Eurostat Production in Construction Index

Europe's monthly pulse-check on whether anybody is building anything - measured as output volume, not sentiment, not permit filings, not invoice values. The index (stable identifier sts_copr_m) tracks changes in the price-adjusted output of construction across NACE Rev. 2 section F, intended to reflect the sector's monthly volume value added: approximated by deflating gross production, or by following inputs of labour and raw materials where direct output measures run thin.

It holds Principal European Economic Indicator status - the EU's shortlist of statistics member states commit to reporting under Regulation (EU) No 2019/2152 - and the completeness shows. 33 geographies sit in one keyed panel: the EU-27 aggregate, euro-area totals, every individual member state, plus Norway, Switzerland, the United Kingdom, North Macedonia, Turkiye and Ukraine. History reaches from January 1990 through June 2026, so any European construction cycle you care to name - the post-unification German boom, the euro crisis, the pandemic whipsaw - is inside a single queryable series.

Datadory delivers the whole panel as analysis-ready rows: coded dimensions rather than prose labels, harmonized definitions, one fully keyed observation per row. Get a sample of this dataset and the first rows arrive with your country filter already applied.

One row per geography per month per activity per adjustment state per unit. Verified readings straight from the archive:

geo      : EU27_2020   (EU-27 aggregate)
indic_bt : PRD         (production volume)
nace_r2  : F           (construction, total)
unit     : I21         (index, 2021 = 100)
s_adj    : CA          (calendar adjusted)
time     : 2026-04     value: 104.1
time     : 2026-05     value: 108.0
time     : 2026-06     value: 110.8

geo      : DE          (Germany)
nace_r2  : F           unit: I21   s_adj: CA
time     : 2026-06     value: 95.7

Five months make the point better than a description could. The calendar-adjusted EU-27 total-construction index climbed from 104.1 in April 2026 to 108.0 in May and 110.8 in June - a sector running almost eleven percent above its 2021 average - while Germany's June print sat at 95.7, nearly five percent below its own base. Same measure, same month, opposite sides of 100. That divergence is precisely why cross-country construction work wants one harmonized panel instead of twenty-plus national releases stitched by hand. Request a sample and you get rows in exactly this shape, extended across whichever countries, activities and years you name.

Eight documented fields define every observation, and each is a coded dimension rather than a prose label - which is why joins never drift between vintages. The definitions below are verified against the delivered data itself, not inferred from documentation.

Geography - 33 geographies in one keyed dimension: the EU-27 aggregate, euro-area aggregates, every individual EU member state, plus Norway, Switzerland, the United Kingdom, North Macedonia, Turkiye and Ukraine. Aggregates and member states ship side by side, so a Europe-wide chart and a per-country breakdown come out of the same query shape with no rollup step of your own.

Temporal - January 1990 through June 2026. The longest national series carry 438 monthly observations; shorter panels exist where reporting starts later. Three and a half decades span several complete construction cycles, recessions included, which is the difference between fitting a model and extrapolating one.

Granularity - country x NACE activity x seasonal adjustment x unit, at monthly reference periods. Six activity members (F, F41, F42, F43, F_CC1, F_CC2), three adjustment states, five units. There is no regional, company or site detail - this is the sector-level instrument, and it pairs naturally with asset-level sources when you need the plants behind the aggregate.

API, files, or your warehouse. Daily, weekly, or hourly.

• Demand forecasting for building products - cement, steel, fittings and equipment planners read the seasonally adjusted index as forward demand for their order books; the workflow is sketched on our demand forecasting page.
• Sub-sector rotation analysis - the F41/F42/F43 split shows whether buildings, infrastructure or specialist trades are doing the moving, which headline construction series flatten away; patterns like this sit on our market researchers use cases page.
• Macro-to-company benchmarking - if the sector index fell 4% and your revenue fell 4%, that is the market, not you; monthly reviews get a defensible denominator.
• Investor and quant work - thirty-five years of seasonally adjusted monthly data makes a clean construction-cycle proxy for backtests against builder and materials returns, as covered on the investors and quants page.
• Cross-country modeling - train on a harmonized 33-geography panel with stable dimension codes rather than reconciling national definitions; worked patterns live on the data scientists use cases page.
• Citable reporting - journalists and academics quote a single official series whose every figure carries the same definition, instead of footnoting a dozen statistical offices.

Market researchers and consultants get the fastest route to a defensible European construction sizing - one harmonized definition, ready-made aggregates, no stitching. Data scientists and ML engineers get a fully keyed, schema-stable cube long enough to train on without augmentation. Investors and quant researchers get a monthly coincident indicator for construction-exposed positions, deep enough to cover several rate regimes. Journalists, academics and students get citable official statistics where the caveats are documented dimensions rather than buried footnotes. Competitive intelligence and product teams get the sector backdrop that separates a rival's bad quarter from an industry downturn. Developers building data products get coded dimensions that survive revisions intact - integration patterns sit on the developers builders use cases page.

Its nearest neighbours in the catalog measure tonnes or assets, not activity. Monthly Global Cement Production Data (CICERO) tracks physical volumes of one material across many economies; the Global Cement and Concrete Tracker inventories 3,884 cement and clinker plants across 171 countries, owners included. This index covers all of construction - every trade in NACE section F - but Europe only, and as an index of change rather than a count of things.

Layered together they answer three different questions: who makes the material, how much of it exists, and what the sector that consumes it is doing this month. The head-to-head continues on our Eurostat production index vs GCCT page, and the full slate sits on our best construction materials datasets ranking.

Three things worth knowing upfront.

First, this is an index, not a quantity. A reading of 110.8 means output sits 10.8 percent above its 2021 average level; it never tells you tonnage, euros or square metres. If your model needs physical volumes, say so in the sample request and we will scope the matching tonnage sources alongside.

Second, treat the series as confirmed history rather than a nowcast. June 2026 is the latest reference month in the current extract, and revisions are part of the deal with any short-term indicator - pin your extract vintage if you re-run comparisons across releases.

Third, the grid is sparse by design. Some country-and-activity combinations return empty value sets - Germany paired with the condensed buildings code is a documented case - so check availability before assuming a complete matrix, and treat absence as a reporting fact rather than a zero.