Robust Multidimensional Spatial Poverty Comparisons in Ghana, Madagascar, and Uganda

Duclos, Jean-Yves ; Sahn, David ; Younger, Stephen D.

Published by Oxford University Press on behalf of the World Bank

关键词: CENTRAL REGION;    CONFLICT;    COVARIANCE MATRIX;    CUMULATIVE DISTRIBUTION;    DECOMPOSABLE POVERTY;   

DOI  :  10.1093/wber/lhj005
RP-ID  :  77512

学科分类：社会科学、人文和艺术（综合）

来源: World Bank Open Knowledge Repository

【 摘 要 】

Spatial poverty comparisons areinvestigated in three African countries usingmultidimensional indicators of well-being. The work isanalogous to the univariate stochastic dominance literaturein that it seeks poverty orderings that are robust to thechoice of multidimensional poverty lines and indices. Inaddition, the study seeks to ensure that the comparisons arerobust to aggregation procedures for multiple welfarevariables. In contrast to earlier work, the methodologyapplies equally well to what can be defined as union,intersection, and intermediate approaches to dealing withmultidimensional indicators of well-being. Furthermore,unlike much of the stochastic dominance literature, thiswork computes the sampling distributions of the povertyestimators to perform statistical tests of the difference inpoverty measures. The methods are applied to two measures ofwell-being, the log of household expenditures per capita andchildren's height-forage z scores, using data from the1988 Ghana Living Standards Study survey, the 1993 NationalHousehold Survey in Madagascar, and the 1999 NationalHousehold Survey in Uganda. Bivariate poverty comparisonsare at odds with univariate comparisons in severalinteresting ways. Most important, it cannot always beconcluded that poverty is lower in urban areas in one regioncompared with that in rural areas in another, even thoughunivariate comparisons based on household expenditures percapita almost always lead to that conclusion.