Abstract/Summary

When analysing asset prices in isolation, the classical asset pricing models only account for the time-series variation of the asset with the factors. However, we show that valuable information can be extracted if we account for spatial dependence across the assets when pricing real estate companies. We extend the factor model in Fama and French (2012) to include a spatial term and estimate a spatial factor model. We model the spatial linkages across real estate company returns using the physical distance between their properties. We find that the spatial factor model is not rejected and the spatial parameter is significant. The spatial factor model performs better than the factor model, substantially improving the model explaining real estate returns. Proximity across the property holdings of real estate companies can be used to model prices for listed real estate companies in addition to size, style and momentum factors. The spatial factor model is then used to disentangle direct and indirect spillover effects through idiosyncratic and market shocks respectively. We find that the spillover effect increases during the global financial crisis and can explain up to 30% of total return variation in the US. UK. and EU markets.