Spatial Dependence, Idiosyncratic Risk, and the Valuation of Disaggregated Housing Data

We investigate whether spatial idiosyncratic risk plays an important role in explaining average housing prices in a representative U.S. market. We discuss a parsimonious hedonic model of demand for differentiated products and derive an equilibrium price function that depends on idiosyncratic risk, among other factors. Empirically, we use a nonlinear spatial regression model and identify a potential measure of idiosyncratic risk from sales data of individual residential properties in Ames, Iowa. The results show that, for our disaggregated housing data, there is a significant volatility interdependence among cross-sectional units because of geographical proximity. In our sample, a 1% increase in idiosyncratic risk, ceteris paribus, is associated with a 0.80% increase in average price of residential properties. We find that accounting for spatial autocorrelation and heteroskedasticity increases the evidence that idiosyncratic risk, which is captured by space-varying volatility, reveals important information about average housing prices. We conclude that using a spatial regression model that allows interaction between property prices and volatility yields strong predictive power.