Scenario analysis is complex and cannot use a shared set of inputs and scenarios for every use. Instead of seeing that as a weakness that undermines comparability, this should be seen as an important part of the process of becoming better prepared to deal with uncertainty as physical impacts and impacts of the climate transition become more visible.
- Regulators and standard setters have made increasing use of scenario analysis as part of stress testing and climate disclosure
- Scenario analysis for climate-related financial risks remains a new area, with considerable uncertainty driven by the complex interaction between economic and environmental systems
- Having a diverse range of scenarios tried by different institutions for different purposes with different inputs will provide more value than having a globally consistent approach that doesn’t match with likely future risk outcomes
Scenario analysis offers significant value through the results it provides as well as the process of undertaking it. This is why regulators as well as the Financial Stability Board have taken an interest in it and why the International Sustainability Standards Board has affirmed that it plans to require scenario analysis as part of the coming ‘baseline’ sustainability standards.
Both economic and climate systems are complex. Scenario analysis tries to take each into consideration, combine the two, and look decades into the future to produce an estimate of the sources and types of climate-related risks that are built into the financial system. More importantly, the structure of the economy and environment, and the relationship between the two, varies from country to country.
This leads to a concern that if scenario analysis isn’t consistently applied with similar models and inputs, then it won’t provide a useful output. This is an understandable concern, but it risks focusing too much on a single metric of ‘climate risk’ rather than recognizing how much is unknown for regulators, financial institutions and investors in terms of how climate risk materializes.
The science of what environmental damage climate change will cause, and what it has already caused, is clear in some ways. We know that reducing emissions is critical to limiting the negative impact of climate change, and that risks increase non-linearly as temperature increases. But that’s only one factor going into climate scenario analysis, and individual risks’ chances of occurring differ significantly in different places.
Just because a specific risk becomes more likely doesn’t mean it will happen. A risk with lower probability of occurrence could materialize while a more likely risk doesn’t materialize. Or both could materialize, and the question becomes how much more vulnerable is the financial system in this situation versus another plausible outcome. The answer isn’t only related to the scientific processes resulting from climate change, but also to the economic, social and policy responses that follow.
Each of these inputs is going to be different for different financial systems, over different time frames and in different economies. Scenario analysis cannot produce a single number of climate risks out of scenario analysis, but the output can still be significantly valuable. For example, consider how one metric, financed emissions, on which the RFI Foundation has done a lot of research, could be used as an input to scenario analysis.
Many scenario analyses — according to a report by the Financial Stability Board published by the Network for Greening the Financial System — are based on a ‘static balance sheet’. Often this creates a situation where the easiest analysis is to focus on a subset of the balance sheet with the highest concentration of customer emissions. If 80% of a bank’s emissions are concentrated in 10% of its financing assets, then it seems straightforward to build a scenario analysis based on the status quo compared to a pathway to reach a country’s Nationally Determined Contribution (NDC) or for global Net Zero by 2050.
However, the financed emissions volumes reveal only a limited amount of information about climate-related financial risks in different scenarios. Often emissions are viewed as being equally costly, but a good metric of risk will provide information about how different emissions pathways in different scenarios will affect repayment probabilities and the likelihood of default.
The problem with this approach is that in most situations there either has to be a substantial impact on the probability of default within a small part of a balance sheet or to be a large degree of leverage for there to be a significant impact in the overall risk to financial stability. A slightly higher risk within a narrow range of assets has to be magnified to have a substantial impact on, for instance, bank capital adequacy.
Broadly speaking, this is the point at which a single metric for climate risk starts to lose value. If the results of the scenario being applied on a narrow range of sectors in a standardized fashion based on tons of CO2 equivalent show limited results, that means that either climate risks aren’t important to focus on, or there may be more context-specific data that need to be included.
Including additional data about how emissions translate into risk for a narrow subset of high-emitting sectors, and then how those sectors are linked into the broader economy and financial sector, will provide a better understanding of the variability of risks in different circumstances. However, it will do so with the cost of comparability because different social, economic and policy assumptions will have to be added.
Sub-national data, such as precise locational data on physical assets, may be needed to further refine the usability of the output from a scenario analysis if a substantial share of assets potentially impacted is long-lived physical assets. All of these improvements in the precision of estimates from scenario analysis come with the cost of the ability to apply the same methodology in a different context.
This should inform how we think about the results of scenario analysis such as the regulatory stress-tests that an increasing number of financial sector regulators are carrying out. The outputs will vary from one to another without necessarily requiring determining one type of scenario analysis to be ‘better’ and another ‘worse’. The fact remains that all of these scenario analyses are modeling complicated situations with substantial uncertainty and will improve over time.
The process of conducting them will also help to improve knowledge about how climate-related risks related to high-emitting sectors, or those with high physical risk exposure, will impact and be transmitted through the economy and financial sector. That will require much more than just a single ‘standardized’ view of scenario analysis. But conducting scenario analysis at the same time as physical and transition risks are elevating will provide benefit for everyone conducting and analyzing the results.
Want to learn more about responsible finance in Islamic markets & Islamic finance? Subscribe to RFI’s weekly email newsletter today!