The short answer
Annual inflation in the euro area rose to 3.2% in August from 2.9% in July. Energy was the main contributor to the latest increase. Services, however, made the largest contribution to the August inflation level. Those findings do not conflict: energy prices rose faster, while services carry much more weight in the consumption basket.
My conclusion is that the latest acceleration was driven by energy. Services’ large contribution and the rise in expectations mean we need to watch whether costs pass through to other prices. So far, there is no evidence of a broad-based wage-price spiral.
Energy drove the August acceleration
In Eurostat’s final data published on September 17 (opens in a new tab), headline inflation was 0.4% month over month and 3.2% year over year in August. The annual rate was revised down from the 3.3% flash estimate published on September 1. Energy prices rose 14.3% year over year, up from 10.3% in July. Energy also increased 2.9% month over month. In other words, energy had a high annual rate and became markedly more expensive within August.
The contribution breakdown shows more clearly what changed between July and August:
| Component | August contribution (percentage points) | Change from July (percentage points) |
|---|---|---|
| Energy | 1.29 | +0.35 |
| Services | 1.43 | -0.12 |
| Non-energy industrial goods | 0.30 | +0.07 |
| Food, alcohol and tobacco | 0.22 | -0.01 |
| Component total | 3.24 | +0.29 |
The headline rate rose from 2.9% in July to 3.2% in August, an increase of 0.3 percentage points. The changes in the component contributions published to two decimal places add up to 0.29 points; the small difference reflects rounding. Energy’s contribution rose by 0.35 points, while services’ contribution fell by 0.12 points. That separates the source of the latest acceleration. This is an accounting breakdown; it does not, by itself, measure the causal effect of energy on inflation.
Services still pose a persistence risk
Energy’s annual price increase was much faster than services’. Even so, services contributed 1.43 percentage points to the August rate, above energy’s 1.29 points. Services have a weight of 468.2 per thousand in the consumption basket, while energy’s weight is only 90.3 per thousand. A component’s price can rise rapidly, but its small basket share can limit its total contribution. A component’s contribution and its rate of price increase are not the same thing.
Annual services inflation fell from 3.3% in July to 3.0% in August. The published, rounded monthly increase in services was 0.0%. That does not mean every service price was unchanged; it means the aggregate monthly rate rounds to zero when shown to one decimal place. Inflation excluding energy and food also eased to 2.4% in August. Services therefore remain the largest contributor to headline inflation, but the August data do not show a fresh acceleration in services.
The distinction is useful: the energy contribution tells us where the latest increase came from, while the services contribution tells us which large component remains within today’s inflation level. A high services contribution is not, by itself, evidence of a persistent wage-price cycle. That would require prices, wages and costs to reinforce one another over several periods.
How does the energy cost pass through to company prices?
Energy costs are felt first through fuel, electricity and production inputs. They can then reach intermediate links such as transport, storage, supply and service delivery. The pass-through does not have to happen in the same month or at the same scale. A business might absorb a cost increase in its profits, have fixed its energy purchase price in advance, change suppliers, shift to a product that uses less energy, or raise its prices. Contract-renewal timing and customers’ ability to absorb a price increase also affect the result. I examine the broader path from oil prices to consumer prices in my guide to oil prices, inflation and markets (opens in a new tab).
To see the mechanism, consider a wholly hypothetical business. It has €100 in revenue and €90 in total costs, including €5 for energy, leaving €10 in operating profit. If its energy bill rises by 20%, costs increase by €1. If the business leaves its prices unchanged while output and everything else stay the same, its profit falls to €9. If it passes the full increase through to prices and raises revenue to €101, costs become €91 and profit remains €10, but its profit margin falls from 10% to about 9.90%.
Full cost pass-through can therefore preserve profit in euros without preserving the profit margin. Actual businesses differ in their energy share, hedging contracts, volumes, product mix, taxes, financing and demand conditions. The example is not about a specific company. It illustrates three possible outcomes for who bears the energy cost: lower profits, higher consumer prices, or some combination of the two. Eurostat or ECB aggregate inflation data do not show which option an individual company chose.
The strongest counterargument: pressure could spread
The persistence concern is that services still make the largest contribution and energy costs may pass through to other prices with a delay. In the European Central Bank’s September projections published on September 10 (opens in a new tab), headline inflation is projected to average 3.0% in 2026, 2.5% in 2027 and 2.1% in 2028. The corresponding projections for inflation excluding energy and food are 2.5%, 2.6% and 2.3%. This path includes a scenario in which core pressure could rise temporarily in 2027; it is not an observed outcome or an inevitable path for prices.
Expectations also need watching. In the ECB’s August Consumer Expectations Survey (opens in a new tab), household median inflation expectations were 3.0% one year ahead, 2.9% three years ahead and 2.5% five years ahead. Compared with July, the increases were 0.1, 0.2 and 0.1 percentage points, respectively. These are household survey medians, not market expectations or professional forecasts. The increase is a signal to watch, but it does not by itself prove that expectations are becoming persistently elevated or that a wage-price cycle has begun. I look separately at how inflation can pass through to rates, currencies and gold in my guide to U.S. inflation (opens in a new tab); its U.S. findings cannot be transferred directly to the euro area.
Wage indicators currently paint a more cautious picture. In her September 28 remarks (opens in a new tab), ECB President Christine Lagarde said recent data had not shown a material wage response to the energy shock. Total compensation per employee rose 3.3% year over year in the second quarter, down from a revised 3.6% in the first quarter. Eurostat’s hourly labor cost index (opens in a new tab) rose 3.1% year over year in the second quarter; wages and salaries increased 3.0%, while non-wage costs rose 3.2%. These measures do not count the same thing: one covers pay and employer social contributions per employee, while the other tracks employers’ cost per hour. The difference reflects their definitions, not a contradiction.
In its September 10 decision (opens in a new tab), the ECB raised its three key interest rates by 25 basis points. The deposit facility rate, which remunerates banks’ overnight deposits with the ECB, rose to 2.50%. The change took effect on September 16. The decision was an institutional response to the inflation outlook at that time, not a commitment to a future rate path. The ECB says its decisions depend on the inflation outlook, underlying trends and the transmission of monetary policy. A central bank cannot change energy supply; its role is to limit the spread of the initial price increase into wages, expectations and other prices.
Interest-rate changes also do not reach the economy immediately or evenly. Borrowing costs, returns on savings and spending decisions are affected over time; business financing and household demand are part of that transmission too. A single rate decision therefore cannot determine by itself how far the energy shock will pass through to prices or what the next inflation report will show. The ECB’s guide to monetary-policy transmission (opens in a new tab) summarizes how the different channels work. The key question here is whether the initial energy-driven move spreads into broader price and wage behavior.
What would change the interpretation?
The persistence case would become more convincing if several indicators strengthened together: services prices reaccelerated across multiple releases, price increases spread beyond energy and food, growth in wages or unit labor costs reaccelerated, and medium- to long-term expectations continued to increase. If that combination appeared, today’s energy move could be turning into a broader inflation problem.
Evidence in the other direction matters just as much. If monthly energy increases slow while services and other non-energy components ease, wage growth continues to lose momentum and expectations stabilize, the persistence concern would weaken. One monthly headline rate cannot determine company pricing behavior or a wage cycle on its own. The assessment should be tested against each new release, without giving too much weight to a single indicator moving in one direction.
Four things I will watch in the October 2 release
Eurostat’s September flash inflation release was scheduled for October 2. Its result is not available at this evidence cutoff. It is more useful to read the new data through four signals than to turn the headline number automatically into an interest-rate or market forecast:
- Energy momentum: Look at monthly and annual energy rates together with how the energy contribution moves alongside headline inflation. A high annual rate and a new monthly jump are not the same information.
- Services and non-energy breadth: Check monthly and annual services movements and inflation excluding energy and food. Services can remain a large contributor while their pace slows; separate contribution levels from momentum.
- Wages and labor costs: Assess compensation per employee, hourly labor costs and, when published, unit labor costs together over several periods, without treating them as the same measure.
- Expectations: Keep household surveys separate from market measures. Rather than treating one survey move as a loss of anchoring, look for persistence and consistency with the other price and wage indicators.
This framework does not automatically tell us the ECB’s next rate decision or the direction of the euro, German bond yields or equities. The October 2 release cannot prove persistence on its own either; it will add another observation on whether the energy-driven increase is spreading beyond energy.
How I will assess margins at European companies
When looking at European companies, I will first separate energy costs as a share of sales from the effect of price increases on sales volumes. Revenue may have risen while the energy bill was climbing; without separating price from volume, we cannot tell how much of the cost a company passed on to customers. When a company has a high energy share and limited pricing power, the same energy shock can put more pressure on operating profit. Fixed purchase prices agreed in advance can delay when that pressure appears.
Cash conversion and debt-refinancing schedules matter alongside the operating margin. More expensive inventory and energy payments can raise working-capital needs; higher interest rates can add a separate cost when debt is refinanced. These pressures vary by company and are not included in the simple operating-profit example above. Reading revenue, operating profit, cash flow and net income together gives more information than looking at a single margin ratio. Inflation data can start this review; assessing actual company outcomes requires financial statements and management’s explanation of price and volume.
Methodology and limitations
Evidence cutoff: 2026-09-30T20:15:00+01:00 (September 30, 2026, 8:15 p.m. London time). The analysis uses Eurostat’s final HICP data for August and retains the revision from the flash estimate to the final figure. The contribution table shows published, rounded percentage-point figures; the component total may not match the headline rate exactly. Contributions are not causal shares. The monthly rates are not seasonally adjusted, and annual rates also include base effects.
The ECB’s September projections were published on September 10, before the August final data were released; they are conditional projections based on energy assumptions. Household surveys are not market expectations. Total compensation per employee and hourly labor costs have different definitions. The company calculation illustrates a hypothetical mechanism only; it does not support conclusions about specific businesses or observed pricing behavior.
Sources and further reading
- Eurostat, annual inflation and component contributions for August 2026, September 17, 2026 (opens in a new tab)
- ECB, monetary policy decisions, September 10, 2026 (opens in a new tab)
- ECB President Christine Lagarde’s European Parliament remarks, September 28, 2026 (opens in a new tab)
- Eurostat, hourly labor costs in the second quarter of 2026, September 16, 2026 (opens in a new tab)
- ECB, August 2026 Consumer Expectations Survey, September 18, 2026 (opens in a new tab)
- ECB, September 2026 staff macroeconomic projections, September 10, 2026 (opens in a new tab)
- ECB, monetary-policy transmission mechanism (opens in a new tab)
- Further reading: What happens when oil rises? Inflation, bond yields, and technology stocks (opens in a new tab) and How Does U.S. Inflation Data Affect Gold and the Dollar? (opens in a new tab).
Not investment advice; for research and educational purposes.
