Nothing distils a highly complicated problem quite like a spreadsheet. This late 1970s computer creation has brought us the analysis of data using clearly defined metrics and mathematical formulae that offer the certainty of finding a right or wrong answer.
This has been music to the banking industry’s ears. But when central banks demand spreadsheets detailing the immense and largely speculative impact of both physical and transitional climate shocks on bank balance sheets, the enthusiasm starts to wane. The data is partial, the metrics are debatable, and there is certainly no right or wrong answer.
The European Central Bank is collecting data from large European banks within its remit for the 2022 supervisory climate-risk stress test. Last month, the Bank of England published the results of its own climate stress test, with participation from banks representing around 70% of lending to UK households and businesses.
The Banque de France, which co-launched its pilot climate stress test back in 2020 with the Autorité de Contrôle Prudentiel et de Résolution (ACPR) as a blueprint for regulators around the world to copy, is due to start its second iteration of the test in 2023.
It is official: climate stress tests are now an annual tradition for the banking sector.
Methodological weaknesses
As with any regulatory exercise, critics have been quick to question both the purpose of the stress tests and their methodological integrity.
A key issue is the 30-year timeframe on which the scenarios are based. This has been determined by the Network for Greening the Financial System and is intended as a prototype tool to model the economic impact of climate change.
Just because the data is insufficient, doesn’t mean the risk isn’t there
Banks are understandably struggling to provide enough data on the impact of climate shocks according to corporate asset location or value-chain emissions over such a long time period. As a result, the best that these exercises can achieve is to provide assumptions on future resistance of bank portfolios based on a partial understanding of climate risk.
Regulators are aware of this. The BoE writes that the inability to capture appropriate and robust data in certain areas are common limitations, and that means that many climate risks are only being partially measured. The fact that banks are still struggling to develop in-house climate-risk assessment frameworks that efficiently quantify the level of risk is something that the ACPR had already recognized last year.
The key is in the name
Does this make climate stress tests a pointless exercise? They certainly don’t claim to provide all the answers. Perhaps they should be better understood as a tool whose purpose is to determine precisely what banks do know, and to stress test that.
Annual stress tests will nevertheless force financiers to look at their internal processes, which can only be a good thing. They are also a useful means for regulators to determine industry-wide progress on integrating climate-related financial risk into banking business models.
As the science progresses and scenarios become more specific, so should the testing methodology. Just because the data is insufficient, doesn’t mean the risk isn’t there.
It is vital that banks are able to move past the data gaps and make operational changes to mitigate those risks. They will need to keep engaging with all stakeholders and make the necessary changes now to get a lower exposure score when the next set of climate stress tests roll around.