Integrating prior information into Gini coefficient estimation: A credibility method
Limin Wen, Ruyi Cao, Yi Zhang · Communication in Statistics- Theory and Methods · 2025
.The Gini coefficient is widely used as a statistical measure of income inequality. Traditional statistical inference methods for the Gini coefficient are typically based solely on population and sample data, with prior information often being neglected. In this article, a Bayesian model is introduced for estimating the Gini coefficient, where the conditional Gini coefficient is modeled as a linear function of the sample. The optimal linear Bayesian estimator for the conditional Gini coefficient is derived and shown to be approximated as a weighted average of moment and aggregation estimations, with weights aligned with the properties of a credibility factor. Additionally, the large sample properties of the linear Bayesian estimator are proven. Finally, the convergence and rate of convergence of the estimator are validated through numerical simulations.