The dependence structure of income distribution

Abstract

This article investigates the dependence structure of income distribution in the US by providing two approaches – one regression-based and the other copula-based – to reveal new information about income dependence. The system of Seemingly Unrelated Regressions (SUR) is estimated for both quintile income shares and mean income growth by controlling for macroeconomic variables, and Kendall's tau statistics are derived for income dependence. Results from less restrictive copula models corroborate the regression-based results. However, income growth models do not support the common claim that the rich are getting richer while the poor are getting poorer. Income dependence patterns do not appear to be affected by business cycles, but Democratic and Republican presidential administrations have drastically different income dependence results.