Most banks are still counting emissions. The ones that moved on have a harder question.
IFRS S2 asks banks to disclose climate risk but does not say how to measure it. Most institutions answered with emissions accounting. The few that went further now face a problem the standard never specified: how physical climate risk enters credit metrics at all.
What happened. The disclosure mandate came first from the TCFD and now from IFRS S2. Both require banks to report climate-related risk. Neither contains a method for quantifying it. Faced with a standard that says what to disclose but not how to compute it, most banks did the reasonable thing: they measured what could be measured. Financed emissions, Scope 1 through 3, PCAF attribution. The numbers are auditable and the regulator accepts them. That is where the majority of the industry stands today.
Why it matters. Emissions accounting answers a transition question. It says nothing about what a flood, a hurricane or a wildfire does to probability of default, loss given default, provisions or capital. The institutions that have moved past the inventory stage, and they are a minority, found that the available guidance is a set of proxies rather than measurements. Canada’s banking supervisor, OSFI, published adjustment factors for transition risk and stopped short of physical. The European Banking Authority (EBA) permits a climate add-on to risk parameters but leaves its size to judgement. The Basel Committee allows a margin of conservatism for climate data gaps and does not say how large. Each of these is a placeholder for a number nobody has derived.
What it means for a credit risk officer. If your institution is still at the emissions stage, that is not a failing. It is the correct response to a standard that did not specify more. If you have moved beyond it, the question in front of you is narrower and harder: how does catastrophic physical risk enter your credit metrics with the information you can actually obtain from a catastrophe model, and no more than that. That question has an answer. My research develops a methodology, grounded in the state of the art, for asking a catastrophe model the right questions so that its output feeds the credit model with exactly the information it needs.