Housing Market Shocks in Italy: a GVAR Approach

In this paper, we use a Global Vector Autoregression (GVAR) model to assess the spatio-temporal mechanism of house price spillovers, also known as “ripple effect”, among 93 Italian provincial housing markets, over the period 2004 − 2016. In order to better capture the local housing market dynamics, we use data not only on house prices but also on transaction volumes. In particular, we focus on estimating, to what extent, exogenous shocks, interpreted as negative housing demand shocks, arising from 10 Italian regional capitals, impact on their house prices and sales and how these shocks spill over to neighbours housing markets. The negative housing market demand shock hitting the GVAR model is identified by using theory-driven sign restrictions. The spatio-temporal analysis carried through impulse response functions shows that there is evidence of a “ripple effect” mainly occurring through transaction volumes.