European chemical industry statistics: the EU27 sector ledger, delivered as rows

The statistical self-portrait of Europe's largest manufacturing exporter, written by the industry's own council. Cefic - Facts and Figures of the European Chemical Industry is the reference record: about twenty indicator families organised into six chapters, annual series from 2005 through 2024, delivered by Datadory as tidy series-by-year rows. The diversified chemicals data hub holds it among 17 primary datasets for the slice.

Scope is deliberately pan-European. Coverage centres on the EU27 aggregate with country breakdowns inside it and deliberate benchmarks outside it - China and the United States recur as comparators in the profile, trade and energy chapters. The six chapters divide the labour cleanly: Profile sets world and EU27 chemical sales by end market, country and structure; Trade Development follows extra-EU27 balances across five segments; Growth & Competitiveness tracks output from 2008 onward alongside capacity utilisation and productivity indices on a 2009=100 base; Our Contribution to EU Industry counts enterprises and SMEs by country; Energy Consumption reports consumption by source beside Europe-versus-US gas prices; Capital & R&I Spending closes with investment trends and R&D expenditure.

Quality holds under the rubric: Cefic scores 8 out of 10 against a catalog-wide mean of 7.81 across the 1,744 datasets Datadory catalogs, with field definitions verified rather than assumed. The 2025 edition fixes the headline picture - EUR 635 bn of turnover, 1.2 million people employed across roughly 31,000 companies, mostly SMEs, a 13% EU27 share of the global market against China's 46%, gas prices roughly triple US levels, a 40% energy-efficiency gain since 1990, and capacity utilisation still 9.5% below pre-crisis (2014-2019) levels.

Three readings from catalog research, exactly as the record carries them:

# Sales sheet - one labelled series row
series label : Sales (Value); Chemicals (20); Mn Euro;
2007         : 21460.724

# Sales sheet - same sheet, coded series row
series code  : A5ASNSL_20
2011         : 13849.319

# Employment sheet - coded series row
series code  : A5ASNSL_20
2005         : 38.766   (thousand persons)

Read the grammar, because it repeats across all ~20 indicator families. A labelled row announces itself - measure, sector code, unit - so Sales (Value); Chemicals (20); Mn Euro; needs no data dictionary to interpret. Many more rows travel under internal codes such as A5ASNSL_20 instead, carrying their meaning in companion decoder sheets rather than in the label itself. Either way the shape is identical: one series per row, one annual value per year column, units declared up front.

Two parsing realities decide whether the file behaves downstream. The label grammar is semicolon-delimited, so naive comma splitting destroys it; and the source layout is wide rather than long - years run across as columns - so most tools want an unpivot first. Datadory ships both fixed: typed long-format rows with decoded labels carried beside their codes, which is why one ingestion pass covers sales, trade, employment, energy and spending alike.

Trade sheets decompose the sector into five canonical segments defined as explicit sums of internal codes: Basic inorganics (2011+2013+2015), Petrochemicals (2014), Polymers (2016+2017+206), Specialties (2012+202+203+205) and Consumer chemicals (204). Partner-country sheets name the customers - United States, China, United Kingdom, Switzerland, Korea, Turkiye and Japan among them - keyed as direction plus country, so a corridor view needs no reshaping.

Four companions cover what one association account cannot, and the division of labour is clean by design:

The pan-European benchmark belongs to Cefic: sales, trade, production, employment, energy consumption, capital spending and R&D as annual series, EU27 plus country splits plus China/US comparisons, 2005 through 2024, presentation charts extending provisionally to 2025.

The official statistical counterpart belongs to Eurostat Production and Consumption of Chemicals by Hazard Class: EU27 tonnage in million tonnes under a constant MIO_T unit, split along two hazard ladders - harm to health and harm to the environment - from hazardous/non-hazardous totals down to most harmful substances and substances of concern, 2004 through 2024 in a 399-observation cube. It scores 8 out of 10, level with Cefic, and its signature move is the descent: within any year you can walk from total tonnage to the tiers where restriction dossiers concentrate first. One caveat travels with it: the current vintage disseminates only the EU27 aggregate, so member-state reads route through companion records.

The national deep-dive belongs to VCI German Chemical Industry statistics (Chemiewirtschaft in Zahlen): twelve chapters covering Germany's chemical-pharmaceutical industry - production value by segment, base-chemical output in tonnes down to chlorine and methanol, producer and trade price indices, sales, employment, foreign trade, energy inputs and R&D - refreshed continuously with Stand dates running December 2025 into August 2026, above an archive of roughly 69 annual editions reaching back to 1955.

Below the aggregate sit two more. UK Chemical Industries Association industry facts turns the UK into twelve regional rows - South East leading exports at GBP 11.3bn, North West GVA at GBP 8.5bn with 27,000 employees, down to Northern Ireland at GBP 1.3bn - beside national headlines of GBP 65.5bn turnover and GBP 61bn exports. And OEC Chemicals and Chemical Products trade profile adds the bilateral mirror: every Chemical Products HS6 product per exporter-importer pair per year, millions of trade-flow rows behind the section, drillable from chapter to heading to product. The table lines all six up.

Grain, history and delivered attributes, side by side - the comparison below is the shortlist for any European chemical workstream:

Read down the quality column and the slice is unusually tight at the top: three records tied at 8 against a catalog-wide mean of 7.81 across 1,744 datasets. What separates them is not score but grain - association benchmark, national account, statistical cube - which is why most teams run two or three together rather than choosing one. For the full ranked shelf including the American regulatory records, see best diversified chemicals datasets.

One habit saves rework across all of them: Cefic revises figures retroactively and states that successive editions are not strictly comparable year over year, while final-year values marked provisional remain subject to restatement. Rebuild trends from a single reconciled delivery rather than splicing numbers across vintages.

Two questions push past the headline totals, and each has a named answer in the catalog.

Which regions carry the industry? The UK CIA record resolves the United Kingdom to twelve rows - the South East export story, the North West employment-and-GVA story, Northern Ireland's GBP 1.3bn - while Cefic's own country sheets do the same for EU member states inside the aggregate. For physical geography rather than flows, chemicalparks.eu European Chemical Parks & Clusters Database registers 76 park and cluster sites across roughly 20 countries: total hectares, free hectares available for new investment, city and country per site, with Germany holding 31 of the 76. It is the site-level complement the macro records deliberately lack - no plant or company dimension exists anywhere in the association or statistical accounts.

How dangerous is what Europe makes? Eurostat's hazard ladder answers with structure rather than adjectives: every tier descends from the hazardous/non-hazardous total through hazardous, health-harmful and environment-harmful classes to the most harmful substances and substances of concern. Because production and consumption print side by side for every class and year, import reliance reads as arithmetic instead of estimation. The vocabulary behind the tiers is pinned in the hazard class glossary entry.

For country-pair detail, OEC's drill path runs from the Chemical Products section through HS chapters, headings and six-digit products to individual exporter-importer corridors, complementing Cefic's top-partner sheets.

Ranked by how directly the European record answers their day job:

1. Market researchers and consultants get a pre-segmented EU27 market sized at EUR 635 bn with country structure attached - segment trade balances bound the sizing exercise, and the German chapters anchor the largest single national market inside it. Their workflow: market researchers use cases.
2. Investors and quant researchers read capacity utilisation 9.5% below pre-crisis levels, the 13%-versus-46% share split against China and gas at roughly triple US prices as model features rather than press-release quotes - margin and capex-timing models need exactly this exogenous spine. Their workflow: investors quants use cases.
3. Data scientists and ML engineers get one series-by-year skeleton that unpivots into a single tidy table across ~20 indicator families, joined by a constant-unit hazard cube whose narrowness is a feature: how much, how dangerous, over time merges cleanly onto anything keyed by geography and year. Their workflow: data scientists use cases.
4. Competitive intelligence and product teams watch partner-country breakouts as demand-shift early warning for European customers, then map where rivals' physical footprint sits using the parks register.
5. Journalists, academics and policy analysts cite figures whose institutional authorship and dated basis survive scrutiny, and read the restriction pipeline in tonnage before headlines arrive. Their workflow: journalists academics use cases.
6. Procurement, supply-chain and site-selection teams rank territories by chemical consumption and trade position before committing coverage, then check hectares and free land at park level rather than guessing at industrial geography.

Because the numbers were never the hard part - the seams are. European chemical statistics arrive in the wild wearing three publisher conventions (an association narrative workbook, a statistical office cube, a national lobby's chapter files), coded rows whose meaning lives in decoder sheets, wide layouts that every tool wants unpivoted differently, units declared inside labels rather than carried as fields, and vintages that revise retroactively. Each seam is small. Together they are why two teams citing European chemical sales can be quoting different years, different boundaries and different segment definitions without either noticing.

Datadory handles the seams upstream of you: labels decoded beside their codes so the join key never depends on archaeology, long-format rows that load once instead of being unpivoted per tool, units carried as fields rather than buried in strings, and vintage handling applied when a trend crosses editions - restated values delivered flagged as-is rather than smoothed over.

Files, feeds, or straight into your warehouse. Daily, weekly, or hourly - your call. Start with a sample: name the indicators, countries, segments and years you need, and the extract comes cut to exactly that shape, field dictionary attached. Scope locks when you request it, and the schema in the sample is the schema you ship against.

Start with the Cefic dataset profile for the full field dictionary and sample rows, then take the national and statistical pairs: the German VCI chapters hold the deepest national archive in the slice, the Eurostat hazard-class profile carries the official tonnage account, and the head-to-head comparison with the EPA Chemical Data Reporting (CDR) Rule settles EU-association-benchmark-versus-US-volume-census line by line.

Widen or specialise from there. The diversified chemicals data hub collects every dataset page in the slice, the ranked best diversified chemicals datasets puts all seventeen primaries on one scorecard, and sibling posts cover the money view of the same complex - US chemical imports by HS code - and its physical twin, EU chemicals production by hazard class.

When you're ready to build, request a sample cut to the indicators, countries and years on your desk this quarter. It arrives with the field dictionary attached - and the schema in the sample is the schema you ship against.