The ECB’s rate decision, staff projections and conference answers can convey different information about future policy. This article shows how to compare each with expectations and earlier communication, while keeping the Governing Council’s decision distinct from the projections and their assumptions.

What the Governing Council actually sets

The Governing Council takes monetary-policy decisions approximately every six weeks, but also meets for other tasks. Identify the actual monetary event rather than treating every Council meeting as a rate announcement. Its assessment concerns the euro area; the policy decision, projections and later discussion account are separate records.

The overnight deposit facility rate, DFR, concerns bank deposits with the Eurosystem. MRO applies to weekly refinancing against collateral; marginal lending, MLF, supplies overnight collateralized credit. The ECB’s March 2024 operational framework specifies DFR as the rate steering the policy stance. Record each instrument and effective date instead of assuming all three mean the same number or every media headline identifies DFR.

Macroeconomic projections are published in March, June, September and December. ECB staff prepare the March and September exercises; Eurosystem staff prepare June and December. These projections contribute to the Governing Council’s assessment. They are not a record of votes, and the Fed’s individual policy projections are not votes either. For the related US framework, see the Federal Reserve’s decision.

Staff projections are an input to the Council’s assessment, not endorsed forecasts or necessarily the Council’s views. Preserve the variable’s definition, period and assumptions; the diagram’s annual-average and December-to-December numbers answer different questions.

Keep the definition beside an annual projection

The word “annual” alone is insufficient. Our example compares the change in the whole year's average index with the change between two Decembers.

Scroll sideways to see the full diagram

Keep the definition beside an annual projectionPrevious year: index one hundred in every month. Current four quarters: 101, 101, 103, 103, constant within each quarter. Average 102 means two percent growth in the annual average. Final December 103 relative to previous December 100 gives three percent. Horizontal lines show the earlier index 100 and current average 102.Annual average ≠ December rateOur invented series • index level99100101102103104Q1Q2Q3Q4101103Earlier year: every month 100Current annual average: 102 → +2%December-to-December: 103 / 100 → +3%
Our model is neither an ECB projection nor HICP data. Each quarter contains three months with the same level: 101, 101, 103, 103. Every month of the previous year is 100. The annual average is a simple mean of twelve months; four equal blocks produce the same result. December represents the final constant block. Steps show constant monthly levels within each block: 101 in the first two quarters, 103 in the last two. Eurostat separates monthly and annual HICP data and links to an annual dataset. The definition of our simple model and its arithmetic are our own.
An annual-average 2% does not mean a December rate of 2%.

Write “change in the annual-average index” or “December-to-December”, rather than just “annual inflation”. When reading an actual projection row, preserve its definition and horizon. Neither example result is a Governing Council decision or a forecast of the euro.

The timing of a meeting day

The ECB normally publishes the decision at 14:15 Frankfurt local time and starts the press conference at 14:45. Use the institution’s dated event calendar and convert from Europe/Berlin to your own zone; the local clock changes between CET and CEST.

The statement and conference are separate information events. Their market reactions and the interval’s liquidity need not follow a fixed pattern of calm, movement and a larger second wave. Check the official event time and read both documents. You can place them in context by learning to read the economic calendar.

Accounts are normally published four weeks after the monetary meeting. Keep decision time, conference time, document-publication date and policy effective date in separate fields. A later conference answer or account was not available at an earlier entry.

Calculate changes versus a dated expectation

Compare what the ECB announces with the expectations recorded beforehand. During the seven-day quiet period, officials avoid remarks that could influence expectations about the forthcoming monetary-policy decision; this is not a ban on every speech. Read the wording and conditions rather than assuming the decision has already been fully priced.

Our hypothetical DFR is 3.00% before the meeting and the recorded forecast is a cut to 2.75%. Three possible announcements illustrate distinct comparisons: 2.75% is −25 basis points from the old level and zero from forecast; a 3.00% hold is zero from the old level but +25 above forecast; 2.50% is a −50-basis-point cut and −25 below forecast. The hold is not a hike. One basis point is 0.01 percentage point, and one assigned forecast does not describe everyone’s expectations.

The hawkish and dovish vocabulary describes a relative interpretation, not a currency outcome. Matching a rate forecast does not exclude new guidance, while a surprising rate does not establish EUR/USD direction. Keep both currencies and other information in the fundamental analysis comparison.

ECB meeting of 14 September 2023 · official policy facts
Decision: three key rates raised by 25 basis pointspublished policy change
Guidance: maintained long enough, those levels would contribute to the inflation targetconditional assessment
Future decisions: incoming data and policy transmission remain relevantno promise of the last hike
No venue-specific EUR/USD path is asserted hereprice claims need separate evidence

The September 14, 2023 statement raised the three key ECB rates by 25 basis points. It judged that the resulting levels, if maintained long enough, would contribute substantially to inflation returning to target, while preserving a data-dependent approach. It did not guarantee that no further hike would occur. A claim about that day’s EUR/USD path requires a separately identified market-price source and observation window.

Preserve timing and operational context

Forecast collection and public release can differ. The ECB SMA calendar separately records questionnaire dates, submission deadlines and aggregate publication dates. For example, its tentative September 2026 round lists submission by 26 August, the Council meeting on 10 September and aggregate publication on 14 September. An earlier collection date does not make the later public aggregate available for a pre-meeting backtest; retain the actual publication record and provisional-calendar status.

Read both the substantive assessment and its conditions. “Data-dependent” does not by itself mean neutral or predict lower volatility; its implications depend on the data and preceding guidance. Compare complete passages rather than assigning a fixed market outcome to a word. The broader comparison is explained in the overview of major central banks.

Compare EUR/USD with euro crosses such as EUR/GBP and EUR/JPY using the same observation window. Both currencies affect each pair, so neither reaction size nor volatility has a guaranteed ranking across those instruments.

Another distinction is operational structure. The 13 March 2024 framework announced an MRO–DFR gap reduction from 50 to 15 basis points, applicable from 18 September 2024. Our separate model holds DFR at a fictional 2.00%: MRO then falls from 2.50% to 2.15%, or 35 basis points, solely through the gap. Those fictional levels are not the actual September rates. Decompose ΔMRO = ΔDFR + Δgap; this arithmetic does not mean refinancing conditions have no economic effect.

Transfer exercise: previous DFR 2.50%, forecast 2.75%, announced hold 2.50%. Change is zero and forecast difference −25 basis points: an absent expected hike, not a cut or guaranteed euro decline.

What to do at the next meeting

  1. Identify the event and forecast. Check the actual calendar, instrument and projections. Retain the expectation source/time; mark it unknown if no pre-release record exists.
  2. Plan exposure and execution. Observe, reduce or retain positions according to a defined budget and actual order terms. Experience measured in months is not a risk threshold.
  3. Separate document windows. Read the decision and later conference answers with their timestamps. Waiting until the conference ends does not ensure safety or trend continuation.
  4. Save contradictory observations. Define the price source and window before measuring, keep other news, and distinguish price change from a transaction after costs. A handful of releases does not establish reversal-prediction skill.