Federal Reserve Vice Chair for Supervision Michelle Bowman said Thursday that the final revisions to the bank stress testing framework will arrive in the coming weeks, and she put specific mechanics on the table: the stress capital buffer will be calculated by averaging each bank's two most recent stress tests, the effective date moves from October to January, and the models themselves will be published with their equations and coefficients. The combination, she said in a London speech, cuts the volatility of capital requirements roughly in half without materially changing the total capital the system holds. The framework banks sued over is being replaced by one they can inspect.
The averaging change is the arithmetic heart of the overhaul. Under the current framework, a bank's stress capital buffer resets each year based on that year's test alone, which means a single adverse drawdown can swing a firm's required capital sharply from one October to the next. Averaging the two most recent results smooths the year-to-year jump. Bowman's claim that the change reduces SCB volatility by half while leaving aggregate required capital levels materially unchanged is the kind of number that gets checked against the first year of data, and it is testable in exactly the way the old framework was not.
The transparency promise that is the actual reform
The disclosure commitments are the deeper change, because they attack the complaint that produced the litigation in the first place. Banks and their trade groups have spent years arguing that the stress test is an opaque model: inputs are known, outputs are known, and the mechanism connecting them is a black box. Bowman said the Fed will now disclose detailed information about the models, including equations, variables, coefficients, assumptions, and limitations, plus the rationale behind decisions, and that the public will be able to comment on material model changes. A regulated firm that can see the model can contest it on the merits. That is the structural fix.
The supervisory uses of stress testing are the part of the speech that will matter most to bank examiners and least to markets. Bowman said stress tests should do more than produce a single capital number: they should identify vulnerabilities to material financial and nonfinancial risks before they emerge. Those supervisory uses will not be formal rulemaking, and the results will not be disclosed. The Silicon Valley Bank reference was explicit: she noted that such analysis could have exposed SVB's interest-rate and uninsured-deposit vulnerabilities as early as the fourth quarter of 2021. The point is that the old test asked one question, and the new framework is supposed to ask several.
What is still under consideration before the final rule
Two additional changes sit between the speech and the final rule, and Bowman described them as live options. The first would freeze firm balance sheets on a specified date before scenarios are released, which is a risk-sensitivity measure aimed at preventing banks from repositioning portfolios to game the test. The second would run two global market shock scenarios on the same as-of date and use the larger loss in the SCB calculation. Both respond to specific criticisms of the current framework, and their inclusion or exclusion will define how far the final rule goes.
The 2027 model proposal is the forward-looking piece. Bowman said she will recommend a third proposal to revise the models used in next year's test, including a new model for noninterest income that captures fee and trading revenue from wealth management, investment banking, and market making. The existing noninterest income model treats diverse businesses as if they are one thing, and the new model is supposed to recognize business diversity. For banks whose earnings mix has shifted toward fees and trading, that model change is not technical; it is the difference between a capital requirement that fits the business and one that does not.
The timeline that makes this a capital event, not a speech
The calendar is the practical news. Bowman expects the stress testing rule finalized within weeks, with reforms to the global systemically important bank surcharge and the Basel III capital framework to follow before year end. That sequence means banks will know the test they face in 2027 before the year is over, which is itself a change from the era when the rules landed after planning cycles had begun. The speech in London was delivered to an audience that includes the banks affected, and the message was that the opaque framework has a closing date.
For the banks, the practical question is whether the new transparency actually stabilizes capital planning. Averaging and disclosure reduce surprise, but they do not reduce the severity of the scenarios themselves, and the supervisory uses Bowman described could surface issues earlier and more often. A framework that publishes its models also publishes its weaknesses, and a bank that can see the model can also see how to argue with it. That is what accountability means, and the Fed has now put a date on it.