Decomposing Asset Risk in Real Estate Cap Rates – The Interplay of Market, Property, and Location Factors

Institutional investors and managers of mixed-asset portfolios face a critical challenge in day-to-day practice: determining the appropriate risk of a specific asset, i.e. the mark-up over the risk-free return. In the context of direct real estate investing, they seek to understand how much of the premium stems from market conditions, property attributes, or location. This study addresses these questions by decomposing and explaining the asset risk of commercial real estate, a vital yet underexplored aspect in today’s volatile market environment. Prior research has identified certain factors that influence commercial real estate risk profiles, but has not successfully quantified their impact magnitudes or accounted for the non-linear and time-varying interactions among them. Using property-level data on U.S. commercial real estate from the early 1970s to 2024, we estimate asset risk premia, i.e. the spread between a property’s cap rate and the risk-free return, as a function of macroeconomic indicators, property characteristics, and location data. Employing eXtreme Gradient Boosting, we uncover non-linear patterns in the data, while SHapley Additive Explanations enhance model transparency by uncovering factor contributions. Our findings reveal how market dynamics and property-specific traits drive risk premia across different economic cycles, offering investors a systematic, data-driven tool for transparent asset risk assessment. This methodology empowers practitioners to make informed decisions in uncertain markets. Academically, the study advances the field by (1) integrating granular property-level and micro-location data, (2) leveraging machine learning techniques, and (3) shedding light on the complexities of private real estate risk decomposition and pricing dynamics over multiple market cycles.