基于空间计量的房地产价格影响因素分析

现有文献大都认为各个地区的房地产价格是相互独立的,而事实上,空间依赖的存在会使得回归结果是有偏的。本文首先利用全局空间自相关指标(Moran’s I统计量)和局部空间自相关指标(LISA指标)检验我国房价的空间自相关,结果显示我国房价存在空间自相关。然后利用1999-2009年的省际面板数据,运用广义空间面板数据模型分析全国范围以及东、中、西部分区域房价,结果表明,在考虑了房价的空间相关性后,我国各区域房价的影响因素已不尽相同,东部地区房价基本完全由空间因素决定,西部地区房价由供给和需求等因素决定,而全国范围及中部地区房价受两者的共同影响,利率和汇率的变化对我国房价无显著影响。

Real Estate Price Impact Factors Analysis Based on Spatial Econometrics

Traditional literatures suppose real estate prices in different regions are independent of each other,but the spatial dependence problem will make the regression results biased.this paper,global spatial autocorrelation Moran’s I indices are calculated and local spatial autocorrelation LISAs are figured.Results show that real estate prices exist spatial autocorrelation in our country.Using the panel data(1999-2009) and general spatial model,we analyse real estate prices of nationwide,eastern regions,middle regions and western regions.we find that space factor determines eastern regions’ real estate price,supply and demand factors determine western regions real estate price,while nationwide and middle regions’ real estate prices are decided by both.Interest rate and exchange rate have no significant effect on real estate prices.