The Federal Reserve published the results of its 2026 supervisory stress test on 24 June, and by the headline measure the exercise was uneventful. All 32 participating bank holding companies remained above their minimum common equity tier 1 requirements while absorbing more than 708 billion US dollars in hypothetical total losses. Aggregate capital fell by 1.6 percentage points.
What makes this year unusual is not the outcome but its consequence. The results will not change a single bank capital requirement. The Board decided in February that the 2026 test would be run and disclosed without feeding into the capital framework, leaving requirements set by the previous year in place until 2027. At the same time, three separate proposals that would rewrite substantial parts of that framework closed for comment in June and remain unfinished. Large US banks are therefore operating under requirements derived from a test that is no longer current, while the rules that will eventually replace them are still being drafted.
What the 2026 test measured
The hypothetical scenario was similar in severity to the prior year. It assumed a severe global recession, a 39 percent decline in commercial real estate prices, a 30 percent decline in house prices, and unemployment peaking at 10 percent, with economic output falling commensurately. These are modelling assumptions, not forecasts.
Projected losses were concentrated in consumer and corporate credit rather than in real estate. Credit cards accounted for roughly 200 billion US dollars of the total, commercial and industrial loans roughly 160 billion, and commercial real estate roughly 75 billion. The distribution matters: it points to household and corporate balance sheets, not property, as the dominant modelled vulnerability.
The Fed identified three factors behind the change from last year. Higher loan losses, driven by larger loan balances and more severe scenario variables, pushed projected capital down. Lower projected unrealised gains on bank securities, a consequence of smaller hypothetical declines in interest rates within the scenario, pushed it down further. Working the other way, higher projected interest income, reflecting recent bank performance and those same smaller rate declines, more than offset the two.
Vice Chair for Supervision Michelle W. Bowman said the results “underscore the strength of the banking system,” adding that public feedback would help improve the test and confidence in it.
Why this year’s results do not bind
In normal years the stress test is not merely informational. It determines the stress capital buffer, the bank-specific layer that sits on top of the 4.5 percent minimum common equity tier 1 ratio. The buffer has a floor of 2.5 percent, and the largest firms carry an additional surcharge for global systemic importance, itself subject to a floor of 1 percent. Together these components set the level of capital a bank must hold before distributions to shareholders are restricted.
By announcing in February that the 2026 results would not flow through to requirements, the Board effectively froze that calculation. The stated reason is methodological: the Fed intends to re-run the exercise in 2027 using loss-estimating models revised in light of public comment. Freezing requirements in the interim avoids setting binding constraints with models the agency has already committed to change.
The practical effect is a two-year gap between the test a bank has just taken and the test that will eventually determine what it must hold.
The rules that will bind are still being written
On 19 March the Federal Reserve, the Federal Deposit Insurance Corporation and the Office of the Comptroller of the Currency jointly requested comment on three proposals to modernise the capital framework. Comments closed on 18 June. None has been finalised.
The first proposal, aimed primarily at the largest internationally active banks, would implement the remaining components of the Basel III agreement and replace the current requirement to run two sets of risk-based capital calculations with a single set. It would also recalibrate the treatment of credit, market and operational risk, with the market risk elements applying only to banks with significant trading activity. Other banks could adopt the approach voluntarily.
The second proposal, covering all but the largest institutions, would realign capital requirements for traditional lending. It would reduce what the agencies describe as disincentives for mortgage lending by changing the treatment of mortgage origination and servicing. It would also require certain large banks, after a transition period, to reflect unrealised gains and losses on some securities in regulatory capital, a change with real consequences for firms carrying large securities portfolios.
The third proposal, from the Federal Reserve alone, targets how systemic risk is measured for the surcharge applied to the largest and most complex banks. It would update the coefficients used under method 2 of the surcharge framework and index them annually to real economic growth and inflation, so the calibration no longer drifts as the economy expands. It would measure certain systemic indicators using average values rather than a single annual date, reducing the incentive to manage balance sheets around measurement points. And it would assign surcharges in increments of 10 basis points rather than 50, softening the cliff effects that currently arise when a firm crosses a threshold.
The agencies expect the package to reduce aggregate capital in the banking system modestly, with a modest reduction for large banks and a moderate one for smaller banks, while leaving levels substantially above where they stood before the 2008 financial crisis.
Why it matters
Capital requirements are the binding constraint on how much a large bank can return to shareholders. A bank that knows its requirement can plan buybacks and dividends with confidence; a bank facing a framework under revision cannot plan with the same certainty beyond the current cycle.
Two elements of the proposals deserve particular attention. Requiring certain large banks to recognise unrealised securities gains and losses in regulatory capital would tie reported capital more closely to interest rate movements, a linkage that became painfully visible during the regional bank failures of 2023. And the shift to 10 basis point increments in the systemic surcharge changes behaviour at the margin: a firm approaching a threshold currently faces a discrete 50 basis point penalty, which creates a strong incentive to manage activity around the cut-off. Finer increments, combined with averaged measurement, are designed to remove that incentive.
For the wider economy, the direction is toward modestly less required capital across the system. Whether that is prudent depends on assumptions about the risks the framework is meant to absorb, and reasonable observers disagree.
Risks and competing interpretations
The March proposals did not command quiet consensus. Separate statements accompanied the release from then-Chair Jerome Powell, Vice Chair for Supervision Bowman, and Governors Michael Barr, Stephen Miran and Christopher Waller. The volume of individual statements is itself a signal that the Board was not uniformly comfortable with the package.
The freeze on capital requirements is open to two readings. One is procedural prudence: it would be odd to impose binding constraints using models the agency has already conceded need revision. The other is that a supervisory tool has been suspended for two years at a point when the framework is being loosened, and that the combined effect is a meaningful reduction in near-term regulatory pressure. Both readings are consistent with the published record.
There is also a question of continuity. The proposals were issued under Powell in March. Kevin Warsh took the oath as Chair on 22 May and delivered his first semiannual monetary policy testimony to Congress on 14 July. The officials who finalise these rules will not be the same group that proposed them, and no statutory deadline compels a particular timetable.
What to watch
Three things will determine how this resolves. First, whether the agencies finalise any of the three proposals in 2026 or let them run into next year, and whether the final texts preserve the calibration set out in March. Second, the design of the revised loss-estimating models that will drive the 2027 test, since those models, not this year’s results, will set the next binding requirements. Third, how bank capital planning responds in the interim, which will show up in distribution announcements rather than in regulatory filings.
Readers following the wider policy backdrop may find our coverage of the Federal Reserve decision and growth data useful context, along with our reporting on how energy prices are feeding into the inflation outlook. Further banking coverage from Capital Pulse Wire is collected in our banking section.
All figures cited are drawn from Federal Reserve publications dated 4 February, 19 March and 24 June 2026, and are stated in US dollars. Stress test loss figures are hypothetical projections under a supervisory scenario, not forecasts.
The rate-setting side of the same institution is examined in our coverage of the July policy meeting.
About the data: Results, participant numbers, aggregate loss figures and capital ratios are from the Federal Reserve 2026 supervisory stress test published on 24 June 2026. Scenario parameters, including the assumed declines in commercial real estate and house prices and the unemployment path, are as specified by the Federal Reserve Board. The decision to run the 2026 test without feeding it into capital requirements is a published Board decision.
Reader note
This article is general information, not personalized financial advice. Read our Financial Disclaimer.