Is the ban leaking? Palm oil, Indonesia's DSI state-export monopoly, and the co-producer blind spot

Trade-flow companion to the price wedge (R72), first EDIBLE-OIL / food-security entry after phosphate and nitrogen fertiliser. This is a DUAL-SCORE / alternative-track signal — never folded into any Tier-1 exposure score. It is also a case the corpus did not yet hold: a control that does not ban or licence the molecule at all, but instead inserts the state as sole legal exporter — a state-trading-enterprise (STE) monopoly that circumvention theory must read differently from every ban/licence row. Research, not investment advice; origin substitution is INFERRED from trade structure and public ownership records, never asserted as illicit re-labelling on any single shipment.

Verdict

On 20 May 2026 President Prabowo signed PP No. 24 Tahun 2026, making PT Danantara Sumber Daya Indonesia (DSI) — a wholly-owned arm of the Danantara sovereign fund — the sole legal exporter ("eksportir tunggal") of Indonesian palm oil (CPO and derivatives), coal, and ferroalloys, with authority to set the export selling price. Transition (mandatory DSI single-window routing) began 1 June 2026; full mandatory channelling lands 1 January 2027 (Sekretariat Kabinet RI; PP 24/2026 gazette, peraturan.bpk.go.id/Details/349945).

Indonesia is the dominant single origin of the world's most-traded vegetable oil — ~48% of world palm-oil export value in 2024 (~US$20 bn), rising toward ~58–62% of physical supply, producing roughly 3× the #2 producer (Statista; worldstopexports; USDA FAS). That is exactly the single-origin dominance that makes an origin worth laundering. But palm oil breaks the antimony template in a way that defines this case:

Sources: USDA FAS oilseeds; MPOB "Overview of the Malaysian Oil Palm Industry 2024/2025"; worldstopexports 2024; Statista.

The would-be transit surger is itself the world's #2 producer. A rise in Malaysian-flagged refined palm oil trips no naïve implausibility alarm — Malaysia genuinely grows and refines palm at giant scale. This is the co-producer blind spot first mapped on Belarusian potash (Russia = #2 potash) and DRC/Rwanda coltan (Rwanda = #2 tantalum): when the neighbour that absorbs the squeezed origin's volume is a co-dominant producer of the same commodity, the volume-implausibility detector is structurally defeated. The fingerprint has to shift from "you can't produce this" to "you can't produce it from your own trees" — the feedstock test (the mode-C signature seen on Russian crude, diamonds, and gold).

The pre-armed leak: Indonesian CPO refined and re-flagged in Malaysia

Malaysia already runs the escape valve, legally and in the open. Its refiners import Indonesian crude palm oil (CPO), refine it, and re-export it as Malaysian processed palm (RBD olein/stearin, oleochemicals) — genuine substantial-transformation that lawfully confers Malaysian processing origin. The flow is real and surging:

Sources: OEC (oec.world) Malaysia palm-oil bilateral trade; MPOB monthly import series; reported in TradeImeX / Vespertool market notes.

Malaysia's own policy pulls this CPO in: the tiered CPO export duty (8.5–10% through 2025, 9% in Aug-2025) is explicitly designed to keep crude oil inside Malaysia for downstream refining (BERNAMA, Budget-2025 export-tax note; ainvest). Indonesian CPO crossing the Strait to feed Malaysian refineries is the same economic logic that fills a mode-C wash — only here both ends are legitimate producers, so the transformed product carries a clean Malaysian flag with an Indonesian core.

The corporate pipe (common-ownership tell — disclosed, not a shell)

The strongest structural tell is that the same vertically-integrated groups own plantations and mills in Indonesia AND refineries in Malaysia — so Indonesian-grown oil can move to a group-owned Malaysian refinery and re-export under the group's own roof, entirely within public corporate structure:

> Wilmar International (Singapore-listed; refineries in both Sumatra/Kalimantan > and Peninsular/East Malaysia), Musim Mas (Indonesian-origin group, refineries > in Indonesia and Pasir Gudang, Malaysia), plus Malaysian majors IOI, KLK and > SD Guthrie holding large Indonesian plantation bases.

Same groups, both flags. This is the disclosed analogue of the antimony Youngsun → Thai Unipet → Youngsun & Essen pipe: a corporate structure that makes origin re-flagging a matter of internal logistics rather than illicit trans-shipment — closer to Nornickel's Zug trading arm in the palladium case (legal, disclosed vertical integration) than to a namesake laundering shell. Entity footprints: companies' own public corporate materials / annual reports; Wikipedia/Down-to-Earth profiles. Presented as the disclosed mechanism, NOT a per-shipment accusation.

Why the DSI control changes the read (the new signature)

Every other corpus row is a control someone routes around. DSI is the inverse: the state captures the export channel itself, and its two design features cut against origin-transparency for a Western risk desk:

1. Price-and-channel opacity. DSI negotiates the contract, books the shipment, receives payment, and sets the export price (PP 24/2026, Pasal 3). Direct price discovery on Indonesian CPO — the world's marginal palm barrel — disappears behind a single state counterparty, widening the incentive to source the same oil through the transparent, duty-managed Malaysian refining channel. 2. A monopoly is the classic under-invoicing trigger. The stated rationale is to close under-invoicing and transfer-pricing capital flight — an admission that mis-declaration on the Indonesian export leg is the baseline problem the state is trying to internalise. A single mandated intermediary with price-setting power is itself a WTO Article XVII (state-trading-enterprise) exposure and a potential Article XI quantitative-restriction vector (per the anchor action).

So the honest verdict is not "a ban is leaking." It is: a novel opacity-control sits on top of a co-producer duopoly whose leak channel is already built, legal, and commonly-owned. The customs data will keep showing legitimate Malaysian palm; the Indonesian core inside it becomes harder to see precisely when the origin state stops publishing transparent direct-export terms.

Why it matters for the buyer

1. Origin data understates Indonesian dependence. A risk team reading "we source refined palm from Malaysia" is partly sourcing Indonesian CPO with a Malaysian processing flag — the feedstock, not the flag, carries the deforestation/EUDR and single-origin-concentration risk. 2. The detector is structurally blind here. No ~0%-capacity phantom flag will appear, because the absorber is a genuine co-producer. The signal lives in the feedstock ledger (Malaysian refined-export volume vs Malaysian own-CPO output, the gap filled by Indonesian imports) and in price-channel opacity, not in an implausible-origin table. 3. It is a leading indicator of a squeeze that has not fired yet. DSI does not restrict volume today; it channels it. But the same architecture that captures the channel can throttle it (Indonesia has done short palm-export bans before, Apr-2022). The moment it does, the cross-Strait refining channel is the pre-positioned escape valve — and the corpus should already be watching it.

Method & honesty rails

• Trade data: USDA FAS oilseeds; MPOB monthly import/export series (HS 1511

palm oil, 1513 palm-kernel); OEC bilateral Malaysia–Indonesia palm-oil trade. Physical vs value shares differ by product mix (CPO vs refined vs oleochemicals); both are shown and labelled.

• Alternative-track only: never touches buyerRelativeScore or the base

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

• Inference, not accusation: the Malaysian refining/re-export flow is **legal

substantial transformation. Nothing here asserts illicit relabelling; the common-ownership structure is disclosed** vertical integration, not a shell.

• Why GATE 0: (a) DSI's binding stage is prospective (full 1-Jan-2027), so

there is no post-control flow table yet — the anticipatory-window problem last seen on SA chromium; (b) the absorber is a genuine co-producer, so the volume-implausibility test is defeated by construction (the potash/coltan co-producer blind spot); (c) the ownership tell is real but disclosed and legal (the palladium-Zug pattern), not a laundering pipe. All figures trace to named free public datasets; the circumvention fingerprint is not yet provable — it is a mapped blind spot and a new control mode, liftable to GATE 1 only if post-2027 data shows Malaysian refined-palm exports rising on Indonesian CPO imports while DSI opacifies the direct line.

Generalises (R72): the co-producer blind spot now spans three axes — fertiliser (potash), conflict-mineral (coltan), and edible-oil (palm) — and the STE sole-exporter is a fourth control mode (after ban, licence, and producer-side quota): a control that manufactures origin opacity rather than a flow to route around.