Reading a credit rating: scales, default and the issuer-pays conflict

A company can have one credit rating while different debts it issues carry different ratings. A secured lender and a junior bondholder may face the same business downturn but have different claims on the assets if the company fails. The letters make sense only once the reader knows exactly what has been rated.

A rating is an opinion about credit risk. It does not establish the price at which a bond is attractive or guarantee repayment. The SEC’s investor bulletin on credit ratings recommends using ratings alongside other information, including the issuer’s disclosures and the terms of the debt.

Begin with the entity or obligation

An issuer rating assesses an entity’s general creditworthiness. An issue rating addresses a particular financial obligation, taking account of features such as seniority, collateral and repayment terms. S&P explains the distinction on its credit-ratings page.

Check the legal borrower, guarantor and security package. A rating attached to a parent company should not automatically be assigned to a subsidiary’s debt. A rating on one secured instrument should not be extended to every creditor. When a report refers to “the company rating,” find the actual rating action and its scope before using it.

The same care applies across agencies. Moody’s description of its long-term ratings includes both the likelihood of default or impairment and expected financial loss. A symbol is inseparable from the definition behind it.

Locate the investment-grade boundary

On the main long-term scales, BBB- at S&P and Fitch and Baa3 at Moody’s are the lowest investment-grade ratings. The next steps down are BB+ and Ba1, respectively. The correspondence is shown in the UK government’s reference table; the agencies’ definitions remain the authority for the meaning of an individual rating.

A borrower downgraded across this boundary is commonly called a fallen angel. The downgrade can affect a portfolio mandate, index eligibility or contractual condition. Its consequences depend on the actual rule and any transition period. A rating action does not create a universal obligation for every holder to sell.

This distinction matters in reporting: specify the affected mandate or contract before attributing a price movement to forced selling. Our credit-spread guide explains the separate market measure.

Default studies need a population and a horizon

S&P’s Guide to Credit Rating Essentials explicitly distinguishes ratings from exact default probabilities. The symbol ranks creditworthiness within an agency’s framework; it is not a universal numerical forecast that can be applied to any borrower.

Historical default studies answer a different question: what happened to a defined population of rated borrowers over a specified period? Before quoting a percentage, record the agency, publication vintage, sample, rating date, horizon and treatment of withdrawn ratings. Check whether distressed exchanges count as default and whether the study covers companies, sovereigns or structured-finance instruments.

A one-year rate cannot be multiplied mechanically to obtain a five-year cumulative rate. A historical average also does not establish the probability that a named borrower will default next year. For the same reason, this guide does not attach timeless default percentages to the rating ladder.

A transition table follows movements between rating categories, including default. It is useful for examining migration, but its interpretation still depends on that study’s population and definitions. The agency’s guide to default and rating-change studies provides the starting point.

Read the rationale, outlook and review status

The rationale tells you which assumptions support the opinion and what could cause a change. The outlook or review status adds information about possible direction; it does not replace the rating or make the next action certain. Check the date of the rating action and whether a more recent action supersedes it, as the SEC’s investor bulletin advises.

For a reader investigating a financing arrangement, the useful question is often contractual: which cash flow supports repayment, which entity owes the debt, and what happens if the central assumption fails? A letter alone cannot answer those questions. The private credit atlas separates managers, funds, insurers, banking facilities and borrowers so those links can be examined individually.

Who pays for the opinion?

Under the issuer-pays model, the entity seeking a rating pays the agency. That creates a potential conflict: the rated entity is also a customer. The US Government Accountability Office examined this and alternative compensation arrangements in its report on rating-agency payment models.

The existence of that incentive does not establish that a specific opinion is wrong. Nor does switching to investor payment remove every possible conflict. Ask what controls apply, whether the assumptions and methodology can be inspected, and whether another evidence source challenges the analysis.

A market spread can provide a different perspective, but it also reflects liquidity, risk appetite and instrument features. It should be analysed alongside the rating, rather than treated as an infallible verdict on the agency’s honesty.

Public and private ratings require different access checks

A public rating is available to a wider audience. Private or confidential ratings have restricted distribution; S&P describes both forms, and Moody’s explains its publication process.

Restricted access limits what an outside reader can verify. It does not, by itself, establish that a private rating is more generous or that the debt is safer. If the rating letter, surveillance terms or underlying assumptions are unavailable, say so. Avoid filling the gap with a public rating on a different entity or instrument.

For an application of those questions, read our analysis of residual-value guarantees in AI infrastructure lending and the guide to assessing a life insurer. The task is to identify the specific promise, its provider and its limitations.

Sources and revision

Sources above include the SEC investor bulletin, GAO’s compensation-model review, agency definitions and the government’s scale comparison. Agency publications are used for the agencies’ own definitions, not as independent validation of their accuracy. The October 8, 2026 revision corrects the description of a rating as a precise default probability, removes default-rate figures without a defined cohort, and qualifies claims about mandatory sales and private ratings.