Information matrix test An application using Pareto’s original income distribution data

Abstract

This article attempts to apply and interpret the correspondent Identification Matrix test (White, 1982) for a Pareto type I distribution. One of the most cited datasets used by Pareto, England’s income distribution in 1843 and in 1879, is used. Our results suggest that an alternative hypothesis, one of random coefficients cannot be accepted, i.e., we are not able to reject the hypothesis that the empirical income of England is a Pareto type I distribution.