ECB Hikes to 2.50% and Lifts 2028 Inflation to 2.1% (10 September 2026): The Projection Revision Was the Decision

The European Central Bank raised its deposit rate by 25 basis points to 2.50% on 10 September, exactly as every economist polled had expected. The number that carried the news was elsewhere. In the new staff projections, headline inflation in 2028 — the final year of the forecast horizon — was revised to 2.1% from 2.0%, and underlying inflation excluding energy and food was marked up to 2.6% in 2027 and 2.3% in 2028, above the headline rate in both years. For a Governing Council that spent the summer arguing the energy shock would wash out on its own, that is a change of position, and it is the justification the 3.3% August headline could not supply by itself.

What actually happened

The decision published at 14:15 CET raised all three key rates by 25 basis points with effect from 16 September 2026.

The stated reason was continuity with June: "The conflict in the Middle East continues to generate inflation pressures, and inflation is set to remain well above target for an extended period." The Council also kept both pieces of language that preserve its optionality — a data-dependent, meeting-by-meeting approach, and an explicit statement that it "is not pre-committing to a particular rate path."

This note was published four days before the meeting and argued that the rate would be the least informative part of it. That held. A hike priced at close to full odds and forecast unanimously cannot move an exchange rate by arriving; the capacity to surprise had already migrated into the projections. So that is where to look.

The revision that decided it

September is one of the four meetings a year that carries a new round of staff projections, and this one moved the lines that matter.

Read the 2028 line first, as this note said to. A final year at 2.0% is a central bank saying the shock passes through and the medium-term path is intact; on that forecast a hike is a hedge against a risk that has not materialised. A final year at 2.1% is a different claim. It is small in magnitude and large in meaning: the staff now expect the energy shock to leave a residue at the end of the horizon, which converts the hike from a precaution into the arithmetic consequence of the forecast.

The growth upgrade compounds it. Tightening into a terms-of-trade loss is a trade-off because higher rates do nothing about the energy bill while adding to borrowing costs in an economy already losing real income. Marking 2027 growth up to 1.4% from 1.2% is the staff saying that cost is more affordable than they thought in June — which makes the tightening easier to justify inside the Council, whatever one thinks of the forecast.

The detail that carries the most information

The most telling number is not the headline revision. It is the relationship between the two inflation lines at the end of the horizon.

That is worth sitting with, because it is the precise thing the Council said it had not yet observed.

From "we are not seeing it" to a forecast that assumes some of it

At the 23 July press conference, where the Council held rates, Christine Lagarde was unusually direct: "Second-round effects: we are not seeing it. Believe me, we are really scrutinising the emergence of second-round effects, but we are not seeing it."

The supporting evidence then was consistent. Compensation per employee had slowed to 3.5% from 3.8%, rising productivity was containing unit labour costs, and most measures of longer-term inflation expectations stood at around 2% while shorter-horizon measures stayed elevated — the standard signature of a supply shock markets expect to pass. Lagarde flagged the two inflation prints the Council would see before September. Those arrived at 2.9% for July and 3.3% for August, hotter on the headline both times, and neither showed the spread she was looking for.

The realised data still does not show it. The August flash estimate had the index excluding energy at 2.2%, unchanged from July, and services — the component that best reflects domestically generated pressure — actually decelerating to 3.0% from 3.3%. What changed on 10 September is not the observation but the forecast: the staff have now written a medium-term path in which underlying inflation does not return to 2%, and the Council has priced that view into the policy rate. The accompanying scenario work is explicit that it examines the conflict's intensity and duration "as well as its indirect and second-round effects."

The distinction matters for anyone reading the euro off this meeting. A central bank tightening against an observed wage-price process and one tightening against a projected one are in different positions. The second is more vulnerable to revision, because the evidence that would overturn it has not been collected yet.

What the August print said, and why composition still rules

Eurostat's flash estimate put euro-area annual inflation at 3.3% in August, up from 2.9%. By component it remained a single-line story.

Every point of the headline's rise came from energy, and the line that most reliably signals home-grown pressure moved the other way. Germany reported national CPI of 2.9% with core at 2.4% and services easing to 2.8%, according to the Federal Statistical Office, and German energy inflation ran at 10.5% against the bloc's 14.3%. The gap between the bloc and its largest member narrowed, and it narrowed on the energy line.

Why hike into a shock you cannot fix

No policy rate lowers the price of a barrel, and the ECB has never claimed otherwise. The case for tightening is not that rates suppress the shock; it is that rates decide whether the shock stays a one-off level shift or becomes an inflation rate.

The transmission runs through behaviour. A jump in fuel and power costs lowers real incomes. If workers and firms accept that as a loss, the price level steps up once and the annual rate drops out of the data twelve months later. If instead they try to recover it — through wage settlements, margin pass-through, indexed contracts — the one-off becomes persistent, and the central bank finds itself fighting a domestic inflation process rather than an import price.

Tightening early is an attempt to make the first outcome the default by removing any doubt about accommodation. The September projections are the Council conceding that the second outcome now carries enough weight to act on.

The scenario map, resolved

This note set out three shapes before the meeting. The first one landed.

The branch that landed was described in advance as the strongest of the three for the euro, because it converts a hedge into a path. That reasoning still applies, with an important qualification: the path was strengthened by a forecast, not by realised underlying inflation, and the Council explicitly refused to commit to it. A rate path that rests on a revision is a rate path that a single soft services print can weaken.

How this reads across the five factors

The Pip Theory meter scores eight currencies on interest rates, growth, positioning, risk sentiment and commodities. This decision touches three of them for the euro, and not all the same way:

• Interest rates: improved. A move to 2.50% alongside an above-target final-year forecast narrows the gap to a Federal Reserve markets began repricing toward tightening after August's 162,000 US payrolls print — the divergence traced in the August US CPI note. The factor is relative, which is why the euro's page has to be read against the dollar's.
• Growth: marginally better on paper, with 2027 marked up to 1.4%. The income squeeze is still real; the staff simply think the economy absorbs it.
• Commodities: unambiguously negative. Energy inflation of 14.3% is a terms-of-trade loss for a large net importer, and the escalation examined in the Hormuz note has kept crude bid.

Bond markets had moved well ahead of the decision. The 10-year Bund set a 15-year high in early September, with the 1 September selloff taking it to 3.36% and the French and Italian 10-years to 4.215% and 4.188%. Whether that is a policy signal or a term-premium and fiscal-supply story — the distinction drawn in the long-end selloff note — determines whether higher euro-area yields support the currency at all. The methodology page explains how the factors combine, and what drives the euro covers the structural version.

What to watch next

The projections have answered the question this note was written around, so the open items shift.

Watch the September flash estimate in early October, which carries a full month of renewed escalation in crude: on current composition the headline gets worse before the underlying picture does, and the Council has now committed itself to a forecast in which the underlying picture matters more. Watch services, which fell to 3.0% in August — it is the single series most capable of contradicting a core projection of 2.6% for next year. And watch the wage trackers and the survey measures of longer-term expectations that Lagarde cited in July, because the September round has effectively written a forecast those measures did not yet support.

The Council has raised rates twice in three months against a price shock it cannot influence, on the argument that doing so keeps the shock from becoming something else. It has now put that argument into its own forecast. What it has not done is claim to see the process it is guarding against.

Educational macro context only — not investment advice.