The rise of non-financial risks

Naeem Siddiqi, senior adviser, risk management, risk research and quantitative solutions at SAS, discusses the effectiveness of stress-testing as a risk management tool in rapidly changing markets, the role of new technologies in developing robust data-rich scenarios and how recent global events have highlighted the significance of non-financial risks

This year will see the introduction of the European Central Bank’s first climate risk stress test. What challenges will firms face in meeting these requirements?

Naeem Siddiqi, SAS: Climate risk analytics are likely to follow a similar lifecycle as other types of risk analytics. The main elements involved must integrate with each other to reduce time to decision, avoid governance issues and reduce deployment risk. These elements include data acquisition, governance, lineage and storing, data preparation, analytics and modelling, model governance/validation, stress-testing and reporting.

Most important is the last mile, which means deploying the results of these analytics into everyday decisions in the bank, such as lending to retail, small and medium-sized enterprises (SMEs) and to corporate customers, setting provisions, regulatory and economic capital, portfolio design, new products, green investments, executing net-zero strategies, and so on. When the entire process is seamless and integrated, the model or calculation generated is used in fast decision-making and creates value.

Given the uncertainty and lack of specifics regarding climate analytics, SAS advises banks to invest in flexible infrastructure that can deal with all the changes to come. As stress-testing for climate risk evolves, a likely outcome is increased reliance on banks’ own internal models. While transition risk can be modelled more generally across geographies and sectors, physical risk breaks down into many varied aspects driven by differences in the basic climate science behind them. Analyses of ‘orderly’ versus ‘disorderly’ versus ‘hothouse’ scenarios will not be granular enough at a regional or local level. Sea levels, hurricane frequency and intensity, rainfall patterns, heatwaves and wildfires all follow different principles of physics and vary significantly geographically. Furthermore, adaptation responses and resources differ by geography, meaning the same type and level of event will have different impacts based on the actions taken by local authorities. Financial institutions will need the flexibility to employ only the factors relevant to their portfolio exposures.

 

The Covid‑19 pandemic, climate change and the war in Ukraine have demonstrated the growing importance and impact of non-financial risks. How will this influence the future of the regulatory stress-testing framework?

Naeem Siddiqi: SAS’s experience supporting governments coping with the Covid‑19 crisis, and evaluating the resulting impacts on local economies, has taught us that the economic effect can vary by geography and sector and may change dramatically as the environment changes. Countries, regions and industries reacted differently to the various waves of Covid-19, and the same will likely be the case for climate change and other related non-financial risks.

Another important aspect observed from this experience is the benefit of analysing environmental factors – meaning anything exogenous that can affect the phenomenon we are trying to understand. When considering Covid‑19, these factors can include geographical or sectoral restrictions and different approaches to vaccinations or treatments. For climate risk, such factors could be about specific climate variables or definitions, such as emissions or sources, or the development of new carbon dioxide removal technologies.

How climate-based tests will be used for capital allocation in the future remains a big question. The lengthier prediction horizons required for climate analyses produce numbers that can vary greatly, and many regulators have expressed concerns regarding the robustness of those estimates. Quite rightly, the regulators are therefore currently focused on exposure concentrations, and process robustness and flexibility rather than results (in effect, on the story of the numbers, not the numbers themselves).

The longer-term issue for regulators will involve weighing the trade-offs between model complexity, transparency and interpretability. The previously mentioned examples (the pandemic, climate change and war) seem more closely related to credit risk than non-financial/operational risk in the sense that a direct connection can be made, at least in theory, to the key elements of credit risk (probabilities of default) and loss given defaults. Unlike legal or cyber risk, these factors have a specific link to default rates at the instrument/type level and so should theoretically be worked into the standard stressed scenarios. However, the question arises of whether that relationship is tight enough to fit within existing credit risk modelling standards or whether they should be treated within the operational risk scope. Are supervisory stress tests going to evolve to accommodate a wider range of (credit) risks, and a broader range of unorthodox scenarios, or does such evolution become too complex for meaningful interpretation and action?

 

What major lessons have been learned so far from climate risk regulatory stress-testing exercises being conducted globally, and what items of interest do you expect out of the findings/results from the 2022 European Banking Authority’s climate risk stress test?

Naeem Siddiqi: Current exercises are largely learning, fact-finding and benchmarking activities, which will identify best practices and inform future guidance. Regulators will learn how banks are overcoming data challenges and ensuring data quality. For example, getting reliable emissions data from small businesses is a concern. One of SAS’s customers is gathering such information by distributing questionnaires to their SME counterparties.

Working jointly with partners and clients, we are finding a divergence in focus worldwide. Asian banks are most concerned with physical risks, as would befit a region susceptible to earthquakes and tsunamis. On the other hand, European banks are placing greater emphasis on transition risks, due primarily to a stronger drive in that region to pursue net-zero goals.

Until regulations mature, we expect these exercises to be much more about banks’ journeys of getting to their final numbers rather than the actual numbers themselves. Understanding what is driving the numbers under certain scenarios and in certain segments – and explaining why – will be very important. Participating banks relying on external/black-box models should prepare now to be able to explain how these models work.

It is also important to establish a clear connection between data, models and systems used for climate risk analyses and those used for existing business-as-usual risk processes – for example, International Financial Reporting Standard 9, known as IFRS 9, the internal ratings-based approach, risk-weighted assets and the Internal Capital Adequacy Assessment Process. Banks operating climate risk activities in silos, disconnected from other risk processes, should prepare to be challenged by regulators and should be considering how to align these areas more closely.

Sponsored content