On February 9, the Federal Reserve Board released the hypothetical scenarios for DFAST 2023.1 The scenarios run from the first quarter of 2023 through the first quarter of 2026.
This year, the capital positions of only 23 large banks will be tested against the DFAST scenarios as part of CCAR. Last year the top 34 banks were tested (two-year stress test cycle).
As advocated elsewhere2 the DFAST scenarios can provide a basis for the stress testing programs in regional and community banks. Of course, in those cases the stress scenarios should be adapted to the geographical footprint and lending profile of the respective banks.
Thus, we next review the 2023 hypothetical scenarios by focusing on key macroeconomic factors and addressing two questions:
- How does the 2023 severely adverse scenario compare with the 2022 severely adverse scenario?
- What do any differences imply for potential impact on the banks’ losses?
Figures 1-6 depict the scenarios for the following six macroeconomic factors:
- Real GDP in Figure 1
- Unemployment rate in Figure 2
- House price index (HPI) in Figure 3
- CRE Price Index in Figure 4
- BBB spread in Figure 5 and
- Stock market index in Figure 6
The 2022 severely adverse scenario is included as a dashed orange line in each of the above figure.s The quarter of the trough/peak in the severely adverse scenario of each macroeconomic factor is also denoted.
More specifically, under the 2023 Severely Adverse Scenario:
- Real GDP bottoms in Q1 2023, at a value significantly lower than the 2022 severely adverse scenario but with a similarly slow trajectory to recovery. This implies a positive effect (increase) on losses compared to the 2022 losses, depending of course on the weighting of Real GDP in the loss forecasts.
- In contrast, Unemployment Rate in the 2023 scenario peaks in Q3 2024. Although both the 2023 and 2022 scenarios share the same peak level, the overall increase in Unemployment Rate is more substantial than in 2022 since it starts from a lower value (3.6% vs. 4.2%). This implies a slightly positive effect (increase) on losses.
- Similarly, HPI in the 2023 scenario bottoms in Q4 2023, at a value much lower than the 2022 scenario. In fact, it is even significantly lower than the trough of the Great Recession. This implies a strong positive effect (increase) on losses, particularly related to residential mortgages.
- CRE price index in the 2023 scenario exhibits a similar recession than in the 2022 scenario. It bottoms in Q4 2024 with a trough slightly lower than in the 2022 scenario. We would therefore expect similar or slightly higher losses from the 2023 scenario than the 2022 one.
- BBB spread widens similarly to the 2022 scenario and peaks in Q4 2023 at the same level. The widening of the spreads starts from a higher level in the 2023 scenario than in 2022 (220 vs. 110). This implies a similar effect on C&I loan losses compared to the 2022 scenario.
- At the same time, the DJ stock market return follows a trajectory like the 2022 scenario and bottoms in Q1 2023 but at a level lower than the 2022 scenario, i.e., 37% vs. 45%. This implies a slightly negative effect (decrease) on losses. Like the 2022 scenario, the stock market return is assumed to have 0% growth throughout the Baseline scenario. This is a significant shift from the Baseline scenario of previous years.
Concluding Remarks
The above analysis indicates that the 2023 scenarios might lead to higher overall losses compared to the ones reported from the 2022 stress test. They would mainly be driven by an increase in losses from Consumer, Residential and CRE loans.
The heightened stress in the CRE price index is an average across CRE asset classes. The impact on CRE from the induced recession has been asymmetric across asset classes, with hotel and retail bearing most of the brunt followed by office.
In fact, in its scenario guidelines the Federal Reserve states: Declines in commercial real estate prices should be assumed to be concentrated in properties most at risk of a sustained drop in income and asset values: offices that may be affected by remote work or hospitality sectors that continue to be affected by reduced business travel.1
What happens if a bank, for example, has higher concentration in CRE retail, CRE Office and/or in urban centers hit more by the pandemic-induced consumer behavior? In such cases, as also argued in our analysis of last year’s scenarios [3], a bank should use scenarios of more granular CRE indices, by asset class and geography, for better accuracy in the loss forecasts and impact on capital. Such scenarios should be conditional on the DFAST severely adverse scenario of the overall CRE index for consistency.
Furthermore, what if a bank has significant exposures in oil & gas or agriculture? For its stress test program, the bank should use scenarios for oil & gas and/or commodity prices that are consistent with the DFAST scenarios.
Thus, to create relevant stress scenarios, a bank would have to adapt the DFAST scenarios to the geographical footprint and portfolio mix of the bank. This is a crucial step that we follow for our clients, given the asymmetry in the impact and recovery from the hypothetical scenario shock across geographies and asset classes.
References
- Federal Reserve Board – Federal Reserve Board releases hypothetical scenarios for its 2023 bank stress tests
- Karakoulas G., How Company-Run and Fed DFAST Projections Measure Up. RMA Journal, October 2022, pp.15-18.
- 2022 Stress Testing Scenarios | InfoAgora
Grigoris Karakoulas is the president and founder of InfoAgora (infoagora.com) that provides risk management consulting, predictive risk analytics, price optimization, scenario generation and CECL/stress testing solutions, and model validation services. Contact him at grigoris@infoagora.com