Additively Decomposable Indices of Poverty

Abstract

Ranking income profiles by summary indices is just a first step in the analysis of poverty. For deeper analysis we should inquire into the factors contributing to poverty using time series or cross-section income data. For carrying out such analysis we need a poverty index that exhibits additive decomposability: for any partitioning of the population, defined along ethnic or geographical or other lines, the total poverty is a weighted average of the subgroup poverty levels. Clearly, this will enable us to calculate a particular subgroup’s contribution to total poverty. Hence this type of poverty breakdown becomes suitable for identifying causal factors of poverty and in choosing and implementing poverty alleviation policies.