Robust Multidimensional Poverty Comparisons with Discrete Indicators of Well-Being

Abstract

This paper provides a method to make robust multidimensional poverty comparisons when one or more of the dimensions of well-being or deprivation is discrete. Sampling distributions for the statistics used in these poverty comparisons are provided. Several examples show that the methods are both practical and interesting in the sense that they can provide richer information than do univariate poverty comparisons.