The use of geostatistics to estimate missing data in a spatial econometric model of housing prices

Abstract

Housing prices have been the subject of many studies, and some of them have tried to determine the influencing structural and location factors through hedonic econometric models. One of the main factors considered in the literature on real estate appraisals is the location of the dwellings. For this reason, this study combines the spatial methodologies of geostatistics and spatial econometrics. On the one hand, this work uses geostatistics to estimate missing data to account for the lack of information in the sampled real estate websites. On the other hand, the explanatory factors of prices are determined through spatial econometrics. The combination of both methods facilitates estimating housing prices in Santa Marta (Colombia), solving the problem of missing data. In the modeling, the problems of spatial heteroscedasticity and multicollinearity are corrected. This combination of methods could be of great interest to company ies and public agencies related to real estate activity, which is sustained by the information available on these real estate websites.