A Cefic tool · CCA WG2
Import Surge Monitor.
Where extra-EU chemical imports surge and exporters undercut intra-EU prices.
Scroll
About
What it is
A monthly rolling read of Eurostat Comext that flags chemical products where
extra-EU import volumes are surging while exporters undercut
intra-EU prices. The quantitative core of the CCA WG2 import-surge methodology.
–
Active alerts this month
–
CN8 products flagged
+5%
Volume-growth trigger, vol up and price down
12m
Rolling window, refreshed weekly
← Back to the monitor
Methodology
How an import-surge alert is computed
This tool operationalises the Critical Chemicals Alliance WG2 Import Surge Monitoring Tool on a live Eurostat Comext cube. It reads every chemical product month by month and raises an alert wherever extra-EU import volumes are surging while the exporter undercuts intra-EU prices. Every threshold and definition below is reproducible from public Eurostat data.
1 · Unit of analysis and scope
Every test is run per CN8 product × extra-EU exporter country, so the same
product can raise several alerts (one per surging origin). The reporter is the
EU27: intra-EU27 trade is the synthetic partner EU27, and every
other partner is extra-EU.
- Product perimeter — the chemicals cube carried by the comext-etl pipeline: Combined Nomenclature chapters 28 to 39 plus 20, 21 and 30. In NACE terms this is division 20 in full and class 21.10 (basic pharmaceutical actives); finished pharmaceutical preparations (21.20) are out of scope.
- Geography — EU27 as importer; all non-EU27 partners as candidate exporters. Eurostat pseudo-partners (confidential, not-specified, bunkers) are excluded.
- Time — a 12-month rolling window ending at the latest month in the cube (the reporting month, selectable in the header).
2 · The rolling windows
Let the reporting month be the latest month available. Three windows are derived from it:
| Window | Definition | Used for |
|---|---|---|
| L12 | the 12 months ending at the reporting month | current volumes and prices |
| P12 | the 12 months immediately before L12 | growth and price-gap baseline |
| 4y ago | the 12 months ending 48 months before | the "volumes 4 years ago" context column |
3 · The three gates
A pair is flagged only if it clears the perimeter filter and then both Step 1 (volumes) and Step 2 (prices). The verdict is a strict AND across all three.
0
Perimeter
Keep only material, non-trivial flows.
Exporter holds ≥ 10% of the product's extra-EU import volume (L12)
Product is > €7.5 mln/year in extra-EU import value (L12)
Both L12 and P12 volumes ≥ 100 t (anti-noise floor)
1
Volume surge
Both conditions must hold.
1a — volume grew ≥ +5%, L12 vs P12
1b — and grew ≥ +5% year-on-year in at least 2 of the 3 previous calendar years
2
Price undercut
Both conditions must hold.
2a — exporter unit value below the intra-EU unit value (L12)
2b — the price gap was stable or widening vs P12
Formulas
Unit value (UV) is value divided by quantity, in EUR/kg. The intra-EU UV is the
EU27 partner's UV for the product.
growth = volume(L12) / volume(P12) − 1 → Step 1a if ≥ +5%
price gap = ( UV_intra − UV_exporter ) / UV_intra → Step 2b if gap(L12) ≥ gap(P12)
verdict = Perimeter AND (1a AND 1b) AND (2a AND 2b)
4 · Reading the columns
| Column | Meaning |
|---|---|
| NACE | 4-digit NACE class from cpa2015, with its label on hover |
| Product / CN8 | product name (full name on hover) and its 8-digit Combined Nomenclature code; the code links into the Comext Trade Radar pre-filtered to the same exporter and window |
| From | the extra-EU exporter country |
| Vol. growth 12m | volume growth, L12 vs P12 (the surge signal) |
| Vol. last 12m / prev 12m / 4y ago | volumes in thousand tonnes for the three windows |
| Δ share (pp) | change in the exporter's share of the product's extra-EU volume, in percentage points |
| Exporter price 12m | change in the exporter's unit value, L12 vs P12 |
| Price vs intra (=100) | exporter unit value relative to intra-EU; negative means undercutting |
| In prev. report? | whether the same pair was already flagged in the previous month's report |
Why the growth % is volume-only. An import surge is, by definition, a
rise in quantity, so the headline growth is volumetric (Step 1). Price is not ignored: it is
captured by the two dedicated price columns (Step 2). There is deliberately no single
value-in-euros growth figure, because it would conflate volume and price and blur the
"volume up, price down" signature the tool is built to isolate.
5 · Thresholds
| Parameter | Value | Rationale |
|---|---|---|
| Volume-growth trigger | ≥ +5% | WG2 proposed value; modest enough to catch early surges |
| Previous-years rule | ≥ +5% in 2 of 3 years | filters one-off spikes; requires a sustained trend |
| Exporter share floor | ≥ 10% | only origins that matter for the product |
| Product value floor | > €7.5 mln/year | drops economically marginal products |
| Volume noise floor | ≥ 100 t / 12m | keeps the growth ratio meaningful |
| Price gap (2b) | gap(L12) ≥ gap(P12) | exporter not becoming relatively dearer |
6 · Data source and refresh
- Source — Eurostat Comext, dataset
ds-045409, EU27 reporter, monthly value and quantity back to 2002. - Pipeline — the comext-etl job loads Comext into PostgreSQL and exports a Parquet cube; this tool reads that cube directly, with no separate extraction.
- Refresh — regenerated weekly (Monday), one new monthly report per run, committed and redeployed automatically. Each month is archived, which is what feeds the "in previous report?" comparison.
7 · Interpretation notes and caveats
- January spike. The previous-years test (1b) uses calendar years, so the three-year window shifts on 1 January. Expect more "new" alerts in January; it is a methodology artefact, not a data error.
- New entrants. A country with essentially no prior volume cannot satisfy the previous-years rule, so pure new entrants fall outside this version by construction.
- Unit values are proxies. Value-over-quantity is an average price, not a transaction price; it can be distorted by product mix within a CN8 code.
- Screening aid, not a determination. An alert flags a pattern worth investigating. It is not, on its own, evidence of dumping or an injury finding.