Trade-based money laundering detection is getting tougher. Focusing solely on shell companies overlooks schemes exploiting legitimate supply chains, leaving your compliance team scanning for the wrong patterns while real risks flow through established business relationships.
Evolving TBML Schemes
Financial institutions have traditionally focused on identifying shell companies as the key to detecting trade-based money laundering (TBML). This approach worked when launderers needed corporate vehicles to hide beneficial ownership. However, recent TBML schemes show a shift: launderers are increasingly using legitimate trading relationships through overseas branches and subsidiaries of multinational companies, where the corporate structure itself isn't suspicious.
The gap between detection capability and scheme sophistication is widening. Your transaction monitoring rules may flag newly formed entities with rapid growth, but they're not set up to catch established companies that pivot sectors or develop complex supply chains without clear commercial reasons.
Key Findings
Legitimate corporate structures facilitate TBML. Investigations have identified multinational companies whose overseas branches developed trading relationships to distribute goods into new markets as part of laundering operations. The corporate entity itself passes compliance checks; the risk lies in the trading pattern.
Unexpected risk indicators. Companies may show rapid growth into existing markets, make significant cash payments to unknown third parties, or receive unexplained payments. The most telling indicator: unnecessarily complicated supply chains with multiple transshipments and no clear commercial justification. For example, an IT company suddenly starts bulk pharmaceutical distribution.
Shell companies are not universal. While shell companies are used to hide beneficial owners and move cash across borders, not all TBML schemes rely on them, especially those exploiting legitimate supply chains.
False positives lead to detection paralysis. General trading companies may trade in multiple commodities, and companies may pivot sectors for valid reasons. Without context, these indicators generate alerts your team can't act on. The challenge is distinguishing genuine risk from normal business evolution.
Current methods miss pattern combinations. A company involved in unrelated sectors might trigger a review. But what about a company showing three indicators at 70% confidence each? Traditional systems evaluate each indicator independently, missing the compounding risk signal.
Implications for Your Team
Your transaction monitoring system was built to catch obvious anomalies like sudden spikes and known typologies. It wasn't designed to recognize that an established electronics distributor's new pharmaceutical supply chain, combined with third-party payments from unrelated jurisdictions, represents TBML risk.
You're facing a detection gap that rule refinement won't close. Adding more indicators increases false positives without improving accuracy. The problem isn't rule sensitivity; it's that TBML schemes now present as combinations of legitimate business activities, each defensible on its own.
Machine learning models can evaluate these indicator combinations, but only if trained on the right patterns. A model that learns "shell company + cash payments + rapid growth = risk" won't recognize "established company + sector pivot + complex supply chain = risk" unless you explicitly train it on that pattern.
Action Steps
Map your detection coverage. Review cases from the past 18 months where you identified TBML risk. Separate them into those involving shell companies and those exploiting legitimate structures. Calculate the ratio. If shell-company cases dominate but don't represent the full threat landscape, you have a coverage gap.
Build a sector-pivot detection layer. Query your database for companies that changed primary business activities in the past 24 months. Cross-reference against payment patterns, particularly third-party payments and cash activity. This manual review will help you understand how sector pivots present in your data before automating detection.
Develop supply chain complexity scoring. For customers engaged in trade finance or international payments, map the number of jurisdictions, intermediaries, and transshipment points involved in typical transactions. Establish baselines by customer segment and industry. Alert when a customer's supply chain complexity increases significantly without corresponding business growth or new market entry.
Train your SAR writers on non-shell TBML narratives. Your Suspicious Activity Reports currently describe shell company structures and cash-intensive businesses. Add narrative templates for legitimate companies showing indicator combinations: sector pivots, supply chain complexity, unexplained third-party payments. Include specific details about why the commercial rationale doesn't support the observed pattern.
Implement relationship-based transaction monitoring. Your current system monitors accounts independently. TBML schemes often involve coordinated activity across multiple accounts, particularly when overseas branches or subsidiaries participate. Build queries that evaluate transaction patterns across related entities, looking for payment flows that don't match stated business relationships.
Test machine learning models on historical false negatives. If you've identified TBML schemes through non-systematic means (examiner findings, law enforcement referrals, customer complaints), use those cases as training data. A model that learns from your actual misses will outperform one trained on generic TBML typologies.
Don't wait for regulatory guidance to mandate these capabilities. By the time TBML detection standards formalize expectations around pattern recognition and relationship monitoring, you'll already be behind.
