基于分组数据基尼系数的一种估算方法  

An Estimation Method for Gini Coefficient Based on Grouped Data

作　　者：张爱国 Zhang Aiguo(School of Economics and Management,Anqing Normal University,Anqing Anhui 246113,China)

机构地区：[1]安庆师范大学经济与管理学院,安徽安庆246113

出　　处：《统计与决策》2022年第2期10-15,共6页Statistics & Decision

摘　　要：基尼系数是测量收入差距的最常用指标,自提出以来学者们演绎出了很多种测算方法,对于离散数据主要有平均差法、几何方法、矩阵法和协方差法等,但这些方法实质上是一致的,只是形式有所不同。文章以上述方法为基础,以分组数据为研究对象,假定各阶层的收入同分布且与总体分布相似,利用“逼近法”来拟合洛伦兹曲线,并在此基础上提出了一种较为简洁的测算基尼系数的方法。对我国部分年份不同群体基尼系数的测算结果表明,所提方法是可行的。The Gini coefficient is the most commonly used index to measure income inequality.Since it was put forward,scholars have deduced many kinds of measurement methods,such as mean difference method,geometric method,matrix method and covariance method for discrete data;these methods are essentially the same,but the forms are different.Based on the above methods,this paper first takes the grouped data as the research object,assumes that the income distribution of each class is the same and similar to that of the population,and then uses the"approximation method"to fit the Lorentz curve.Finally,on this basis,the paper proposes a relatively simple method to calculate the Gini coefficient.The calculation results of Gini coefficient for different groups in some years in China show that the proposed method is feasible.

关 键 词：基尼系数 分组数据 逼近法 同分布 

分 类 号：C813[社会学—统计学]