Bayesian Model Averaging for Spatial Autoregressive Models Based on Convex Combinations of Different Types of Connectivity Matrices

There is a great deal of literature regarding use of nongeographically based connectivity matrices or combinations of geographic and non-geographic structures in spatial econometric models. We focus on convex combinations of weight matrices that result in a single weight matrix reflecting multiple types of connectivity, where coefficients from the convex combination can be used for inference regarding the relative importance of each type of connectivity in the global cross-sectional dependence scheme. We tackle the question of model uncertainty regarding selection of the best convex combination by Bayesian model averaging. We use Metropolis–Hastings guided Monte Carlo integration during MCMC estimation of the models to produce log-marginal likelihoods and associated posterior model probabilities. We focus on MCMC estimation, computation of posterior model probabilities, model averaged estimates of the parameters, scalar summary measures of the non-linear partial derivative impacts, and their associated empirical measures of dispersion.