A deeper look at impacts in spatial Durbin model with sphet

Abstract

This article focuses on the estimation of the spatial Durbin model and associated relative impacts implemented in the R library sphet. Specifically, the current version of the library offers two ways of performing inference: one based on drawing samples from a multivariate normal distribution, and the other based on an analytical formula. The performance of these two methods is compared using an extensive Monte Carlo experiment. As an illustration of the kind of analysis that can be performed with sphet, the article also presents an empirical application looking at economic growth of Italian provinces.