Sanctions screening is not just checking a name against a list. It is a complex mix of regulatory rules, technology, and human judgment. Trade finance deals with documents, dual-use goods, and multiple parties, which makes the job even harder.
Miss one detail, and a bank can face huge fines. This guide breaks down the core components of an effective screening program.
Trade finance involves checking parties, vessels, and goods against dynamic sanctions lists. It is not a one-click job.
Automated systems help, but human analysis is needed to understand complex ownership structures and trade patterns.
| Regime | Issuing Body | Primary Focus | Extraterritorial Reach |
|---|---|---|---|
| OFAC Sanctions | US Treasury | Country-specific & List-based (SDN) | High (USD clearing) |
| EU Restrictive Measures | European Council | Targeted asset freezes & sectoral bans | Limited (EU persons) |
| UN Sanctions | UN Security Council | Global peace & security threats | Global (Member states) |
| UK Sanctions | OFSI (UK Treasury) | Post-Brexit autonomous listings | Moderate (UK persons) |
A transaction might be legal in one country but illegal in another. Banks often apply the strictest standard to avoid risk.
A German exporter ships machine parts to Dubai. The end-user is in Iran. The EU might allow it, but the US bank handling the dollar payment blocks it immediately.
The Screening Process Flow
Screening happens at multiple touchpoints in the trade lifecycle. It starts when a letter of credit is issued and doesn't stop until the vessel arrives.
You can't just screen the applicant. You must screen the beneficiary, the shipping company, the notify party, and even the crew.
| Stage | Data Screened | Method | Risk Window |
|---|---|---|---|
| Advising/Issuing | Applicant & Beneficiary names | Automated fuzzy matching | Onboarding |
| Document Presentation | Bill of Lading (shipper, vessel) | Manual & Automated | High |
| Settlement | Intermediary banks | SWIFT GPI screening | Medium |
| Post-Financing | Goods end-use & transshipment | Periodic review | Continuous |
Bill of Lading checks are critical. You must verify that the vessel is not blacklisted and that it won't pass through a sanctioned port.
A bank approved a shipment on the vessel “Blue Star.” A week later, the bank finds out the vessel changed its name to avoid sanctions. The ship was seized, and the goods were lost.
Common Red Flags in Trade
Criminals use trade to move money and goods illegally. They exploit the complexity of documents. The Financial Action Task Force (FATF) has listed several warning signs.
Some red flags are obvious, like a shipment to North Korea. Others are subtle, involving complex corporate structures.
Double invoicing, over/under-shipment, and the use of shell companies in free trade zones are major risks.
Discrepancies between the goods described and the actual container weight raise immediate suspicion.
| Red Flag Category | Specific Indicator | Why It Matters | Response Protocol |
|---|---|---|---|
| Entity Structure | Complex shell company chains | Hides Ultimate Beneficial Owners (UBOs) | Enhanced Due Diligence |
| Routing | Transshipment to high-risk jurisdictions | Goods may be diverted to sanctioned areas | Stop Shipment / Escalate |
| Goods | Dual-use or military components | WMD proliferation risk | Export License Check |
| Financial | Third-party payments unrelated to trade | Sanctions evasion or TBML | Block funds / Report |
If a customer asks to change the destination port at the last minute, that is a huge warning sign. Always check the end-use of the goods carefully.
A food trading company suddenly starts importing high-tech refrigeration units. The price is three times the market rate. This is a classic case of value transfer to a sanctioned individual.
The Technology Landscape
Manual screening is impossible today. Banks use advanced regtech software that scans thousands of lists. But technology alone is not enough.
Fuzzy logic helps catch misspelled names. For example, “Kaddafi” vs. “Gaddafi.” Without this, you miss hits.
| Task | Technology (AI/Robotics) | Human Analyst | Optimal Balance |
|---|---|---|---|
| List Filtering | 100% automated | 0% | Machine only |
| False Positive Review | Pre-discounting (70%) | Final decision (30%) | Mixed |
| Ownership Research | Graph analytics (40%) | Source verification (60%) | Human-led |
| Adverse Media | Web crawling (80%) | Relevance check (20%) | Machine-led |
The biggest problem is alert fatigue. If the system generates 5,000 alerts a day, an analyst might miss the real threat.
An analyst ignored a true match because the system had generated ten false alarms for “Mr. Smith” earlier that morning. The real Mr. Smith was a designated terrorist.
If the threshold is too low, you drown in alerts. If it is too high, you miss matches.
Regular tuning of fuzzy matching (e.g., 85% similarity vs 95%) is essential for an effective program.
Ownership and Control Challenges
The “50% Rule” is a classic trap. OFAC says if a sanctioned person owns 50% or more of a company, that company is blocked. But what about 49%?
Aggregation is also key. If one blocked person owns 25% and another blocked person owns 25%, the combined control by sanctioned individuals reaches 50%.
A bank financed a company in Malta. No single sanctioned shareholder owned 50%. But three cousins on the Specially Designated Nationals (SDN) list each owned 17%. The bank missed the aggregate rule and violated sanctions.
Key Takeaways
| Key Point | What It Means | Action Item |
|---|---|---|
| Strictest Regime Applies | Overlapping global rules create compliance traps | Always screen against US, EU, UK, and UN lists |
| Documentary Focus | The Bill of Lading is the prime document | Screen vessel IMOs and port call histories |
| Aggregate Ownership | Multiple small stakes equal a blocked entity | Calculate total sanctioned ownership percentages |
| End-Use Verification | Goods can be diverted to restricted parties | Investigate when goods don't match the buyer's business |
| Human overrides tech | AI reduces noise, but cannot interpret context | Never fully automate the release of a positive match |