关键词: 贫困, 不平等, 分位数回归, 反事实分解

Abstract: Based on income distribution function, this paper discusses the relationship among income growth, inequality and poverty using CHNS (China Health and Nutrition Survey) data from 1990 to 2005. The sample is divided into two parts, i.e. 1990-1999 and 1999-2005. The results of mean regression and O-B decomposition show that revenue decision mechanism is the dominant factor of income increase. Through quantile regression and M-M decomposition, we find that revenue decision mechanism is a main reason to force income distribution curve towards the right. In the first stage, the expansion of income gap is mainly due to the changes in labor force characteristics. However, the reason is not clear in the next stage. Finally, we apply the Shapley decomposition to the poverty analysis, and find that revenue growth can lead to a significant reduce poverty and income gap can intensify the poverty. In the two stages, growth effects always make the greatest contribution. But, the sequence of allocation effects and poverty line changing effects is alternated in the different stages.

Key words: Poverty, Inequality, Quantile Regression, Counterfactual Decomposition