The European Central Bank raises its deposit rate 25 basis points to 2.50% at 12:15 UTC today. Nobody disagrees. Euronews puts market odds at near-certainty; ING calls it an “insurance rate hike” and notes 2.50% is still inside the ECB’s own estimate of neutral. When a decision is that agreed, the decision is not the release. What has not been decided is what the new staff projections say about inflation the ECB itself attributes to supply, and which of two numbers — both called core, twenty basis points apart — the Governing Council chooses to lean on. Those arrive with Christine Lagarde at 12:45 UTC. EUR/USD traded an eleven-pip range in Tokyo waiting for it.
2.50% is the least informative number on today’s sheet
EUR/USD did 1.1629 to 1.1640 through the Tokyo session on Fisco’s reading, 1.1630 to 1.1638 on DZH’s. Eleven pips, on decision day. Yesterday, thirty hours before the same decision, we wrote that the pair had done thirteen pips in Tokyo. It has now done less on the day of the event than on the day before it. That is not the market holding its breath. That is the market having already spent the information.
The forward curve says the same thing in a different unit. InvestingLive puts roughly 47 basis points of ECB tightening in the price through year-end — today’s 25 included — and about 80 basis points through June next year. FXStreet has the terminal deposit rate near 3.00% in 2027. So today is priced, December is already priced as most of a hike, and the destination is priced too. There is no version of a 25bp move to 2.50% that surprises anyone.
Which means the tradeable content of the next several hours sits entirely in the projection document and the press conference. FXStreet’s own preview is blunt about it: the ECB is expected to keep “full flexibility” and Lagarde to restate the reaction function rather than signal a path, while the new staff projections “send mixed signals” with core-inflation scenarios that could read dovish. A hike delivered alongside dovish core scenarios is not a hawkish day, whatever the headline says.
Two numbers are both called core and they are twenty basis points apart
Euro-area August inflation, as Euronews reports it: headline 3.3%, up from 2.9% in July. Energy 14.3%, up from 10.3%. Services 3.0%, down from 3.3%. Core 2.4%, down from 2.5%.
We published 2.2% for core yesterday, from Eurostat’s flash. Both figures are right. Eurostat’s core-excluding-energy measure and the broader core measure vendors carry are different baskets, and the gap between them is about twenty basis points. This is not a contradiction to be resolved by picking the one you prefer. It is a live problem for anyone with an inflation threshold coded into a filter, because a threshold at 2.30% classifies today’s euro area as above target on one series and below it on the other.
It matters today specifically because the projections are the release. If the Governing Council builds its 2027 and 2028 profile off the measure that is falling, the accompanying language can be soft while the rate goes up. If it leans on the headline that went from 2.9% to 3.3% in a month, it cannot. Read which series the projection tables anchor before you read the direction of the arrows.
The ECB is hiking into a shock its own economists call supply
Euronews cites ECB research attributing roughly 90% of the rise in energy inflation between January and May to supply factors, driven by geopolitics rather than demand. FXStreet’s preview states the same figure more broadly, as energy supply accounting for 90% of the 2026 inflation episode so far. We use the narrower, dated version, because it is the one we can point at a window.
A central bank raising rates cannot do anything about a tanker. It can only compress the demand side of an economy whose demand side is, on the same August data, cooling — services down to 3.0%, core down on whichever measure you use. And it does so across a currency area where August inflation ran 4.5% in Spain, 2.9% in Germany and 2.7% in France. One rate, a 180-basis-point spread.
The oil input has also stopped cooperating with the hawkish story. Trading Economics has Brent at $100.42, down 0.78% on the day, and WTI at $95.54, down 0.53%. Still extraordinary levels, but the direction into the decision is down, not up. Our own Tokyo note yesterday flagged that this oil-risk correlation had flipped sign inside a day. It has not stopped flipping.
US PPI lands in the fifteen minutes between the decision and the press conference
Release timing is the part most schedules get wrong, so here it is in UTC. German final HICP landed at 06:00 and confirmed the flash at 2.9% year on year, unchanged from July and exactly consensus — a non-event by construction, and duly a non-event. ECB decision 12:15. US August PPI and initial jobless claims 12:30. Lagarde 12:45. US existing home sales and wholesale inventories 14:00. EIA natural gas storage 14:30. EIA weekly crude 16:00. The 30-year bond reopening 17:00.
PPI consensus is +0.4% month on month headline and +0.3% core, against priors of 0.0% and +0.2%; year on year, 5.3% and 4.6% against 4.7% and 4.2%. That resolves a vendor disagreement we have tracked for three days — Train2Invest had +0.4% while Trading Economics had 0.0% — in favour of +0.4%, with FXStreet’s event pages and TradingKey’s preview now agreed. Energy is the expected driver, which is the same shock the ECB is hiking into, arriving fifteen minutes later on a different continent.
The practical consequence is that between 12:15 and 12:45 UTC, EUR/USD is being priced by two independent information sources at once, and neither is finished. A euro-area statement lands, then a US inflation print lands on top of the reaction to it, then the person who wrote the statement starts talking. Any system that reads the 12:15 move as the ECB signal will be holding a position whose thesis was overwritten at 12:30 and re-overwritten at 12:45. If you carry a news filter, this is one window, not three.
What this does not tell you
It does not tell you whether the ECB hikes. It is priced, not delivered, and 12:15 UTC has not happened yet. Everything above is a preview.
We have no option-expiry levels for today. InvestingLive’s 10 September New York cut page was not published at the time of writing, and the only version we could reach was last year’s. Yesterday we built a whole piece on the 1.1615 and 1.1600 expiries pinning EUR/USD; today we have nothing of the kind, and we would rather say so than reprint yesterday’s strikes as though they were live. Traders Web FX’s order pages update roughly once a day and we do not publish levels from them unless the stamp is fresh.
We have not read the projection document. Nobody has. The characterisation of it as “mixed signals” with possibly dovish core scenarios is FXStreet’s expectation, and the “insurance hike” framing is ING’s, quoted by Euronews. The market-pricing figures — 47 basis points to year-end, 80 to June — are InvestingLive’s single reading, not a consensus of curves.
And it does not tell you what to do with EUR/USD. FXStreet’s technicals have the pair just below the mid-1.1600s near a one-and-a-half-week high, with 1.1709 the recent swing high, 1.1625 the 23.6% retracement, and support stacked at 1.1580–1.1570 where the 38.2% retracement meets the 200-period EMA at 1.1582, RSI around 58. That is a description of an eleven-pip range with a two-hundred-pip envelope drawn around it. It is not a signal, and if reading it changes which trades your system takes today, you are discretionary and your backtest does not describe you.
Related
- FX events calendar — the 12:15, 12:30 and 12:45 UTC cluster and the rest of the week
- Signals
- EA presets — news-window and filter settings
