北京都市区土地价格梯度估算与解析

Estimating the Land Price Gradients in Metropolitan Beijing

文章简要梳理了国内外地产价格梯度案例研究的脉络,采用特征价格模型、非参数估计方法研究了北京都市区住宅、商业土地出让价格梯度与土地开发强度的变化,解析了新时期城市空间发展的特征。研究表明:住宅和商业土地出让价格梯度随时间呈现扁平化倾向,从土地价格视角说明了北京城市空间持续的郊区化扩展趋势。2009年以来住宅土地价格梯度为正,表明北京住宅空间结构特征不完全符合单中心城市假设,商业地块则较好地体现出市场机制下价格梯度的变化规律。线性模型在刻画住宅土地开发强度的空间变化方面不典型,但却较好地揭示了商业土地随到城市中心距离的增加,开发强度逐步降低的特征,2005年以来这种特征有所缓和。非参数估计方法由于不需要预设函数形式,有效避免了城市中心预设和模型选取引致的估计偏误,应用非参数方法中的核估计拟合北京住宅土地出让价格、开发强度与距离之间的曲线,能够描绘出变量之间的复杂关系,更符合北京都市区的实际。

The paper gives a brief review about the empirical studies that have been conducted on property price gradients. Based on the classic urban monocentric model, the urban land leasing price-distance relationship is estimated with the aids of the hedonic price model. It is shown that both commercial and residential land price gradients become flattening from 2005 to 2011. It indicates the suburban expanded trends of construction land in metropolitan Beijing. The residential land price curve took a reverse trend since 2009, which violates the traditional monocentric bit- rent curve. The reason lies in two aspects： firstly, not all the residential land parcels are leased through open auction or tender; secondly, there are influential sub-centers in the outskirts of Beijing. Contrary to the residential land, the commercial land price curve fits the negative gradients well during the time. A floor-distance linear regression model is built to explore the change of land development intensity from urban core to urban fringe. The negative floor-distance coefficients of commercial land are captured which demonstrates the land development intensity is decreasing from urban center, but it is flattening during the time. The coefficients of residential land are not significant, illustrating linear regression model does not fit well. As the choice of functional form may result in inefficiency of estimation, a nonparametric analysis is proposed. The advantage of nonparametric estimation is that no priori assumptions on the urban spatial structure and functional form are required, so it is more flexible than the parametric estimations. The results show that the new method reveals more reasonable information about residential land price curve and floor-distance curve in metropolitan Beijing.