The spatial diffusion of regional housing prices across U.S. states

•I estimate spatial impulse response functions from a dynamic spatial model.•Spatial IRFs measure diffusion of state-level housing prices across states.•Spatial diffusion of housing price persists for up to four years after a shock.•Spatial diffusion estimates are robust to issues with cointegration.•Persistence of spatial diffusion suggests transmission mechanism for price bubbles.

In this paper I estimate the spatial diffusion of housing prices across U.S. states over a period from 1975 to 2011, showing how long and to what magnitude state-level housing prices are affected by a price shock emanating from surrounding states. I capture the spatial diffusion of regional housing prices with impulse response functions estimated directly from a single equation spatial autoregressive model. In addition, I compare and contrast spatial impulse response estimates over sub-periods. Results show that for the 1975 to 2011 period spatial diffusion of housing prices is statistically significant and persistent across states. This conclusion is robust to whether the model is estimated in levels or in first differences controlling for cointegration. Sub-period estimation, too, suggests the magnitude and persistence of spatial diffusion may be more pronounced after 1999 than before.