Bayesian Model Averaging in the Context of Spatial Hedonic Pricing: An Application to Farmland Values

Since 1973, British Columbia created an Agricultural Land Reserve to protect farmland from development. In this study, we employ GIS-based hedonic pricing models of farmland values to examine factors that affect farmland prices. We take spatial lag and error dependence into explicit account. However, the use of spatial econometric techniques in hedonic pricing models is problematic because there is uncertainty with respect to the choice of the explanatory variables and the spatial weighting matrix. Bayesian model averaging techniques in combination with Markov Chain Monte Carlo Model Composition are used to allow for both types of model uncertainty.