基于回归方程的基尼系数分解

Decomposition of Gini Coefficient Based on Regression Equation

本文提出了基尼系数的等价算法:收入份额法。该算法不仅方便基尼系数按组群分解和来源分解,而且可以进行边际效应分析。收入份额法对变量进行基尼系数的回归方程分解,可以直接利用基尼系数按来源分解的结果。同时,由于截距项、误差项对因变量基尼系数具有重要的影响;因此,本文给出了综合处理截距项、误差项和回归变量关于因变量基尼系数贡献的具体计算公式,最后还进行实证研究。

This paper promotes an equivalent calculated method of Gini coefficient： Income Share Method, which can conveniently decompose Gini coefficient by subgroups and by factors, and carries out an analysis of marginal effect. Income Share Method applying to Gini coefficient decomposition based on regression equation can directly use the result from Gini coefficient decomposition by different sources. According to Both of intercept item and error item have important effects on Gini coefficient, the formula to calculate contributions of regression factors to Gini coefficient of dependent variable have been put forward. At last an applying example is submitted.