Three Rating Events Look Alike. Only One Reaches the Contract Language.

A credit rating headline can describe three different events, and the word most readers carry away is the same in all three. One is a change in the agency's stated view of where a rating may go. One is a formal notice that the rating is under examination. One is a change in the rating itself. Only the third alters the symbol that contract clauses, index rules and program eligibility tests are written on.

The distinction is not a matter of taste. It is written into the Securities and Exchange Commission's rules for rating agencies, into the construction rules of the bond indices that force buying and selling, and into the federal capital and liquidity regulations that describe what contractual clauses are allowed to do. A trader who reads a rating headline without knowing which of the three occurred is reading a sentence with the verb removed.

The sections below work outward from the rulebook: what the SEC counts as a rating action, what a written threshold looks like, why index deletion is a calendar event rather than a news event, and what the regulators say about clauses that read a rating directly.

The Regulator's List of Rating Actions Is Shorter Than the News Cycle's

Rating agencies registered with the SEC are known as nationally recognized statistical rating organizations, or NRSROs. The SEC's Office of Credit Ratings reported on ten of them in its staff report covering calendar year 2024, with outstanding ratings grouped into five categories: financial institutions, insurance companies, corporate issuers, asset-backed securities and government securities.

Rule 17g-7 under the Securities Exchange Act, codified at 17 CFR 240.17g-7(a), defines what a "rating action" is for the purposes of the disclosure obligations it imposes. The definition is a closed list:

• the publication of an expected or preliminary credit rating before the publication of an initial credit rating;
• an initial credit rating;
• an upgrade or downgrade of an existing credit rating, including a downgrade to, or assignment of, default;
• an affirmation or withdrawal of an existing credit rating, if the affirmation or withdrawal is the result of a review.

An outlook change does not appear on that list. Neither does the act of placing a rating under review or on a watch list. What appears is the resolution — an affirmation or a withdrawal that comes out of a review, or a move in the rating itself. The regulation treats the announcement that an agency is looking at something as a different category of event from the announcement of what it decided.

The same rule's paragraph (b) requires each NRSRO to publish rating histories in an interactive data file using XBRL format, updated no less frequently than monthly. The action types carried in that history are the initial rating, an upgrade, a downgrade, a withdrawal, and an entry recording that a rating was outstanding. The machine-readable record of an agency's work is a record of symbols and dates.

One caveat belongs here rather than buried later. Each agency publishes its own definitions document setting out how it uses the words outlook and watch, including the time horizon it attaches to each. Those documents were not retrievable for this piece, so no horizon is asserted anywhere in it. Anyone sizing a position around a stated review period should read that agency's own definitions.

What a Written Threshold Looks Like When a Federal Program Writes One

The clearest public examples of a rating threshold in a live document are the Federal Reserve's 2020 facility term sheets, published in full and unambiguous about what is being tested.

The Secondary Market Corporate Credit Facility term sheet dated 28 July 2020 required that the issuer "was rated at least BBB-/Baa3 as of March 22, 2020, by a major nationally recognized statistical rating organization," and that an issuer rated by multiple major NRSROs "must be rated at least BBB-/Baa3 by two or more NRSROs as of March 22, 2020." It then handled downgrades separately: an issuer that met that test on the cut-off date but was later downgraded "must be rated at least BB-/Ba3 as of the date on which the Facility makes a purchase."

The Municipal Liquidity Facility term sheet dated 11 August 2020 used the same architecture with different levels. States, cities and counties had to be rated at least BBB-/Baa3 as of 8 April 2020 by two or more major NRSROs, with a post-downgrade floor of BB-/Ba3 at the time of purchase. Multi-state entities faced a higher bar at A-/A3 as of the same date, with a floor of BBB-/Baa3 at purchase.

Three structural features recur in private documents as well as public ones.

• The test is on the symbol, on a date. Every clause names a rating level and a moment at which it is measured. None references an outlook, a watch listing, or an agency view about direction.
• Multiple agencies are combined by rule, not by judgment. Where more than one agency rates the issuer, the threshold must be met by two or more of them. A single agency's action does not resolve the test.
• Crossing down and staying eligible are separate questions. The fallen-angel provisions distinguish the level required to qualify from the level required to remain, and the two are not the same number.

The boundary itself is published by the agencies. S&P Global Ratings states that investment grade ratings "include those rated BBB- or higher" and that speculative grade ratings "range from BB+ down to D (default)." The Federal Reserve term sheets pair BBB- with Baa3 and BB- with Ba3, which is how the two scales line up in a federal document. Fitch's own scale document was not retrievable, so no Fitch symbol is asserted here beyond the fact, discussed below, that index rules count Fitch ratings alongside the other two.

The Index Rule Is a Calendar, Not a Headline

Forced selling by index-tracking portfolios is the mechanism most often invoked when a downgrade is discussed. The rules governing it are more specific than the story usually is.

The ICE BofA US Corporate Index, whose rules are set out in the series notes published by the Federal Reserve Bank of St. Louis on FRED, requires that securities "have an investment grade rating (based on an average of Moody's, S&P, and Fitch)," carry more than one year of remaining maturity, and have a minimum amount outstanding of $250 million. The corresponding high yield index requires a below investment grade rating on the same averaged basis, an investment grade rated country of risk, and a minimum amount outstanding of $100 million.

The timing rules matter more than the size rules for anyone trading around a rating event. The index "is rebalanced on the last calendar day of the month, based on information available up to and including the third business day before the last business day of the month." And explicitly: "Issues that no longer meet the criteria during the course of the month remain in the Index until the next month-end rebalancing at which point they are removed from the Index."

Two consequences follow. The date a downgrade is published and the date a bond leaves the index are different dates, and the gap can be most of a month. And a downgrade published after the information cutoff does not reach that month's rebalancing at all. The mechanical flow a rating event is supposed to cause is scheduled, and the schedule is public.

The averaging rule adds a third. Because eligibility runs on an average of three agencies' ratings, one agency moving a name a notch below the boundary does not necessarily move the composite across it. Whether it crosses depends on where the other two sit, which is knowable before any announcement.

Where the Price Step Sits on the Ladder

Option-adjusted spreads by rating band give a cross-sectional picture of what the market charged for each rung on one particular day. On 20 August 2026, the daily closes of the ICE BofA indices retrieved from FRED were as follows, in percent: AAA 0.42, AA 0.61, single-A 0.69, BBB 1.00, BB 1.63, single-B 2.94, and CCC and lower 10.35.

Expressed as steps between adjacent bands, in basis points: AAA to AA 19, AA to A 8, A to BBB 31, BBB to BB 63, BB to B 131, B to CCC and lower 741.

The step at the investment grade boundary was roughly twice the step from single-A to BBB on that day — 63 basis points against 31. That is consistent with the boundary carrying something the other rungs do not, which is what the eligibility rules above describe. It is also modest next to what happens further down: the single largest step on the ladder by a wide margin sits between single-B and CCC, where the question stops being index eligibility and starts being repayment.

Three limits on reading this chart are worth stating plainly. It is one day's snapshot, not a distribution. Each bar is a different population of issuers with different maturities and sector mixes, so the gap between two bars is not an estimate of what happens to a given bond when its rating moves. And option-adjusted spread already embeds whatever the market priced in advance — the one quantity a trader reacting to an announcement cannot observe separately.

Clauses That Read the Rating, in the Regulator's Own Words

Contract terms that respond to credit standing are not folklore. Federal capital regulation names them in order to exclude them.

Under the Federal Reserve's capital rule, an additional tier 1 instrument qualifies only if "the instrument does not have a credit-sensitive feature, such as a dividend rate that is reset periodically based in whole or in part on the Board-regulated institution's credit quality" (12 CFR 217.20(c)(1)(ix)). The tier 2 criteria carry the same requirement in broader terms: the instrument "has no credit-sensitive feature, such as a dividend or interest rate that is reset periodically based in whole or in part on the Board-regulated institution's credit standing" (12 CFR 217.20(d)(1)(vii)).

The same rule addresses coupon step-ups from the redemption angle. An additional tier 1 instrument must have "no maturity date" and must not "contain a dividend step-up or any other term or feature that creates an incentive to redeem" (12 CFR 217.20(c)(1)(iv)), and a tier 2 instrument "must not have any terms or features that require, or create significant incentives for" early redemption (12 CFR 217.20(d)(1)(iv)).

A regulation that prohibits a feature is evidence the feature exists in the instruments it is written about. Coupons that reset on credit standing, and step-ups that create redemption incentives, are recognised mechanics — recognised well enough that supervisors barred them from capital instruments.

The liquidity rules describe the collateral side. The liquidity coverage ratio requires an outflow amount equal to "100 percent of all additional amounts of collateral the Board-regulated institution could be contractually required to pledge or to fund under the terms of any transaction as a result of a change in the Board-regulated institution's financial condition" (12 CFR 249.32(f)(1)). The regulation assumes, as a baseline, that contracts contain condition-triggered collateral obligations and that they are funded in full.

These citations establish the shape of the mechanism, not its presence in any particular security. No regulation reveals whether a specific bond carries a step-up or a specific agreement carries a collateral trigger. That sits in the offering document and the indenture, which for registered securities are filed publicly. Reading the instrument is the only way to know.

What Would Invalidate This

The frame in this piece treats the rating symbol as the thing contracts test, and outlooks and watch listings as information that changes nothing mechanically. Several conditions break that.

• A contract written on outlook or watch. Nothing prevents two parties from drafting a covenant that references a negative outlook or a watch listing rather than a rating level. The SEC rule cited above governs what agencies must disclose, not what private parties may write. If a specific document does this, the frame does not apply to that document.
• The composite does not move. Where index eligibility runs on an average of three agencies, one agency's downgrade may leave the composite on the same side of the boundary. In that case the mechanical consequence described here does not occur at all.
• The instrument is not index-eligible to begin with. A bond below the $250 million or $100 million size minimums, or inside one year to maturity, is outside the rules discussed here regardless of its rating.
• The repricing already happened. Spreads can move on the outlook change, the watch listing, or the news that preceded both. Where that has occurred, the rating change may pass with little visible effect, and this sequence explains the plumbing rather than the price.
• Program and index rules change. The Federal Reserve term sheets quoted here were written for 2020 facilities, and index rules are revised by their administrators. Both show how thresholds are drafted, not what is eligible today.
• The issuer is thinly rated. Where only one agency rates a name, the two-or-more-agency logic in the term sheets and the three-agency averaging in the index rules have nothing to operate on.

Concrete Framework

1. Classify the event before anything else. Determine whether the announcement is an outlook change, a placement on watch or review, or a change in the rating symbol. Only the third is a rating action in the sense of 17 CFR 240.17g-7(a).
2. Write down the current symbol from each agency that rates the name. The composite, not any single agency's view, is what index rules test. Record all of them, not just the one that moved.
3. Locate the nearest written boundary. Identify whether the name sits at BBB-, one notch above, or already below, and how many notches separate the current composite from the investment grade line.
4. Check the size and maturity screens. Confirm the issue clears the index minimums — $250 million for the investment grade index, $100 million for high yield — and has more than one year remaining. If it does not, index mechanics are irrelevant to it.
5. Put the calendar on paper. Note the announcement date, the third business day before the last business day of the month, and the last calendar day. Determine whether the event landed before or after the information cutoff.
6. Read the instrument, not the summary. If contractual consequences matter, open the offering document and search for interest rate adjustment provisions, collateral posting obligations and rating-linked put rights. Regulations establish that such clauses exist; only the document shows whether this one has them.
7. Separate what is priced from what is scheduled. Compare the spread the name already carries against the band levels for its current and prospective rating. Where it already trades near the lower band, the scheduled index event is the remaining variable, not the credit view.
8. Define the invalidation before entry. Decide in advance which observation would show the frame is wrong here — a composite that does not cross, a cutoff that is missed, a document with no rating-linked clause — and size accordingly.