The Effect of Spatial Interpolation on the Hedonic Model: a Case of Forest Pest Damages

In the hedonic model, when an investigator wishes to merge property sale data with spatial data on an environmental amenity, one problem encountered in the matching process is that the environmental data are usually limited. A challenging task is how to scale site and time specific environmental data to all property sales within a defined geographical area. Inverse distance weighting, Kriging and splines are three commonly used geo-statistical methodologies to make spatial interpolation. In this study, we investigate the effect of these spatial interpolation methods on the estimation of a hedonic model in the context of an invasive forest pest, the hemlock wooly adelgid. Our results indicate a statistically significant relationship between hemlock health and residential property values at the 0.1 km level. Comparing through different interpolation methodologies, Kriging provides the most reliable interpolation results which provides us a useful tool to scale up our analysis from specific sample sites to broad geographical area.