After the 2023 Basel III endgame proposal drew intense criticism for raising capital requirements well beyond what many in the industry considered justified, US regulators went back to the drawing board. On March 19, 2026, the Federal Reserve, the OCC and the FDIC released a re-proposal that rescinds the 2023 version and replaces it with a recalibrated package. The comment period closed on June 18, 2026.
Headlines focused on capital: according to the agencies’ estimates as summarized by law firms including Debevoise & Plimpton and Hogan Lovells, the combined effect of the re-proposal, a revised GSIB surcharge proposal and stress capital buffer changes would reduce common equity tier 1 requirements for the largest firms by roughly 5%. But for credit risk, finance and technology teams, the more important story is operational. Moving from internal models to a more granular standardized framework changes what data must be captured, how exposures must be classified and how calculation engines must work. This article explains the main credit risk changes and sets out a practical readiness playbook. It reflects the proposal text as published; banks should check the latest status of rulemaking before finalizing plans.
What the Re-Proposal Changes, in Brief
The package has two main parts. First, it revises the generally applicable standardized approach that most US banks use today. Second, it introduces an expanded risk-based approach, or ERBA, for Category I and II institutions, generally the global systemically important banks and firms with $700 billion or more in assets. Smaller firms may opt in.
Under ERBA, the existing advanced approaches, which allow banks to use internal models to estimate credit risk parameters, would be replaced by a single revised standardized approach in which risk weights depend on supervisory factors rather than bank models. ERBA also includes a revised market risk framework, a new standardized approach for credit valuation adjustment risk, and a standardized operational risk charge that, compared with 2023, drops the internal loss multiplier.
Key credit risk elements highlighted in the law firm analyses include:
- Corporate exposures: under ERBA, investment-grade corporate exposures could receive a 65% risk weight instead of 100%. The 2023 restriction to companies with publicly traded securities has been dropped, and banks would need to identify investment-grade obligors through qualifying internal rating systems and define which grades count. Under the revised standardized approach, corporate exposures would move from 100% to 95%.
- Residential real estate: risk weights would depend on loan-to-value ratios and on whether repayment depends on cash flows generated by the property. Modified loans would not qualify for the preferential treatment, and junior liens would receive a 100% risk weight.
- Commercial real estate: lower-LTV exposures to strong obligors could fall well below 100%, while non-qualifying and higher-risk structures could rise to as much as 150%.
- Retail: ERBA retail risk weights would be conformed to the Basel framework, and the 2023 granularity limit would be removed.
- Securitization: the risk-weight floor for senior securitization exposures could fall to 15% from 20%, while resecuritizations and non-performing loan securitizations would face 100% floors.
- Other changes: a broader definition of “commitment” to capture more arrangements that could create future credit exposure, a flat 250% risk weight for mortgage servicing assets in place of the current deduction approach, and a 90% risk weight for “other assets.”
For Category III and IV firms, the standardized proposal would require recognition of most elements of accumulated other comprehensive income in CET1, with a five-year transition, which will make capital ratios more sensitive to interest rate movements for banks with large available-for-sale portfolios. No effective date has been proposed; the agencies requested comment on timing and transition.
Why This Is a Data Problem Before It Is a Capital Problem
It is tempting to treat lower aggregate capital as a reason to relax. That would be a mistake. Granular standardized approaches are, in many ways, more data-hungry than the models they replace, because every risk weight depends on specific attributes of the exposure, the obligor and the collateral, and those attributes must be complete, accurate and auditable.
Consider what the proposed treatment implies in practice:
- Real estate risk weights require current, defensible loan-to-value data, a reliable flag for whether repayment depends on property cash flows, lien position and modification history for every relevant loan.
- The 65% corporate risk weight depends on internal rating systems that qualify under the rule and a documented mapping of grades to investment-grade status. Weaknesses in rating governance will translate directly into higher capital.
- A broader commitment definition means more off-balance-sheet arrangements must be identified, classified and fed into calculations, including ones that may sit outside traditional lending systems.
- AOCI recognition for Category III and IV firms links capital ratios to securities valuations, requiring closer integration between treasury, finance and regulatory reporting data.
Every attribute that is missing or unreliable typically forces a conservative fallback, which increases risk-weighted assets. In other words, data quality now has a direct and quantifiable capital cost.
Under a granular standardized framework, data quality is capital. Every missing loan-to-value, unclear lien position or unmapped rating grade pushes exposures into higher risk weights.
A Readiness Playbook for Credit Risk Teams
Even without a final effective date, the direction of travel is clear enough to act on. The following eight-step playbook is designed to deliver value whatever the final calibration turns out to be.
- Run a proposal-based impact analysis. Recalculate risk-weighted assets under the re-proposal using current portfolio data, by business line and product. Identify where capital rises, where it falls and which assumptions drive the result.
- Build an attribute inventory. List every data attribute the new calculations require, such as LTV, property cash-flow dependency, lien position, modification status, obligor grade, commitment type and collateral eligibility. Map each to its system of record.
- Measure data completeness and quality. For each attribute, measure coverage and accuracy across the portfolio. Quantify the capital impact of gaps by comparing results with and without conservative fallbacks.
- Strengthen rating system governance. If you expect to use the investment-grade corporate treatment, review rating methodologies, grade definitions, override processes and validation evidence so that they will withstand supervisory scrutiny.
- Map off-balance-sheet exposures. Identify arrangements that could fall under the broader commitment definition, including those in trade finance, treasury and specialized lending systems.
- Assess calculation engines. Determine whether current regulatory calculators can support the new exposure classes, parallel runs under old and new rules, and transparent drill-down from capital numbers to individual exposures.
- Plan parallel runs. Design a period of parallel calculation and reconciliation well before any compliance date, with clear ownership for breaks between risk, finance and reporting.
- Integrate with capital planning. Feed revised risk weights into pricing, limit setting, portfolio strategy and stress testing so that business decisions reflect the new capital economics.
Architecture Choices: Strategic Fix or Tactical Patch
Banks face a familiar choice: build tactical workarounds that produce compliant numbers on time, or use the regulatory change to modernize the risk data architecture. In our experience, the best answer is usually a staged hybrid.
Characteristics of a sustainable target state
- A single, governed source of exposure, obligor and collateral data shared by risk, finance and regulatory reporting, rather than separate extracts reconciled manually
- Attribute-level data lineage from source systems to capital figures, so that supervisors and auditors can trace any number back to its origin
- Data quality rules embedded in pipelines, with thresholds, ownership and remediation workflows rather than one-off cleansing exercises
- Rule-based calculation engines that are configurable, versioned and able to run multiple regimes in parallel
- Reporting layers that support both regulatory submissions and internal analytics, including scenario analysis for capital strategy
Modern lakehouse platforms and open table formats make it more practical to build this kind of shared foundation without replacing every source system. The key is to treat regulatory data as a product with defined owners, quality standards and consumers, rather than as a by-product of reporting.
Alignment with international regimes
Banks with operations outside the United States must also reconcile US rules with other Basel implementations, such as the EU’s CRR3, which began applying in January 2025, and the UK’s Basel 3.1 framework. Differences in exposure classes, real estate treatment and output floor mechanics mean that a configurable, multi-regime calculation layer is far more sustainable than separate jurisdiction-specific builds.
Where Machine Learning and Automation Fit
The re-proposal reduces reliance on internal models for capital calculation, but it does not reduce the value of analytics in credit risk management. Several applications are particularly relevant to readiness work:
- Data quality and anomaly detection: statistical and machine learning techniques can flag implausible LTVs, inconsistent lien data or suspicious rating changes far faster than manual sampling.
- Document extraction: AI-assisted extraction from loan agreements, appraisals and facility documents can fill attribute gaps for legacy portfolios where data was never captured in structured form.
- Classification support: models can help classify complex exposures, such as identifying whether repayment depends on property cash flows, with human review for borderline cases.
- Capital optimization analytics: scenario tools can show how portfolio mix, collateral and structuring choices affect capital under the new rules.
Any such tools should be governed proportionately. The revised interagency model risk management guidance issued in April 2026 (SR 26-2) takes a risk-based approach, so a data quality model used to prioritize remediation may need lighter oversight than one that directly determines a regulatory classification.
Turning New Capital Economics Into Business Decisions
Capital rules change the relative profitability of products, clients and structures. When risk weights move, so does the return on equity of a mortgage segment, a corporate lending relationship or a securitization position. Banks that wait for a final rule before updating pricing and portfolio models risk making a year or more of decisions on outdated economics.
Credit risk and finance teams should work together on a few practical steps:
- Update risk-adjusted pricing tools to show returns under both current rules and the re-proposal, so relationship managers can see which deals improve or deteriorate
- Review concentration and limit frameworks, particularly for commercial real estate and investment-grade corporates, where the proposed treatment changes most
- Reassess collateral and structuring strategies, since LTV bands and lien position now carry more direct capital consequences
- Evaluate securitization and risk transfer programs in light of the proposed changes to senior securitization floors
- Brief business leaders early so that strategic plans and budget cycles reflect the likely direction of capital requirements
The objective is not to optimize aggressively against a proposal that may change, but to understand sensitivities and avoid being surprised. A bank that knows which portfolios are most affected can move quickly once the final calibration is known.
A Worked Example: A Commercial Real Estate Portfolio
Consider a generic large regional bank that may opt into or be affected by the new framework, with a sizable commercial real estate book. A first impact analysis shows a mixed picture: some low-LTV loans to strong sponsors could attract lower risk weights, but a significant share of the portfolio defaults to higher weights because loan-to-value data is outdated and the flag for property cash-flow dependency is missing for many older loans.
The bank launches a targeted remediation program. Appraisal data is refreshed for the largest exposures, and AI-assisted extraction pulls key terms from loan files to populate missing attributes, with credit officers reviewing every extracted value above a materiality threshold. Data quality rules are embedded in the loan origination system so that new loans cannot be booked without the required fields.
The bank then re-runs the impact analysis. Capital requirements for the portfolio fall compared with the first run, not because risk changed but because the data finally reflects it. Equally valuable, the bank now has lineage from each capital figure to the underlying loan attributes, which strengthens its position in supervisory reviews and improves pricing decisions for new originations.
Pitfalls and Outlook
Common pitfalls
- Treating the re-proposal as final and hard-coding parameters that may change in the final rule
- Running impact analyses on idealized data rather than on actual portfolio data with its gaps
- Leaving off-balance-sheet and non-lending systems out of scope until late in the program
- Building a regulatory calculator that cannot explain its results at the exposure level
- Separating capital readiness from business strategy, so that pricing and portfolio decisions lag the new economics
What to watch
The industry will be watching for the final rule, its effective date and transition arrangements, and any changes to the calibrations described above. Banks should also expect continued attention to stress testing, the GSIB surcharge and liquidity rules, all of which interact with risk-based capital. Whatever the final details, the shift toward granular, attribute-driven capital calculation is unlikely to reverse.
Conclusion
The Basel III endgame re-proposal is good news for the capital position of most large US banks, but it is not a reason to stand still. It moves the burden of proof from models to data: banks will need complete, accurate and traceable exposure information to earn the risk weights the rules allow. Credit risk teams that use the remaining time to inventory attributes, fix data quality at source, modernize calculation engines and connect capital to business decisions will not only be ready for the final rule. They will also have a stronger credit risk foundation for everything that follows.
How Hanoi Technology Can Help
Hanoi Technology helps banks assess capital rule impacts, remediate credit risk data and build traceable regulatory calculation and reporting pipelines.