID, IFPRI, OPHI (Oxford) Measures by Alkire-Foster

62 Five domains of empowerment A woman s empowerment score shows her own achievements

63 Who is empowered? A woman who has achieved adequacy in 80% or more of the weighted indicators is empowered

64 How is the Index constructed? Five domains of empowerment (5DE) A direct measure of women s empowerment in 5 dimensions WEAI is made up of two sub indices 5DE = (1-M 0 ) GPI = (1-P 1 ) Women s Empowerment in Agriculture Index (WEAI) All range from zero to one; higher values = greater empowerment Gender parity Index (GPI) Women s achievement s relative to the primary male in hh

65 Formula 5DE = H e + H d A e = (1- H d A) H e is the percentage of empowered women H d is the percentage of disempowered women A e is the average absolute empowerment score among the disempowered GPI = H p + H w R p = (1- H w I) H p is percentage of women with gender parity H w is the percentage of women without gender parity R p is the women s relative parity score compared to men = (1-I) H e + H d = 100% H p + H w = 100%

66 Bangladesh Pilot results H e = 31.9% of women are empowered Disempowered women have adequate achievements in A e =60.7% of domains H p = 59.8% of women enjoy gender parity Households without gender parity have a I = 25.2% empowerment gap between the woman and man

67 Thank you