中国住户调查数据收入变量的比较

Comparisons on the Income Measurements in Chinese Household Survey Data

随着对中国社会经济相关问题的关注程度日益提高,相应的住户调查数据也在逐渐增多,并且各套数据在调查方式、覆盖目标人群,甚至研究主题设计上都具有较强的相似性。本文对研究中常用的几套中国住户调查数据在样本人群结构偏差、消费与收入之间的关联性以及收入分布与收入不均等程度解释因素的差异性进行了比较。研究发现各数据都存在一定程度的人口结构偏差,但这种偏差对于收入水平和收入不均等程度的影响不明显;在一些数据中,消费和收入之间的关联性比较弱,并且与变量的度量误差对回归结果的影响具有相似性;不同数据来源的对数方差难以从变量结构特征和系数特征中得到解释。

With more attention to social and economic issues during economic transition in China, more and more household surveys have been collected. There are many similarities among the surveys in terms of the survey method, targeted population groups, and even research topics design, but the findings and conclusions drawn from the various surveys might be very different. Therefore, this paper attempts to make some comparisons of demographic characteristics of sample population structure deviation, relationship between income and consumption, and differences in the explanatory factor of income distribution and income inequality. The findings indicate that,(1) almost all the surveys are seriously biased in the composition of'gender-age-educational attainment', although the readjustment on the demographic composition makes little change in the level of income and income inequality;(2) there is very weak correlation between income and consumption in some survey datasets, which seems to be related to the measurement error on income and consumption variables;(3) the inequality measures(variance of log income) is hard to be explained by the demographic composition and coefficient effect by regressed income function among the surveys.