Physical Climate Risk in Commercial Real Estate

Physical climate risks pose a threat to real estate values around the world. The rising frequency and intensity of extreme weather events, such as hurricanes and floods, have led to an increased vulnerability of the property sector to physical climate risks. While insurance currently serves as a safeguard against certain risks, the advent of climate change may result in elevated premiums, more rigorous underwriting standards, and potential uninsurability in high-risk regions. This study aims to examine whether physical climate risks at the property level are reflected in the equity returns and bond risk premia of U.S. REITs. Central to the analysis is the use of physical climate risk scores. First, we obtain data on REIT asset coordinates and other property information from S&P Capital IQ Pro. Second, the data on REIT portfolios is imported into climate risk tools to obtain asset-level climate risk scores for each hazard. These physical climate risk assessment tools are used in the real estate industry by institutional investors and reinsurance companies. In addition, for robustness, we aim to use publicly available data on physical climate risk from the Federal Emergency Management Agency (FEMA). Moreover, control variables such as bond characteristics, macroeconomic indicators, equity market factors, and REIT balance sheet data are used to capture additional drivers of REIT equity returns and bond risk premia. The methods include cross-sectional OLS regression analysis on the one hand and Artificial Neural Network (ANN) models, combined with the explainable AI method Shapley Additive Explanations (SHAP) on the other hand. The latter approach is used to decompose the variation in risk premia, thereby providing further insights into the influence of climate risks. This approach not only captures potential non-linear relationships but also serves to verify the robustness of the regression results. Furthermore, the study compares outcomes across REITs operating in different regions to assess whether geographic diversification mitigates the impact of climate risks on bond pricing. Additionally, the robustness of climate risk scores generated by different tools is evaluated to determine their reliability in financial decision-making. In conclusion, these findings contribute to the understanding of how market participants price physical climate risks and provide guidance for REIT managers in assessing the impact of climate risks on the cost of equity and public debt.