Kampala's gold-trade cases: pricing counterparty risk

A matter listed at Buganda Road Chief Magistrate's Court on 31 July 2026 concerns an alleged $1.5 million payment, roughly Shs5.6 billion, made against a promised 16-tonne gold consignment to Dubai that court documents say never shipped. A separate charge concerns nuggets held without a dealer's licence. The accused has denied both charges, remand was granted while inquiries continue, and nothing has been proven; the presumption of innocence applies in full.

We are not covering this as a crime narrative, and we take no view on guilt. Our interest is narrower and, for anyone deploying capital into East African metal, more useful: at what point in that transaction chain did verification fail, and which of those steps is cheap to fix before money moves.

Ugandan authorities have confirmed that a group of private firms active in the mineral trade is under inquiry, with several matters before the courts and trading licences under review. The composition of that group is the analytically interesting part. It is dominated by service companies: logistics agents, security providers, self-described processors and general trading vehicles. Producers are scarce.

That distribution tells you where the risk actually sits. An entity with no ground, no plant and no reserve statement cannot deliver metal from production, whatever its documentation implies. Establishing which of those three a counterparty has is a one-afternoon exercise, and it screens out most of the exposure before commercial terms are ever discussed.

In a legitimate offtake, the buyer's money moves after metal is independently weighed and assayed, usually against documentary credit. Where the sequence inverts, and cash is required for storage, escort, clearance or a tax deposit ahead of any verified metal, the economics of the transaction have already changed: the buyer is funding the counterparty's working capital against an asset nobody has confirmed exists.

Scale is what makes the inversion plausible. Regional export volumes are genuinely large, and a headline consignment therefore does not read as absurd to a first-time buyer who has only seen the trade statistics. Familiarity with the numbers is not familiarity with the paperwork, and the two are routinely confused.

International counterparties price jurisdictions rather than nuance. Each proceeding of this kind widens the diligence budget, lengthens the time needed to open a banking relationship and shifts the opening posture of offtake talks toward suspicion, for every operator in the belt regardless of conduct. That is a genuine cost-of-capital effect and it is borne collectively.

Which is why documentation has become an asset rather than an overhead. An operator able to evidence a live licence, surveyed ground, installed plant and a custody trail is separated from the disputed cases by records rather than by assertion. Our analysis of LBMA responsible sourcing sets out why accredited refiners now make that evidence a condition of purchase.

One: confirm the instrument at source with the Directorate of Geological Survey and Mines, and note which instrument it is. Prospecting, exploration, mining, dealer and export licences are separate grants, and holding one implies nothing about the others. Two: appoint your own laboratory, at a facility you nominate, with your representative present at weighing and sampling.

Three: refuse any pre-verification payment, whatever it is called. Four: establish whether the counterparty produces metal at all, or merely handles it. Five: assume no state body stands behind a private mineral sale. Neither the central bank nor the mines ministry acts as escrow agent or guarantor, whatever an intermediary suggests.

Arrest counts are a poor indicator. The outputs that would change how we underwrite Ugandan counterparties are narrower: licences actually cancelled, cancellations published, and convictions recorded. Earlier announced investigations have not produced a public record of revocations at scale, which is why the market discount has not moved.

The cheapest available remedy is also the most testable one: a searchable public register of current dealer, export and mining licence holders, updated at source. If that appears, verification stops depending on private introductions and the jurisdictional discount should narrow. We will revisit this analysis against those outputs rather than against further arrests.