Status and the correct comparison
On March 19, 2026 the banking agencies released a revised capital package. It includes a Basel framework for the largest banking organizations, changes to the standardized approach for other banks, and a Federal Reserve G-SIB surcharge proposal. Comments were due June 18, 2026. The materials reviewed for this September 27 article describe proposals, not a final effective capital rule. [1]
Do not carry forward a headline estimate from the 2023 proposal or a 2024 speech as though it measures the March 2026 text. Equally, do not describe all proposed capital reductions as the effect of Basel alone. The components move in different directions, and separately proposed stress-test changes alter the combined picture.
The September 18 speech described expected action at that time. The Fed subsequently finalized separate stress-testing rules on September 30, as detailed below. That action does not itself finalize the March Basel III, standardized-capital or G-SIB proposals. [8, 9]
September 30 final stress-test changes: adjacent to Basel III
On September 30, 2026, the Federal Reserve finalized separate stress-testing rules requiring annual public input on scenarios and material model changes and averaging stress-capital declines from two consecutive annual tests for participating firms. Averaging begins with the 2028 calculation, using the 2027 and 2028 tests, for buffers effective January 1, 2029; the dividend add-on remains annual. The stress capital buffer rule itself takes effect December 1, 2026. The first January 2028 buffers remain unaveraged. [9, 10]
The Fed estimates approximately 50% less year-to-year capital-requirement volatility without a material aggregate capital effect. A separate fee-income model revision remains proposed, with comments due December 1, 2026. These stress-testing actions do not finalize the March capital proposals. The March combined-impact estimates below remain historical proposal scenarios rather than estimates of this final package. [9, 11]
A capital change can alter the mix of financial services
Capital supports several activities inside a bank, including lending, trading and operational risk. A change in one requirement can alter their relative attractiveness even when the aggregate requirement moves little. That makes portfolio composition and the location of the binding constraint more informative than a single industry-wide relief estimate.
The wider question is which useful services become easier or harder to supply at a sustainable price. A bank may respond through balance-sheet growth, different asset choices, retained earnings or distributions. None of those responses follows mechanically from a proposed reduction in one risk-weight calculation.
Who would face which framework?
The Basel proposal principally addresses Category I and II firms. It would replace the existing overlapping credit-risk calculations with a single expanded approach. The standardized proposal addresses other banking organizations. Significant trading activity can bring additional firms within proposed market-risk requirements; the fact sheet identifies $5 billion of trading activity or 10% of assets as relevant thresholds. [1, 2]
Recommended scoping starts with the legal entity, prudential category and activity thresholds. A regional lender’s exposure to the standardized changes is different from a large trading bank’s exposure to market-risk reform. Consolidated parent results also need reconciliation to subsidiary constraints before anyone estimates capital available for distributions or loan growth.
What changes economically
The agency fact sheet describes more differentiated treatment of mortgages and other exposures, revisions to operational and trading risk, changes in mortgage-servicing-asset treatment and a five-year AOCI transition for Category III and IV firms. The G-SIB proposal refines score bands and measurement. These are distinct mechanisms, not a uniform percentage haircut to every loan. [2]
The proposed standardized rule is the right place to test exposure eligibility, attributes and assigned weights. Portfolio classification and data quality will matter. A generic business-loan label is not sufficient evidence that an exposure receives a favorable category. [3]
Analytically, implementation can increase reporting and data costs even where required capital declines. Mortgage attributes, borrower characteristics, commitments and trading positions need traceable mappings. A lower aggregate requirement does not prove that a particular card, mortgage or warehouse portfolio benefits.
Customer pass-through depends on competition and demand
A lower capital cost can create room for a lower price, but customers receive that benefit only if the institution changes its offer. Competition, customer bargaining power and the availability of attractive business all influence the result. Weak demand can leave additional capacity unused even when a bank is willing to expand.
Conversely, higher requirements can lead to repricing or a smaller position without proving that an activity is socially undesirable. The policy debate weighs the resources absorbed by intermediation against the resilience those resources provide. Assess actual prices, availability and performance rather than equating regulatory cost with the full public cost or benefit.
Credit cards: the drawn balance is only part of the calculation
The proposed expanded approach assigns 75% to qualifying regulatory retail exposures that are not transactor exposures, 45% to transactor exposures, and 100% to other retail exposures. It also proposes a 10% credit conversion factor for unused unconditionally cancelable commitments, versus zero under the current rule. These provisions belong to the expanded approach and its eligibility definitions; do not apply them mechanically to every bank or consumer loan. The proposal itself notes that lower retail weights can be offset by the new unused-line charge. [7] For a credit facility, the proposed transactor definition requires full repayment at each scheduled repayment date during the preceding 12 months. Promotional zero-interest installment balances do not qualify merely because no payment was yet required. [7]
Illustrative credit-risk-only comparison: assume a qualifying revolving card has a $10,000 limit and $2,000 drawn. Under an assumed current standardized treatment of a 100% drawn risk weight and zero conversion on the remaining $8,000, is $2,000. Under the proposed 75% weight and 10% conversion, the exposure amount is $2,000 + $800 = $2,800, producing $2,100 of RWA. The lower weight produces 5% more RWA in this particular low-utilization example.
The figures below hold eligibility and the contractual line fixed. They exclude operational-risk capital, buffers, leverage constraints and any change in borrower behavior. They are calculations under stated assumptions, not a forecast for any issuer.
Scroll horizontally to see all columns.
| Hypothetical $10,000 line | Drawn / unused | Current assumed credit RWA | Proposed credit RWA |
|---|---|---|---|
| Qualifying revolver; 20% utilized | $2,000 / $8,000 | $2,000 | $2,100 at 75% × ($2,000 + 10% × $8,000) |
| Qualifying revolver; 40% utilized | $4,000 / $6,000 | $4,000 | $3,450 at 75% × ($4,000 + 10% × $6,000) |
| Qualifying transactor; 20% utilized | $2,000 / $8,000 | $2,000 | $1,260 at 45% × ($2,000 + 10% × $8,000) |
Why installment lending and unused card limits respond differently
Under the revolver assumptions above, credit is unchanged at approximately 23.1% utilization. The equation is 0.75 × [u + 0.10 × (1 − u)] = u, where u is the drawn share of the line. Below that level, the unused commitment can outweigh the reduced drawn weight. This is an analytical crossover, not an optimal credit-limit target: reducing a useful line can affect customer , , spending and competitive position, and must respect applicable notice and other requirements.
A fully disbursed closed-end installment loan has no remaining revolving line in this simplified comparison. If it qualifies for the assumed 75% category, $10,000 outstanding generates $7,500 of credit RWA. But classification, treatment, collateral and legal-entity scope still need verification. The contractual amortization schedule affects future exposure; longer duration also changes credit, funding and interest-rate risk. A capital calculation should not substitute for lifetime loss estimation or an affordability assessment.
For a card business, maintain a bridge from accounts and limits to drawn exposure, unused commitments, assigned categories and total RWA. Keep transactor classification and repayment history reproducible. Then add the applicable operational-risk calculation and capital stack. Portfolio averages can hide a concentration of low-utilization lines or customers migrating between repayment behaviors. An apparent reduction at launch may reverse as utilization and losses change.
Market liquidity and bank resilience need compatible evidence
Trading-related capital can affect the willingness to hold inventory and intermediate customer transactions. A claim that lower requirements improve market functioning should identify which activity changes and how execution or available improves. Higher trading volume alone does not establish more dependable liquidity during stress.
A balanced review considers normal-market service and the ability to absorb losses when conditions deteriorate. The March package remains a proposed framework in the checked agency release, and separately proposed changes must retain their own assumptions. Final text and bank-specific implementation will determine how the mechanisms interact. [1]
Reading the impact estimates without mixing denominators
Historical March 2026 proposal scenario: The Federal Reserve staff memorandum estimates that, for Category I and II firms, the Basel component increases required capital by 1.4%, while the G-SIB component reduces it by 3.8%, producing a combined 2.4% reduction. Including separately proposed stress-test changes produces a 4.8% reduction. For Category III and IV firms, the standardized package including AOCI produces an estimated 3.0% reduction, or 5.2% including the stress changes. [4]
These are aggregate percentage changes in estimated required capital, not percentage-point reductions in regulatory capital ratios, individual-bank forecasts or already available cash. Different mixes of assets, AOCI and binding constraints can produce substantially different outcomes.
These March estimates assumed separately proposed stress-test changes. They are retained as historical scenario analysis and are not recalculated estimates of the September 30 final stress-test rules. [4, 9, 10]
Scroll horizontally to see all columns.
| Comparison | Staff estimate | Interpretation |
|---|---|---|
| Category I/II: Basel component | +1.4% | Increase in estimated required CET1 |
| Category I/II: Basel plus G-SIB | −2.4% | Combined package, excluding separate stress changes |
| Category I/II: also include proposed stress changes | −4.8% | Broader combined scenario |
| Category III/IV: standardized plus AOCI | −3.0% | Aggregate; individual bank outcomes differ |
Worked example: capital efficiency is not lower credit loss
Illustrative calculation using hypothetical weights, not a classification conclusion: a $100 million exposure pool at a 100% risk weight produces $100 million of . At 90%, it produces $90 million. With an assumed 10% capital target, the associated capital falls from $10 million to $9 million. The difference is $1 million of capital capacity, not $10 million of new cash.
At an assumed 12% annual cost of equity, that $1 million represents $120,000 of annual capital-cost sensitivity. A 20-basis-point deterioration in annual losses on the $100 million pool is $200,000 and would outweigh it. Funding, expenses and taxes further affect returns. The example explains why capital relief cannot justify weaker underwriting by itself.
Nor is the $1 million necessarily distributable. Leverage requirements, stress losses, subsidiary requirements or a management buffer may be binding. A bank must identify the actual limiting constraint and rerun its capital plan before turning a risk-weight calculation into a lending or payout decision.
Why policymakers disagree
Vice Chair Bowman supports the package as better matching requirements to risk, addressing overlap with stress testing and reducing incentives for traditional lending to migrate outside banks. Her argument emphasizes calibration and credit availability. [5]
Governor Barr dissented. He argues that the combined reductions are not justified, that some departures weaken the international framework and that the interaction with leverage and stress changes matters. He supports some individual elements, including more risk-sensitive features and AOCI recognition, while opposing the package as a whole. [6]
The practical disagreement is about the resilience purchased by an additional dollar of capital versus its effect on intermediation. My assessment is that aggregate release estimates are insufficient to settle it. Evidence should examine stress losses, funding fragility, correlated exposures and whether lower requirements actually translate into durable lending capacity.
Translate a rule scenario into a bank-specific decision
Recommended analysis uses four separate columns: current requirements; the March proposal alone; the proposal with separately assumed stress-test changes; and a downside operating scenario. Reconcile every column to the same balance-sheet date. AOCI recognition, risk weights, operational risk and surcharge changes should each have their own bridge so benefits and costs are not counted twice.
Hypothetical constraint test: a bank holds $110 million of usable capital. Assume its internal risk-based target falls from $100 million to $90 million under a proposal, while a different binding capital requirement remains $105 million. Usable headroom remains $5 million, not $20 million. The capital types, legal entities and denominators must be compatible before using this simplified maximum-constraint comparison. Management buffers and projected losses can further reduce headroom.
Price the marginal loan after identifying that constraint. Compare expected interest and fees with funding, credit loss, servicing, fraud and capital cost over a consistent period. A warehouse-backed installment portfolio may be limited by or before regulatory capital becomes binding. A deposit-funded card issuer may instead be constrained by stress losses or funding concentration. This is why an aggregate industry capital estimate is a starting point for analysis, not a credit-growth forecast.
Implementation priorities and watch points
Recommended priorities are a versioned current-versus-proposed engine, an exposure-data gap assessment, reconciliation to regulatory reports and stress testing at both consolidated and subsidiary levels. Keep allowance assumptions, regulatory capital effects and economic expected loss distinct so the business case does not double count benefits.
Watch final rules, transition dates, changes to the standardized and trading calibrations, the G-SIB methodology and separate stress-test decisions. This article should be revised when authoritative final text changes the scenario. Evidence that would strengthen the credit-growth thesis includes bank-specific headroom under all binding constraints and an attractive risk-adjusted loan pipeline. Weak demand or worsening could absorb the benefit even if the eventual rule reduces capital requirements.
Sources
- Federal Reserve and banking agencies: March 19, 2026 capital proposalsOfficial releaseBack to text: ↑1↑2↑3
- Agencies: March 2026 capital proposal fact sheetOfficial release · PDFBack to text: ↑1↑2
- Federal Register: March 27, 2026 standardized-approach proposalOfficial sourceBack to text: ↑1↑2
- Federal Reserve staff memorandum: scope and estimated capital effectsOfficial source · PDFBack to text: ↑1↑2
- Vice Chair Bowman: supporting statement, March 19, 2026Official releaseBack to text: ↑
- Governor Barr: dissenting statement, March 19, 2026Official releaseBack to text: ↑
- Federal Register, Category I/II expanded capital approach proposal, 91 FR 14952; March 27, 2026; retail exposures and commitment conversion factorsOfficial sourceBack to text: ↑1↑2
- Federal Reserve Vice Chair Bowman, The Final Chapter on Modernizing Bank Regulatory Stress Testing; September 18, 2026; speech, not final ruleOfficial sourceBack to text: ↑
- Federal Reserve: September 30, 2026 final stress-testing rules and separate proposal announcementOfficial releaseBack to text: ↑1↑2↑3↑4
- Federal Register: final capital-plan and stress capital buffer rule, published October 2, 2026Official sourceBack to text: ↑1↑2
- Federal Register: proposed fee-income model revision, published October 2, 2026; comments due December 1Official sourceBack to text: ↑