Is the ban leaking? Russian seafood after the Feb-2024 anti-wash clause

Trade-flow companion to the price wedge (R72). This is a DUAL-SCORE / alternative-track signal — never folded into any Tier-1 exposure score. The divergence between "who the controlling state says it stopped buying from" and "who suddenly started selling to the buyer" IS the signal. Research, not investment advice; origin-relabelling is INFERRED from statistical implausibility plus public records, never asserted as smuggling on any single shipment.

Verdict

Every other circumvention case in this corpus answers "the ban leaked, and here is the flag it leaked under." This one does not. Russian seafood is the corpus's cleanest negative result, and the reason is structural rather than lucky: the United States wrote the anti-transformation clause into the control text, in advance, instead of adjudicating it years later.

• The direct line died and stayed dead. US frozen-crab imports from Russia:

31,739 t (2021) → 11,872 t (2022) → 0 t in 2023, 2024 and 2025. Not "near-zero" — a literal zero on all three annual lines, on all four seafood HS codes pulled.

• The wash route was the *real* channel, and it was legal. A peer-reviewed

study of the pre-determination window found 96% of Russian seafood consumed in the US arrived from a country other than Russia, primarily China (Gephart et al., npj Ocean Sustainability, 27-Feb-2025). The 2022 ban on Russian-origin product therefore touched ~4% of the actual dependency.

• The Feb-2024 anti-wash clause bit — visibly, and only on the US line. US

imports of Alaska-pollock fillets from China: 26,446 t (2022) → 21,031 t (2023) → 14,081 t (2024) → 9,938 t (2025), −62% from the peak, with the break landing in the first year the clause was in force. Over the same span China's pollock-fillet exports to all destinations went up (174,674 t in 2021 → 194,078 t in 2024). Chinese reprocessing capacity did not shrink; the US share of it was legislated away.

• No phantom flag replaced it. The lost US volume was not re-flagged through

a zero-capacity origin — it largely stopped being imported at all (US pollock-fillet imports from the world: 28,817 t in 2022 → 12,174 t in 2025).

• The ownership layer returns N — and that is itself the finding. No

namesake-shell chain of the Youngsun → Thai Unipet → Youngsun & Essen type appears here, because until 21-Feb-2024 the route required no concealment. See "When the wash is lawful, the ownership tell disappears" below.

The control, in two stages

Sources: Executive Order 14068 (signed 11-Mar-2022; published Federal Register 15-Mar-2022, FR doc. 2022-05554); E.O. 14114 and the OFAC "Prohibitions Related to Imports of Certain Categories of Fish, Seafood, and Preparations Thereof" determination of 22-Dec-2023; OFAC General Licence 83.

Enforcement runs through an importer self-certification collected in ACE: the declaration must be filed for every entry of salmon, cod, pollock or crab regardless of the country of origin shown on the entry — the rule stops asking where the box was packed and starts asking where the fish was caught.

The direct line (US imports from Russia)

Source: UN Comtrade, reporter USA, annual, net weight; committed artifact `data/intelligence/seafood-russia-circumvention.json`.

Russia was a third of the US frozen-crab book in 2021. Two years later it was nothing. That part of the control worked immediately — and, by the study above, it was also the small part.

The input line (what China buys from Russia)

Mode-C cases are diagnosed on the input line, not the finished-good origin line. This one is unusually easy to read, because China reports it directly.

Source: UN Comtrade, reporter China, annual, net weight. China's 2025 annual lines were not yet published at pull time.

Two readings, and they point in opposite directions:

1. The feedstock line never closed. Russia still lands ~483 kt of pollock and ~94 kt of cod into China every year — more than thirty times the 14 kt US fillet line it once fed. The US did not shut the flow; it removed itself from the flow. The molecules went to the EU, the UK, Brazil, Korea and Chinese domestic plates instead. 2. Crab is the one line with a genuine surge. China's Russian-crab intake ran +866% from 2021 to 2023 (1,115 → 10,774 t) — precisely the window between the direct ban and the anti-wash clause, when a Chinese-processed Russian crab was still a lawful US import. It then fell back to 6,417 t in 2024, the year the clause bit. That is the shape of a route that opened, was used, and was shut.

The output line, and the diversion cross-check

The critical discriminator is whether the US line fell because China had less to sell or because the US stopped being allowed to buy.

Source: UN Comtrade — US-reported imports and China-reported exports, annual, net weight. China-reported 2025 not yet published.

China's total pollock-fillet output was +11% in 2024 versus 2021 while its US-bound volume was −42% over the same span. Brazil alone took +585%. There is no supply story here; there is a policy story. The clause did not destroy the wash — it redirected it away from the one buyer that outlawed it.

Crab: substitution, not relabelling

The 31.7 kt hole Russia left in the US frozen-crab book was filled by producers that genuinely have the resource, not by a zero-capacity flag:

Canada and Norway are Tier-1 crab nations with real Barents/North-Atlantic harvest — a surge from either is plausible, not a fingerprint. Non-Russian origins are +14.5 kt on 2021, so roughly 46% of the 31.7 kt Russia vacated was genuinely replaced; the remaining ~17 kt is demand destruction and price, and it never came back. The US frozen-crab line from China (2,970 t in 2025, the highest since 2019) is the residual to watch — it is small, it is flat, and post-Feb-2024 any Russian-harvested crab inside it is prohibited no matter where it was picked.

When the wash is lawful, the ownership tell disappears

The two-layer detector's second layer — a transit-country exporter sharing ownership, name-root or address with an entity in the controlled origin — returns nothing here. That is not a failure of the search; it is a property of the mode.

Between March 2022 and February 2024, routing Russian pollock through a Qingdao or Dalian plant and shipping it to the US as product of China was legal. Nobody needs a namesake shell to do a legal thing. The ownership fingerprint in the antimony, magnesium and plywood cases is a signature of evasion; it is absent wherever the transformation genuinely conferred origin under the rules in force. Layer 2 of the detector is mode-dependent, and its absence is informative rather than inconclusive.

What the corporate record does show is the other move available to a sanctioned principal — divestment rather than concealment:

> Gleb Sergeevich Frank — designated by OFAC under E.O. 14024 on > 24-Mar-2022 — had already resigned as chairman of the boards of Russian > Crab Group and Russian Fishery Company (RFC) on 17-Mar-2022, and his > 70.09% stake in RFC and 70.1% stake in Russian Crab were sold to five named > senior executives (Yevgeny Orlov, Anton Chertkov, Alexander Sapozhnikov, > Savely Karpukhin, Stanislav Aksenov).

Sources: OFAC SDN designation of FRANK, Gleb Sergeevich, 24-Mar-2022 (Russia designations action, E.O. 14024); Interfax, "U.S.-sanctioned Gleb Frank sells controlling stakes in Russian Fishery, Russian Crab Group, resigns as chairman". Reported as the public record of an ownership change — not an assertion that any subsequent shipment was unlawful.

The resignation preceded the designation by a week. The physical asset — quota, vessels, plants — did not move. This is the fisheries analogue of the divest-and-continue pattern, and it is why an entity-list-only exposure screen would have cleared RFC cargo from 25-Mar-2022 onward while the boats kept fishing. The species-and-harvest-water test in the Seafood Determination is immune to it; a beneficial-ownership test is not.

Mode C⁰ — the pre-emptively voided transformation wash

This case adds a fourth cell to the corpus's transformation-wash taxonomy, and it is the only one that is forward-looking:

The practical value is that C⁰ is detectable at announcement, not at autopsy. When a new control lands, the single highest-information sentence is whether it attaches the prohibition to a harvest/extraction fact (Russia-flagged vessel, Russian waters, wholly-or-in-part) or to a tariff heading. Attach it to the fact and the wash is priced out on day one; attach it to the heading and you are on the plywood clock, waiting years for a circumvention finding while the route runs.

Note the cost of C⁰, visible in the Brazil column: it protects the buyer's border without touching the producer's revenue. The molecules are not stranded, they are re-addressed. A C⁰ control is a decoupling instrument, not a pressure instrument — and reading it as pressure will misprice the target's cash flows.

Why it matters for the buyer

1. A "successful" control looks identical to a lost market. A risk team reading "US pollock-fillet imports −58% since 2022" could conclude a supply shock. There was none: Chinese output rose. The exposure moved from supply risk to substitution cost, and only the input line distinguishes them. 2. Screen the harvest, not the flag — and not just the entity list. The Frank divestment shows an ownership screen going green while the physical chain is unchanged. Any exposure model keyed on stated origin or on designated-party ownership alone is compromised for this control class. 3. The producer is doubling down, not retreating. The anchor action — Russia's Order No. 3959-r of 23-Dec-2025, RUB 1.3bn from the reserve fund to complete four crab-catching vessels, inside a 13-vessel / ~RUB 6.4bn investment-quota programme — is state capital being committed to crab capacity three years after the premium US market closed. That is a bet on the non-sanctioning demand side (China, Korea, Japan) absorbing it, and it is the producer-side confirmation that the flow was re-addressed rather than destroyed.

Method & honesty rails

• Trade data: UN Comtrade free public preview API, annual, net weight, HS

030614 / 160510 / 030475 / 030494 (US-reported imports), HS 030367 / 030363 / 030614 (China-reported imports), HS 030475 / 030614 / 160510 (China-reported exports). Fetcher: scripts/py/iptm/fetch_seafood_russia_circumvention.py; artifact: data/intelligence/seafood-russia-circumvention.json.

• Known data limits. China's 2025 annual lines were unpublished at pull time,

so every 2025 figure is US-reported. Mirror statistics between US and Chinese reporting differ; each table states its reporter. Net weight is product weight, not live-weight equivalent, and fillet weight is not comparable to headed-and-gutted feedstock weight — the input and output tables are read for direction and break-point, never subtracted from one another. No processing yield factor is applied anywhere in this case.

• A zero is a reported zero. The 0 t lines are absent/zero partner rows in

Comtrade, which is consistent with a total prohibition but does not by itself exclude misdeclared entries.

• Enforcement is attestation-based. The prohibition is policed by importer

self-certification, not by catch-document or genetic verification. Gephart et al. (2025) conclude existing traceability tools — SIMP covers roughly half of US seafood imports — are insufficient to verify harvest origin. The volume collapse documented here is consistent with the clause biting; trade data cannot rule out residual mislabelled entries, and the flat US–China crab line is exactly where such a residual would sit.

• Inference, not accusation. No shipment, company or person is asserted to

have violated the prohibition. The Frank/RFC ownership change is reported as published by OFAC and Interfax.

• Alternative-track only. This never touches buyerRelativeScore or the base

exposure — it sits beside them, exactly like the China–West price wedge.