Study On Gender Income Gap Of Urban Residents In China Based On Bayesian Quantile Regression

For the gender income gap problem of urban residents in China,we establish the Bayesian quantile regression model to analyze the factors that affect the gender gap in different income levels by using CGSS2013 data;and apply counterfactual decomposition method to analyze the cause of gender gap in different income level;finally,we put forward some related proposals of narrowing the gender income gap of urban residents based on the above conclusion.Chapter 1,we introduce the research background and significance,research present situation at home and abroad,as well as the research route,main content and innovation.Chapter 2,we first elaborate the basic theory of gender income gap to determine the variables;and then specify the source and preprocessing of the data;finally,we describe the statistical characteristics of the research variables by using the methods of descriptive statistics and kernel density estimation.The result shows: the gender income gap is improved,but there is a big gap between men and women in each income level,and the gap present “ceiling effect”;women's education is greatly improved,but women is still significantly lower than men in working experience.Chapter 3,according to chapter two of the basic theory of the gender income gap and the statistical characteristics of the research variables,we set up the quantile regression model.In order to improve the accuracy of the model parameter estimation,we estimate them by using the Bayesian estimation method which based on the MCMC algorithm,and the Bayesian quantile regression model is obtained.Chapter 4,based on the model from chapter three,we first use all samples to analysis and find that men's income is 30%-38% higher than women's in each income level.Therefore,we analysis the men and women's samples respectively based on the same method.The result shows: the men's base wage is higher than women's in each income level;the educational returns of women is higher than men 1-4 percentage point except for 0.7 and 0.9 levels;men and women's experience time and income present inverted u-shaped relationship in each level;the duration of income growth for men is more stable than women,and it is higher than that of women 2-3 years;in addition to 0.4 level,the returns of 4 years experiences for men is higher than women;both of men and women's returns of 30 years experiences present a “Matthew effect”;the area returns of women is more stable and higher than men's,but the area returns of men is significant higher than that of women.According to the above conclusions,we put forward related suggestions.Chapter 5,we further use Oaxaca and quantile decomposition methods to analyze the causes of gender income gap.the result shows: based on the average,personal characteristics differences are the main causes of gender income gap;based on the whole income distribution,gender gap almost completely caused by gender discrimination in the low-end and high-end;in the middle,some is caused by personal characteristics difference,while the other is caused by gender discrimination;women are discriminated against in terms of base wage,experience and sector ownership but not in education and the region;the distribution of education years has a great influence on income,and the distribution of the experience,the region and sector ownership has little effect on it.Chapter 6,we make a summary and put forward some related suggestion;for this research insufficiency,we propose the prospect.