We present a wide class of empirical poverty indices including the most proposed in the literature. By the modern theory of empirical processes indexed by functions, we establish a uniform functional central limit theorem for this class of poverty estimators. From this result, we derive a generalized Wald-type test for comparing in a robust manner two income distributions by using a large spectrum of poverty indices. We conduct a simulation experiment which supports well our asymptotic results. The methodology of test is then applied to a real data set for analyzing poverty in four areas in Sénégal.

poverty indices, poverty orderings, uniform central limit theorem, general empirical process, generalized Wald test.