Study on the Parameter Estimation Methods for Location Model based on Spatial General Equilirbium

基于空间一般均衡的位置模型参数估计方法研究

• 批准号: 08455235
• 负责人: SHIMIZU Eihan
• 金额: $ 1.92万
• 依托单位: University of Tokyo
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (B)
• 财政年份: 1996
• 资助国家: 日本
• 起止时间: 1996 至 无数据
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-08455235/
• 关键词: location model; spatial econometrics; parameter estimation; multicollinearity; spatial autocorrelation; 一般均衡; 空間相互作用; 混合方程式

In order to evaluate the economic impacts of sets of interdependent major interregional infrastructure investment, such as transportation projects, location models based on Spatial Computable General Equilibrium (SCGE) models have been developed. One important issue remaining is the parameter estimation. Although Ordinary Least Squares Method (OLS) is widely used, it is not enough to estimate parameters of these location models. Aside from the problem caused by the simultaneous equations system, multicolinearity, spatial autocorrelation, spatial heteroscedasticity are also difficult ones. The purpose of this study is to describe these issues and methods for parameter estimation of the models. Firstly, we reviewed counter-measures for them, especially multicollinearity, and spatial autocorrelation and we described the theoretical relationships among them. Subsequently, we applied these methods to a land price model and location-choice model which are typical models for economic analysis of infrastructure planning and are also sub-models of the location model based on SCGE.We also did a practical comparison of them. This developed a kind of manual which provides a deeper understanding of regional modeling.

为了评估一系列相互依存的主要区域间基础设施投资（例如交通项目）的经济影响，开发了基于空间可计算一般均衡（SCGE）模型的区位模型。剩下的一个重要问题是参数估计。尽管普通最小二乘法（OLS）被广泛使用，但不足以估计这些定位模型的参数。除了联立方程组带来的问题外，多重共线性、空间自相关、空间异方差也是困难的问题。本研究的目的是描述这些问题和模型参数估计的方法。首先，我们回顾了应对措施，特别是多重共线性和空间自相关，并描述了它们之间的理论关系。随后，我们将这些方法应用到地价模型和区位选择模型中，这两个模型是基础设施规划经济分析的典型模型，也是基于SCGE的区位模型的子模型，并对它们进行了实际比较。这开发了一种手册，可以更深入地了解区域建模。