A rating compresses a repayment story
A commercial loan file can contain financial statements, collateral records, calculations, projections and management explanations. A risk rating compresses that evidence into a category that other people can use. Its value lies in comparability across time and exposures, not in the label by itself. A number that looks precise can still conceal inconsistent judgment.
The OCC’s rating framework distinguishes regulatory problem-credit categories from the internal grades banks use to differentiate stronger loans. Its common categories include special mention, substandard, doubtful and loss; pass encompasses exposures outside those categories. [1] An internal grade of four or six therefore has no universal meaning across institutions. Even the direction of an internal scale can vary.
A rating is also different from a consumer credit score or a public bond rating. It concerns a particular institution’s credit-risk assessment and may reflect information unavailable outside the lending relationship. This article explains the framework, without inferring confidential grades for named borrowers or treating public financial ratios as a substitute for a complete credit review.
Borrower strength and facility protection are different dimensions
A business may owe several obligations with different collateral and priorities. Its operating deterioration affects the ability to pay all of them, but recovery after failure may differ. A first-priority claim on readily saleable assets can have a different loss profile from an unsecured claim, even when both depend on the same enterprise for ordinary repayment.
A hypothetical company with $1 million of annual cash available for debt service and $800,000 of required payments has a simplified coverage ratio of 1.25 times. If available cash falls to $650,000 while payments remain unchanged, that ratio falls to about 0.81. The deterioration concerns repayment capacity. A pledge of additional collateral may reduce a lender’s eventual shortfall, but does not itself restore the missing operating cash.
Conversely, reliable cash flow can coexist with weak documentation or uncertain lien priority. The point is not that collateral is irrelevant; it answers a different part of the credit question. A useful analysis distinguishes the likelihood of repayment difficulty from the severity of loss if that difficulty occurs, rather than forcing both into a single undifferentiated narrative.
Weakness is not synonymous with a realized loss
Special mention identifies potential weaknesses deserving attention but is not an adverse classification. Substandard concerns well-defined weakness that jeopardizes repayment. Doubtful adds that full collection is highly questionable on existing facts. Loss concerns amounts whose continued treatment as bankable assets is unwarranted, even if some future recovery remains possible. Substandard, doubtful and loss are classified categories. [1]
These categories describe different states of uncertainty. A warning about emerging weakness need not imply imminent default. An identified repayment problem need not mean the entire balance will disappear. A write-off does not necessarily mean that collection activity has ended or that every possible recovery is zero.
The terminology can be especially confusing in public discussion because criticized and classified are sometimes used loosely. The exact definition attached to a reported figure matters. A total including special mention is broader than one restricted to classified exposures. Comparing the two as if they measured the same severity would exaggerate or understate a change without any loan’s condition actually changing.
Cash collection, delinquency and accounting have separate clocks
Payment status records whether scheduled amounts have arrived. A risk assessment asks whether the repayment story remains credible. Accounting recognition determines what income and asset amounts are reported under the applicable framework. These clocks can move together, but there is no logical requirement that they change on the same day.
Consider a hypothetical borrower that remains current by drawing down a cash reserve while its operating losses deepen. Timely payments are real, yet they do not demonstrate that the reserve can support debt service indefinitely. Alternatively, a short administrative payment delay at a financially strong borrower does not necessarily mean the business’s long-term repayment capacity has collapsed.
A rating category consequently cannot be mechanically equated with a fixed provision percentage, a particular number of days past due or a amount. Those measurements answer different questions and rely on their own facts and requirements. Combining them can be informative; substituting one for another removes information. The examples here illustrate the distinction and do not prescribe an accounting treatment for a specific loan.
A recovery example shows why balances need context
Suppose a hypothetical troubled $10 million exposure has collateral expected to produce $7 million net of selling costs under one supportable scenario. Another $1 million might be recovered if a pending claim succeeds, and $2 million appears unsupported. This is a scenario analysis, not a supervisory classification assignment. The timing, legal priority, valuation reliability and other repayment sources still matter.
The exercise shows why calling the full $10 million a loss can be wrong even when the credit is seriously impaired. It also shows why calling the loan fully protected because gross collateral once appraised at $11 million can be equally misleading. Gross value, net recovery, availability of proceeds and the time required to receive them are separate quantities.
If the asset takes two years to sell, the cost of waiting can further alter the economics. Maintenance, taxes, legal expenses and deterioration might reduce proceeds, while interest-rate changes affect the value of delayed cash. A collateral estimate is therefore a set of assumptions about conversion to cash, not merely a price attached to an object.
Migration explains more than a single snapshot
Imagine a hypothetical $500 million portfolio with $25 million in weaker categories at the beginning and end of a year. The unchanged 5% share might suggest stability. But suppose $20 million of the original weak loans were repaid or charged down and $20 million of previously stronger loans deteriorated. The ending stock conceals substantial movement.
The reverse pattern is possible too. A weaker-loan percentage can fall because strong new originations expand the denominator, even if no troubled exposure improves. Amounts, proportions, upgrades, downgrades, repayments and exits therefore describe different features of portfolio change. A single ratio cannot communicate them all.
Migration analysis also depends on consistent standards. A new rating policy could move loans between categories without a corresponding change in borrower economics. That does not make the revised policy wrong; it creates a comparability issue. Distinguishing changed facts from changed classification practice is essential to interpreting a trend accurately.
Independent review tests whether labels remain credible
The 2020 interagency guidance describes independent, ongoing credit risk review and communication to management and boards. It addresses review scope, qualifications, findings and follow-through, while allowing arrangements suited to an institution’s size and complexity. [2][3]
The economic rationale is an incentive problem. The person closest to a borrower may have the richest information and also the strongest attachment to the original lending decision. A separate review can challenge assumptions without discarding relationship knowledge. Independence is useful when it produces an informed second assessment, rather than a ritual approval based on the first assessment’s conclusions.
Technology can flag stale statements, missed or unusual financial changes, but a flag and a rating are not identical. An unexplained variance could represent deterioration, an accounting change or a data mapping error. Automated extraction can accelerate attention; judgment still connects the signal with the borrower’s actual repayment sources and the facility’s legal and economic protection.
The information value depends on what changed and why
A rating system creates a common language for uncertainty. Its strongest use is explaining how a loan’s repayment prospects have changed and which evidence supports that conclusion. Its weakest use is treating a category as a deterministic forecast of loss or as an excuse to stop examining the underlying business.
The framework also has limits outside the institution. Public disclosures may aggregate grades or omit the information required to reproduce them. A reader can analyze the disclosed trend and definitions without claiming access to the underlying confidential assessment. More forceful conclusions require more evidence, not merely a more elaborate model.
The OCC handbook and interagency guidance provide the supervisory foundation cited here. The worked numbers are invented analytical examples, not published portfolio statistics or official grade assignments. They demonstrate why repayment strength, loss protection and the timing of recognition remain related but distinct questions.
Sources
- OCC, Rating Credit Risk handbook, regulatory definitions and rating architecture, especially printed pages 15–19Official source · PDFBack to text: ↑1↑2
- OCC Bulletin 2020-50, Interagency Guidance on Credit Risk Review SystemsOfficial sourceBack to text: ↑
- Interagency Guidance on Credit Risk Review Systems, 85 FR 33278, June 1, 2020Official source · PDFBack to text: ↑