中国城镇地区性别工资差距的分布变化：2002～2008

Distributional Changes in the Gender Wage Gap in Urban China, 2002~2008

本文采用CHIP2002年和2008年数据，分析不同性别工资差距分解的结果以及工资增长率分解的结果，以此来研究影响性别工资差距变化的因素。本文利用Oaxaea—Blinder的分解方法来分解整个分布的工资差异，并且为消除潜在模型（非线性）带来的估计误差，采用重新赋予权重的方法和半参数估计的方法来进行敏感性分析。研究发现不同性别的工资差距的均值在2002～2008年存在加大的趋势，对于工资分布顶端的劳动者，这一效应更加明显，对女性存在着明显的“玻璃天花板”效应；对于工资分布底端来说，不同性别的工资差距却在进一步缩小。对工资增长率的分解可以看出，男性的工资增长存在着明显的“马太效应”，女性的工资增长呈现“U”型曲线。

This paper investigates changes in the gender wage gap between men and women across the wage distribution using China Household Income Project Survey（CHIP） data. We take advantage of a newly-de- veloped Oaxaca-Blinder-type decomposition for unconditional quantile regression models by Firpo et al. （2009） to decompose wage differentials across the entire distribution. We show that this approaches allows a decomposition of both changes in the gender wage gap and gender differentials in wage growth across the distribution if the underlying regression model is assumed to be linear. A reweighting procedure is em- ployed to across the consequences of this assumption and a semi-parametric decomposition method is used to perform a sensitivity analysis. We find that the mean of gender wage gap tended to increase in 2002- 2008, but there is a clear ＂glass ceiling＂ effect in the top of wage distribution, and for the bottom of the gen- der wage gap is further reduced. The result of decomposition of the wage growth rate means a Clear male wage growth ＂Matthew Effect＂, the women wage growth presents a ＂U＂ shaped curve.