Robust Multidimensional Poverty Comparisons

We investigate how to make poverty comparisons using multidimensional indicators of well-being, showing in particular how to check whether the comparisons are robust to aggregation procedures and to the choice of multidimensional poverty lines. In contrast to earlier work, our methodology applies equally well to what can be defined as “union”, “intersection” or “intermediate” approaches to dealing with multidimensional indicators of well-being. When one of two indicators is discrete, our methods specialize to those that have previously been developed to deal with household composition heterogeneity. To make these procedures of some practical usefulness, the paper is also the first to derive the sampling distribution of various multidimensional poverty estimators, including estimators of the “critical” poverty frontiers outside which multidimensional poverty comparisons can no longer be deemed ethically robust. The results are illustrated using data from a number of developing countries.