住房特征对房价条件分布的影响:分位回归估计

Impact of housing attributes on the conditional distribution of housing price:Quantile regression estimation

传统的Hedonic住房价格模型采用OLS估计,该方法仅能解释住房特征对住房价格条件均值的影响,因而存在一定的局限性.文章采用分位回归的方法估计了住房特征对住房价格五个分位点条件分布的影响,研究发现大部分住房特征隐含价格在不同分位点存在显著差异,并对分布规律及原因进行了讨论.结果表明,与OLS估计相比,分位回归估计可以提供更全面的信息.

Traditional hedonic housing price model uses OLS estimation, this method can only explain the impact of housing attributes on housing conditional mean price, therefore has some limitations. The paper applied quantile regression method to estimate the effects of housing attributes on the housing price at five representative quantiles of the conditional distribution, the study finds most of the implicit price of housing attributes at different quantiles are significant differences, and followed by a discussion on distribution rules and the cause. The results show that quantile regression estimation can provide more comprehensive information than OLS regression estimation.