我国城镇居民消费行为区域异质分析——基于动态面板分位数回归视角

An Empirical Analysis on the Heterogeneity in Consumer Behavior of Urban Residents in China——Based on Quantile Regression for Dynamic Panel Data

笔者采用面板模型识别中的变系数变截距检验方法将全国31个省份划成5个区域,使得区域内的模型变为面板变截距模型,然后采用动态面板分位数回归方法对我国城镇居民的消费行为进行区域异质分析。研究结果表明,采用面板模型识别中的变系数变截距检验方法在区域划分中具有一定的优势,目前我国城镇居民消费行为存在显著的区域异质性。

This paper analyzes the heterogeneity of consumer behavior of urban residents in China by the method of dynamic quantile regression for panel data proposed by Antonio and Galvao （2011） together with the panel data identification method testing whether it has variable intercepts or variable coefficients by dividing the 31 provinces into five area. Results show that the panel data identification method has some advantages to divide regions and there has a demonstration effect in Chinese consumption behavior of regional hetero- geneity between five rezions currently.