Bank for International Settlements publishes methodology to incorporate physical climate risk into banks’ IRB credit risk models

The Bank for International Settlements published a working paper proposing a practical way to embed physical climate risk into banks’ credit risk modelling and internal ratings-based capital calculations. The approach extends the one-factor Vasicek asymptotic single risk factor framework by adding a physical-risk component while preserving portfolio invariance, aiming to make physical-risk adjustments usable in portfolio credit loss measurement, provisioning and risk-weighted assets. Physical risk is modelled as an externally parameterised tail event that triggers a jump down in a borrower’s asset value with probability q, which in turn affects both probability of default and loss given default, including an explicit channel for climate-driven impairment of recoverable assets. The paper derives a generalised portfolio loss distribution and a generalised risk-weighted asset formula that keeps the structure of the existing IRB calculation but replaces the conditional probability of default with a climate-adjusted version and adds a multiplier reflecting climate-related LGD uplift. Implementation requires banks to produce baseline and climate-adjusted PDs, an external statistical estimate of q from meteorological models, and LGD estimates with and without climate damage. An illustrative example for a secured investment-grade exposure in Mobile, Alabama shows model risk-weighted assets increasing by roughly 8% to 16% under stated parameterisations, rising to around 20% under more conservative LGD assumptions.