The spatial diffusion of housing prices across Chinese cities

We examine the spatial diffusion of housing prices across 152 Chinese cities from 2005 to 2018 through both geographical proximity and socio-economic similarity. We decompose socio-economic similarity into two components: (i) fundamental similarity, measured by similarity in calibrated equilibrium housing prices determined by population, wages, land supply and construction costs; and (ii) non-fundamental similarity, measured by similarity in deviations of observed housing prices from their calibrated equilibrium levels. Using a dynamic spatial panel model, we quantify the magnitude and persistence of housing price shocks. The results show that both geographical and socio-economic diffusion effects are statistically significant and persistent. Neglecting socio-economic linkages may overstate the importance of geographical proximity. Within the socio-economic channel, non-fundamental similarity generates larger and more persistent diffusion effects than fundamental similarity, although fundamental similarity remains an important transmission mechanism. High-tier cities exhibit substantially stronger cross-city diffusion than low-tier cities. Finally, in a leapfrogging subsample that excludes geographically and socio-economically proximate cities, diffusion effects become substantially weaker: non-fundamental effects become statistically insignificant, while fundamental effects remain only weakly significant.