Bill Dudley, former president of the Federal Reserve Bank of New York, is making a simple argument that somehow keeps needing to be repeated: fix the safety net before you take it down.
In a Bloomberg Opinion column published on September 9, Dudley laid out the case that US regulators need to develop a significantly stronger resolution regime before they even think about loosening bank capital requirements and supervision. The timing is pointed, given that multiple regulatory agencies are currently doing exactly the opposite, proposing lighter capital rules while resolution frameworks remain largely unchanged from the ones that fumbled through the 2023 banking crisis.
The regulatory paradox taking shape
On one hand, US regulators proposed changes in March 2026 that would reduce capital requirements for large banks on a net basis. On the other hand, the mechanisms for dealing with a bank that actually fails remain stubbornly inadequate. The collapse of Silicon Valley Bank’s parent company resulted in nearly $20 billion in FDIC losses. Those losses weren’t absorbed by SVB’s investors, the people who theoretically accepted the risk. They were recovered through higher assessments on other banks.
The FDIC, for its part, appears to be moving in the same direction as the capital rule changes. In June 2026, the agency proposed raising the asset threshold for resolution plan submissions from $50 billion to $100 billion, while also reducing the documentation burden on covered institutions. Fewer banks would need to maintain detailed plans for their own orderly wind-down, and those that do would face lighter requirements.
For context, SVB had roughly $209 billion in assets when it failed. It was above the $50 billion threshold. But the resolution still went sideways, suggesting that having a plan on paper and having a plan that actually works are two very different things.
Why resolution matters more than capital ratios
A well-functioning resolution regime is supposed to solve this problem. When a bank fails, its equity holders get wiped out, its bondholders take losses, and the broader financial system keeps functioning. That’s the theory. In practice, the SVB episode and the Credit Suisse crisis in Switzerland showed that regulators often lack the tools, or the will, to impose those losses cleanly.
The UK offers an interesting counterpoint. British authorities have explicitly linked credible resolution regimes to lower benchmark capital requirements, estimating that a functioning resolution framework could justify a reduction of approximately 5 percentage points in capital buffers.
Dudley appears to endorse a version of this logic, but with a critical caveat. The UK approach only works if the resolution regime is actually credible. Cutting capital requirements first and promising to fix resolution later reverses the order of operations in a way that leaves the financial system exposed.
Dudley’s prescription, building resolution tools that foster prudent risk management, limit contagion, and uphold market discipline, reads like a checklist of everything that went wrong during those episodes.