INEQUALITY AND DEVELOPMENT - A CRITIQUE

In this paper we subject the influential cross-section estimation of the inequality-development relationship by Ahluwalia to detailed scrutiny. We investigate the robustness of his estimates to variations in functional form, and find that different functional forms - between which the data cannot choose - lead to widely differing shapes for the inequality-development relationship. We also examine his data set which consists of income distributions for 60 developing and developed countries, and find that the distributions are not comparable with respect to income concept, population unit, and survey coverage. We construct a minimally consistent data set to reestimate the relationship, and show that the preferred functional form displays a reversal of the commonly accepted U-hypothesis.