Leaving Poverty Behind – Using Multidimensional Poverty Measures to Tackle Poverty and Achieve the Sdgs in Asia and the Pacific

A Guidance Note

Empowered lives. Resilient nations.

Leaving poverty behind – Using multidimensional poverty measures to tackle poverty and achieve the SDGs in Asia and the Pacific

A Guidance Note

Empowered lives. Resilient nations.

Leaving poverty behind – Using multidimensional poverty measures to tackle poverty and achieve the SDGs in Asia and the Pacific A Guidance Note

Leaving poverty behind: using multidimensional poverty measures to tackle poverty and achieve the SDGs in Asia and the Pacific

This is the second in a series of two reports prepared to support UNDP country offices in Asia and the Pacific in their efforts to end poverty in all its forms, everywhere. The first report takes stock of the calculation and use of multidimensional poverty measures in countries of the region. This accompanying guidance note backs the practical application of multidimensional poverty measures as a policy tool.

ISBN: 978-974-680-433-2

Author: Bishwa Nath Tiwari, UNDP Bangkok Regional Hub

Editing: Gretchen Luchsinger

Design: www.minhdesigns.com

Published by: United Nations Development Programme Bangkok Regional Hub 3rd Floor, United Nations Service Building Rajdamnern Nok Avenue, Bangkok 10200, Thailand. Email: [email protected] Tel: +66 (0)2 304-9100 Fax: +66 (0)2 280-2700

Disclaimer:

The views expressed in this publication are those of the author and do not necessarily reflect the views of the United Nations, including the United Nations Development Programme or UN Member States. Reproduction and dissemination of material in this report for educational or other non-commercial purposes are authorized without prior written permission from the copyright holder, provided that the source is fully acknowledged.

iv Contents –

Lists of Figures and Boxes vi 1.3 Coordinating policies across sectors and levels of government 15 Acknowledgements vii 1.4 Developing multisectoral Acronyms viii programmes that reflect interconnected deprivations 15 Part I Background of 1.5 Tracking SDG target 1.2 on Multidimensional Poverty halving poverty in all its forms 15 Measurement and Applications 1 1.6 Informing resource allocation at the subnational level 16 What is multidimensional poverty? 1 2. Measuring multidimensional poverty Multidimensional poverty and from a census survey for targeting and the SDGs 4 tracking deprivations at local level 16 What is the Multidimensional 2.1 Targeting at the lower level of Poverty Index? 5 administrative units 16 Deprivations in Asia and the Pacific: 2.2 Targeting at the household Who has been left behind and why? 8 and individual levels for the How can multidimensional poverty effective implementation of measures help in ending poverty? 9 social protection 17 Development of a management Part II Multidimensional information system 19 poverty measures as policy Communication and advocacy 19 tools—fostering uses in Asia Institutional framework 20 and the Pacific 11 Conclusion 20 Current applications of References 22 multidimensional poverty measures 11 How to promote the uses of Endnotes 25 multidimensional poverty measures 14 1. Continue measuring the national MPI using national household surveys 14 1.1 Complementing income-based poverty measures 14 1.2 Setting national goals and targets 15

v Lists of Figures and Boxes –

Figures

Figure 1: Overview of the uses of multidimensional poverty measures 12 Figure 2: Identifying and monitoring deprived households 18

Boxes

Box 1: Well-being and poverty have many aspects 2 Box 2: Counting approaches: the Alkire-Foster method as a prominent measure of multidimensional poverty 4 Box 3: Asia and the Pacific—home of half of the multidimensionally poor people 7 Box 4: The DHS and MICS sample designs provide robust MPI estimates at higher levels of administrative units 13

vi Acknowledgements –

This guidance note was commissioned Valuable inputs and insights came from by UNDP and produced by the UNDP Sabina Alkire, Director of the OPHI, and Bangkok Regional Hub, Inclusive Growth Pedro Conceicao, Director of UNDP’s Team. It was written by Bishwa Nath Human Development Report Office in Tiwari. New York. At the UNDP Bureau for Policy and Programme Support in New York, The guidance note draws on previous Mansour Ndiaye, Renata Rubian and reports, including the Review of MPI Serge Kapto helped in sharing useful Measurement and Uses in Asia and information and in liaising with relevant the Pacific, produced by the UNDP colleagues from other bureaus. Bangkok Regional Hub, and How to Build a National Multidimensional Poverty A note of appreciation goes to Balazs Index, prepared by UNDP and the Oxford Horvath for a peer review of the report, Poverty and Human Development and to Scott Standley, Hannie Meesters, Initiative (OPHI). James Chaco and Jaco Cilliers from the Bangkok Regional Hub for their oversight The document benefited from reviews and support. and valuable inputs from colleagues in UNDP’s Regional Bureau for Asia and Cedric Monteiro and Mahtab Haider from the Pacific Economist Network, including the Bangkok Regional Hub have been Balazs Horvath (UNDP China), Matthew kind enough to review the guidance note Johnson-Idan (Pacific Office), Richard and develop communications briefs to Marshall (UNDP Cambodia), Swastik Das promote it. Szewun Tong, a UN Volunteer, (UNDP India), Tandin Wangchuk (UNDP supported a review of literature. Bhutan), Haniza Khalid (UNDP Malaysia) and Rima Artha (UNDP Indonesia). Thanks are due to Phatsaraphorn Champasri from the Inclusive Growth Outside the region, valuable comments Team of the Bangkok Regional Hub, who were received from Ivan Gonzalez de supported the procurement of services Alba in the Regional Bureau for Latin for editing and publishing the document, America and the Caribbean, Mihail and to Supaporn Tampirak, who assisted Peleah in the Regional Bureau for Europe in ensuring the quality of the report, and Central Asia, and Elena Danilova- including through organizing a peer Cross at UNDP’s Istanbul Regional Hub. review.

vii Acronyms –

DHS Demographic and Health Survey GDP Gross domestic product MICS Multiple Indicator Cluster Survey MPI Multidimensional Poverty Index MPPN Multidimensional Poverty Peer Network OPHI Oxford Poverty and Human Development Initiative PPP Purchasing power parity SDGs Sustainable Development Goals UNDP United Nations Development Programme UNICEF United Nations Children’s Fund

viii Part I – Background of Multidimensional Poverty Measurement and Applications

What is income continues to be a widely used measure of poverty, it cannot capture multidimensional the full spectrum of deprivations poverty? that make people poor. Monetary measures are therefore not enough to guide policies to eradicate poverty as Poverty is a complex, multidimensional envisaged by Sustainable Development phenomenon encompassing both Goal (SDG) 1, on ending all forms of income and non-income deprivations. poverty everywhere, or to contribute Apart from the paucity of income, to the achievement of all other SDGs. various non-income deprivations that The goals require a multidimensional can affect people’s well-being include approach that includes non-income inadequate education, lack of health deprivations. care, low-quality housing, lack of water and sanitation, or work in a hazardous There are at least three reasons to environment, among others. Although use a multidimensional approach to

1 measuring poverty. First, poverty and At the international level, the Stiglitz- well-being are multidimensional. This Sen-Fitoussi Commission in 2008 reality has been increasingly accepted and the Atkinson Commission in 2016 both nationally and internationally. described well-being and poverty At the national level, the happiness as multidimensional (Box 1). The approach, with multiple dimensions and latter recommended that the global indicators, has advanced in countries Multidimensional Poverty Index (MPI) like Bhutan and been endorsed by the be used to complement the income United Nations General Assembly.1 poverty line of US $1.90 per day New Zealand’s move towards a new (purchasing power parity or PPP). “well-being budget” that emphasizes citizen well-being over economic A second important reason to use growth is another noteworthy effort.2 a multidimensional approach is the commitment to the SDGs. With the adoption of the 2030 Agenda for

Box 1: Well-being and poverty have many aspects

The Stiglitz-Sen-Fitoussi Commission Later, the Atkinson Commission constituted by former French created by the World Bank argued President Nicolas Sarkozy held that that poverty is multidimensional well-being is multidimensional. and suggested a non-monetary Based on academic research and a multidimensional poverty measure number of concrete initiatives, the covering the following dimensions: Commission identified the following nutrition, health status, education, key dimensions that should be housing conditions, access to taken into account: material living work and personal security. The standards (income, consumption Commission also recommended and wealth), health, education, complementing income poverty personal activities including work, measures with a global MPI. The political voice and governance, World Bank subsequently accepted social connections and relationships, some of the recommendations of environmental conditions (present the Atkinson Commission. and future), and economic as well as physical insecurity.

Source: Stiglitz et al. 2009, World Bank 2017.

2 Sustainable Development at the United context. To distinguish between the Nations in 2015, 193 UN Member States two, this guidance note refers to the committed to achieving 17 global goals global MPI and national MPI. It focuses by 2030, promising to “leave no one primarily on the latter. behind” and to “endeavour to reach the furthest behind first”. SDG 1 addresses The main motivation for producing multidimensional poverty, aiming to the guidance note is to foster the “end poverty in all its forms everywhere”. application of multidimensional poverty Under the goal, target 1.2 is to reduce measures, in particular the MPI, as at least by half the proportion of men, policy tools in Asia and the Pacific. In women and children of all ages living in line with UNDP’s global launch of its 5 poverty in all its dimensions according national MPI handbook in 2019, the to national definitions. Progress organization’s Bangkok Regional Hub on this target depends on taking a reviewed the measurement and uses multidimensional approach to assessing of multidimensional poverty measures, and tracking poverty. in particular the MPI, in Asia and the Pacific.6 The review found that more Finally, a third reason for taking a than half of the countries in Asia have multidimensional approach is to been measuring multidimensional measure non-income deprivations, poverty, mainly through the MPI.7 But such as child mortality, primary school only a few countries are actively using completion rates or undernourishment, multidimensional poverty measures for that do not highly correlate with levels developing policies and programmes, and trends of income poverty.3 By a and defining resource allocations specific income threshold, a person can explicitly to target poverty.8 be income poor, but not necessarily multidimensionally poor. Or he/she can This guidance note complements the be income rich, yet multidimensionally national MPI handbook and builds on poor.4 the findings of the regional review. It has two main parts. The first part To measure multidimensional poverty, introduces the MPI, followed by a brief the United Nations Development note on deprivations in Asia and the Programme (UNDP) and the Oxford Pacific, and how the MPI can serve Poverty and Human Development as a policy tool to address these. The Initiative (OPHI) launched the MPI in second part offers some guidelines to the 2010 Human Development Report, foster the uses of multidimensional allowing comparison across countries. poverty measures for planning and Several countries have subsequently programming, and for targeting at local adapted the MPI to their national level.

Part I – Background of Multidimensional Poverty Measurement and Applications 3 Multidimensional gender equality (SDG 5), clean water and sanitation (SDG 6), and potentially poverty and the SDGs reduced inequalities (SDG 10). In turn, eradication of all forms of poverty also The 17 SDGs are strongly requires achieving good health, quality interconnected. The first SDG on education, clean water and sanitation, ending all forms of poverty has and reduction of inequality, among links with all of the other goals.9 For other advances. example, progress in achieving SDG 1 has direct and indirect positive As other examples of links, promoting impacts on health and well-being economic growth, full employment (SDG 3), quality education (SDG 4), and decent work for all (SDG 8) are

Box 2: Counting approaches: the Alkire-Foster method as a prominent measure of multidimensional poverty

The most widely used most widely used income poverty multidimensional poverty measures measure. The resulting measure since the 1970s have been aggregates information to reflect ‘counting approaches’, with most poverty in a way that is robust, and applications using a headcount can be broken down by regions ratio. While a headcount is easy to and groups. It shows who is poor, understand and communicate, it where they live, and the dimensions does not provide enough motivation and indicators in which they are for policymakers to focus on deprived. reducing the deprivations of the poorest of the poor. To identify the poor, the Alkire- Foster method counts overlapping In 2007, Sabina Alkire and James or simultaneous deprivations that Foster created a new method a person or household experiences for measuring multidimensional in various indicators of poverty. poverty. It uses a counting The indicators might have different approach to identify ‘who is weights. People are identified poor’ by considering the range of as multidimensionally poor if the deprivations that people suffer, weighted sum of their deprivations combined with the Foster-Greer- is greater than or equal to a poverty Thorbecke methodology, the cut-off.

4 necessary for poverty eradication. What is the Building resilience to environmental shocks and sustainably managing Multidimensional environmental resources (SDGs Poverty Index? 12 to 15) can reduce vulnerabilities to poverty. Peaceful and inclusive The MPI is part of an influential class of societies (SDG 16) depend on poverty measures advanced by Alkire strengthening social cohesion through and Foster based on the ‘counting’ forging a sound social contract and approach (Box 2).10 Since the launch of reducing poverty and inequalities. the MPI in 2010, UNDP and OPHI have measured a global version of the index for more than 100 countries.

Having identified who is poor, between one-fifth and one-third the Alkire-Foster method then of the cut-off (between 0.20 summarizes information to to 0.33) are near-poor who are show the deprivations the poor vulnerable to poverty. experience as a proportion of all •• Intensity of poverty (A): the possible deprivations. The simplest average percentage of (weighted) measure in the class of Alkire- indicators in which poor people Foster measures is the MPI, which are deprived. In simple terms, is the product of two components: this means how intense •• The incidence of poverty (H): the multidimensional poverty is, on percentage of people who are average, for those who are poor. identified as multidimensionally poor, or the multidimensional This product (H x A) is called the adjusted headcount ratio, or M in poverty headcount. The poverty 0 line or cut-off for the global MPI is the Alkire-Foster method, or the one-third or more of the weighted MPI. The measure is rigorous, easy to disaggregate or to ‘unpack’ down indicators (c ≥0.33). This also includes those who are in severe to the indicator level, and useful for poverty or deprived in half or policy. It is flexible, which makes it more than half of the weighted adaptable to different contexts. indicators (c ≥0.5). Those lying

Source: Alkire and Foster 2007, 2009 and 2011.

Part I – Background of Multidimensional Poverty Measurement and Applications 5 MPI measurement entails at least the prioritize those furthest behind. An MPI following steps: closely aligned with the SDGs:

•• Select dimensions and indicators •• Measures poverty in multiple forms •• Set deprivation cut-offs for each and dimensions. indicator •• Complements monetary poverty •• Set weights for each dimension or measures. indicator •• Addresses multiple SDGs and their •• Decide the poverty cut-off indicators, simultaneously. •• Compute the incidence and intensity •• Directly reflects interlinkages across of poverty or the MPI indicators at the household level. •• Can be disaggregated by region, Details on these steps appear in the age, disability status, and other national MPI handbook.11 Four factors sociodemographic characteristics to that are critical for measurement are the identify the poorest, and to leave no number of dimensions and indicators, one behind. the indicator cut-off, indicator weights •• Can be nationally specific or and the aggregate poverty cut-off. For internationally comparable. A national example, a higher weight given to MPI reflects poverty according to indicators in which most households are national definitions but cannot be deprived will produce a higher poverty compared across countries. Regional rate. Similarly, a lower poverty cut-off and global MPIs permit comparisons (above which people are deprived) and cross-learning across countries. produces a higher poverty rate. If the •• Can be used as a tool for improved poverty cut-off decreases from 33 governance.13 percent to 20 percent of the weighted indicators, for example, the proportion of For policymakers, the MPI provides people who are deprived in 20 percent strong evidence of the causes of of indicators will be higher than those poverty and deprivation, and how each deprived in a third of the indicators. dimension and indicator contributes to Through the selection of dimensions overall poverty. The most useful feature and indicators, a country’s MPI can of the Alkire-Foster method is that it reflect key poverty-related SDG allows the addition of any dimension priorities. It can guide the response or indicator as well as disaggregation to the SDG call for integrated, by geographical areas, population multisectoral policies, and for groups and indicators. This supports the disaggregating indicators by regions, formulation of context-specific national and population groups to identify and development plans.

6 Since the MPI provides both the associated with traditional ‘headcount headcount and intensity of poverty only’ poverty measures that can result in a single measure, it can indicate in moving people from just below ways to reduce poverty by both to just above the poverty line, while decreasing the number of poor people leaving the poorest behind.14 and by alleviating their deprivations. For example, five deprivations of the The global MPI provides disaggregated poor can be cut to four or less and statistics on the main contributors finally to zero over time. The measure to poverty in three equally-weighted of intensity also helps planners to dimensions: health, education and focus on households or individuals left standard of living. Ten indicators include furthest behind, and fulfil one of the nutrition, child mortality, years of major pledges of the 2030 Agenda. schooling, school attendance, drinking water, sanitation, cooking fuels, Capturing reductions in both the electricity, housing and assets. As the incidence and intensity of poverty global MPI compares multidimensional is one of the MPI’s key strengths, poverty across countries (Box 3), the creating useful incentives to reduce dimensions and indicators have been both the number of poor people and the same since 2010, though with the number of deprivations. The index some refinement in the specification of thus overcomes a well-known problem indicators.

Box 3: Asia and the Pacific—home of half of the multidimensionally poor people

In 2019, the global MPI was Of the 1.3 billion people in poverty, calculated for 101 countries. half were children under age 18, It indicated that 1.3 billion and a third were under age 10. people—23.1 percent of the total Two-thirds of multidimensionally population—are multidimensionally poor people lived in middle-income poor with overlapping deprivations countries. In 10 selected countries in health, education and living where changes were analysed over standards. About 661 million time, deprivations declined faster (49.9 percent) were in Asia and among the poorest 40 percent of the Pacific, with 548 million (41.3 the population than among the total percent) in South Asia alone. population.

Source: UNDP and OPHI 2019, UNDP 2019a.

Part I – Background of Multidimensional Poverty Measurement and Applications 7 Unlike the global MPI, a national MPI •• South Asia alone was home to can have different dimensions and almost half of the world’s population indicators as well as a different poverty of illiterate people in 2016. cut-off, although several countries have •• An estimated 673 million people (9 adhered, to the extent possible, to the percent of the global population) still global index’s three dimensions and 10 practised open defecation in 2017, indicators as well as the poverty cut-off. the majority of them in South Asia. One distinct difference between the •• In 2018, 46 percent of young national and global MPIs is that the women in Central and South Asia, latter is purely a non-income measure, 22 percent in East and South-east whereas national MPIs may not be. Asia and 16 percent in Oceania Some countries, such as Malaysia were not in education, training or and Mexico included income in their employment, compared to the global national MPI, but most countries find it average of 30 percent. more convenient and more powerful to have distinct, complementary poverty •• 12 percent of employed people measures.15 in Central and South Asia and 20 percent in Oceania were living below $1.90 a day in 2018, higher than the Deprivations in Asia global poverty rate of 8 percent. and the Pacific: Who •• Only 16 percent of people in Central and South Asia and 59 has been left behind percent in East and South-east Asia were covered by at least one and why? social protection measure in 2016, compared to the global average of Despite significant progress on poverty 45 percent. reduction and human development •• Globally, the share of national output throughout Asia and the Pacific,16 some used to remunerate workers is 661 million women, men and children declining, implying rising inequality. still live in multidimensional poverty.17 In 2017, labour received only 45.8 The region lags behind on several percent of GDP in Central and South fronts, including: Asia, compared to 51.4 percent •• 38 percent of children under 5 in globally. Oceania, 32 percent in Central and •• Of the over 1 billion people living in South Asia, and 13 percent in East urban slums globally in 2018, 370 and South-east Asia were stunted million were in East and South-east (inadequate height for age) in 2018, Asia, and 227 million in Central and compared to 22 percent globally. South Asia.

8 •• In 2018, only 21 percent of people in The link between economic growth and Oceania, 39 percent in Central and non-monetary dimensions of poverty South Asia, and 49 percent in East is less strong and even non-existent in and South-east Asia had convenient some cases.19 If income is not used for access to public transport, compared well-being, overlapping deprivations in to 53 percent globally. health, education, assets, livelihoods •• Only 64 percent of mountainous and other areas prevail, requiring areas in Central and South Asia programmes that specifically target and 71 percent in East and South- these non-income deprivations to 20 east Asia were covered by forests, achieve all forms of poverty reduction. grasslands or shrublands, and With often wide development gaps croplands in 2017, compared to 76 remaining in Asia and the Pacific, percent globally on average. governments have a strong incentive to invest in non-monetary components These wide gaps in SDG-related of well-being and expand inclusive priorities leave a large chunk of social policy measures, in line with the the population in Asia and the aspirations of the 2030 Agenda and the Pacific vulnerable to a complex commitment to leave no one behind. mix of challenges stemming from Progress depends on policy tools such socioeconomic deficits and risks as as multidimensional poverty measures well as environmental deterioration. to effectively target programmes to poor people and to monitor their How can deprivations. Multidimensional poverty measures such as the MPI multidimensional can guide investments in building the poverty measures help human capabilities of poor individuals as well as in improving housing, in ending poverty? assets and infrastructure for poor households and communities. Since Eradicating extreme poverty remains multidimensional poverty measures a global challenge. Half of the world’s can be disaggregated across areas, nations are not on track for achieving dimensions or population groups, they SDG 1.18 Economic growth can can help coordinate policies across contribute to poverty alleviation, as sectors and levels of government.21 shown by many studies. But it will not address multidimensional poverty independent of other interventions.

Part I – Background of Multidimensional Poverty Measurement and Applications 9 10 Part II – Multidimensional poverty measures as policy tools— fostering uses in Asia and the Pacific

Current applications multidimensional poverty measures can help in reforming policies and of multidimensional institutions and directing resource poverty measures allocation (Figure 1). In supporting regular tracking of progress, they can also help in phasing out programmes, Multidimensional poverty measures as has been the case in Viet Nam, can be used in each phase of a where an annual review assesses planning cycle, from initial design people in overlapping poverty and the of programmes, which requires need for State support.22 the identification of beneficiaries or target groups and decisions Effective use of a multidimensional on programme components, poverty measure at the local level to assessments of programme depends on the level of disaggregation. effectiveness and impacts during and For example, when an MPI is at the end of implementation. Using computed at the district level, it can

11 Figure 1: Overview of the uses of multidimensional poverty measures

Multidimensional poverty measures

Planning Implementation Evaluation

Management information system

Analysis, communication and advocacy

OBJECTIVES Reform in policy, institutions and resource allocation

help in identifying deprived districts, surveys such as Pakistan’s Social and accordingly inform choices about and Living Standards Measurement resource allocation and programmes for Survey or Viet Nam’s Household Living districts most in need. Standards Survey. National sample surveys are designed to provide Global and national MPIs currently use robust estimates of indicators only sample survey data, and are limited down to the regional or provincial to providing estimates at the national level or at most the district level in level, or for rural and urban areas, and some countries, however (Box 4). This for the subnational level to a limited constrains MPI measurement at a degree. Data typically come from lower level of administrative units and national household surveys such as limits its wider application as a policy the Demographic Health Survey (DHS) tool. and Multiple Indicator Cluster Survey (MICS) or other national household

12 Box 4: The DHS and MICS sample designs provide robust MPI estimates at higher levels of administrative units

Since 1984, the DHS Program has The United Nations Children’s Fund provided technical assistance to (UNICEF) launched the MICS as an more than 400 surveys in over integrated survey over two decades 90 countries. This has advanced ago, and has carried out more than global understanding of health and 300 surveys in over 100 countries population trends in developing including Pakistan and Thailand, countries by collecting nationally generating data on key indicators representative data on fertility, family of the well-being of children and planning, maternal and child health, women, and helping shape policies gender, HIV/AIDS, malaria and for improving their lives. In 2019, nutrition. The programme operates in MICS surveys were conducted in 15 countries of South and South-east Fiji, Kiribati, Nauru, Samoa, Tonga Asia, including the Philippines and and Tuvalu, which will help ease data Pakistan. constraints in the Pacific.

The main objective of the 2017 DHS The 2017 MICS for Pakistan provided in the Philippines, for example, estimates for a large number of was to provide current estimates indicators on the situation of women on fertility levels, marriage, sexual and children at the provincial level, activity, family planning, nutrition, in urban and rural areas, and in 3 childhood mortality, maternal and regions (Gilgit, Baltistan and Diamer) child health, violence against women, and 10 districts. In Thailand, the and HIV/AIDS and other sexually primary objective of the 2015-2016 transmitted infections. It covered the MICS was to produce statistically national level, urban and rural areas, reliable estimates of indicators at and 17 administrative regions. the national level, for urban and rural areas, and for five regions The 2017-2018 DHS in Pakistan yielded (Bangkok, Central, North, North-east representative estimates at the and South). The survey included national level, and for urban and rural oversampling for 14 provinces to areas, four provinces (Punjab, Sindh, prepare separate reports for each. Khyber Pakhtunkhwa and Baluchistan), A multistage, stratified cluster two regions (Azad Jammu and sampling approach was used for the Kashmir, and Gilgit Baltistan), the selection of the survey sample. Islamabad Capital Territory and the Federally Administered Tribal Areas. In total, there were 13 second-level survey domains. Source: USAID n.d. and UNICEF n.d.

Part II – Multidimensional poverty measures as policy 13 tools—fostering uses in Asia and the Pacific How to promote 1. Continue measuring the uses of the national MPI using multidimensional national household poverty measures surveys

To foster the deeper and broader The first approach adheres to the current application of multidimensional practice of measuring multidimensional poverty measures, current MPI poverty based on national household measurement practices should extend survey data. This approach can be used beyond the provision of national and at the national level for: provincial estimates, and identify overlapping deprivations at a much 1.1 Complementing disaggregated level, including at the household level, as has been done income-based poverty in Viet Nam. Most social protection measures and poverty programmes require data on households and their members. Measures of income poverty and Yet MPI measurement based on a multidimensional poverty can national sample survey is generally not complement each other. Countries useful for targeting social protection such as Afghanistan, Bhutan, Malaysia, or other State support programmes Nepal, Pakistan, Philippines, Thailand to individuals or households, given and Viet Nam have successfully used 23 data limitations. Therefore, alternative this approach. The midterm review th ways of collecting data and identifying of the 11 plan of Malaysia (2016- deprivations should be pursued. This 2020) advanced several initiatives to guide suggests two complementary increase the income and purchasing approaches for multidimensional power of the bottom 40 percent of poverty measurements, described households based on income (the so- on the following pages. These can be called B40) and reduce socioeconomic implemented sequentially, if necessary. inequalities. These initiatives are aimed at enhancing the capability of households to increase their employability, boost entrepreneurship and uplift productivity. Accompanying measures to enhance social protection are intended to cushion households

14 facing socioeconomic uncertainties. estimates. This can guide both vertical Using the MPI strengthens the delivery and horizontal coordination, including and monitoring of these initiatives.24 across sectoral ministries of education, health, employment and housing, as in Colombia.28 1.2 Setting national goals and targets 1.4 Developing Only income poverty used to feature multisectoral programmes as the development target of developing countries. In recent years, that reflect interconnected a multidimensional poverty headcount deprivations target has started appearing in national The decomposition of the MPI by development plans, including in Bhutan, dimensions and indicators provides Nepal, Pakistan, the Philippines and Viet useful information for formulating, Nam, among others. The Philippines’ implementing and monitoring 2014-2016 plan pledged to reduce the integrated programmes. Tracking incidence of multidimensional poverty implementation can help in course by 16 to 18 percent.25 The Government corrections and improvements. of Nepal aims to bring down extreme The proportion of overall poverty poverty (PPP $1.90 per day) to 4.9 explained by a dimension can percent and multidimensional poverty provide an indication of the share of to 10 percent of the population the budget to be allocated to that by 2030.26 The recently approved particular dimension compared to other National Framework for Sustainable dimensions or sectors. Development Goals in Pakistan set a federal target for reducing multidimensional poverty to 19 percent 1.5 Tracking SDG target 1.2 of the population by 2030, down from on halving poverty in all its the current level of 38.8 percent.27 forms

1.3 Coordinating policies Some countries have started using a across sectors and levels national MPI to track indicator 1.2.2 of SDG target 1.2, which seeks to reduce of government at least by half the proportion of men, women and children of all ages living in A national MPI can help coordinate poverty in all its dimensions, according policies across sectors and levels of to national definitions, by the end of government, based on disaggregated

Part II – Multidimensional poverty measures as policy 15 tools—fostering uses in Asia and the Pacific 2030. For example, Pakistan’s national targeting communities, households SDG framework adopts its 2014-2015 and individuals for social protection national MPI estimate as the base and other essential forms of support. value for tracking indicator 1.2.2.29 The This complementary approach could Government plans to compute the MPI draw on novel ways to gather data and every other year, with the next round of information at the local level, as has measurement scheduled for 2020. been done in Mexico and Viet Nam. These countries have collected data on multidimensional poverty indicators 1.6 Informing resource from all households in selected poor allocation at the communities and identified their subnational level deprivations. Taking such an approach involves the following broad steps National MPIs have served as tools for to identify and target communities, resource allocation at subnational levels households and individuals: in countries such as Bhutan, Pakistan and Viet Nam. Bhutan, for example, has used an MPI since 2013 as one of five 2.1 Targeting at the lower criteria for allocating national resources level of administrative units to local governments. The MPI is given high importance with a 45 percent Targeting at the village, intradistrict weight.30 or block level should start with an analysis of evidence presented by the national MPI. Depending upon budget 2. Measuring availability, some districts with high levels of multidimensional poverty can multidimensional poverty be chosen for targeted assistance. from a census survey Once these are selected, a census survey can be conducted if the districts for targeting and tracking are small. Otherwise, they can be deprivations at local level divided into subdistricts or blocks to identify the poorest of these.31 While the current practice of MPI Measurement and identification can measurement should continue to take place in at least two ways: provide data for planning and monitoring (i) Integrating data sources. Integrating at the national level as described in the two or more data sources can preceding section, a complementary provide a disaggregated estimate mechanism should be designed for of deprivation on some indicators

16 of well-being or deprivation. For protection programmes and other example, a national household survey measures to leave no one behind. can be combined with a census survey to provide a disaggregated 2.2 Targeting at the estimate of deprivations through the small area estimation method.32 In household and individual addition, a household survey, such levels for the effective as a MICS, can be combined with implementation of social administrative data, such as from health institutions. Data integration protection tools can help estimate deprivations To better understand deprivations in at a very low level of administrative households, a census survey can be units—for example, at the subdistrict conducted in the beginning at the or block level or if possible at the household level in villages or blocks gram panchayat (village committee) identified as poor in the preceding level. process. The survey can also collect (ii) Collecting background data and information about individual members information. This method can to inform the targeting of individual be applied when the use of the benefits.33 This process involves several small area estimation method is steps, including listing all households, not possible given the absence of conducting household surveys a recent census and/or national for identifying and verifying poor household (sample) survey. Various households, and collecting detailed checklists and questionnaires can information for State support (Figure be used to collect information 2). The process can be conducted at lower administrative levels annually but only for monitoring and about community assets such as tracking deprived households and their infrastructure, including roads, members. An annual review can help to electricity, markets, banks and define appropriate budget allocations. industrial establishments. The process generally entails organizing Implementation of these steps requires a consultation with various building consensus on indicators and governmental representatives across seeking necessary information at the the sectors of the administrative unit community, household or individual as well as non-governmental groups. levels. While the national MPI can Information on community assets guide the selection of indicators, data helps define deprived blocks or for some impact or outcome level communities most in need of social indicators may be difficult to obtain

Part II – Multidimensional poverty measures as policy 17 tools—fostering uses in Asia and the Pacific Figure 2: Identifying and monitoring deprived households

Listing of Prepare a complete list of all the households of deprived areas households (blocks/villages).

Survey & classification Collect data and identify poor and non-poor households. of households

Verification of deprived Verify the classification of households by organizing community households meetings, and publishing the list of poor and non-poor households.

Approval of the list by Submit consolidated list of poor households at a higher level for higher authority review and approval.

Survey of deprived Conduct a detailed survey of poor households including their households members for State support.

from each household. This may require •• Revise and finalize the list of chosen changes in the specification and/or indicators. inclusion of some proxy indicators •• Decide a cut-off at the individual specific to the local context. Steps to indicator level and the aggregate level. take include: Prepare instruments and guidelines Select indicators for targeting •• Develop instruments to collect data •• Review the current national MPI on indicators at the household and indicators, and if necessary, suggest individual levels. proxies for some multidimensional poverty indicators, taking into •• Prepare guidelines and instructions account the process (time, to collect data from forms/ complexity) of data collection. questionnaires or directly through using a tablet or mobile phone, so •• Organize a stakeholder workshop that no additional efforts are required on indicator selection and weights, for data entry. stressing selection criteria, the desirable properties of indicators, reasons for inclusion or exclusion, etc.

18 •• Develop computer programmes prevent the information from being for data entry and processing to misused in ways that could cause harm. categorise households into poor, near-poor and non-poor. •• Develop a summary form at the Communication and village or block level to prepare a advocacy synthesis of results. •• Pilot and refine the instruments. Communication should be part of multidimensional poverty measurement initiatives from the very beginning. It Development of should build, ideally, on the political will a management demonstrated by top national leaders in deciding to develop a poverty measure. information system Communication messages should be tailored to specific audiences. For To institutionalize multidimensional instance, technical and methodological poverty measurement and effectively messages need to be conveyed to use data to track poverty over statisticians and the central statistical time at various levels, including organization, whereas messages to among households and individuals, policymakers might emphasize how governments should establish a data poverty measures can play a vital role repository, such as an integrated and in accelerating progress across the user-friendly management information SDGs. system. It should link data collected from various communities, and The business community has been facilitate and organize the flow of an enthusiastic and powerful partner information, allowing speedy use by of multidimensional poverty reduction analysts, policymakers, programme in Latin America; NGOs can also be implementors and other stakeholders, stalwart allies in poverty reduction. including to track the deprivation status Among citizens, public messaging and of households and individuals. Annual incorporation of MPI materials into tracking can help governments define higher level education courses could the provision of social protection be vital. or other State support to deprived individuals and households. Privacy protections for individual and household data should be in place in order to

Part II – Multidimensional poverty measures as policy 19 tools—fostering uses in Asia and the Pacific A communication strategy on Conclusion multidimensional poverty measurement can be part of an overall national Inclusive social protection and poverty communication strategy of a programmes are vital for achieving government. As noted in the national SDG 1 as well as the 2030 Agenda and MPI handbook, three communication the SDGs as a whole, which call on plans should cover specific stages of all countries to leave no one behind, the process. The first plan should be and focus first on those furthest implemented during the design of a behind. Identifying the poorest or most multidimensional poverty measure. deprived people is a necessary first Another should be prepared for its step in this direction. launch, and the last should cover the period after the launch to showcase A multidimensional poverty measure how the measure is providing useful provides not only information on who is information to guide public actions.34 poor, but also on aspects of well-being where people are most deprived. As an important example of these measures, Institutional an MPI can provide essential insights framework for developing inclusive policies, addressing overlapping deprivations and accelerating the reduction of The institutional framework for poverty development gaps. Using such a tool measurement should be embedded for resource allocation by sectors or in existing national systems. Creating administrative units can ensure that a separate or a parallel system could more resources go where deprivation otherwise undermine their integrity. is greatest. The MPI can also help Specific modalities depend on in tracking progress on SDG target the institutional structures of local 1.2, and in monitoring and evaluating governments and provincial or central poverty programmes. Thus, it is an governments. In Viet Nam, for example, important planning and policy tool. commune leaders are ultimately responsible for generating data with The usefulness of a multidimensional the support of volunteers and the poverty measure also depends on commune chairperson. Data are sent to how much disaggregated information districts and then provinces for approval it provides. While a national MPI and analysis, and use in making budget computed from national household allocations. surveys using the Alkire-Foster method can offer disaggregated estimates

20 of deprivation by indicators and dimensions, it cannot provide poverty estimates at very low administrative levels. Therefore, a national MPI should be accompanied by other subnational poverty identification measures. Using multidimensional poverty measures at various levels—national, subnational, household and individual— is necessary for designing sound poverty programmes, ensuring efficient targeting and effective implementation, tracking deprivation status, and monitoring and evaluation. A wider application of multi-dimensional poverty measures can help achieve the goal of eradicating all forms of poverty everywhere.

Part II – Multidimensional poverty measures as policy 21 tools—fostering uses in Asia and the Pacific References –

Alkire, Sabina, and James Foster. Chatterjee, Mihika, Richard Colin 2007. “Counting and Multidimensional Marshall, Nguyen Bui Linh and Pham Poverty Measures.” OPHI Working Minh Thu. 2014. Multidimensional Paper 7. Poverty in Ho Chi Minh City: Insights from Survey Data for a Future Policy ———. 2009. “Counting and and Service Delivery Agenda. Study Multidimensional Poverty conducted by UNDP. Measurement.” OPHI Working Paper 32. Datt, Gaurav. 2010. “Distribution- sensitive Multidimensional Poverty ———. 2011. “Counting and Measures with an Application to India.” Multidimensional Poverty Working paper 1/17. Monash Business Measurement.” Journal of Public School, Monash University, Australia Economics 95(7-8), 476-487. and OPHI.

Blanca, David Le. 2014. Towards Drèze, Jean, and Amartya Sen. 2013. Integration at Last? The Sustainable An Uncertain Glory: India and its Development Goals as a Network of Contradictions. Princeton: Princeton Targets. New York: United Nations University Press. Department of Economic and Social Affairs, Division for Sustainable Duflo, Esther. “Women’s Development. Empowerment and Economic Development.” Journal of Economic Bourguignon et al. 2010. “Millennium Literature 50(4). Development Goals: An Assessment.” In Ravi Kanbur and Michael Spencer, Ellsmoor, James. 2019. “New Zealand eds., Equity and Growth in a Globalizing Ditches GDP For Happiness And World. Washington, DC: World Bank. Wellbeing.” Forbes, 11 July.

22 Foster, James. n.d. “Multidimensional National Statistical Office of the Poverty Measurement Has Followed Government of Thailand and UNICEF an Interesting Path.” Dimensions (United Nations Children’s Fund). Interviews. MPPN. 2016. Thailand Multiple Indicator Cluster Survey 2015/2016. Final report. Government of Malaysia. 2018. “The Bangkok. Mid-Term Review of the Eleventh Malaysia Plan, 2016-2020: New OPHI (Oxford Poverty and Human Priorities and Emphases”. Ministry of Development Initiative). 2015. Economic Affairs. “Measuring Multidimensional Poverty: Insights from Around the World.” Government of Pakistan. 2018. “Sustainable Development Goals: Rogan, Michael. 2016. “Gender National Framework.” Ministry of and Multidimensional Poverty in Planning, Development and Reform, South Africa: Applying the Global Planning Commission. Multidimensional Poverty Index (MPI).” Social Indicators Research 126(3): pp. Independent Group of Scientists 987–1006. appointed by the Secretary-General. 2019. Global Sustainable Development Royal Government of Bhutan. 2013. Report 2019: The Future is Now— Eleventh Five Year Plan (2013 – 2018) Science for Achieving Sustainable – Volume 1: Main Document. Gross Development. New York: United National Happiness Commission. Nations. Sachs, Jeffrey et al. 2019: Sustainable MPPN (Multidimensional Poverty Peer Development Report 2019. New Network). “Nepal.” https://www.mppn. York: Bertelsmann Stiftung and the org/paises_participantes/nepal/. Sustainable Development Solutions Network. ———. “Philippines.” https://www. mppn.org/paises_participantes/ Sigal, Samuel. 2019. “Forget GDP— philippines/. New Zealand is Prioritizing Gross National Well-being.” Vox 18 June. ———. 2018. “OPHI 10th Anniversary”. Dimensions. Special Issue, June. Stiglitz, Joseph E., Amartya Sen and Jean-Paul Fitoussi. 2009. Report by ———. 2019. MPPN side event. United the Commission on the Measurement Nations Statistical Commission, 50th of Economic Performance and Social Session. Progress. www.stiglitz-sen-fitoussi.fr.

References 23 United Nations. 2017. Guide on Poverty UNICEF (United Nations Children’s Measurement. New York and Geneva: Fund). n.d. “Multiple Indicator Cluster United Nations Economic Commission Surveys.” http://mics.unicef.org/. for Europe. USAID (United States Agency for United Nations. 2019. The Sustainable International Development). n.d. “DHS Development Goals Report 2019. New Program.” https://dhsprogram.com. York. World Bank. 2017. Monitoring Global United Nations General Assembly. Poverty: Report of the Commission on 2011. “Happiness: towards a Holistic Global Poverty. Washington, DC. Approach to Development.” Resolution 65/309, 19 July.

UNDP (United Nations Development Programme). 2011. Human Development Report 2011— Sustainability and Equity: A Better Future for All. New York.

———. 2019a. Human Development Report 2019—Beyond Income, Beyond Averages, Beyond Today: Inequalities in Human Development in the 21st Century. New York.

———. 2019b. Multidimensional Poverty Measures as a Policy Tool for Achieving the Sustainable Development Goals: A Review of MPI Measurement and Uses in Asia and the Pacific. Bangkok.

UNDP (United Nations Development Programme) and OPHI (Oxford Poverty and Human Development Initiative). 2019. How to Build a National Multidimensional Poverty Index (MPI): Using the MPI to Inform the SDGs. New York.

24 Endnotes –

1 United Nations General Assembly 9 The goals on sustainable consumption resolution 65/309 dated 19 July 2011 and production, inequality, poverty, and invited the UN Secretary-General growth and employment each have to seek the views of UN Member links with 10 other goals or more. For States and relevant regional and detail, see Blanca 2014. international organizations on the 10 Datt 2010. pursuit of happiness and well-being, 11 UNDP and OPHI 2019. and to communicate such views to the General Assembly at its 66th session 12 Recent efforts have been made to for further consideration. Resolution use a multidimensional approach to 66/281 subsequently proclaimed the measure the gender poverty gap. The 20th of March as the International multidimensional gender poverty gap Day of Happiness. The resolution is similar to the poverty gap measured recognizes the relevance of happiness by the conventional money-metric and well-being as universal goals approach in several national poverty and aspirations in the lives of human lines. For details, see Rogan 2016. beings around the world. 13 UNDP and OPHI 2019. 2 Sigal 2019, Ellsmoor 2019. 14 UNDP 2011, p. 50. In the case of 3 Bourguignon et al. 2010. income poverty, the poverty gap measure also addresses poverty 4 OPHI 2015. intensity, along with the poverty 5 UNDP and OPHI 2019. headcount estimates, but this is not 6 UNDP 2019b. simple to interpret. Intensity in the 7 The metadata of SDG indicator 1.2.2 Alkire-Foster method is about the have not yet been defined. While number of indicators an individual countries are selecting dimensions and is classed as deprived in, not how indicators based on their needs and badly an individual is deprived in any data availability to estimate national single indicator. Thus, an individual MPIs, several have not made changes who was just below the cut-off in to the global MPI dimensions and three indicators would be classed as indicators. having more ‘intense’ poverty than an individual much farther below the cut- 8 UNDP 2019b. off in two indicators. 15 UNDP and OPHI 2019.

25 16 The region of Asia and the Pacific 31 This process should not leave out the varies depending on context, but it poor in relatively rich districts. generally includes East Asia, South 32 Small area estimation is a widely used Asia, South-east Asia and Oceania. indirect estimation technique for micro- 17 All data presented in this section come level geographic profiling. Traditional from United Nations 2019 and UNDP sample surveys designed for national 2019a. estimates do not provide large enough 18 Sachs et. al. 2019. samples to produce reliable direct estimates for small areas such as 19 Bourguignon et al. 2010, Independent counties or villages. The use of valid Group of Scientists appointed by the statistical models can provide small Secretary-General 2019. area estimates with greater precision; 20 Drèze and Sen 2013. however, bias can occur due to an 21 MPPN 2018. incorrect model or failure to account 22 UNDP 2019b. for informative sampling. OPHI has been assessing to what extent the 23 Chatterjee et al. 2014. method is sufficient to guide policy. 24 Government of Malaysia 2018, MPPN In addition, it is also exploring other 2019. According to James Foster, methods. theoretically, a natural joining of 33 China’s ‘Accurate Targeting income poverty and multidimensional Programme’, for example, targeted poverty could happen either by about 90 million persons on the basis merging the two approaches (a person of livelihood and monetary poverty, the is poor if she or he is either income two no-worries (food and clothing), and or multidimensionally poor) or by an the three guarantees (education up intersection between the two, leading to 9 years, access to health care and to something like the definition that housing). has emerged in Mexico (a person is poor if she or he is both income and 34 UNDP and OPHI 2019. multidimensionally poor). At present, income and multidimensional poverty measures are mainly kept separate for sound data reasons. See Foster n.d. for more details. 25 See: MPPN, “Philippines,” at: https:// www.mppn.org/paises_participantes/ philippines/. 26 See: MPPN, “Nepal,” at: https://www. mppn.org/paises_participantes/nepal/. 27 Government of Pakistan 2018, p. 8. 28 UNDP and OPHI 2019. 29 Government of Pakistan 2018. 30 OPHI 2015, Royal Government of Bhutan 2013.

UNDP works in about 170 countries and territories, helping to achieve the eradication of poverty, and the reduction of inequalities and exclusion. We help countries to develop policies, leadership skills, partnering abilities and institutional capabilities, and to build resilience in order to sustain development results.

United Nations Development Programme Bangkok Regional Hub 3rd Floor, United Nations Service Building Rajdamnern Nok Avenue, Bangkok 10200, Thailand. Email: [email protected] Tel: +66 (0)2 304-9100 Fax: +66 (0)2 280-2700 – www.undp.org