The Fed Will Average Two Years of Stress Test Results Into Bank Capital Requirements. It Will Also Start Publishing the Models.
Two final rules published on 2 October change how the stress capital buffer is calculated and how much of the supervisory stress test the public gets to see before it runs.

The short answer
- The Federal Reserve will average the stress capital decline component from two consecutive annual stress tests, symmetrically, rather than using a single year's result.
- The Board estimates this cuts year-over-year volatility in the requirement from an average of 64 basis points to 47 basis points, while leaving the average buffer itself close to unchanged at 3.80 percentage points against 3.83 today.
- The annual effective date of the stress capital buffer moves from 1 October to 1 January, lengthening the gap between the data behind a test and the capital requirement it produces from about 11 months to about 18.
- A companion rule commits the Board to publishing model documentation each year by 15 May, proposed scenarios from 10 January with at least a 30-day comment period, and material model changes by 31 August of the preceding year.
The Federal Reserve published two final rules in the Federal Register on 2 October that together rework the supervisory stress test: one changes how the test sets a bank's capital requirement, the other changes how much of the test the public sees before it is run. Both were adopted after a comment process, and both take effect this autumn.
Averaging, and what it is meant to fix
The stress capital buffer sits on top of a large bank's minimum capital requirement. Its size is driven by how far the bank's capital falls in the Board's annual severely adverse scenario. Because that scenario changes every year and the models behind it change too, the resulting requirement has moved around a good deal from one year to the next.
Under the final rule, the Board will take the stress capital decline components from two consecutive annual supervisory stress tests and average them symmetrically. The Board's own estimate is that this reduces year-over-year volatility in the requirement from an average of 64 basis points to 47 basis points.
It is a volatility change, not a level change. The Board puts the average stress capital buffer requirement under the new framework at 3.80 percentage points, against 3.83 percentage points under the current one.
The dividend add-on is not averaged
The stress capital buffer also includes an add-on covering four quarters of planned common dividends. That component survives the rule intact and continues to be updated annually. It is not subject to the two-year averaging.
The calendar moves, and the lag grows
The annual effective date of the stress capital buffer requirement shifts from 1 October to 1 January, which gives firms an additional quarter between notification and compliance. The trade-off is in the other direction: the Board says the gap between the data a test is run on and the capital requirement it ultimately produces lengthens from about 11 months to about 18 months on average.
- Final rule effective: 1 December 2026
- FR Y-14 reporting form revisions: 30 June 2027
- Results averaging first applies: 1 January 2029, following the 2028 stress test
- Firms covered: large bank holding companies, savings and loan holding companies, U.S. intermediate holding companies of foreign banking organisations, and nonbank financial companies supervised by the Board
The Board received 16 comments. Commenters generally supported reducing volatility; some opposed averaging outright, and others argued over whether the averaging should be symmetric or asymmetric and whether the dividend add-on should be retained at all.
The second rule: showing the work
The companion rule, on transparency and public accountability of the supervisory stress test models and scenarios, is the larger document of the two at 64 Federal Register pages. It commits the Board to a published annual calendar of disclosure rather than leaving the models proprietary.
- Proposed scenarios published for comment from 10 January, after the 31 December jump-off date, with a minimum 30-day comment period.
- Final scenarios published by 28 February.
- Model documentation disclosed by 15 May each year.
- Material model changes published for public input by 31 August of the year preceding the stress test they would apply to, with the Board responding to substantive comments before adoption.
The rule defines a material model change with a threshold rather than a judgement: a change that would affect any individual firm's post-stress capital ratio by 20 basis points or more, or the average across firms by 10 basis points or more.
The 2027 cycle, dated
- 5 January: Category IV firms decide whether to participate
- 10 January: proposed scenarios published
- 28 February: final scenarios released
- 30 April: capital plan submission deadline
- 15 May: final model documentation disclosed
- 30 June: stress test results published
- 30 September: final stress capital buffer notification
This rule takes effect on 2 November 2026.
Sources
- Modifications to the Capital Plan Rule and Stress Capital Buffer Requirement (final rule, 91 FR 62636) — Federal Reserve System / Federal Register
- Enhanced Transparency and Public Accountability of the Supervisory Stress Test Models and Scenarios (final rule, 91 FR 62870) — Federal Reserve System / Federal Register
Spotted an error? Tell our corrections desk.
How this article was produced
- Responsible desk:
- Analysis & Opinion
- Published:
- 5 Oct 2026, 05:14 UTC
- Last updated:
- 5 Oct 2026, 05:14 UTC
- Verification:
- Figures and quotations checked against primary sources under our fact-checking policy and editorial standards.
- Independence:
- No advertiser or affiliate partner had any involvement in this article — see editorial independence and how we make money.
- Corrections:
- Report a factual error.
This article is general financial information and journalism, not personalised financial, investment, tax or legal advice.
