Spatial Autoregressive Models for House Price Dynamics in Italy

This paper elaborates a Spatial Autoregressive and Spatial Error Model (SAR-SE Model) to investigate the Italian house price dynamics. House prices in real terms have been modelled for the period 1995-2008 in all the 103 Italian provinces along with affordability ratio, persistency term, some social-economic variables and credit market variables. One of the key results of this paper, is the evidence on house price spatial autocorrelation, verified through the Baltagi, Song and Koh (2003) LM test. On the contrary, no evidence of housing price overvaluation has been found, in comparison with the fundamental values determined by interest rates, households income, rents, employment and construction cost.