The Federal Reserve on September 30, 2026, adopted two final rules and issued a separate proposal intended to make its annual bank stress tests more transparent, more risk‑sensitive for trading‑heavy firms, and less volatile in their impact on certain capital requirements over time. The announcement came at 13:00 UTC (09:00 EDT) and underscores that stress‑test scenarios reflect severe, hypothetical conditions rather than an economic forecast. [1]
Key points
- Two final rules: annual public input and a revised framework that includes two global market shocks for trading‑heavy banks in 2027. [1]
- Certain firms’ stress capital buffers (SCBs) will be based on a two‑year average of supervisory results beginning in the 2028 cycle. [1]
- A proposed model for noninterest income is out for comment and is not final. [1]
- The Federal Reserve estimates roughly 50% lower year‑to‑year volatility in some requirements, without materially changing aggregate capital. [1]
- Scenarios are severe and hypothetical; they are not predictions of a recession or market downturn. [1]
What happened
The central bank finalized two changes to the supervisory stress‑testing program and, separately, released a proposal for comment related to modeling noninterest income. The final rules formalize an annual process to solicit public input on the design of hypothetical scenarios and on any material changes to supervisory models, and they set out how the program will operate for the 2027 cycle. In that cycle, banks with substantial trading activity will be evaluated against two distinct global market shocks, with the higher of the two loss outcomes determining the market‑shock component of their results. [1]
In parallel, the Board adopted a modification to how certain firms’ SCBs are set. Rather than hinging on a single year’s supervisory test results, the relevant SCB amount will reflect the average of the two most recent annual supervisory stress‑test outcomes. This averaging approach will begin with the 2028 capital‑planning cycle. The Federal Reserve characterizes the resulting reduction in volatility as an estimate, not an observed outcome, and emphasizes that it does not expect a material change in the aggregate level of capital across the banking system from this change. [1]
The proposal, distinct from the final rules, addresses how the Federal Reserve models noninterest income — for example, fee‑based revenue streams — within the supervisory framework. That proposal is not yet effective. It will be open for public comment for 60 days starting from its publication in the Federal Register, after which the Board will evaluate comments and decide on next steps. [1]
Why it matters
These actions are designed to improve the predictability and stability of capital requirements that stem from supervisory stress testing, while preserving the program’s resilience objectives. For banks, averaging consecutive supervisory results when setting applicable SCBs can mitigate the influence of idiosyncratic features in a single year’s scenario or model calibration, supporting steadier capital planning and execution. The Board estimates that this approach could reduce year‑over‑year volatility in certain requirements by about half, even as overall system‑wide capital remains materially unchanged. That estimate speaks to intent; it is not a backward‑looking measurement of realized experience. [1]
For firms with large trading books, the introduction of two alternative global market shocks — and the use of whichever produces the larger loss — is meant to increase the risk sensitivity of the test to a broader range of plausible market dislocations. By evaluating outcomes across two severe configurations, the framework aims to guard against narrow dependence on any single market narrative without implying that either path is more likely. As the Federal Reserve reiterates, the scenarios are tools for assessing resilience under hypothetical stress, not predictions about the economy or markets. [1]
Background and context
The supervisory stress test is a forward‑looking assessment designed to evaluate whether large banks could withstand severe yet plausible economic and financial conditions. The Federal Reserve’s announcement clarifies that, beginning with the 2027 cycle, the framework will explicitly test trading‑heavy firms against two distinct global market shocks and will bind those firms to the higher modeled loss. This sits alongside a new, formalized annual opportunity for public input on scenario design and on material supervisory model changes, embedding a predictable cadence for engagement around how the tests are constructed. [1]
Separately, the Board adopted a methodological change that affects how certain firms’ SCBs are set, replacing reliance on a single year’s supervisory outcome with an average of the two most recent results. The aim is to reduce sensitivity to year‑to‑year noise while maintaining the overall resilience provided by capital buffers. The Board estimates roughly a 50% reduction in volatility from this change and does not anticipate a material effect on aggregate capital; these are forward‑looking estimates contained in the rulemaking materials. [1]
In the same release, the Board issued — but did not finalize — a proposal on modeling noninterest income. The proposal will undergo a standard public‑comment process of 60 days after publication in the Federal Register, and no requirements change until any final rule is adopted. [1]
What remains uncertain
Several elements will be determined through forthcoming supervisory communications and annual publications. The precise macroeconomic paths and the specifications of the two global market shocks that will apply to trading‑heavy firms in 2027 will be finalized through the normal scenario‑release process, now explicitly paired with an annual public‑input step. The realized effect of two‑year averaging on SCB paths will depend on future supervisory results and on how bank business models interact with the scenarios over time. These outcomes cannot be known in advance and will be observed as the framework is implemented. [1]
Uncertainty also surrounds the ultimate shape of the proposed noninterest‑income model. Stakeholders have 60 days after the proposal appears in the Federal Register to submit views, data, and analysis. After that period, the Board may modify, finalize, or decline to adopt the proposal based on the comment record. Until any final action is taken, existing requirements remain in force, and the proposal does not itself alter capital standards or methodologies. [1]
What happens next
The next formal milestone will be the proposal’s publication in the Federal Register, which will start the 60‑day comment window. In parallel, the program will transition to the updated framework elements in the upcoming cycle: an institutionalized annual request for public input on scenario design and on material supervisory model changes, and preparations for the 2027 test that will include two global market shocks for firms with significant trading activity. For those firms, planning and measurement processes will need to account for both hypothetical shocks, with the higher of the two loss outcomes determining the market‑shock component of results. [1]
Beginning with the 2028 capital‑planning cycle, certain firms’ SCBs will reflect the average of their two most recent supervisory test results, operationalizing the objective of reducing year‑over‑year volatility if the Board’s estimate bears out in practice. Throughout, the Federal Reserve reiterates that stress‑test scenarios are hypothetical and severe by design and should not be interpreted as forecasts of the economy or of market performance. [1]
