Residents Income Factors Quantile Regression To Counterfactual Decomposition

In this paper, based on the CHNS survey data, we describe the urban household per capita income distribution in China, and find that the income obeys a right-tailed distribution. It is difficult to choose a good parametric method to fit the distribution. Naturally, nonparametric kernel density estimation is a more appropriate one. Meanwhile, we give a description about the income quantile changes and income gap in brief, and consider that the gap is expanding ceaselessly due to the Matthew effect.Then, we combine the nonparametric and semi-parametric theory and quantile regression analysis to study on the individual income distribution, construct the least-square linear regression, the linear quantile regression, the nonparametric quantile regression, and the semi-parametric quantile regression one by one, disscuss the best model to fit the data in this paper, and conclude that the gender, education and the urban-rural household registration have a parametric influence on the income, while the age and the region economy give a nonparametric one. In terms of every single factor, the effects on the different income quantiles show a large divergence. The semi-parametric quantile regression analysis provides an all-round display of the income structure and its change rules caused by the factors mentioned above.Finally, in the empirical analysis based on the quantile regression, we employ the counterfactual analysis method with the traditional index theory together, to decompose the gap of the income distribution into several factors, not only on the time dimension, but on the cross-sectional one. We separate three effects from the income change, which are aroused by the coefficients, the covariates, and the residuals respectively. We get a relatively thorough decomposition of the income change and draw a rational conclusion. Especially, the decomposition conclusions on the two cross-sectional dimensions deserve to be mentioned. We successfully separate the gender discrimination and the rural discrimination from the income gender difference and urban-rural difference respectively, and take deep account about the different effects on the different income quantiles.These conclusions have certain significance in deep study on the income gap and the income allocation policy in China. Meanwhile, they also can provide certain reference to make decisions about the tax, social security and the people’s livelihood system.