Spatial dependence of housing values in Northeastern Italy

This study aims to test whether spatiotemporal dependence exists in residential property values in Northeastern Italy, or it is insignificant once accounting for exogenous explanations. A peculiar feature is that the analysis focuses on aggregate housing values, at the level of municipalities, rather than individual housing units. Based on the literature on spatial data analysis, a general serial and spatial model is used. Along with several demographic and socioeconomic predictors, two time lags and two spatial lags of the dependent variable are considered. As far as the results are concerned, serial dependence is more important than spatial dependence. Moreover, there are clues that the exogenous explanations cause the spatial effects to fade away, which would support the claim of the skeptics on the occurrence of spatial spillovers, namely, that peer and neighborhood effects are mostly a matter of neglected, underlying influences. However, some spatial dependence is intrinsic in the analyzed data and failing to consider it may lead to biased estimates. The results provide a deeper understanding of how housing values possibly spread across space. They are also suitable for predictive purposes, especially as regards the locations characterized by weak market conditions.