The Basel III Scorecard: 200 Public Banks Ranked

On March 19, 2026, the Federal Reserve, FDIC, and OCC jointly released three notices of proposed rulemaking that together represent the most consequential overhaul of U.S. bank capital requirements since the original Basel III framework took effect in 2013. The three NPRs cover the revised standardized approach to credit risk (applicable to all banking organizations), the expanded risk-based approach (applicable to Category I and II firms), and revisions to the G-SIB surcharge methodology.

The proposals create two parallel frameworks. Category I firms (the eight U.S. G-SIBs) and Category II firms (non-GSIB banks with $700 billion or more in total assets, or $75 billion in cross-jurisdictional activity) would be subject to the expanded risk-based approach. This framework adds investment-grade corporate distinctions, retail exposure tiers (transactors vs. revolvers), and an operational risk charge. All other banking organizations—including Category III firms ($250–$700 billion in assets) and Category IV firms ($100–$250 billion)—would use the revised standardized approach, which introduces LTV-based residential mortgage risk weights, modestly recalibrates corporate and CRE exposures, and revises credit conversion factors for off-balance-sheet commitments.

We built a full-balance-sheet model scoring 200 publicly traded U.S. banks on these proposals. The model uses actual RC-R Part II regulatory data from Q4 2025 call reports as the denominator—real risk-weighted assets, not estimated—and HMDA loan-level origination data to construct bank-specific residential mortgage LTV distributions. It decomposes each bank’s estimated RWA change into five components: residential mortgages, other loans, other assets, off-balance-sheet commitments, and securitization exposures. The model calibrates within 0.3–0.6 percentage points of the agencies’ own aggregate estimates published in NPR Table V.5.

The Proposals: What Changes

The centerpiece of the revised standardized approach is residential mortgages. The current framework applies a flat 50% risk weight to all first-lien residential mortgage exposures regardless of borrower leverage. The proposed rule replaces this with six LTV-based buckets drawn from NPR Table III.1 (pp. 32–33):

Beyond mortgages, the proposals touch nearly every exposure class on a bank’s balance sheet:

For Category I and II firms (the eight GSIBs and any non-GSIB bank above $700 billion in assets), the expanded risk-based approach adds further distinctions: investment-grade corporate exposures at 65% risk weight, credit card transactors at 45%, revolvers at 75%, and an operational risk capital charge. These additional provisions drive significantly larger RWA reductions for the G-SIBs. Category III banks ($250–$700 billion), including U.S. Bank, Capital One, PNC, Truist, and TD Bank, do not receive these benefits—a distinction that materially affects their scores.

The commitment CCF change is directionally mixed. Long-term commitments (>1 year) see their CCF drop from 50% to 40%—a reduction. But short-term commitments (<1 year), currently at 20%, double to 40%. Banks with heavy short-term revolving facilities can actually see RWA increase from this component.

The Model

Our model decomposes each bank’s estimated change in risk-weighted assets into five components:

The denominator is actual risk-weighted assets from Schedule RC-R Part II—the same figure banks report to their regulators. This eliminates the estimation error inherent in constructing RWA from asset-side data. For the 67 banks with sufficient HMDA origination volume, we use a three-year blend (2022–2024) of loan-level combined LTV ratios to compute bank-specific residential mortgage risk weight changes. The remaining 130 banks receive the NPR average residential mortgage delta of −30%.

The model calibrates against the agencies’ own aggregate estimates. NPR Table V.5 reports the weighted-average RWA change across all affected institutions by exposure category. Our model’s aggregate results fall within 0.3–0.6 percentage points of those figures across all five components.

The Scorecard: 200 Banks

The chart below ranks all 200 banks by their estimated change in total risk-weighted assets under the proposals. More negative values indicate greater RWA reduction. Color coding reflects the magnitude: green for the largest beneficiaries (≤ −14%), blue for strong beneficiaries (−14% to −10%), orange for moderate beneficiaries (−10% to −6%), and red for banks with the most modest relief (above −6%).

What Drives the Score: Component Attribution

The stacked bar below decomposes the RWA change for the 15 largest banks by assets. Five colors correspond to the five model components. The “other loans” component—which captures C&I, CRE, consumer, and credit card exposures—dominates for the G-SIBs because the expanded risk-based approach grants investment-grade corporate treatment (65% RW) and tiered retail treatment (transactors 45%, revolvers 75%). The contrast with Category III banks is stark: Capital One’s “other loans” component flips to a large positive (+9.9pp) because its massive credit card book receives only a 90% risk weight under the standardized approach—versus the 45%/75% tiered treatment available to G-SIBs. Morgan Stanley Private Bank is the outlier on the other end: its score is driven almost entirely by residential mortgages, reflecting an ultra-low-LTV wealth management lending book.

Note that commitments can be positive. The Bank of New York Mellon (+1.1pp) and State Street (+1.4pp) see RWA increase from commitments because their off-balance-sheet books are dominated by short-term facilities whose CCF doubles from 20% to 40%. Morgan Stanley Private Bank (+2.6pp) shows the same effect. For lending-heavy banks like Wells Fargo (−3.1pp) and PNC (−3.2pp), the long-term commitment CCF reduction from 50% to 40% dominates.

The Winners

TrustCo Bank (−20.9%) is the largest beneficiary in the scorecard. TrustCo is a thrift: 86% of its risk-weighted assets sit in residential mortgages. Because the revised standardized approach applies the full LTV-based schedule to this concentrated resi book, TrustCo’s RWA falls by more than a fifth. Third Federal Savings and Loan (−16.1%) follows the same pattern—a pure-play residential mortgage lender.

The G-SIBs cluster between −14% and −16%. Citibank (−15.9%), Bank of America (−15.8%), Wells Fargo (−15.8%), and JPMorgan Chase (−14.5%) all benefit broadly from the expanded risk-based approach. The combination of investment-grade corporate treatment, tiered credit card risk weights, and residential mortgage relief produces double-digit RWA reductions across the board. U.S. Bank (−11.5%), as a Category III institution, receives the revised standardized approach rather than the expanded framework, resulting in a meaningfully smaller benefit than the G-SIBs despite its size.

Wealth management banks rank higher than their size would suggest. Charles Schwab Bank (−10.3%) and UBS Bank USA (−13.7%) hold large residential mortgage portfolios originated at very low LTVs—the typical borrower is a high-net-worth client pledging a home worth multiples of the loan balance. Morgan Stanley Private Bank (−14.0%) is the extreme case: residential mortgages account for −14.9pp of its total −14.0% score (offset slightly by a +2.6pp increase from short-term commitment CCFs).

The LTV Distribution: Why It Matters

The chart below shows the LTV distribution of residential mortgage originations for 15 selected banks, based on three years of HMDA data (2022–2024). The differences are stark. East West Bank originates 61% of its dollar volume below 60% LTV. USAA, serving active military and veterans via VA loan programs, originates 59% above 90% LTV. Under the proposed risk weight schedule, East West’s residential mortgage book would see a blended risk weight reduction far exceeding the NPR average, while USAA’s book would see a net increase in risk-weighted assets on the resi component.

The Other End: Credit Card Banks, Custody Banks, and CRE Lenders

American Express National Bank (+14.3%) and Capital One (+8.2%) are the two banks in our scorecard whose risk-weighted assets would actually increase under the proposals. Both are overwhelmingly credit card banks. Under the revised standardized approach, credit card exposures receive a flat 90% risk weight—only a 10% reduction from the current 100%. Meanwhile, the doubling of the short-term commitment CCF from 20% to 40% adds substantial RWA on their large unfunded card lines. The net effect is negative: the modest credit risk relief is more than offset by the commitment CCF increase. Only banks subject to the expanded risk-based approach (GSIBs) receive the far more favorable tiered treatment—transactors at 45%, revolvers at 75%—which would deliver meaningful relief on card portfolios. Capital One and American Express, as Category III institutions, do not qualify.

The Bank of New York Mellon (−2.4%) and State Street (−3.3%) also sit near the bottom. These are custody and asset-servicing banks with tiny lending books relative to their total risk-weighted assets. The proposals primarily affect credit risk on lending exposures; custody banks have minimal lending to benefit from. Their short-term commitment books actually increase RWA under the new 40% CCF, partially offsetting the modest relief they receive on other assets.

Bank OZK (−8.3%) is an instructive case from the lending side. Despite having a moderate score overall, OZK’s asset mix constrains its benefit: the bank is concentrated in CRE and construction lending, where risk weight reductions are minimal (CRE: 100% → 95%) or nonexistent (construction stays at 150%). Its residential mortgage book is small, limiting the LTV-based relief available to it.

Banks with heavy short-term commitment books also lag. The 40% CCF on short-term commitments—up from 20%—adds RWA for banks that serve as backup liquidity providers, maintain large unfunded revolving credit facilities, or hold significant short-term trade finance commitments. This partially offsets the gains they receive on lending exposures.

The HMDA Advantage

Call reports tell you that a bank holds $50 billion in residential mortgages. They do not tell you whether those mortgages were underwritten at 50% LTV or 95% LTV. Under the current flat-50% risk weight regime, the distinction is irrelevant for capital purposes. Under the proposed LTV-based schedule, it is the single most important variable for the residential mortgage component.

RC-R Part II gives us the bucket sizes—how much of a bank’s total RWA sits in residential mortgages, C&I, CRE, and so on. HMDA gives us the LTV distribution within the residential mortgage bucket. The combination produces a bank-specific residential mortgage RWA change that ranges from −36.7% to −11.6% across the 67 banks with sufficient HMDA coverage, versus a flat −30% assumption for banks without it.

Twenty-five banks shift by more than 0.5 percentage points in their total score due to HMDA data, and 12 shift by more than 1 percentage point. Banks with concentrations at the extremes of the LTV distribution—either very low (wealth management) or very high (VA/FHA lenders)—see the largest HMDA-driven score adjustments. For the 130 banks without sufficient HMDA coverage, we apply the NPR average residential mortgage delta of −30%, which is a conservative central estimate.

What Comes Next

The proposals carry a 90-day comment period, with a deadline of June 18, 2026. Industry feedback will be extensive; the American Bankers Association, the Bank Policy Institute, and every major financial trade group have already signaled detailed responses. Finalization is not expected before late 2026 or early 2027.

PMI treatment is a key open question. Whether private mortgage insurance can reduce the effective LTV for risk-weight purposes has material implications for banks originating above 80% LTV. The current proposal is ambiguous on this point, and comment letters will likely press for clarification. If PMI is recognized, the benefit to high-LTV lenders would be significant.

A note on scope: we excluded Sallie Mae (predominantly student lending, not captured by the proposal’s credit risk categories) and consolidated subsidiary charters to one entry per holding company where applicable.