Trade-Based Money Laundering (TBML)

The most prevalent TBML technique involves deliberately misrepresenting the price of goods on invoices to transfer value between parties.

• Over-invoicing: The importer pays more than the goods are worth, transferring excess value from the exporter to the importer.
• Under-invoicing: The exporter ships goods at a price below fair market value, allowing the importer to sell at market price and retain the difference.
• This manipulation requires collusion between buyer and seller and exploits the difficulty customs agencies face in verifying real-time global commodity pricing.

A single shipment of goods is invoiced multiple times using different financial institutions, allowing the same trade to justify numerous cross-border payments.

• Each invoice appears legitimate when viewed in isolation.
• The technique exploits the fragmentation of trade documentation across jurisdictions.
• Detection requires entity resolution to link disparate invoices to the same underlying shipment, often through container numbers or bill of lading cross-referencing.

Also known as ghost shipments, this technique involves creating complete documentation for goods that never physically exist or move.

• Fraudsters fabricate bills of lading, packing lists, and customs declarations.
• The absence of physical goods makes verification nearly impossible without port-level inspections.
• Often combined with shell corporations in high-secrecy jurisdictions to further obscure the paper trail.
• Machine learning models can detect phantom shipments by identifying anomalies in shipping route data and inconsistencies between declared cargo weight and vessel capacity.

The exporter ships fewer goods than declared on the invoice, with the importer accepting the shortfall and paying the full invoiced amount.

• The excess payment represents the transfer of criminal proceeds disguised as a legitimate trade loss.
• Common in high-volume, low-value commodities where individual item counts are difficult to verify.
• Detection requires statistical analysis of shipment weight-to-value ratios compared to industry benchmarks for the declared commodity class.

Goods are deliberately misdescribed in terms of quality or type to justify a price that does not reflect their true market value.

• Example: Used textiles declared as new designer garments.
• Example: Low-grade agricultural products invoiced as premium organic produce.
• This technique exploits the subjectivity of quality assessments and the limited inspection capacity of customs authorities.
• Anomaly detection algorithms can flag shipments where the declared value per unit significantly deviates from the statistical norm for that harmonized system code.

Criminals exploit the regulatory gaps and limited oversight within free trade zones to manipulate trade documentation and obscure the origin or destination of goods.

• Goods can be repackaged, relabeled, or commingled with legitimate shipments within the zone.
• Multiple layers of transshipment through FTZs break the audit trail and complicate chain-of-custody verification.
• The lack of harmonized data-sharing between zone operators and national customs agencies creates a significant visibility gap that TBML networks systematically exploit.

The second stage of money laundering where illicit proceeds are separated from their source through complex financial transactions. In TBML, layering occurs through multiple cross-border shipments, circular trade between shell companies, and repeated invoice manipulation across jurisdictions. Each trade transaction adds a layer of apparent legitimacy while obscuring the audit trail. Investigators trace these layers by reconstructing the transactional chain from origin to destination, identifying discrepancies in declared values at each step.

A legal entity with no significant assets or active business operations, used as a vehicle to obscure beneficial ownership in TBML schemes. Shell companies appear as legitimate importers or exporters on trade documentation but exist solely to fabricate transactions. Key red flags include:

• Registration in offshore secrecy jurisdictions
• No physical presence or employees
• Directors who serve on hundreds of similar entities
• Rapid dissolution after a few high-value transactions Entity resolution systems link these shells to their true controllers by analyzing shared addresses, phone numbers, and nominee directors across corporate registries.

The computational process of disambiguating and linking disparate data records that refer to the same real-world entity. In TBML investigations, entity resolution connects:

• Multiple shell companies to a single beneficial owner
• Trade documents to sanctioned vessels or ports
• Phone numbers and email addresses across seemingly unrelated importers
• Addresses shared by hundreds of registered entities Advanced systems use fuzzy matching, graph embeddings, and probabilistic record linkage to pierce corporate veils and reveal the hidden networks behind fraudulent trade documentation.

The technique of mapping and examining relationships between entities to identify hidden connections and collusion patterns in TBML. Analysts construct financial graphs where nodes represent companies, individuals, and vessels, while edges represent transactions, shared addresses, or familial ties. Key patterns detected include:

• Circular trade loops where goods move between related parties without economic purpose
• Hub-and-spoke structures with a central orchestrator controlling multiple shell importers
• Sudden dense clustering of new entities around a single port or commodity Graph neural networks automate this analysis at scale, flagging anomalous subgraphs for investigator review.

The final stage of money laundering where illicit funds re-enter the legitimate economy. In TBML, integration occurs when:

• Over-invoiced payments are deposited into seemingly legitimate business accounts
• Proceeds from undervalued exports are reinvested in real estate or luxury assets
• Phantom shipment payments are commingled with genuine business revenue The challenge for detection systems is distinguishing integrated criminal proceeds from legitimate trade income. Behavioral profiling and peer group analysis compare a company's declared trade activity against industry norms to identify integration anomalies.

The automated analysis of unstructured news and public data sources to identify negative information linking trade entities to financial crime. For TBML, screening engines scan:

• Shipping registries and port authority bulletins for seized contraband
• News articles about customs fraud investigations
• Regulatory enforcement actions against freight forwarders
• Sanctions advisories related to dual-use goods and strategic commodities Natural language processing models extract entities, classify risk sentiment, and link adverse mentions to existing customer profiles, providing critical context for transaction monitoring alerts.