A Bayesian Copula-Based Modelling For Income Concentration And Inequality Indexes

In this paper we model the dependence between the pre-tax share of income of the top 1% and the post-tax Gini coefficient, we find strong evidence of the existence of a nonlinear dependence and well represented by Joe's copula. We estimate the parameter of the model in a frequentist and Bayesian way and we build conditional curves allowing us to conclude that from a high threshold imposed in the Gini coefficient (high income inequality) the probability of share of income of the top 1% exceeding high thresholds, becomes almost constant. Thus, it seems that for high thresholds, the share of income of the top 1% is mainly associated with intervals of values of the Gini coefficient.