第一个主题：经济增长
第19篇文章：Ying, L. G. (2000). Measuring the spillover effects: SomeChinese evidence.Papers in Regional Science, 79(1), 75–89.

从文章学到什么？

一是，怎么使用Moran指数确定辐射的范围。

二是，怎么使用Moran指数测度是否存在辐射效应。

三是，文章对既有研究的梳理很清晰。

辐射距离的确定

I start by searching for an appropriate value for thedistance d. This value should represent the distance at which spatialassociation is maximized and beyond which there is an overall cessation ofspatial effects. Because the spatial association maximizes at d = 800kilometers according to a diagnostics of spatial autocorrelation using globalMoran I.

如何定义增长极

In defining the “growth pole”, Perroux (1950) argued thatone should not distinguish a propulsive industry simply by its relative rate ofgrowth. In a similar way, Higgins (1983) also suggested that if the “propulsiveregion” was too small and too unsophisticated in its economic activities, itwould not generate significant spread effects of any kind.

关于MoranI 的显著性

The normal assumption may not be appropriate according toAnselin (1995). For this specific case, an asymptotic Wald test for normalityis computed for the local Morans. The result turns out to be highlysignificant, revealing a non-normal, highly skewed distribution of the observedIi. In order for the test to be used in the specific and more realisticbackground, an empirical random permutation approach is conducted.

评议：

空间计量具有不可识别性。