Software for Bayesian Spatial Model Comparison

Taking a Bayesian perspective on model comparison for crosssectional and static panel data models considerably simplifies the task of selecting an appropriate model. A wide variety of alternative specifications that include various combinations spatial dependence in lagged values of the dependent variable, spatial lags of the explanatory variables, as well as dependence in the model disturbances have been the focus of a literature on various statistical tests used by practitioners to distinguishing between alternative specifications. LeSage and Pace (2009) make a theoretical argument that implies the task of model selection can be simplified by focusing on only two model specifications, one reflecting theoretical situations involving global spillovers (the spatial Durbin model, SDM) and the other theoretical scenarios involving local spillovers (the spatial Durbin error model, SDEM). LeSage (2014) extends this theoretical argument to the case of static panel data models. MATLAB software functions for carrying out Bayesian cross-sectional and static spatial panel data model comparisons is described here along with a number of illustrative applications.