Did Namibia's raw-lithium ban hold?

Alternative-track (R72) — never folded into any Tier-1 exposure score. Research, not investment advice. Origin relabelling is inferred from statistical implausibility plus public reporting, never asserted for any single shipment.

Verdict

• On the direct line, yes. China's recorded imports of HS 253090 (the

basket that carries crushed lithium ore / spodumene) from Namibia went 105,570 t (2023) → 13,684 t (2024) → 638 t (2025): −99%. The 2023 peak is pre-ban shipments plus a ministerial exemption window (30,000 t permit to Xinfeng, Nov–Dec 2023), not a leak.

• The exemption, not a transit flag, was the pressure valve. The ban

(Cabinet, 6-Jun-2023) is ministerial, not statutory, and allows "small quantities" at the Minister's discretion. Five months in, the Ministry of Mines and Energy let the one company whose exports triggered the ban ship 30,000 t of crushed ore through Walvis Bay.

• **Regional volume-implausibility fingerprint: present, but not attributable

to Namibia. South Africa has no lithium mine production (USGS MCS 2025 lists none) and no concentrator (CRU, quoted in Oxpeckers). Yet South-Africa-labelled HS 253090 into China went 63,500 t (2022) → 972,832 t (2024), taking it from 1.5% to 11.9% of China's volume in the basket. Public reporting ties that surge to Zimbabwean** ore on the Beitbridge→Durban road, not to Namibian ore. We cannot apportion it.

• Common-ownership tell: Y (consignee-affiliate), registry not traced.

Namibia's own customs books its 253090 exports to Hong Kong (69,513 t 2022, 50,889 t 2023), while China books the same flow as Namibia-origin. This split is consistent with the documented consignee, Tangshan Xinfeng HongKong Limited, a presumed affiliate of the exporter Xinfeng Investments (Global Witness). Export-declared unit values are 2–5× below China's import-declared values.

Table 1 — the banned origin, both mirrors (HS 253090, annual)

Source: UN Comtrade public API (preview), reporter China (156) imports and reporter Namibia (516) exports, HS 253090, max-per-corridor (no summing across partner2/mode/customs rows). Namibia had not reported 2025 at fetch time.

How to read it:

• Partner split. Namibia records the flow as going to Hong Kong, and China

records it as arriving from Namibia. That is ordinary consignee-vs-origin bookkeeping, but here it lines up with a named Hong Kong affiliate consignee.

• Valuation gap. In 2022, 69.5 kt left Namibia at $97/t and 72.8 kt

arrived in China at $519/t. A 5× CIF/FOB gap is far too large to be freight. It is consistent with under-declaration at export, which is a royalty and transfer-pricing question rather than circumvention of the ban. Flagged, not scored.

• 2023 volume gap. China recorded ~2× what Namibia recorded (105.6 kt vs

50.9 kt). Part of this is shipping lag (late-2022 loadings landing in 2023). The rest is an open question, not evidence of anything.

• 2024 shift. The Namibia-side unit value jumps to $855/t, which is consistent

with a move from crushed ore to concentrate. The ban exempts processed material, so that is the intended outcome.

Table 2 — the implausible surger (HS 253090 into China)

Same source, reporter China (156) imports; South Africa's own export mirror (reporter 710) in the last column.

• Why this is implausible: South Africa has no lithium mine output (USGS

MCS 2025 world-production table: none listed; "other countries" = —). Namibia itself is USGS e2,700 t Li. Take a deliberately low 1% Li₂O grade, under half of CRU's reported 2–2.5%: 954,385 t of ore then carries ≈4,400 t Li. That is more than Namibia's entire estimated mine output, labelled "South Africa".

• Unit values ($155/t China-side, $62/t SA-side in 2024) fit low-grade

crushed ore in a collapsed price year, not concentrate.

• Corroboration: CRU (Cameron Hughes, via Oxpeckers, 7-Apr-2025) says

"about 147,000 tons of ore was shipped to China" from South Africa in September 2024 at "2% to 2.5% lithium content": "We are still unclear where all of this came from."

• 2025 mirror divergence: South Africa reports 1.02 Mt to China, but China

records only 181 kt from South Africa. One hypothesis is that China began re-assigning origin (to Zimbabwe?) or that its 2025 data is partial. This is an open question, not an interpretation we rely on.

Ownership chain (Namibia leg)

This matches the Youngsun → Thai Unipet pattern from the antimony case in shape: a namesake affiliate sits between the controlled producer and the buyer. Here, though, the affiliate is the consignee, not a transit exporter, so it explains the partner split in Table 1. It does not show any post-ban leakage.

Transmission chain

Uis pegmatite (Namibia) → Long Fire claims → Xinfeng crushing → Walvis Bay → [paper consignee: Tangshan Xinfeng HongKong Ltd] → Caofeidian → Tangshan Xinfeng Lithium Industry (Hebei) conversion.

After the ban, this line is severed on the record. The regional ore that still reaches Chinese converters arrives increasingly under a South African flag that has no mine behind it.

What the Tier-1 score misses

• A buyer screening lithium feedstock by declared origin sees "South Africa". That

looks like a non-sanctioned, non-Chinese, investment-grade jurisdiction with no ban. The fingerprint says the ore is Southern-African pegmatite (most plausibly Zimbabwean) routed around two producer-side ore bans.

• Producer-side ban geometry (cf. cobalt, nickel, bauxite, and the Zimbabwe

lithium case). The risk owner is the converter and the royalty-collecting state, not a Western buyer being cut off. The practical exposure is provenance and due diligence (EU Batteries Reg. supply-chain due diligence, Art. 24 CRMA), not supply loss.

• Cross-reference: the Zimbabwe case

(2026-lithium-zimbabwe-concentrate-export-ban-circumvention-transshipment.md) concludes "no phantom flag — bulk concentrate can't re-flag". That holds for concentrate. For crushed ore in 2023–2024, Table 2 shows a phantom flag of ~1 Mt/yr. That case's reading should be qualified accordingly.

Caveats

• Basket pollution: HS 253090 is "mineral substances n.e.s.", not lithium

alone. We do not have China's 8-digit lithium-ore split. The SA line's lithium identity rests on (a) timing, (b) ore-grade unit values, and (c) CRU's statement, not on a lithium-specific code.

• Inference, not proof: individual South-African-flagged shipments may be

entirely legitimate, for example genuine SA exploration material or declared transit misrecorded as export.

• Attribution: no public source ties the SA surger to Namibian ore. The

two bans overlap (Zimbabwe raw-ore ban Dec-2022 under SI 213/2022; Namibia Jun-2023), and geography (Beitbridge→Durban) and reporting both point to Zimbabwe.

• Data lags: Namibia 2025 is unreported, China 2025 is provisional, and

monthly onset timing could not be pulled (Comtrade rate limit at fetch). This is a labelled gap.

• Ownership: every hop is press/NGO-sourced. Namibia BIPA and the Hong Kong

Companies Registry were not queried, so this case stays GATE 0.

Sources

• UN Comtrade public API, HS 253090, reporters 156 / 516 / 710, annual

2019–2025 (fetched 2026-09-25).

• USGS, Mineral Commodity Summaries 2025 — Lithium (world mine production

table): https://pubs.usgs.gov/periodicals/mcs2025/mcs2025-lithium.pdf

• Global Witness, "The lithium rush in Africa" (14-Nov-2023):

https://www.globalwitness.org/en/campaigns/natural-resource-governance/lithium-rush-africa/

• African Mining Market, "Xinfeng receives green light to export 6 000 tonnes

of unprocessed lithium ore" (6-Nov-2023): https://africanminingmarket.com/xinfeng-receives-green-light-to-export-6-000-tonnes-of-unprocessed-lithium-ore/17177/

• VOA, "Namibia restores Chinese company's export permit despite bribery probe"

(16-Nov-2022): https://www.voanews.com/a/namibia-restores-chinese-company-s-export-permit-despite-bribery-probe-/6837427.html

• Oxpeckers, "On the trail of lithium smugglers in Southern Africa"

(7-Apr-2025): https://oxpeckers.org/2025/04/lithium-smugglers/

• Anchor action: docs/iptm/actions/2023-06-06-namibia-critical-minerals-export-ban.md