The capital proposal targets volatility in large-bank buffers. The separate SVB review agrees that supervisors acted too slowly, while disputing how much post-2018 regulatory tailoring contributed to the failure.
Federal Reserve Vice Chair for Supervision Michelle Bowman says the Board is preparing final changes that would average the two most recent annual stress-test results when setting a large bank's stress capital buffer. She estimates the change would cut buffer volatility roughly in half without materially changing aggregate capital requirements.
The Fed is also considering frozen balance-sheet assumptions in parts of the test and two global-market-shock scenarios. Implementation of the revised buffer framework is expected to move from October 1 to January 1.
For bank management, two-year averaging can make dividends and repurchases easier to plan because one unusually harsh annual test has less immediate effect. The trade-off runs both ways: an older weak result can remain in the requirement for longer.
The SVB reviews agree on delay and disagree on attribution
Bowman also released initial findings from an externally led review of the Federal Reserve's supervision of Silicon Valley Bank. The review points to vulnerabilities that were visible before the March 2023 failure, including securities losses larger than capital, a deposit base that was roughly 94% uninsured and weak readiness to borrow from the Fed's discount window.
The new review says supervisors knew or should have known about serious weaknesses by March 2022 and failed to force prompt corrective action. It also argues that post-2018 regulatory tailoring should not carry the explanatory weight assigned to it in the Fed's 2023 internal review.
The earlier review led by Michael Barr reached a partially different conclusion. It also found that supervisors were too slow, but said the Fed's tailoring framework and less assertive supervisory stance impeded effective oversight.
The chronology matters. The 2026 review does not erase the 2023 record; it changes the institutional debate over why the supervisory failure persisted.
Capital formulas and supervision solve different problems
Smoother stress buffers can reduce year-to-year noise in large-bank capital distributions. SVB remains the counterpoint: standardized capital tests do not replace institution-specific judgment about liquidity, interest-rate exposure and deposit concentration.
Mechanically, the reform offers lower volatility in required buffers. The supervision debate remains less settled, because the Fed's two reviews agree on missed action while assigning different weight to the rules and supervisory culture that produced it.
Sources
- Federal Reserve, September 18 stress-test remarks: https://www.federalreserve.gov/newsevents/speech/bowman20260918a.htm
- Federal Reserve, September 18 SVB review remarks: https://www.federalreserve.gov/newsevents/speech/bowman20260918b.htm
- Federal Reserve, April 2023 SVB review: https://www.federalreserve.gov/newsevents/pressreleases/bcreg20230428a.htm