基于贝叶斯面板平滑转换模型的房价阈值效应研究

Bayesian inference on Price of the Real Estate Threshold Effect Based on Panel Smooth Transition Regression Model

针对经济变量之间普遍存在的非线性关系,导致线性模型拟合失效的问题,构建面板数据平滑转换模型,刻画变量之间关系的非对称性。采用贝叶斯方法进行模型的参数估计,避免非线性最小二乘算法难以收敛,参数估计不确定。通过分析模型结构,选择参数先验分布,设计相应的Metropolis-Hasting-Gibbs混合抽样算法,据此估计模型参数;在此基础上,利用省域面板数据分析房价阈值效应问题。研究结果表明:参数的动态迭代轨迹收敛,MH-Gibbs混合抽样算法能够准确地估计模型各参数,解决了非线性最小二乘无法收敛的问题,证明了贝叶斯面板数据平滑转换模型的有效性;同时也验证了房价波动的阈值效应以及房价与城市化、城乡收入差距之间的非线性关系。

For non-linear relationship between the prevalence of economic variables, resulting in failure of the linear model fitting problems, panel data smooth transition regression models are established . bayesian method is used to address uncertain risk of parameters estimation caused by common estimation algorithm which is difficult to converge. Based on the analysis of model statistic structure and the selection of parameters prior, the Metropolis-Hasting within Gibbs sampling method is utilized to estimate model pa- rameters, predicting parameters in use of Monte Carlo Markov Chain. The empirical research applies Bayesian panel data smooth model to analyze the data in Chinese provinces. The research outcomes indicate that the iteration traces of parameters are convergent, and the Metropolis-Hasting within Gibbs sampling method estimates parameters accurately, resolving the problem difficult to converge, showing the effectiveness of Bayesian panel smooth transition model. Furthermore, the existence of threshold effect in the price of the Real Estate has been certificated.