Font Size: 

The SDPD (Spatial Dynamic Panel Data) models have been proposedin the socio-econometric literature to analyze spatio-temporal data. In this paper weconsider a particular version of such models, where the set of spatial units is assumedto be partitioned into clusters and the parameters of the model are assumed to beconstant within clusters and not constant across clusters. We propose a mutiple testingprocedure that helps to choose the best model for a dataset by testing a given partitionof clusters assumed under the null hypothesis.