Unsafe or Unsound Practices,
Matters Requiring Attention
Bank Supervision, Materiality, and the Threshold for Corrective Intervention
The rule redraws the boundary between supervisory knowledge and compulsory correction.
Examiners may still identify emerging weakness. What changes is the point at which that information acquires institutional force. Material financial harm and legal noncompliance become the principal conversion mechanisms that determine when supervisory knowledge can compel remediation.
The result is a system that can know a weakness exists before the supervisory architecture has decided that the weakness is consequential enough to require repair.
Final OCC/FDIC rule. This analysis examines supervisory architecture rather than post-effective-date outcomes. GAO separately identified a Congressional Review Act timing issue concerning the stated November 2 effective date; that chronology is discussed below without treating it as a determination that the rule is legally ineffective.
Four supervisory categories carry different institutional force
When does information acquire the power to compel change?
The rule does not primarily change what supervisors are allowed to see. It changes when what they see acquires institutional force.
Weak signal → observation, violation, MRA, or unsafe practice
An examiner can identify an emerging weakness before the weakness satisfies the conditions for compulsory remediation. Classification therefore becomes the hinge between knowledge and obligation.
Voluntary response → tracked remediation → enforcement
Different classifications activate different institutional powers: management discretion, correction of a violation, board-level remediation and validation, or formal enforcement.
For prudential concerns, the MRA standard asks whether a practice contrary to prudent operation could reasonably be expected, under current or reasonably foreseeable conditions, to materially harm the bank's financial condition or create material risk to the Deposit Insurance Fund. An actual banking or banking-related legal violation can also independently support an MRA.
The agencies therefore preserve broad informational feedback while narrowing compulsory corrective feedback. A weakness may remain visible to supervisors, influence supervisory judgment, and recur across examinations without automatically producing board presentation, an action plan, mandatory correction, or examiner tracking.
Supervisory concern and compulsory remediation become less tightly coupled
Deficient practice could itself carry more corrective force
Prior OCC guidance treated deficient practices involving governance, internal control, risk management, or substantive legal compliance as matters that could require board-level attention and tracked correction.
Prudential weakness must cross a materiality threshold
Policies, process, documentation, control, and governance weaknesses remain observable, but prudential weaknesses need the specified relationship to material financial harm before they may become an MRA.
Constraint and relocation happen together
Examiners must use objective facts and sound reasoning and explain the basis for material findings. Yet prudent operation, foreseeability, materiality, and tailoring remain contextual and institution-specific.
Visibility is no longer equivalent to compulsory standing
The key structural transition is no longer simply whether the examiner sees a problem. It is whether that problem can cross the rule's financial or legal threshold and activate mandatory correction.
What shape does the new supervisory architecture create?
The scorecard measures structural capacity, not whether the rule is normatively desirable or whether it will prevent or contribute to future bank failures. The central pattern is the distance between a highly capable sensing system and a more conditional corrective system.
Eight functions
Select a function to see what its placement is anchored to in the final rule.
The defining feature is a sensing-to-obligation gap: informational feedback remains strong while compulsory correction becomes more conditional.
Informational feedback and corrective feedback are no longer the same thing
An examiner can identify and communicate a weakness without activating a mandatory action plan, board presentation, tracking, or validation. The ability to sense remains broader than the ability to compel correction.
Accountability becomes categorical
Once a finding crosses the MRA threshold, the remediation architecture becomes serious: responsible parties, commitments, deadlines, examiner follow-up, validation, and closure. Below that line, management receives substantially more discretion.
Power is constrained and relocated
The rule narrows what examiners may compel while preserving consequential interpretive authority over prudent operation, foreseeability, materiality, and tailoring. Bank agency expands most clearly below the compulsory threshold.
Care remains structurally peripheral
The rule's core protected objects are bank financial condition and the Deposit Insurance Fund. Consumer or depositor harm enters directly when law makes it cognizable or when it becomes relevant to safety and soundness.
The strongest mechanisms
Material financial harm converts knowledge into obligation
What can the system's unit of account not see?
Governance, process, documentation, and control weaknesses can remain visible while failing to acquire compulsory status unless they satisfy the material-financial-risk standard or an applicable legal violation.
Different classifications activate different powers
What kind of entry confers standing for compulsory intervention?
An observation, other violation, MRA, and unsafe practice are different institutional rooms. A concern can enter the supervisory record without entering the corrective system.
The system can record a weakness without maintaining a repair loop
What remains documented after the institution is told?
Supervisory observations can preserve institutional knowledge while deliberately omitting mandatory action plans and tracking. Observational memory and remedial memory become separable.
The actor changes how materiality is measured
Do identical weaknesses carry the same supervisory weight everywhere?
Materiality and remediation expectations vary with size, capital, complexity, activities, and risk. The agencies also excluded institution-affiliated parties after recognizing that bank-level materiality could distort how individual misconduct is measured.
Not strongly established: The Cliff and The Trap
Both are worth monitoring, but the more defensible finding in the final rule is the gap between sensing and compulsory correction rather than a discrete threshold cliff or compounding trap.
Local procedural integrity can coexist with relational failure
The relevant relation is not simply supervisor to bank. It is emerging weakness → examiner perception → classification → supervisory communication → management or board reception → corrective obligation → follow-up → validation → escalation.
Every node can comply with its own rule. An examiner can correctly see a weakness. The observation can be properly documented. Management can be legally free to decline the suggested enhancement. The examiner can properly decline to track it. The weakness can later become materially consequential. No single node necessarily violated its own procedure, yet the relation as a whole can fail to convert knowledge into timely correction.
The new rule places more responsibility on the transition between observation and MRA. That transition becomes the point at which a weak signal either acquires enough relational force to change institutional behavior or remains informational.
Can the system preserve an effective relation between sensing and correction when it deliberately loosens the obligation connecting them below the materiality threshold?
Why escalation cannot be treated as an abstract problem
Supervisors saw weaknesses before failure
The Federal Reserve's 2023 review concluded that supervisors did not fully appreciate SVB's vulnerabilities and, after identifying problems, did not ensure correction quickly enough. SVB was Federal Reserve-supervised, so it is not a direct test of this OCC/FDIC rule. It demonstrates why the sensing-to-escalation relation matters.
GAO identified delayed escalation across both failures
GAO found regulators had identified concerns at SVB and FDIC-supervised Signature Bank before failure but did not escalate actions in time. Its later work recommended clearer procedures for timely escalation and highlighted the absence of earlier noncapital triggers.
The agencies' efficiency case is coherent
The rule is designed to reduce low-value supervisory micromanagement, concentrate examiner and board attention on material threats, and lower the compliance burden associated with MRAs. The agencies expect fewer MRAs and argue that fewer, more consequential findings can improve supervisory focus.
The MRA threshold also remains anticipatory. “Could reasonably be expected” under current or reasonably foreseeable conditions is intentionally less demanding than the “likely” harm standard governing an unsafe-practice finding.
The empirical question is therefore not whether fewer MRAs occur. The agencies expect that. The meaningful question is which findings disappear from MRA status, and what happens to them afterward.
The decisive evidence is which findings disappear — and what happens next
MRA composition
Track the categories of concerns that move out of MRA status: governance, internal control, cyber, liquidity, interest-rate, operational, compliance, or other weaknesses.
Observation → MRA migration
Measure how often observations recur, intensify, or later cross the materiality threshold, and how much time passes before compulsory remediation begins.
Board visibility
Examine whether weaknesses kept below MRA status still reach boards through other channels and how board attention changes when presentation is no longer a supervisory requirement.
Ratings and unremediated observations
Test how supervisory ratings incorporate weaknesses known to examiners but not formally tracked to correction.
Remediation speed
Compare whether fewer, more material MRAs close faster and more effectively than under the prior framework.
Failure / loss linkage
Identify whether later material losses are preceded by documented observations that remained below the compulsory threshold.
GAO identified a Congressional Review Act timing issue
The rule was published September 1, 2026. GAO reports that the Senate received it September 4 and notes that the stated November 2 effective date is less than 60 days after Senate receipt, even though the CRA generally delays a major rule for 60 days from the later of publication or congressional receipt.
GAO does not resolve the legal consequence of that chronology. This analysis therefore treats the timing issue as an unresolved procedural question rather than a determination that the rule is invalid or legally ineffective.
What would show whether the architecture is working?
How much early-warning value survives when supervisors can identify weakness but cannot require or track remediation below the MRA threshold?
What evidence should show that an observation has become reasonably foreseeable material harm before an examiner may convert it into compulsory correction?
How consistently will examiners apply institution-specific materiality when the same weakness can carry different supervisory weight across banks?
Will dynamic tailoring reduce supervisory noise, or make escalation too dependent on local judgment to create reliable early intervention?
If a weakness remains an observation until harm becomes material, who bears the cost accumulated during the interval?
The rule makes coercive supervision more answerable while loosening one early corrective coupling
The 2026 OCC/FDIC rule creates a more explicit threshold architecture for bank supervision. Unsafe-practice findings, MRAs, other violations, and supervisory observations now carry more clearly differentiated consequences. That can make coercive supervision more explainable and contestable for regulated institutions while focusing mandatory remediation on legally cognizable or materially financial risks.
The same architecture creates a larger space in which supervisors can know about a weakness without possessing a duty — or, in the case of an observation, the authority — to require board attention, an action plan, tracked remediation, or closure. The rule therefore does not remove sensing. It changes the institutional conversion from sensing to obligation.
The success of the design depends on whether supervisors can distinguish low-value process criticism from emerging material risk early enough for the anticipatory MRA threshold to matter.
Download
The designed PDF contains the complete Applied Research analysis and preserves the Answerability Index in the report's print layout.
Download the full analysis (PDF) →
How to cite
Methodology
This analysis uses the Answerability Index, Taxonomy of Structural Harm, and DFTR relational analysis to examine how supervisory knowledge moves through classification toward corrective obligation.
The Index scores are structural coordinates supplied for this analysis. They are not empirical performance measures or an overall grade. Historical bank failures are used as counterevidence about escalation risk, not as proof of how this rule will perform.
Primary sources & source notes
OCC & FDIC — Final rule, 91 FR 56004 →
OCC Bulletin 2026-40 — Final rule summary →
OCC Bulletin 2026-41 — MRA implementation procedures →
GAO B-338727 — CRA major-rule review →
GAO-24-106974 — timely escalation of supervisory action →
Federal Reserve — Review of supervision and regulation of Silicon Valley Bank →
Analysis by Melissa Cosgrove for Toward Reconstitution. The Answerability Index™ and the Dynamic Fractal Theorem of Relation © 2026 Melissa Cosgrove. This is structural and interpretive analysis of a federal supervisory rule, not legal or financial advice and not a claim about post-effective-date outcomes that have not yet been observed.