The Bias of the Gini Coefficient Due to Grouping : Revisiting First-Order Corrections

We propose a first order bias correction term for the Gini index to reduce the bias due to grouping. The first order correction term is obtained from studying the estimator of the Gini index within a measurement error framework. In addition, it reveals an intuitive formula for the remaining second order bias which is useful in empirical analyses. We analyze the empirical performance of our first order correction term using income data for 15 European countries and the US, and show that it reduces a considerable share of the bias due to grouping