Housing prices are influenced by a variety of physical, tax, proximity, neighbourhood and local attributes. Hedonic specifications are able to measure the influence of dwelling size and age (and other physical characteristics) on housing prices.

This paper examines the influence that local housing market characteristics have on price prediction accuracy using hedonic housing price equations with single-family transactions for Guarda city.
The regression analysis employs cross-section data; in this instance specialized methods of spatial econometrics are used to avoid potentially biased results and faulty inference.
An alternative to a spatial autoregressive process to avoid the complexity of definition of the spatial weight matrix can be expressed in the form of autocorrelation tests. So we propose the serial correlation LM test – specifically the Breusch-Godfrey Lagrange Multiplier test – as a proxy of spatial correlation tests.