Category Archives: OPHI Working Papers

Diverging Identification of the Poor: A Non-random Process. Chile 1992–2017

This paper investigates the degree of association in the identification of the poor between the standard monetary FGT measure and the Alkire-Foster Multidimensional Poverty Index. For this purpose, we use a measure of redundancy between the two poverty measures (R0). In Chile, over the past 25 years, R0 has declined at a rate of 1.5% per year. The decline is unimportant during the 1990s, a decade of rapid economic growth, while it is notable thereafter, in a period characterized by modest economic growth and the progressive introduction and deepening of social policies. The conditional correlation between socioeco­nomic and demographic characteristics with R0 is examined at the province and household levels. After controlling for household non-eligibility across some of the indicators of the Multidimensional Poverty Index, we find that the divergence in the identification of the poor can be explained by improve­ments in education, increasing urbanization, and a reduction in the household size. Consequently, the divergent identification of the poor seems to be a real process, which is not randomly distributed across the population. On the basis of our results, we hypothesize that this divergence is a general phenomenon that tends to occur in countries undergoing demographic transition, urbanization, and progress in educa­tion. If so, and given the fact that poverty alleviation strategies are adopted partly on the basis of poverty statistics, the diverging identification of the poor might have distributive consequences for the poor themselves.

Citation: Klasen, S. and Villalobos, C. (2019). ‘Diverging identification of the poor: A non-random process. Chile 1992–2017’, OPHI Working Paper 128, University of Oxford.

The State of Multidimensional Child Poverty in South Asia: A Contextual and Gendered View

Many poverty measures identify a household as poor or non-poor based on the achievements of all its members. Using the household as the unit of identification has the benefit of enabling a poverty measure to draw on information about persons of different ages and genders, and in different life situations. However, it also loses individual information because this is summarized at the level of the household.  For example, the underlying microdata contain additional information on individual children. As a consequence, gendered and intrahousehold inequalities, for instance, are not evident even when data for them exist. This paper proposes methods to augment a household multidimensional poverty index (MPI) by applying individual-level analyses to the same dataset, and analysing these alongside the matrix of deprivations underlying an MPI. In particular we scrutinise (i) what proportion of deprived children live in multidimensionally poor households; (ii) what proportion of deprived children are girls or boys; and (iii) what proportion of deprived children live in households in which other children are not deprived in that same indicator. We also observe (iv) what other deprivations deprived and poor children experience in addition to the focal deprivation. Finally, we study what proportion of people live in households where children of different ages experience two different child deprivations concurrently. More complex analyses can also be undertaken that combine information on the deprivation status of more than one eligible member, and we illustrate this to identify pioneer children, who completed six years of schooling although adults in their household have not. Overall, this study provides a prototype methodology that can be mainstreamed into subsequent national and global MPI analyses in order to shine a light on child poverty multidimensionally. We illustrate the methodology with analyses of the global MPI for seven countries in South Asia.

Citation: Alkire, S., Ul Haq, R. and Alim, A. (2019). ‘The state of multidimensional child poverty in South Asia: a contextual and gendered view’, OPHI Working Paper 127, University of Oxford.

The Role of Inequality in Poverty Measurement

The adjusted headcount ratio, or MPI, is widely used by countries and international organizations to track multidimensional poverty and coordinate policy. Several characteristics have encouraged its rapid diffusion: applicability to ordinal data, ease of communication, a practical identification of the poor based on multiple deprivations, and a dimensional breakdown that informs and coordinates policy. Sen (1976) and others have argued that poverty should also be sensitive to inequality among the poor. This paper provides a new axiom that embodies this perspective in the multidimensional context and defines an M-gamma family containing a range of measures satisfying the axiom. Like the FGT or P-alpha class of monetary measures, it has three main members: the headcount ratio to evaluate the prevalence of poverty, the adjusted headcount ratio to account for its intensity, and the “squared count” measure that reflects severity and inequality among the poor. We note that any inequality sensitive measure must violate the dimensional breakdown axiom and investigate Shapley decomposition methods as an alternative. Unfortunately, these methods can yield counterintuitive result; however, the squared count measure avoids this critique and its Shapley breakdown reduces to an easy to compute formula that supplements the traditional breakdown for the MPI with information relevant to inequality among the poor. An example from Cameroon illustrates our method of using M-gamma measures in tandem to evaluate multidimen­sional poverty while accounting for inequality and dimensional contributions.

Citation: Alkire, S. and Foster, J. (2019). ‘The role of inequality in poverty measurement’, OPHI Working Paper 126, University of Oxford.

Measuring Autonomy: Evidence from Bangladesh

The search for rigorous, transparent, and domain-specific measures of empowerment that can be used for gendered analysis is ongoing. This paper explores the value-added of a new measure of domain-specific autonomy. This direct measure of motivational autonomy emanates from the ‘self-determination theory’ (Ryan and Deci, 2000). We examine in detail the Relative Autonomy Index (RAI) for individuals, using data representative of Bangladeshi rural areas. Based on descriptive statistical analyses, we conclude that the measure and its scale perform broadly well in terms of conceptual validity and reliability. Based on an exploratory analysis of the determinants of autonomy of men and women in Bangladesh, we find that neither age, education, nor income are suitable proxies for autonomy. This implies that the RAI adds new information about the individuals and is a promising avenue for further empirical exploration as a quantitative yet nuanced measure of domain-specific empowerment.

Citation: Vaz, A., Alkire, S., Quisumbing, A., and Sraboni, E. (2019). ‘Measuring autonomy: Evidence from Bangladesh’, OPHI Working Paper 125, University of Oxford.

Does the Hunger Safety Net Programme Reduce Multidimensional Poverty? Evidence from Kenya

The purpose of this research is to evaluate the short-term impact and long-term sustainability of Kenya’s Hunger Safety Net Programme (HSNP). Difference-in-difference and propensity score matching estimations are used to determine the impact of programme participation on the household multidimen­sional poverty index (MPI). We found that programme participation reduced the MPI significantly, which is mainly driven by the food insecurity dimension, and that the reduction in poverty is due to the reduction in the incidence and intensity, the latter in particular, of poverty among the ultra-poor households. Our analysis of the political economy of Kenya suggests that, while the government is making progress in the institutionalisation of social protection, weaknesses in the implementation and financing of the programme, as well as the short-term focus of impact evaluation, may undermine the programme’s poten­tial to help build a strong state that is accountable for the eradication of poverty.

Citation: Song, S. and Imai, K.S. (2018). ‘Does the Hunger Safety Net Programme reduce Multidimen­sional poverty? Evidence from Kenya’, OPHI Working Paper 124, University of Oxford.

Does Aid Reduce Poverty?

Fifty years of literature on aid effectiveness has so far proven inconclusive. Two main challenges still require some attention. The first is to properly identify the causal effect of aid on poverty alleviation. To address it, I exploit differences in the number of years countries have been temporary members of the United Nations Security Council as an instrument for the average amount of economic aid disbursed by the United States. The second is to obtain reliable data on poverty, which I confront by using multidimensional poverty data from the Oxford Poverty and Human Development Initiative (OPHI). For a sample of 64 developing countries, I estimate a significant relationship between higher amounts of aid received during the period 1946–1999 and lower Multidimensional Poverty Index (MPI) between 2000 and 2014. On the contrary, the relationship does not seem to be significant when poverty is measured from an income perspective. Alternative measures of poverty could help improve the understanding of the relationship between development aid and poverty alleviation and might also contribute to improved targeting for aid disbursements.

Citation: Milovich, J. Y. (2018). ‘Does aid reduce poverty?’, OPHI Working Paper 122, University of Oxford.

Assessing Deprivation with Ordinal Variables: Depth Sensitivity and Poverty Aversion

The challenges associated with poverty measurement within an axiomatic framework, especially with cardinal variables, have received due attention during the last four decades. However, there is a dearth of literature studying how to meaningfully assess poverty with ordinal variables, capturing the depth of deprivations. In this paper, we first propose a class of additively decomposable ordinal poverty measures and provide an axiomatic characterisation using a set of basic foundational properties. Then, in a novel effort, we introduce a set of properties operationalising prioritarianism in the form of different degrees of poverty aversion in the ordinal context, and characterise relevant subclasses. We demonstrate the efficacy of our methods using an empirical illustration studying sanitation deprivation in Bangladesh. We further develop related stochastic dominance conditions for all our characterised classes and subclasses of measures. Finally, we elucidate how our ordinal measurement framework is related to the burgeoning literature on multidimensional poverty measurement.

Citation: Seth, S. and Yalonetzky, G. (2018). ‘Assessing deprivation with ordinal variables: depth sensitivity and poverty aversion’, OPHI Working Paper 123, University of Oxford.

The New Global MPI 2018: Aligning with the Sustainable Development Goals

Early in 2018, the United Nations Development Program’s Human Development Report Office (HDRO) and the Oxford Poverty and Human Development Initiative (OPHI) agreed to adjust and unify their methodologies on poverty measurement and consider indicator improvements, in order to better monitor the Sustainable Development Goals (SDGs).

This paper sets out the specifications of a joint global Multidimensional Poverty Index first published in 2018, which is an internationally comparable measure of acute poverty that captures the multiple deprivations poor people experience with respect to health, education and living standards.  It builds on the original MPI launched in 2010, and an innovative MPI launched in 2014. The best features of both of these are subsumed in the joint global MPI 2018, which also reflects new data possibilities to better align the global MPI to the Sustainable Development Goals.

Because the objective of revising the MPI to create a more credible and legitimate measure of multidimensional poverty that enables comparisons across countries using existing data was challenging to realize, the paper first sets out five key principles for a global poverty measure related to data coverage, communicability, comparability, disaggregation, and robustness.

Drawing on expert interventions, a global consultation, empirical trials, and these principles, the paper then explains conceptually the motivation and nature of adjustments that were made to five of the ten included indicators. It also recognizes desirable changes that could not be made due to data constraints – for example including data on the environment, work, and security, or on intrahousehold inequalities. And it identifies key issues for future research related to household composition and the use of land and livestock variables.

Citation: Alkire, S. and Jahan, S. (2018). The New Global MPI 2018: aligning with the Sustainable Development Goals’, OPHI Working Paper 121, University of Oxford.

Evaluating the Effects of Housing Interventions on Multidimensional Poverty: The Case of TECHO-Argentina

The objective of this paper is to evaluate the effect of the NGO TECHO’s emergency housing programme on multidimensional poverty. It employs a quasi-experimental ‘pipeline’ evaluation design and is based on household survey data from 34 informal settlements in Buenos Aires, Argentina. The aim is to demonstrate the additional insights that can be gained from using a multidimensional framework based on the Alkire and Foster (2011) method to evaluate a programme’s impact. The results indicate that the programme reduces both the incidence and the intensity of poverty and causes the multidimensional poverty measure to fall by more than half. The magnitude of the effect is greater for the households that initially were the poorest. Privacy, interpersonal relations and psychological health are the dimensions that contribute the most to explaining the decline in multidimensional deprivation.

Citation: Mitchell, A. and Macció, J. (2018). ‘Evaluating the effects of housing interventions on multi­dimensional poverty: the case of TECHO-Argentina’, OPHI Working Paper 120, University of Oxford.

The Research Agenda on Multidimensional Poverty Measurement: Important and As-yet Unanswered Questions

The application of multidimensional poverty measures is proliferating, in part due to the emphasis in Goal 1 of the Sustainable Development Goals (SDGs) on ending poverty in all its forms and dimensions.  This paper first traces the emergence of a priority for non-monetary poverty measures in key texts that consid­ered then set out the SDGs. It then outlines some vital and feasible research questions on a sub-set of fascinating empirical topics on counting-based multidimensional measures. The topics covered here relate to the SDGs’ focus on measuring the multidimensional poverty of men, women, and children. Building on the existing literature, fascinating questions remain in terms of measurement design (the selection of dimensions, indicators, cutoffs, and weights), the analysis of multidimensional poverty measures, their application to child poverty and their implementation using gendered data. In each of these areas, it is expected that significant breakthroughs are possible.

Citation: Alkire, S. (2018). ‘The research agenda on multidimensional poverty measurement: important and as-yet unanswered questions’, OPHI Working Paper 119, University of Oxford.