Anti-circumvention and third-country diversion detection

Anti-circumvention and third-country diversion detection uses trade, customs, financial, corporate and logistics data to identify transactions that may evade sanctions or export controls. It is an analytical technology category, not one official product or legal finding.

Detection methods

Authorities and firms compare commodity codes, quantities, values, routes, counterparties, addresses, payment chains and end-use statements. Sudden trade growth, unusual unit values, new intermediaries or a route through a hub can generate a lead. Trade and customs data analytics systems and Entity-resolution and graph-analysis software help connect records that use different names or identifiers.

The Common High Priority Items list identifies goods of concern in Russia-related diversion analysis. BIS also publishes diversion-risk letters, address-based Entity List controls and certification guidance. These sources guide screening and diligence; they do not make every transaction involving a listed item or hub country unlawful.

Evidentiary boundaries

A red flag supports inquiry, not a conclusion. Reexports can reflect legitimate redistribution, demand shifts, statistical reclassification or parallel supply chains. Nostro pre-funding and third-country clearing arrangements can obscure a payment path without establishing the destination of goods.

Export-control end-use and end-user verification technology can add documents, site checks, serial data and post-shipment evidence. Even then, a false document, inaccessible site or inconsistent address must be assessed under the governing rule and procedure.

Editors should separate a risk alert, transaction hold, licence requirement, administrative denial, customs seizure, enforcement settlement, charge and adjudicated violation. Each carries a different standard and attribution.

False positives and outcomes

Controls should use commodity specificity, materiality thresholds, temporal comparison and human review. Address normalisation and beneficial-ownership analysis reduce missed links but can also merge unrelated parties. Firms need escalation, correction and release procedures for lawful trade.

As at 30 July 2026, current authority guidance and item lists controlled the screening setting. Success should be measured through verified diversions prevented, investigations supported or compliance improved, not alert volume alone. Publication should state the dataset, period, anomaly, legal nexus, corroboration and outcome.

Data and model governance

Customs codes can be broad, revised or reported differently across jurisdictions. Mirror statistics may disagree because of valuation, timing, transshipment or country-of-origin rules. Analysts should preserve the original unit and flag discontinuities before calculating an anomaly.

Entity models need versioned identifiers and explainable links. Shared addresses, directors or telephone numbers can be meaningful, but can also arise from corporate-service providers or crowded commercial sites. Confidence levels and manual review should be recorded.

Financial data show value movement rather than physical end use. Shipping records show a route but may not identify the final consignee. Combining sources strengthens an inquiry only when dates, parties and goods align.

Operational controls should include feedback from cleared alerts and confirmed cases. Without it, a model can repeatedly burden lawful trade or miss adaptation. Performance measures should report precision, recall where estimable, review time and verified outcomes, while protecting investigative information and personal data.

Sources

1. US Bureau of Industry and Security, guidance on letters identifying transaction parties presenting diversion risk.
2. US Bureau of Industry and Security, Common High Priority Items certification guidance.
3. UK Government, Common High Priority Items guidance (accessed 30 July 2026).
4. European Commission, guidance for EU operators on Russia sanctions circumvention.