Reviewing the Federal Reserve scenario analysis requirements
Derek Jun, Managing Director, Head of Climate Risk, TIAA & Nuveen
Below is an insight into what can be expected from Derek’s session at Climate Risk USA 2023.
The views and opinions expressed in this article are those of the thought leader as an individual, and are not attributed to CeFPro or any particular organization.
Can you summarize the Federal Reserve scenario analysis requirements?
The scenario analysis has three major sections:
- Physical risk scenario analysis
- Transition risk scenario analysis
- Qualitative responses.
For the physical risk module, the focus portfolio is the residential and commercial real estate lending portfolios. The focus metrics are the impacts to the one-year probability of default and loss given default, assuming the portfolio has suffered damages from the hazards projected in these scenarios as of the year 2050.
The physical risk scenarios considered include a moderate RCP 4.5 emission pathway scenario, a more extreme RCP 8.5 scenario, and an RCP 8.5 scenario without insurance as a potential mitigant. There is a common shock element that prescribes the analysis of the northeast region against hurricane risks, as well as an idiosyncratic shock element where the banks can choose their own hazard and the region to analyze.
For the transition risk module, the portfolio of focus is the corporate and commercial real estate lending portfolios. The focus metric is again the impact on PD and LGD from the projected macroeconomic variables assumed in the climate scenarios for each year from 2022 through 2032. The scenarios considered are the NGFS Current Policies and Net Zero by 2050 scenarios.
Lastly, the Fed exercise requires a discussion of the bank’s climate risk management framework and governance as well as a robust discussion around methodologies, results, and any lessons learned from performing the physical risk and transition risk scenario analysis.
While the Federal Reserve is focusing on large banking institutions at this stage, it may be possible for similar exercises to be rolled out for the broader banking sector in the future. As the exercises evolve from the current exploratory objective of assessing capabilities into fulfilling a more macroprudential objective, we might see increased standardization of methodologies and assumptions that the participants can use.
Can you outline the challenges of benchmarking results of climate scenario analysis?
While the Fed scenario analysis exercise has elements of standardization, several elements suggest that the aim for the Federal Reserve is not to use this exercise as a means to assess the macroprudential risks to US banks from climate risks, but rather to assess the state of capabilities for banks to assess and manage climate risks.
First is that the exercise is not exhaustive in covering all of the bank’s potential exposures to climate risks. Second, as observed in the Physical risk module, there is flexibility for banks to choose where and how to apply an idiosyncratic shock. Third, there is plenty of flexibility throughout the exercise for using overlays and adjustments, including interpolations and data transformation. And lastly, the exercise does not require banks to calculate expected losses.
Considering these points, it is unlikely that banks will be able to use the results of scenario analysis to benchmark their own exposures or vulnerabilities to climate risks. However, one could gather the level of sophistication and capabilities in modeling climate risks.
How do requirements from the Fed, ECB, and PRA differ? How will these differing regulations affect institutions?
The key distinction between the Fed exercise and the climate stress testing exercises by the European Central Bank and the Prudential Regulation Authority of the Bank of England is the scope and objective. As discussed above, the scope of the Fed exercise is limited to certain aspects of a bank’s portfolio and does not cover all the potential risks.
The ECB and PRA exercises, however, are intended to be exhaustive in terms of scope. The objective is also different in that the Fed exercises appear to be more exploratory in nature to assess large banking institutions’ capabilities, governance, and methodologies in modeling climate risks. For the ECB and PRA, the stress tests are designed to measure the banking sector’s resilience to climate risks and to inform climate policy action by the government ultimately.
How can firms overcome data limitations of severe stress scenarios?
We need to recognize that everyone is dealing with the same data challenges. The scenario analysis exercise from the Fed is good in that it will galvanize institutions, large and small, to start thinking about the risks and to start building up the capabilities to assess and manage the risks. I think the Fed will be sensitive to these data challenges, as it can be implied from their focus on assessing “risk measures” as opposed to “expected loss” in the exercise.
Lastly, given data challenges, it will be important for any user of scenario analysis to focus on the relative risks that exist within a portfolio as opposed to any nominal levels of risk or loss that may be modeled.
How can we define when a normal event becomes a climate type analysis?
A normal event becomes a climate type analysis when it is studied within the context of larger climate scenarios. Climate analysis typically involves long-term time horizons, typically 30 years or more, and employs large data sets. Most importantly, while the study of normal events employs historical data to inform future likelihood of events, climate analysis is based on projections and forecasts.
As an example, the likelihood and damage from a hurricane to real estate values can be studied as a normal event in the context of historical data. Climate type analysis might employ damage functions informed by historical data but apply higher degrees of frequency or severities to simulate what may happen in a changing climate.