Measuring urban quality of life using multivariate geostatistical models

Urban quality of life (QOL) is usually measured through an index defined as the estimated value of a set of urban amenities. However there is an increasing awareness that omitted variables might seriously undermine the method's ability to accurately estimate QOL. Here we extend the hedonic approach using a multivariate geostatistical model to address the omitted variable bias by identifying the latent common factors responsible for the spatial distribution of the amenities. A new QOL index is then defined as a function of the latent factors whose implicit prices are estimated through hedonic regressions. Our methodology is shown on a data set of individual-level property transactions from the city of Vicenza. As a result we obtain the spatial distribution of QOL calculated according to the new index.