统计研究

• 论文 • 上一篇    下一篇

组合分布在我国居民收入分布拟合中的应用研究

刘洪 王超   

  • 出版日期:2017-06-15 发布日期:2017-06-20

The Application of Composite Distribution in the fit of Chinese Resident Income Distribution

Liu Hong & Wang Chao   

  • Online:2017-06-15 Published:2017-06-20

摘要: 本文使用CHNS家庭微观调查数据,发现我国城乡居民家庭人均收入对数呈现明显左偏特征。为了适应这种需要,根据组合分布函数的构造思想,首次提出使用反向帕累托-对数正态组合分布对居民收入进行建模,并提出使用极大似然估计方法对参数进行估计。此后,使用变量替换方法导出我国城乡居民收入原始数据的确切分布,并在此基础上,测算出我国城乡居民收入的基尼系数。实证研究发现,农村居民家庭人均收入基尼系数要高于城镇居民,就基尼系数发展态势而言,我国城乡居民家庭收入基尼系数在2007-2009年间发生了转变,由之前的持续增加转为下降。

关键词: CHNS, 居民收入, 左偏, 反向帕累托-对数正态组合分布, 基尼系数

Abstract: Based on the fact that the log household income per capita data from China Health and Nutrition Survey (CHNS) show negative skewness. I In order to meet this requirement, we propose Reverse Pareto-Lognormal composite distribution (RPLC) based on the composite distribution theory. The maximum likelihood method is used to estimate the parameters. Empirical results show this model can fit the log household income per capita well, then we derived the exact distribution of household income per capita by using changing variable method, the corresponding Lorenz curve, and the Gini coefficient. Our results show the Gini coefficients of the rural household are higher than the urban household, and reach its peak between 2007 and 2009, which is from a continued increase to a decline.

Key words: China Health and Nutrition Survey (CHNS), Resident Income, Negative Skewness, Reverse Pareto-Lognormal Composite Distribution (RPLC), Gini Coefficient