Physical Risk: Assessment Methodology
Climate change impacts financial systems both directly and indirectly (ref. this report by the World Bank; Fig. XX). Direct damages refer to immediate to short-term physical destruction, including destruction of housing, critical infrastructure, and means of production which results in direct financial loss. Indirectly, climate hazards affect the financial sector via impacting the macroeconomic environment (see section 2.1 Transmission channels).
Direct damages, primarily at the asset level, are typically assessed using a traditional risk assessment framework where risk is defined as the product of hazard, exposure, and vulnerability (ref. IPCC report). This is often done through a catastrophe modeling approach which integrates climate hazard information with exposure and vulnerability metrics to estimate the financial cost of physical damage (or “losses”) to a geographically specific portfolio of physical assets.
A catastrophe model typically includes three main components: a hazard module, which assesses the extent and intensity of climate hazards; a vulnerability module, which relates hazard to physical damage; and a financial module that translates physical damage to financial losses. The key output of a catastrophe model is the distribution of possible losses, expressed in financial terms, to the portfolio. (Footnote here: Note that variations in catastrophe modeling exist. For example, the World Bank’s report on Moroccan banking (ref. here) presents a different structure, where the financial module is omitted and an exposure module is included instead. Similarly, in the report by Oxford (ref. here), exposure is treated as an input). Scenario analysis (see section 2.4), often implemented using Monte Carlo simulations as seen in models like CLIMADA, is also usually embedded in catastrophe models where the events are modeled thousands of times with slight modifications each time to reflect the overall likely distribution of losses due to inherent uncertainty in those estimates. See the data section for lists of open-source catastrophe models.
The indirect impacts are estimated by macroeconomic models which capture how the direct damages propagate and interact with other sectoral and macroeconomic variables over time. Macro-economic information, such as GDP, employment, household consumption, and trade flow, and the country specific socioeconomic data is needed to xx. Examples of models are listed in Table 1 (add table here). Page 29 of this also gives advice on the choice of macroeconomic models.
Chronic physical risk is usually neglected by asset-level assessment. In the NGFS scenarios, chronic risk is determined using simple damage functions.