Spatial Statistics Applied to Commercial Real Estate

Portfolio theory shows that diversification can enhance the risk-return trade-off. This study uses the absolute location of commercial real estate property along with spatial statistics to address the inherent problem of determining geographical diversification based upon a set of economic and property-specific attributes, some of which are unobservable or must be proxied with noise. We find that commercial real estate portfolios exhibit statistically significant spatial correlation at distances ranging from adjacent zip codes to neighboring metropolitan areas. Given the common structure of dependence found in the data series, we discuss feasible strategies for obtaining diversification within direct-investment real estate portfolios.