Banks, Capital & Financial Stability • October 5, 2026
The Fed Just Rewired How America Tests Its Biggest Banks
The new framework opens scenarios and material model changes to public input, adds two market shocks, and begins averaging capital results in 2028.
Executive Takeaways
- The Federal Reserve finalized two rules intended to make its annual large-bank stress test more transparent and less volatile.
- Beginning with the 2027 test, the public will be invited to comment annually on scenarios and material model changes.
- Banks with large trading books will face two global market shocks, with the larger loss used for each firm.
- Starting in 2028, stress capital buffers will use the average of the two most recent tests; the Fed estimates roughly 50% less year-to-year volatility and no material change in aggregate capital requirements.
What Changed—and When
The Federal Reserve’s September 30 package changes both the process and the math behind the annual supervisory stress test. The first final rule requires annual public input on hypothetical scenarios and material model changes, updates the scenario-design framework, adopts the models for the 2027 exercise, adjusts the calendar, and changes the global market shock. Firms with large trading books will be evaluated against two market shocks, and the larger loss will determine that component of each firm’s result. The second rule changes the stress capital buffer calculation: beginning in 2028, eligible firms’ results from the two most recent annual tests will be averaged. The delay is deliberate. The Fed says the averaging should begin only after the calculation can rely on models that have incorporated public feedback.
The Case for Stability
The strongest argument for averaging is that a bank’s capital plan should not swing sharply because a single annual exercise produces an unusually high or low modeled loss. Capital requirements shape dividends, share repurchases, lending capacity, balance-sheet decisions, and investor expectations. A two-year average can make those decisions more stable while retaining risk sensitivity. The Fed estimates the package is likely to reduce year-over-year volatility in requirements by about 50 percent and says it is not expected to materially change aggregate capital requirements. That last qualification matters: less movement does not automatically mean less capital systemwide, although individual firms can still see meaningful changes as their portfolios and modeled losses evolve.
The Risk of Making the Test Too Predictable
Transparency also creates a genuine tradeoff. Public comment can expose weak assumptions, improve accountability, and let outside experts challenge the framework. But excessive predictability can encourage banks to optimize portfolios for the published test rather than for the underlying risk. Governor Michael Barr dissented, warning that disclosure and procedural limits could make models less dynamic and more vulnerable to gaming. Governor Lisa Cook supported the rule while emphasizing that scenarios must remain genuinely stressful and that exploratory, nonbinding scenarios can preserve dynamism. The two-shock market component is one defense: a bank cannot rely on a single market path. The long-run test is whether the framework continues to identify emerging losses before markets do.
A transparent test earns confidence only if it remains severe enough to reveal the losses banks would rather not find.
The distinction between a stress scenario and a forecast is essential. A supervisory stress test asks how a bank would perform under a deliberately severe hypothetical path; it does not predict that path will occur. In the 2026 exercise, the scenario included unemployment peaking at 10 percent, house prices falling about 30 percent, commercial real-estate prices falling 39 percent, and severe financial-market volatility. All 32 tested banks remained above minimum common-equity tier 1 capital requirements even after more than $708 billion in projected total losses. Those results were not used to reset capital requirements because the Board had decided to wait for models incorporating public input.
For households and small businesses, the relevant outcome is not the elegance of a model. It is whether the banking system can absorb losses and continue extending credit when unemployment rises, asset prices fall, and borrowers come under pressure. If requirements are erratic, banks may react with abrupt balance-sheet adjustments. If requirements are too low, the system may lack adequate loss-absorbing capacity. If tests are too predictable, firms may appear resilient inside the model while risk migrates elsewhere. The policy challenge is therefore to combine stability, severity, transparency, and supervisory judgment rather than maximizing any one of them in isolation.
The business-model question remains open. Alongside the final rules, the Fed requested comment on proposed revisions to the model that projects noninterest income, such as fee revenue. A universal model can miss differences among custody banks, trading-heavy firms, card lenders, and more traditional commercial banks. A more granular model may improve fairness and risk sensitivity, but additional complexity can also make validation harder. Comments are due sixty days after Federal Register publication, and the final design will matter for how well the test reflects the real engines of each institution’s revenue.
Confirmed facts versus analysis: the effective dates, annual public-comment process, two global market shocks, two-year averaging method, and estimated 50 percent volatility reduction come from Federal Reserve materials. The judgment that the framework must balance stability, severity, transparency, and supervisory judgment is RedWaveBrief analysis. Governor Barr’s concerns and Governor Cook’s safeguards are presented as their stated positions, not settled outcomes.
Three scenarios frame the next two years. In the base case, the 2027 process becomes more transparent, model criticism improves at the margin, and 2028 averaging smooths buffers without materially lowering aggregate capital. In the upside case, public scrutiny finds weaknesses early, dual shocks expose trading risks, and investors gain confidence in predictable but credible requirements. In the downside case, firms optimize for disclosed models, emerging risks sit outside the annual scenario, and smoother buffers mute a warning that should have been louder.
Action Checklist: bank investors should track each firm’s stress capital buffer and management payout plans; business owners should watch credit standards rather than assume smoother rules guarantee easier loans; households should keep deposits within applicable insurance limits and avoid treating any test as a zero-risk certificate; boards should challenge whether internal scenarios are broader than the supervisory test; policymakers should publish evidence on model performance, not just process changes. Choose Our Next Deep Dive: trading-book shocks, capital buffers, bank fee-income models, or deposit resilience. Ask the Analyst: send the bank-risk question you want examined.
How to Read This Development
Readers should resist two common mistakes. The first is treating an official announcement as proof that every projected benefit has already arrived. The second is dismissing a serious program because execution is not immediate. Public policy moves through stages: announcement, award or rulemaking, contracting, implementation, measurement, and revision. Each stage produces different evidence. A disciplined reader asks what has actually occurred, what remains conditional, who bears the cost, and which public record can verify the next milestone.
That framework also separates national strategy from partisan theater. America benefits when infrastructure is reliable, trade commitments are enforceable, and financial markets are resilient. Those goals do not require blind faith in an administration or reflexive hostility to it. They require transparent metrics, clear accountability, and a willingness to update conclusions when new evidence arrives. Our scenarios therefore describe conditions, not certainties, and our practical checklist is designed to help readers follow the evidence.
What Could Change the Conclusion
A later contract, regulatory filing, shipment report, construction update, market statistic, or official revision could materially change this assessment. We will treat those records as higher-value evidence than anonymous speculation. Readers should also distinguish nominal totals from inflation-adjusted value, capacity from energy produced, planned purchases from delivered goods, and trading volume from economic output. Those distinctions prevent impressive numbers from doing more work than the underlying facts support.
Finally, this analysis is general information, not individualized investment, legal, tax, or financial advice. Decisions should reflect personal time horizons, cash needs, risk tolerance, and independent professional guidance where appropriate.
A Practical Accountability Standard
We use five questions to judge the next update. First, is the metric observable in a public record rather than available only as a talking point? Second, does it measure an outcome—capacity delivered, goods shipped, trades cleared, costs reduced—instead of an activity such as meetings held or dollars announced? Third, is there a deadline and a responsible institution? Fourth, can outsiders compare the result with a prior baseline? Fifth, does the evidence identify who gains, who pays, and what risks remain? A development that passes all five tests deserves more confidence than one supported only by broad assurances.
Timing also matters. Short-term market reactions can reflect positioning, headlines, and expectations rather than the eventual economic effect. Medium-term evidence usually comes from contracts, regulatory records, operational statistics, and audited results. Long-term judgment requires comparing the promised national benefit with total cost and opportunity cost. We therefore avoid declaring victory or failure from a single day’s price move. The useful question is whether the evidence is moving in the direction promised.
For readers making decisions now, preserve flexibility. Do not rely on one policy announcement for a major purchase, concentrated investment, hiring plan, or retirement decision. Build a base case that can tolerate delays, identify the data that would justify greater confidence, and write down the condition that would prove the thesis wrong. That simple discipline turns a news headline into an accountable decision process.
We will revisit the thesis when the responsible agencies publish the next measurable milestone. If the official record conflicts with an earlier claim, the record—not the rhetoric—will control our update.
What to Watch
- 2027 scenarios and models: the first full exercise under the new transparency framework.
- Noninterest-income proposal: whether the Fed better captures different fee-based business models.
- 2028 buffer averaging: the first capital requirements based on two recent supervisory tests.
Sources & Methodology
- Federal Reserve — final stress-test transparency and capital rules
- Federal Reserve — 2026 stress-test results
- Governor Lisa Cook — statement on the final rule
- Governor Michael Barr — dissenting statement
Primary official materials were reviewed directly. Facts and published estimates are identified as such; interpretation and scenarios are RedWaveBrief analysis. Accessed October 5, 2026.