The job vacancy rate reads 2.6% and 2.8% for Germany in the same quarter

Eurostat published two job vacancy rates for Germany in the fourth quarter of 2025. Its dissemination API returns 2.6 %. Its own news release of 18 March 2026 ranks Germany at 2.8 %. Same country, same quarter, same dataset, same activity aggregate — and a gap of 0.2 percentage points that is not an error, a revision or a correction. One figure is seasonally adjusted and the other is not. ¹

That is the first thing an employer should know about this rate. It exists in two published versions, both are official, and the figure alone does not say which of them is on the page.

One dataset, two series

The figures come from Eurostat dataset jvs_q_nace2, queried for indicator JVR, activity aggregate B-S, size class TOTAL, at quarterly frequency. The seasonal-adjustment flag is a dimension of that query like any other. Set it to SA and the same request that produced Germany’s 2.6 % produces 2.9 % for Austria and 4.0 % for the Netherlands. Set it to NSA and Germany becomes 2.8 %, Austria 2.8 % and the Netherlands 3.9 %. ¹

The same quarter, ranked twice

Job vacancy rate, 2025-Q4: seasonally adjusted and unadjusted

Country	Seasonally adjusted	Not seasonally adjusted
Netherlands	4.0	3.9
Belgium	3.5	3.5
Austria	2.9	2.8
Germany	2.6	2.8
Hungary	2.0	2.0
Czechia	1.9	1.9
Slovakia	1.0	1.0
Poland	0.8	0.7
Bulgaria	0.8	0.8
Romania	0.6	0.6

Six of these ten countries read identically on both series. The four that do not are enough to change the answer to a question an employer might actually ask.

The rate is a share, and the extract does not carry its denominator

The job vacancy rate is a ratio, not a count. It cannot be converted into a number of open posts without knowing what sits underneath it, and the API response used here does not supply that: the query returns indicator JVR and a single value per country, with no accompanying series of occupied or vacant posts.

We are therefore not in a position to restate the formula from the evidence behind this piece. The published series we retrieved states the indicator code and the value; it does not state the denominator. Any employer who needs the arithmetic — rather than the ranking — has to go to Eurostat’s methodological documentation for the job vacancy collection rather than to the number itself.

What the arithmetic of a share does establish is that a country’s rate carries no information about the size of its labour market. The Netherlands at 4.0 % and Romania at 0.6 % differ by a factor of about 6.7 on the adjusted series for 2025-Q4, but that ratio compares two proportions, not two quantities of unfilled work.

The quarterly clock

The series moves on a quarterly cycle, and the headline moves with it. Eurostat’s news release covering 2025-Q4 was published on 18 March 2026; by the June 2026 extraction of its Statistics Explained article on job vacancy statistics, the headline had rolled forward a quarter. ^{1 2}

Job vacancy rate, euro area and EU, not seasonally adjusted

Quarter	Euro area	EU
Q1 2025	2.4	2.2
Q4 2025	2.2	2.1
Q1 2026	2.3	2.1

Read as a year-on-year comparison, the euro area is down 0.1 points and the EU is down 0.1 points on the same quarter of 2025. Read as a quarter-on-quarter comparison, the euro area is up 0.1 points and the EU is flat. Both readings are correct, and for the euro area they point in opposite directions — which is the reason a single quoted rate without its quarter is close to useless. ²

What is changing underneath the series

The dataset that supplies the country detail above is closed at the end. Its own label reads “Job vacancy statistics by NACE Rev. 2 activity - quarterly data (2001-2025)”, it was last updated on 20 March 2026, and the last period it returns is 2025-Q4. The Q1 2026 figures in the second table are therefore not available from it. We have not established, from the evidence behind this piece, which successor dataset now carries the quarterly country series. ¹

The rate also measures something narrower than the term “labour shortage” implies. It is not a cost measure: the Netherlands, top of the 2025-Q4 vacancy table, recorded hourly labour costs of €47.9 in 2025, and Romania, bottom of it, €13.6. We draw no causal inference from that, in either direction, and none is available from these series. ³

Before quoting the number

Four things need to be fixed in writing whenever this rate is used in a board paper, a shortage argument or a pay case, because each of them changes the figure.

The series. Adjusted or unadjusted. The country table in Eurostat’s news release of 18 March 2026 is unadjusted; if the figure came from a database query, the flag was set explicitly, whether or not anyone noticed.

The quarter. The headline changes four times a year, and the article-style pages that carry it are rewritten in place rather than dated and archived, so a link alone does not pin the number down.

The activity aggregate. The figures here are B-S with all size classes. A different aggregate is a different rate for the same country and quarter.

The geographic aggregate. Euro-area figures changed composition for 2026. For 2025-Q4 the twenty- and twenty-one-country aggregates both read 2.2 %, so nothing turns on it in that quarter — but that coincidence should not be assumed to hold in the next one.

Anyone maintaining an internal series should also record the date the query was run. Eurostat’s article pages carry an extraction date rather than a publication date, and the dataset stamps its own last update; neither is visible in the number once it has been copied into a slide.