At the Source: Eurostat's flash estimate on euro area GDP, read in full

The document is the “Euro indicators” release published by Eurostat on 14 August 2026, accessible in full on the website of the Union’s statistical office (full text).

1. What it says

In the second quarter of 2026, seasonally adjusted gross domestic product increased by 0.4% in the euro area and by 0.5% in the European Union compared with the previous quarter. Over the same period, employment rose by 0.1%.

These are two quarter-on-quarter variations, that is, measured against the immediately preceding three months and not against the same quarter of the previous year. The distinction is not pedantry: on such small values, the sign and magnitude change depending on the comparison chosen.

2. The limits the document declares itself

The release does not hide its own nature, and the caveats are in the text, not in a footnote.

First: coverage. The flash GDP estimates are based on data from member states representing 99% of the GDP of the euro area and the EU; those on employment cover 95% of euro area employment and 92% of EU employment. The remainder is estimated, not measured — and the margin on employment is four to eight times wider than that on output.

Second: the sequence of publications. A preliminary flash estimate for the same quarter had already come out on 30 July 2026, built on data from only 19 member states. The 14 August release updates it; it is therefore not new data, it is more complete data.

Third, and this is the decisive clause: the figures may be revised. Regular estimates are expected on 7 September and 20 October 2026, and the next publication in the series is set for 13 November 2026. One tenth of a point published in August can move twice before autumn is well underway.

3. How it is discussed, and what is left out

A release of this kind almost always circulates reduced to a single percentage. What gets lost along the way are precisely the three elements just listed: that the figure is provisional, that the employment component rests on narrower coverage, and that two revision dates are already on the calendar.

The difference between the two perimeters is also lost. Euro area and European Union do not coincide — 0.4% against 0.5% — and the one-tenth gap tells us that the part of the Union outside the single currency grew slightly faster in the quarter. Reporting “Europe grows by 0.4%” means, literally, citing one of the two figures as if it were the other.

4. Why it matters

Because decisions worth far more than a decimal point are built on these decimals. The same Eurostat series puts euro area annual inflation at an expected 2.9% in July 2026, according to the flash estimate of 31 July, up from 2.8% in June, with the energy component at 10.0% (31 July release); the complete July figure is scheduled for 19 August 2026, while the previous release on June prices dates from 17 July (HICP June release).

Moderate growth in output, near-stagnant employment, prices rising due to energy: these are the three variables that make up the macroeconomic picture underlying discussions of national budgets and monetary decisions. This newspaper draws no operational indications from this, neither on the economy nor on the markets: it merely notes that the figure destined to be cited for months is, by declaration of the very body that produces it, subject to two revisions already scheduled.

The first check is on 7 September 2026, when the regular estimate will replace the one published on 14 August.