知识溢出,地理邻近与区域异质性的空间关系

摘要

基于修正的Griliches-Jaffe知识生产函数,构建了未考虑和考虑区域异质性情况下知识溢出与地理邻近关系的空间计量模型;结合全域Moran's I指数、Moran散点图以及LISA集群地图,对31个省份和直辖市知识产出的全域和局域空间相关性进行探讨;运用普通最小二乘法(ordinary least square,OLS)、空间滞后模型(SLM)和空间误差模型(SEM),对比研究区域异质性在知识溢出与地理邻近关系中的作用.最后,总结了区域异质性、不同地理邻近区间以及知识溢出的关系差异.%Two spatial econometrics models are established to estimate the relationships between knowledge spillovers and geographical proximity based on the modified knowledge production function of Griliches-Jaffe. One model takes regional heterogeneity into account while the other not. Furthermore, the global and local s-patial correlation of knowledge output in 31 provinces are discussed by combining Global Moran's I, Moran scatter plots and LISA cluster maps. Moreover, applying ordinary least square, spatial error model and spatial lag model, the paper also compares the effects of regional heterogeneity in the relationship between geo-graphical proximity and knowledge spillovers. Finally, the different relationship among regional heterogeneity, geographical proximity and knowledge spillovers are summarized.