Integrating Physical Climate Hazards into Credit Risk: A Multi-Risk Modelling Approach

Assessing how physical climate hazards affect borrower solvency and portfolio resilience remains a critical challenge for financial institutions. Existing approaches often focus on single hazards analysis or top down macroeconomic frameworks. Here, we present a practical, scalable framework that enables central banks and financial institutions to quantify loan-level exposure to multiple physical hazards, and to translate those exposures into asset-level financial impacts and, ultimately, into portfolio expected loss estimates.  

This multi-risk, micro-level modelling framework, developed jointly together with an emerging markets central bank and a european decelopment agency, maps asset locations to four hazards: floods, heat, drought, and wildfire. It combines established natural catastrophe and climate-impact methods with new, tractable procedures to convert hazard intensity into yield, revenue, and profit shocks. These shocks are then propagated through a Merton-type credit risk model to produce loan- and portfolio-level expected loss estimates. The entire workflow is implemented in an R Shiny application, allowing users to build custom multi-year, multi-hazard scenarios, upload portfolio data, and directly analyse impacts across firms, sectors, and regions.

This framework has been initially designed and calibrated for the profile of a single country. However, its modular structure enables straightforward scaling to new datasets, additional hazards, and new regions. We believe this setup can be particularly valuable to stakeholders and financial institutions, especially those in developing economies, to advance physical risk assessment and understanding, as well as future regulatory exercises.