空间断点回归分析框架及其应用

Spatial Regression Discontinuity Analysis Framework and Its Application

空间断点回归方法是将通过实验研究经济学的方法向地理空间推广,是利用地理信息实现因果关系推断的前沿方法。文章在反事实潜在结果的理论基础上,重点梳理了基于地理空间不同实验场景下,能够识别因果效应的重要假设及相应模型的设定方法和估计方法,构建了空间断点回归方法的基本分析框架,已有相关文献都可归结为此框架下的应用。重点讨论了如何将不同地理信息应用于识别假设和模型设定,并分析了不同地理信息使用方法和使用效果的区别与联系。基于此框架,对重庆市升级为直辖市的经济效应进行实验设计、统计推断和评价。

Spatial regression discontinuity,which extends the method of studying economics by experiment to geographic space,is a frontier method that uses geographic information to achieve causality inference.On the basis of the counterfactual potential outcome theory,this paper focuses on the important assumptions used to identify causality effect and setting methods and estimation methods of the corresponding model based on different experimental scenarios in geospatial space,and constructs the basic analysis framework of spatial regression discontinuity method.The existing literature can be attributed to the application under this framework.Then,the paper focuses on how to apply different geographic information to identification assumptions and model settings,and analyzes the differences and connections between different geographic information usage methods and effects.Finally,based on this framework,the paper carries out the experimental design,statistical inference and evaluation of the economic effects of Chongqing’s upgrading to a municipality directly under the Central Government.