The Sovereign Rating Calendar Is Public a Year Ahead. Downgrades Still Reprice Bonds.

A sovereign credit rating downgrade arrives packaged as news. The wire copy carries surprise, the letter grade moves one notch, and commentary treats the action as fresh information about a government's ability to pay. Almost none of that framing survives contact with how the process works.

In the European Union, the dates on which sovereign ratings may be published are fixed in advance and disclosed publicly. Under Regulation (EU) No 462/2013, a rating agency must publish, at the end of December, a calendar for the following twelve months setting a maximum of three dates on which it may issue sovereign ratings and related outlooks. Those dates must fall on a Friday. Publication must occur after the close of business of trading venues in the Union and at least one hour before they reopen. Deviation from the calendar is permitted only where the agency needs it to discharge specific obligations, and the deviation must come with a detailed explanation.

The calendar is not a secret. What follows is what remains once the surprise is stripped out.

The Surface Reading, and the Part of It That Survives

The standard interpretation says markets price fiscal deterioration continuously while agencies review periodically, so the downgrade is lagging confirmation of something already in the price. The French case in 2025 is an unusually clean test, because the review dates were published in advance and the market's own repricing is separately observable.

Through late August 2025, the 10-year French OAT traded at roughly 65 basis points over the German Bund on 15 August. It reached 70 basis points on 25 August, then widened to 82 basis points on 27 August, the widest of the range since 2024, and settled near 80 basis points on 2 September. Every one of those moves happened before any agency acted. The scheduled 2025 review dates for France were 12 September for Fitch, 24 October for Moody's, and 28 November for S&P.

Fitch acted on its scheduled date, lowering France to A+ from AA- with a stable outlook, citing political fragmentation that weakened the capacity to deliver fiscal consolidation. The accompanying projections were a deficit of 5.5% of GDP for 2025 and a debt ratio reaching 121% of GDP by 2027 against 113.2% in 2024. None of those figures were novel on the day. The direction had been visible in the spread for weeks.

That much of the conventional reading holds. The agency confirmed a direction the market had already established. The problem is what the conventional reading does next, which is to conclude that the announcement is therefore inert. It is not, and the reason has almost nothing to do with information.

The Structural Layer: Timing Is Scheduled, but Not Fixed

The first crack in the pure-lag story is the deviation clause. S&P did not wait for 28 November. It downgraded France to A+/A-1 on 17 October 2025, with the outlook moved to stable, in an unscheduled action, and French bond futures slipped on the news. Moody's, on its scheduled 24 October date, affirmed at Aa3 and moved the outlook to negative rather than cutting.

Three points follow from that sequence, and they matter more than the letters themselves.

• The calendar removes date uncertainty, not action uncertainty. Participants knew when each agency could speak, not what it would say, and in one of three cases not the date either.
• Agencies do not move together. Within five weeks, one cut on schedule, one cut off schedule, one held with a negative outlook. Treating "the rating" as a single variable discards most of the signal.
• An off-calendar action is itself a message. Choosing to deviate, and filing the required explanation, communicates urgency that a scheduled action of identical content does not.

What the Empirical Record Says About Anticipation

Two central bank research strands are worth separating here, because they are frequently collapsed into a single claim they do not jointly support.

An ECB working paper studying 24 EU countries on daily data from January 1995 through October 2010 found two-way causality between sovereign ratings and government bond yield spreads within a one-to-two-week window. Past spread changes helped predict rating changes, and rating changes helped predict spreads. That is the empirical backbone of the lag argument. But the same study found that negative rating events were associated with spread increases of roughly 0.08 percentage points for yields and 0.13 percentage points for CDS, that announcements were essentially not anticipated at one-to-two-month horizons, and that a country downgraded within the prior six months traded at spreads roughly 0.5 percentage points wider than similarly rated peers without a recent downgrade, with that penalty fading over about six months.

Read together, those findings say the market anticipates the direction over a horizon of days to weeks and does not fully anticipate the event over a horizon of months. Those are different claims, and only the first supports the headline version of the lag thesis.

A BIS working paper covering 55 advanced and emerging economies from January 2005 through December 2012, with 1,221 rating announcements filtered to 759 non-overlapping transitions, asked whether the price response had faded after the financial crisis. It had: the mean cumulative abnormal CDS response to downgrades in a two-day window fell from about 6.8% pre-crisis to about 2.2% in the 2010 to 2012 window. Faded is not gone. Announcements still produced statistically significant CDS responses, and the strongest reactions came from transitions out of a stable or developing status rather than out of an already-negative watch.

The Real Transmission Channel Is the Rulebook

Here is the part most coverage omits. A downgrade can move a price that already reflects the deterioration because ratings are wired into binding rules. Those rules do not care what a portfolio manager believes. They specify a letter and an action.

Index membership

The FTSE World Government Bond Index sets entry for a new market at A- by S&P and A3 by Moody's. Exit occurs below BBB- by S&P and Baa3 by Moody's. A market failing the minimum is removed at the next monthly rebalance, with a provisional treatment window for downgrades landing after the monthly fixing date and a cut-off at 5:00 p.m. New York time on the second-to-last business day of the month. Passive mandates benchmarked to that index do not deliberate. They sell on the rebalance date.

Capital treatment

Bank and insurer capital rules map agency ratings onto credit quality steps. Where three agencies rate the same issuer, EU rules do not take the worst one; the highest and lowest are discarded and the middle assessment governs. This is why France remained at credit quality step 1 after the September 2025 Fitch cut, with the other two agencies still higher, and only moved to step 2 once S&P cut in October. One downgrade changed nothing in the capital stack. The second one did.

Collateral eligibility

The Eurosystem's minimum credit quality requirement for marketable assets is BBB-, which is credit quality step 3, with asset-backed securities treated separately. The 2020 collateral easing package illustrates how consequential that line is: to prevent pandemic-era downgrades from mechanically shrinking the collateral pool, the Governing Council grandfathered assets that met the standard on 7 April 2020 so that they stayed eligible even two notches below the threshold, down to credit quality step 5, equivalent to BB, with haircuts applied according to their actual ratings.

The exception that swallows the rule

Against all of that, exposures to EU member state central governments denominated and funded in the domestic currency carry a 0% risk weight under the bank capital regulation, and receive no spread-risk capital charge under the insurance regime, regardless of the rating assigned. A euro area bank holding euro-denominated sovereign debt of a euro area state sets aside no capital against default on that position whether the letter reads AA or A+. For the largest single class of holders, three downgrades in twelve months changed the capital requirement by nothing at all.

What Markets and Coverage Consistently Miss

First, the notch matters only where a rule sits. A cut from AA- to A+ crosses no index threshold, no collateral floor, and no investment-grade boundary. A cut from BBB- to BB+ crosses several at once. Treating those as the same category of event, because both are "one notch," is the single most common analytical error in this area. The distribution of consequences is not smooth; it is concentrated at a handful of letters.

Second, spread and level are separate channels. The French spread stood near 69 basis points on 22 May 2026, tighter than its pre-downgrade August 2025 peak of 82 basis points, after three rating actions in between. Over the same stretch the 10-year OAT yield reached 4.12% on 18 August 2026, the highest since October 2008. The relative credit premium compressed while the absolute cost of borrowing rose. Confusing the two produces the recurring claim that a downgrade "raised borrowing costs" when the general level of rates did most of the work.

Third, outlook changes can outweigh downgrades. On 6 February 2026 Moody's affirmed Indonesia at Baa2 and moved the outlook to negative from stable, citing reduced predictability in policymaking and spending increases without matching revenue. The rating did not change. The IDX Composite fell 130 points, or 1.5%, to 7,976 in Friday morning trade, capping a third consecutive weekly decline of about 4%. Fitch followed on 4 March 2026, holding the BBB rating and cutting the outlook to negative on comparable reasoning. Two affirmations, no letter change, and a sustained repricing. If the letter were the operative variable, none of that should have happened.

Where the Analogy Breaks

Several conditions break the framework above, and they should be stated plainly rather than buried.

• Markets outside the EU calendar regime. The three-dates-per-year Friday calendar is an EU regulatory construct. A sovereign rated by agencies acting outside that perimeter, or rated by domestic agencies, has no equivalent public schedule, and the anticipation argument weakens accordingly. The May 2025 move of the United States from Aaa to Aa1 with a stable outlook was not governed by that calendar.
• Thin and captive markets. The lead-lag evidence rests on liquid secondary markets with continuous price discovery. Where a sovereign's debt is held predominantly by domestic banks under regulatory or political pressure, the spread is not a free-floating opinion and cannot lead anything. In that setting the rating may carry more information than the price, not less.
• Discrete events rather than trends. The framework describes gradual fiscal deterioration. It does not describe a default, a moratorium, a currency redenomination, or a sudden restructuring announcement. Those arrive faster than any review cycle and the rating follows mechanically within days.
• Private contracts. Ratings sit inside bilateral collateral agreements, repo eligibility schedules, and institution-specific mandates. Those are not observable in advance, so the rulebook channel can only be partially mapped from outside.
• Sample limits. Both studies cited end well before the present period. The declining announcement effect documented through 2012 need not have continued, and the ECB estimates come from a sample dominated by one European crisis episode.

There is a legitimate opposing reading of the same facts, and it deserves equal weight. If the announcement effect were purely mechanical, an outlook change that crosses no threshold should be close to a non-event. The February 2026 Indonesian sequence says otherwise. Agency commentary may aggregate governance and institutional assessments that are genuinely costly for outside investors to replicate, with the market reacting to that aggregation rather than to the letter. On that reading, ratings do not lag; they compress a diffuse judgment into a discrete, tradeable signal. The available evidence does not decisively separate the two explanations.

What to Watch Next Week

• Published review dates. Each agency's sovereign calendar is public. Check whether a sovereign under fiscal scrutiny has a Friday date approaching, and whether that date exhausts the agency's allotment for the year.
• Off-calendar filings. An unscheduled sovereign action, with its required explanation, is a higher-information event than a scheduled one. These are worth flagging in real time rather than in review.
• Spread versus level. Track the sovereign-to-benchmark spread and the outright yield as two series. Divergence between them separates credit repricing from a general rate move.
• Distance to the nearest rule. Count the notches to A-/A3, to BBB-/Baa3, and to the next credit quality step boundary. Zero or one notch is a different situation from three.
• Middle-of-three positioning. Where three agencies rate a sovereign, note which one currently sits in the middle. Only a move by that agency, or a move that changes which agency occupies the middle, alters the regulatory mapping.
• Outlook and watch changes. Affirmations with outlook cuts have moved markets, so they warrant the same attention as rating changes.

Concrete Framework

A monitoring checklist that follows from the above, in the order the questions should be asked.

1. Locate the date. Before treating an action as a surprise, confirm whether it fell on a published calendar date or was a deviation. The two carry different weight.
2. Separate letter from threshold. Write down the pre- and post-action ratings from all three agencies. Then check each against A-/A3, BBB-/Baa3, and the applicable credit quality step boundary. If none is crossed, expect the flow effect to be small regardless of headline volume.
3. Identify the middle rating. For regulatory capital purposes the middle of three governs. Confirm whether the action changed it.
4. Check index calendars, not just index rules. Removal happens at a scheduled rebalance, not on the announcement date, and the cut-off timing decides which month it lands in.
5. Look backwards at the spread. Chart the sovereign-to-benchmark spread over the prior 60 trading days. If it widened materially before the action, the informational content of the announcement is correspondingly lower.
6. Decompose the yield move. Split any post-action yield change into benchmark and spread components. Attribute only the spread component to the rating action, and only tentatively.
7. Ask who is forced to sell. List the holder categories subject to a hard rule at the crossed threshold. Where the answer is none, the mechanical channel is absent and any move is discretionary repositioning.
8. Set a decay expectation. Research on European sovereigns found the post-downgrade spread penalty fading over roughly six months. Treat persistence beyond that as separate, ongoing deterioration rather than a lingering rating effect.
9. Record the counterfactual. Note in advance what would falsify the framework: a threshold-crossing downgrade that produces no flow, or a non-crossing downgrade that produces a large one. Both occur, and both are informative.

The lag is real in direction and unreliable in timing, and what transmits a downgrade into a price is more often a rulebook than a revelation. The framework is weakest where it is most often applied: a single-notch move in a highly rated sovereign, where the letter changes, the headlines follow, and nothing in the plumbing moves.

This article analyses market structure and regulatory mechanics. It is not investment advice and is not a recommendation regarding any security or issuer.