【摘要】In this note, we investigate the finite-sample properties of Moran’s I test statistic forspatial autocorrelation in tobit models suggested by Kelejian and Prucha. We fill a void in the theoretical literature by investigating the finite sample properties of this test statistic in a seriesof Monte Carlo simulations, using data sets ranging from 49 to 15,625 observations. We find thatthe test is unbiased, has considerable power and approximates the asymptotic normal distribu-tion even for medium-sized sample sizes, empirically confirming the theoretical results ofKelejian and Prucha. However, some caution is needed, since the statistic turns out to besensitive to misspecification in the form of heteroscedasticity. In such instances the test over-rejects the null hypothesis, mistaking heteroscedasticity for spatial autocorrelation.

【关键词】Spatial econometrics; spatial tobit; Moran’s

【文献来源】Pedro V. Amaral; Luc Anselin.Papers in Regional Studies.2014（4）