The Federal Reserve finalized changes to its large-bank stress tests on September 30, setting up a debate over how to make the exercise more predictable without making it easier to prepare for a narrow set of risks. The package changes public consultation and future capital calculations; it is not an announcement that a particular bank has passed a new test.
Stress tests ask how banks would fare in hypothetical adverse conditions. For households and businesses, their relevance is the resilience of institutions that provide credit and payment services. The results are one part of supervision, rather than a promise that every future shock has been anticipated.
What is changing
Under the Board’s final rules, the Fed will seek annual public input on scenarios and significant model changes. Large trading firms will face two market-shock components, with the larger firm-specific loss used in the result. Starting in 2028, eligible firms’ stress capital buffer calculations will average their two most recent annual tests.
The Fed estimates the package will reduce annual volatility in capital requirements by about half, without materially changing aggregate requirements. Those are the Board’s expectations, not measured outcomes from an implemented multiyear system. A separate fee-income model change remains a proposal with its own comment process.
Why officials disagree
Governor Michael Barr opposed the final rule, arguing that greater predictability and procedural constraints could weaken the tests’ ability to respond to emerging risks. His concern is that banks could adjust their portfolios to do well against known models while underlying vulnerabilities remain.
Governor Lisa Cook’s separate statement emphasized the information supervisors gain from stress testing and the need to preserve sufficiently demanding scenarios. She also highlighted the value of exploratory scenarios that need not directly determine regulatory capital.
The disagreement concerns the design of a safeguard. Transparency can make assumptions easier to challenge, while adaptability can help a test examine an unfamiliar problem. Neither a more stable capital number nor a more complicated model, by itself, demonstrates that the banking system is safer.
Read the next results in context
When future test results arrive, look first at the scenario, the participating firms and the period being assessed. A loss estimate under one hypothetical recession is not a forecast that the recession will occur. Likewise, a change in a capital requirement can reflect a changed method as well as changes in a bank’s balance sheet.
A simple comparison illustrates the problem. If a school changes the examination and a student’s score changes, the score alone cannot tell you how much the student’s knowledge changed. Bank testing is more complex, but the need to distinguish the measurement from the measured institution is similar.
For everyday banking decisions, this announcement does not set a new savings rate, loan price or account condition. Those remain questions about the actual product and institution. The practical next step for readers following the policy is to track implementation dates, proposed changes and published results separately, rather than treating the entire package as already in effect.
This is distinct from the contingency planning covered in our report on large-bank resolution plans.
Illustrative financial calculation; no bank stress-test result is pictured. Photo: olia danilevich / Pexels.
