Poverty, Inequality and Stochastic Dominance, Theory and Practice: The Case of Burkina Faso

In this paper we provide a set of rules that can be used to check poverty or inequality dominance using discrete data. Existing theoretical rules assume continuity in incomes or in percentiles of the population. In reality, with the usual household surveys, this continuity does not exist. However, such a discontinuity can be exploited to test for stochastic dominance. This paper also proposes stochastic dominance conditions that check for the statistical robustness of the inferred rankings. The methodology of this paper is illustrated using Burkina Faso's household surveys for the years of 1994 and 1998.