On Iterated Lorenz Curves with Applications: The Multivariate Case

TL;DR

This paper extends the analysis of iterated Lorenz curves to the multivariate case, showing convergence of marginals to a power-law distribution with golden ratio exponent and independence in the limit.

Contribution

It generalizes previous univariate results to multivariate Lorenz curves, demonstrating convergence properties and independence of marginals in higher dimensions.

Findings

Marginals converge to a power-law distribution with exponent equal to the golden section.

Iterated Lorenz curves lead to independent marginals in the multivariate case.

The convergence is uniform under reasonable restrictions.

Abstract

It is well known that a Lorenz curve, derived from the distribution function of a random variable, can itself be viewed as a probability distribution function of a new random variable [4]. In a previous work of ours [26], we proved the surprising result that a sequence of consecutive iterations of this map leads to a non-corner case convergence, independent of the initial random variable. Namely, the limiting distribution follows a power-law distribution. In this paper, we generalize our result to the multivariate setting. We do so using Arnold's type definition [4] of a Lorenz curve, which offers the greatest parsimony among its counterparts. The situation becomes more complex in higher dimensions as the map affects not only the marginals but also their dependence structure. Nevertheless, we prove the equally surprising result that under reasonable restrictions, the marginals again converge uniformly to a power-law distribution, with an exponent equal to the golden section. Furthermore, they become independent in the limit. To emphasize the multifaceted nature of the problem and broaden the scope of potential applications, our approach utilizes a variety of mathematical tools, extending beyond very specialized methods.