摘要
本文讨论了中国居民收入差距估计中所可能出现的各种偏差,并对收入定义、样本权重结构、抽样偏差以及地区间货币购买力差异进行调整,试图得到比较客观的收入差距衡量指标。本文利用了帕累托分布修正住户调查抽样偏差对收入差距指标的影响。本文的估计结果表明,中国居民收入的基尼系数达到了较高的程度。高收入人群样本的偏差导致了城镇内部收入差距的严重低估,也导致了城乡之间收入差距和全国收入差距的较大程度的低估。
This paper discusses the potential biases in inequality estimation in China, such as the income definition, reweighting, sampling bias, and regional differences in purchasing power parity. We try to eliminate such biases and reestimate the inequality measure, Gini coefficients for urban and national. We correct the sampling bias by estimating the up-tail shape of income distribution by use of Pareto distribution and income information for the riches from Forbs list, Hurun list and compensation for CEO. The findings in the paper indicate China is more unequal if the sampling bias were eliminated. Resulting from the sampling bias in up-tail of income distribution, inequality measures within urban, between urban and rural, and national wide are seriously underestimated.
出处
《经济研究》
CSSCI
北大核心
2011年第4期68-79,共12页
Economic Research Journal
