What trade data doesn't show

Gap 1 · Processing concentration

Flag of origin is not the same as where the material was processed

Trade statistics record the country that shipped the consignment — not the country where the material was actually refined or processed. For critical minerals, that distinction is the whole story.

Rare earths

China refined ≈90% of the world's rare earth oxides and ≈92% of permanent magnet alloys in 2023–2024. EU domestic processing share: <1%.

Lithium

China processed ≈60–65% of the world's lithium chemicals (carbonate and hydroxide) in 2024 — even though Australia and Chile dominate mining.

Cobalt

China refined ≈75% of global cobalt in 2024 — primarily into sulfate and chemicals for cathode active material — even though ≈70% of mined cobalt comes from the DRC.

Natural graphite

China produced ≈80% of the world's mined natural graphite and ≈95% of spherical / coated anode-grade graphite in 2024.

Nickel chemicals

China processed ≈70% of class-1 nickel chemicals (battery-grade nickel sulfate) — much of it from intermediate products shipped from Indonesia.

A European battery cell maker buying lithium hydroxide "from Australia" or nickel sulfate "from Korea" sees a tier-1 supplier. The chemical may have been refined or partially processed in China before re-export. Trade data shows the last hop; the processing geography is upstream of it.

Gap 2 · Beneficial ownership

The country flag tells you who shipped it. Not who owns it.

Indonesian nickel is the canonical case. The material is Indonesian by trade-data classification; the processing capacity is overwhelmingly Chinese by ownership, finance, and offtake.

Indonesia share of global mined nickel

≈55% (2024 USGS estimate, up from ≈30% in 2020).

Indonesia share of refined nickel growth

≈100% of global growth in class-1 and class-2 refined nickel capacity since 2021 has come from Indonesian RKEF and HPAL plants.

Chinese capital share

Independent estimates (IEEFA, AidData, S&P Global Market Intelligence) place Chinese-financed or Chinese-controlled capacity at ≈70–85% of Indonesia's nickel processing footprint (Tsingshan, CATL, Huayou, GEM, Zhejiang Huayou Cobalt). Trade data records the output as "Indonesian".

CRMA implication

The Critical Raw Materials Act's 65% single-country dependency cap (Art. 5) targets concentration at any relevant processing stage. "Country" defined as flag of origin would understate exposure; defined as beneficial ownership, several materials already breach the cap today.

Battery Regulation due-diligence (Art. 47–53) reaches into the supply chain — but only if the buyer reaches further than the import declaration. Ownership graphs are not in Eurostat or UN Comtrade.

Gap 3 · Mirror data divergence

Reporter and partner trade statistics often disagree by 10–40%

When the exporter's customs record and the importer's customs record describe the same shipment, they should match. They routinely don't — and for sensitive flows the gap is structural, not random.

Myanmar–China rare earths

Chinese imports of rare earth concentrates from Myanmar persistently exceed Myanmar's reported exports — by multiples in some years. Concentrates flow across the Kachin State border largely outside formal documentation.

Russia post-2022

EU imports of aluminum, fertilizers, and refined metals from Turkey, Kazakhstan, the UAE, and Georgia rose sharply after Russian sanctions. Mirror data with Russia indicates significant transshipment that is not visible in tier-1 EU import statistics.

DRC cobalt

Hand-mined (ASM) cobalt enters formal supply chains via Lubumbashi-based trading houses; total DRC export tonnage is consistent with reported figures, but the LSM/ASM split is not visible in either reporter or partner trade data.

Mirror discrepancies are useful as a signal of where flows are being obscured — not as a substitute for ground-truth supplier data. UN Comtrade documents the discrepancies; reconciling them requires field-level intelligence.

Gap 4 · Artisanal and informal flows

ASM cobalt is mixed into formal supply before it reaches a port

Artisanal and small-scale mining produces a meaningful share of DRC cobalt. It enters formal supply chains through Chinese-owned trading houses well before any commodity export figure is recorded.

ASM share of DRC cobalt

Estimates range from ≈15% to ≈30% of DRC output, depending on year and methodology (OECD: 15–30%; Amnesty earlier studies: up to 30%). Exact figure is structurally unknowable from trade data alone.

Aggregation point

ASM cobalt is purchased at négociant level, sold to processing entities at Lubumbashi (predominantly Chinese-owned), then exported as formal cobalt hydroxide. The provenance distinction is lost at this aggregation point.

Due-diligence implication

Battery Regulation Art. 48 requires identification of risks in the upstream — including ASM contamination. A trade-flow view that stops at DRC-to-importer is structurally incapable of meeting the OECD-aligned five-step framework.

Gap 5 · Latency

Trade data describes the past; supply risk moves in weeks

Even at its fastest, formal trade reporting lags the underlying flow by months. Critical-material risk events — export controls, sanctions, plant outages, project halts — materialise faster than any official series refreshes.

Eurostat COMEXT

Monthly bilateral trade, published ≈2–3 months after the reference month.

UN Comtrade

Country submissions take 3–6 months at best; some emerging-market reporters lag further. Cross-checked mirror data is later still.

USGS Mineral Commodity Summaries

Annual publication in early February, covering the prior calendar year.

China export controls

Gallium and germanium licensing (effective Aug 2023), graphite (effective Dec 2023), tungsten and antimony (effective 2024) — material price and availability impacts preceded full data visibility by 3–9 months.

Indonesia ore-export bans

2014 and 2020 nickel ore bans were predictable from policy signalling (legislation, ministerial statements) months before trade statistics confirmed the shift.

Trade data is necessary for trend confirmation. It is structurally unsuited to early-warning. The signal that an export restriction or supply disruption is coming sits in policy documents, capex announcements, and field-level intelligence — not in last quarter's import returns.

Gap 6 · Forward-looking signal

Emerging suppliers are not in the trade data — that's exactly what makes them emerging

A producer that already shows up in import statistics at meaningful volume is no longer an opportunity. The buyers who source from them already have established positions. Forward-looking supply intelligence lives upstream of trade data, not in it.

Argentina lithium

Brine-based projects (Olaroz, Cauchari-Olaroz, Sal de Vida) have been visible in financing, permitting, and offtake announcements for 3–8 years before reaching meaningful export tonnage.

Australia rare earths

Lynas Mt Weld and several junior heavy-rare-earth projects (Iluka Eneabba, Arafura, Australian Strategic Materials) were reportable from project pipeline data years before any non-China rare-earth export volumes mattered.

Mozambique graphite

Balama, Nicanda Hill, Caula projects each had multi-year lead times in geological and financing data prior to commercial production.

Indonesia mid-stream

Shift from ore export to processed nickel intermediate (MHP, matte) was visible in capex announcements and SOE policy 2–4 years before showing up in trade-flow data.

Trend-spotting in critical materials lives in capex pipelines, permit filings, financing rounds, and geological surveys — sources that have multi-year lead time over trade statistics. Predictive supplier intelligence is built from these signals, not retrospective import returns.