Estimating Multi-Modality and Inflection in Wealth Disparity

Abstract

This paper demonstrates how kernel density estimation and non-convergent inflection point identification contribute towards the study of wealth disparity. Measuring the dispersion of non-convergent paired inflection points taken from density estimates of the log of net wealth reveals multi-modal sub-distributions. The methodology overcomes shortcomings in the Gini measure of inequality and demonstrates an analytical framework for examining economic inequality. This approach is applied to U.S. data as well as estimating bi-modality within German incomes revealing sub-populations with extreme variations of economic, financial and demographic variables through identifying inflection points as the density is smoothed through increased iterations of the bandwidth.

Keywords: density derivatives, varied bandwidths, combined density estimates, inflection points, multi-modality, economic inequality

JEL Classification: C10, C14