This paper examines the properties of Moran's I test for spatial error autocorrelation when endogenous variables are included in the regression specification and estimation is carried out by means of instrumental variables procedures (such as two stage least squares). We formally derive the asymptotic distribution of the statistic in a general model that encompasses endogeneity due to system feedbacks as well as spatial interaction (in the form of spatially lagged dependent variables).
Testing for Spatial Error Autocorrelation in the Presence of Endogenous RegressorsLuc Anselin and Harry H. Kelejian ,
1-2( 20 )
International Regional Science Review