The European Central Bank raised interest rates on September 10 for the second time this year, lifting the deposit facility rate to 2.50 percent, the main refinancing rate to 2.65 percent and the marginal lending rate to 2.90 percent, effective September 16. The vote was unanimous. Christine Lagarde called the decision “a no-brainer.” The bank’s stated reason is that “the conflict in the Middle East continues to generate inflation pressures, and inflation is set to remain well above target for an extended period.” Alongside it, the bank published projections showing inflation at 2.1 percent in 2028 - above its own two percent target, and higher than the 2.0 percent it forecast in June. That decimal is where the story actually lives.
What the Bank Did
The move was well telegraphed and followed June’s increase - the first rise since 2023, and the first by any major central bank in response to the war. Bloomberg and the Guardian framed it the same way: a second increase, aimed at a price shock the bank did not cause and cannot fix.
The tone was not gradualist. Lagarde told the press conference the decision was unanimous and - her word - “a no-brainer,” adding that the council had focused entirely on the current meeting and had not debated the future path at all. The statement said it twice: the Governing Council is “not pre-committing to a particular rate path,” and will decide “meeting by meeting.” The decision restates the mandate in one line: two percent, over the medium term.
Markets heard a door left open. Aviva’s rates desk concluded more hikes are coming, “and potentially more than one,” while cautioning that with two delivered and more than two priced, things “may well have gone too far.” A JP Morgan strategist was blunter: “One hike is not a ceiling.” A Deutsche Bank client survey found no consensus, splitting between a 2.75 percent peak, a hold, and a terminal rate of 3 percent. CNBC carries the full range.
The Case That the Bank Was Right
Let us give the hike its strongest form, because it has one.
- June’s increase responded to a genuinely deteriorating outlook, and the bank moved before its peers.
- Short-horizon inflation expectations are elevated. Households notice fuel and electricity prices more than any other, and 2021-22 left a memory of what happens when a visible shock spreads into wages.
- De-anchoring is an institutional concern, not a straw man. A bank that appears indifferent to a three-year high in the headline rate invites a more expensive correction later.
- The economy held up better than expected. The September projections revised growth up for 2026 and 2027, and unemployment is projected to fall to 5.9 percent by 2028, a new historical low.
- The bank rejects the soft label outright. The minutes of the June meeting insist the adjustment “should not be seen as an insurance hike but rather as a decision that was robust across the baseline outlook and the full range of alternative scenarios.”
That is a real case - for vigilance, and for being seen to act. Notice the last item, though: the refusal of the insurance label matters later.
What the Index Is Actually Made Of
Now the composition. Eurostat’s flash estimate for August put euro area inflation at 3.3 percent, up from 2.9 percent in July, the highest in three years. Take the index apart and the shape of the problem is obvious:
- Energy: 14.3 percent, up from 10.3 percent in July.
- Services: 3.0 percent, down from 3.3 percent.
- Non-energy industrial goods: 1.2 percent, up from 0.9 percent.
- Food, alcohol and tobacco: 1.2 percent, unchanged.
- Core inflation, excluding energy, food, alcohol and tobacco: 2.4 percent, down from 2.5 percent.
- Inflation excluding energy alone: 2.2 percent in both July and August, flat.
Energy alone rose 2.9 percent in the month. Read that list twice. The components a policy rate is designed to reach - services, core, the domestic pressures that build from demand - are flat or falling. The component that accelerated is priced in dollars, in Rotterdam, and in the Strait of Hormuz.
The average is not the experience. The household at the till buys fuel, electricity and food, not an index, and the shock lands hardest on those with the least room to adjust - the pattern traced in The Cantillon Effect - Who Gets the New Money First. The index averages. The household does not.
The Bank’s Own Research Says the Same Thing
On September 1, nine days before the hike, ECB staff published an analysis of what is actually driving this inflation. The finding deserves more attention than it got.
“Adverse energy supply factors accounted for around 90% of the increase in energy inflation between January and May 2026,” the authors write. Headline inflation rose 1.5 percentage points over those months, from 1.7 percent to 3.2 percent, almost entirely on energy supply shocks. Monetary policy contributed minus 0.1 percentage points; fiscal policy minus 0.2.
The paper then sets out the principle the September decision ran into: “Supply-side shocks, by contrast, push inflation and output in opposite directions, demanding a more measured monetary policy response.” The 2021-22 surge had a real demand component and justified forceful, persistent tightening. This episode does not, and it has so far justified a response that is “more gradual and flexible.”
That is the bank’s own framework, in its own voice - an argument for restraint, not proof that September was wrong. A precise warning about what the hike can achieve.
The Real Story Is the 2028 Line
The decision made the headlines. The September staff projections were the event.
June had inflation averaging 3.0 percent in 2026, 2.3 percent in 2027 and 2.0 percent in 2028. September kept 2026 at 3.0 percent and raised the other two: 2.5 percent in 2027, and 2.1 percent in 2028. Core inflation follows the same shape, from 2.5 percent this year to 2.6 percent and then 2.3 percent.
A 2.1 percent projection for the final year of the horizon is not a rounding error. It tells the public that on the bank’s central forecast, assuming the energy shock fades, inflation is still above target three years out. That is what turns a hedge into a path: with the terminal year above target, every meeting becomes a decision about another step.
So what moved the path? The bank says: higher assumed energy inflation in 2027, and stronger core inflation in both years, “driven by the better outlook for economic activity and somewhat higher wage growth.” Two of the three reasons are the price of energy, which the bank cannot set, and stronger growth - ordinarily a reason for relief, not alarm.
Meanwhile the second-round effects the bank fears - the wage and price spiral - remain, in its own assessment, “contained,” and “broadly unchanged compared with the June 2026 projections.” The persistence story did not change. The forecast did, mostly for supply-side reasons.
One line in the same document deserves a mention. The projections expect energy inflation to rise again in 2028 with the introduction of the EU Emissions Trading System 2. A legislated increase in the price of energy, written into the same index the bank now manages with interest rates. The state raises the price. The bank raises the rate to answer it. The household pays both - the mechanism set out in The Second Inflation Wave: The Good News Is Over.
What a Rate Can and Cannot Do
A policy rate is a demand instrument. It changes the price of borrowing and the exchange rate, and over time it moves expectations. It cools spending. It cannot drill a well, reopen a shipping lane, refine a barrel or move gas.
That is why supply shocks are the hardest case for any central bank. The staff paper says it plainly: supply shocks push inflation and output in opposite directions, so tightening does not remove the price increase - it removes demand against it. Inflation falls back, if it falls back, partly because the economy weakens. The cost is real output, paid at the end of the chain: the mortgage that resets, the credit line that reprices, the fixed income that does not move.
If around 90 percent of the energy move is supply, the instrument is aimed at the demand the bank can reach, against a price set elsewhere. Nothing decided on September 10 changes a barrel of oil. Readers who want the mechanics underneath can start with Money & Inflation - What They Actually Are and The Information Problem - Why Central Planning Fails.
Which leaves the honest defense of the hike: credibility. Expectations are fragile, the memory of 2022 is fresh, and a bank that looks passive through a visible price shock may find the next shock more expensive to contain. That is a serious argument - but about the institution, not the inflation. And the bank has refused that framing, insisting June was analysis rather than insurance.
Both claims cannot hold at once. Either the hikes are fighting inflation, in which case the composition evidence is awkward, because the inflation a rate can reach is falling anyway. Or they are protecting credibility, in which case the insurance label is accurate and the bank has denied it. The confusion is not a talking point invented by critics. It is in the statements.
Three Banks, One Shock
The Federal Reserve decides on September 16, with markets pricing roughly a 60 percent chance of a hike, a president threatening to halt trade with surplus countries unless rates come down, and US inflation at 3.4 percent. The Bank of England decides September 17, expected to hold at 3.75 percent for a sixth consecutive meeting after last splitting 6-3. Oil has traded above 105 dollars a barrel. Three central banks, one supply shock, and none of them able to lower the price of oil.
The Lens
Next time a central bank raises rates into a supply shock, ask what the hike is actually for - the inflation, or the institution’s credibility.
Then ask the second question, because the first is usually answered in a way that cannot be checked: which line of the index is actually moving? If the part that is accelerating is the part no interest rate can price, the tool is aimed at the wrong number and the bill arrives elsewhere - at the household, at the small firm, at the economy’s capacity to grow.
The bank’s own projections admit as much. The inflation it cannot touch is the inflation it is chasing. The number in the headline is always the past, and the price of a barrel is not set in Frankfurt.