Modelling transport and real-estate values interactions in urban systems

Abstract

This article presents hedonic Multiple Linear Regression models (MLR), spatial autoregressive hedonic models (SAR), Spatial autoregressive hedonic in the Error term Models (SEMs) and spatial Durbin hedonic models (SDMs) to estimate house price variations in metropolitan areas as a result of changing environmental and accessibility conditions. The goodness of fit of the different models has been compared along with a series of hypotheses about the performance of the specifications considering spatial relationships between observations. The case study for such analysis was the urban area of Santander (Spain). The models which considered spatial dependence between observations offered a greater degree of fit in a scenario showing strong spatial correlation in MLR residuals. The SEM model combined with a Queen-Contiguity matrix provided a good fit to the data and at the same time presented significant parameters with theoretically coherent signs. This model estimated increases of 1.8% for each additional transit line present in the areas of housing, as well as a reduction of 1.1% in their prices for each additional minute in travelling time to the Central Business District. Closeness to the train stations, however, implied reductions in house prices.