Related News

MonetaGo sets sights on “trade finance OS” with beefed-up board

BB Energy Asia, Komgo and SGTraDex complete first digital borrowing base facility

US threatens anti-money laundering crackdown on offshore lenders

Going digital: Banks and authorities face shifting TBML threat

Aurubis alleges criminal activity after copper, precious metals invoices manipulated

Financial institutions are increasingly using network analytics and anomaly detection tools to identify potential collusion between parties in a trade transaction, as warnings grow over the misuse of trade finance for money laundering and fraud. 

Trade misinvoicing, where the stated value of goods being sold is higher or lower than in reality, has long been flagged as a means of moving illicit funds across international borders. 

More recently, researchers have warned that instruments such as standby letters of credit (LCs) are being exploited by criminals, who deliberately violate the terms of an agreement to trigger a payment from a bank that is underpinned by apparently legitimate documentation. 

Though these two methods are different, with trade misinvoicing typically used in open account transactions rather than documentary trade finance deals, they share one common factor: collusion between the buyer and seller. 

“We have seen that exact scenario,” says Alexon Bell, chief product officer and executive director at technology firm Quantexa, which aggregates data from corporate registries to identify networks of connected individuals and companies. 

“You have two separate companies – company A and company B – but whose directors are also directors of company C together. Then, companies A and B can over-invoice, under-invoice and ship goods, and neither party is going to complain,” he told GTR on the sidelines of last month’s Sibos event in Amsterdam. 

“In one case we identified this possibility, and a few months later they were flagged for suspected over-invoicing. Finding that social collusive link is very significant.” 

The same principle applies to trade transactions that are arranged primarily as a means of obtaining liquidity from banks, for instance through back-to-back or circular trades. Such structures are legitimate and widely used in the commodities market, but have increasingly attracted scrutiny after a flurry of fraud scandals in 2020. 

In some cases, Bell says, related companies – such as subsidiaries of larger corporates – set up circular transactions where they are “essentially trading with themselves”. 

“By doing this, they’re using trade finance as a kind of credit,” he says. “Pretty much universally, the banks say no to that, but if you’re not got the corporate registry data, how do you know those two subsidiaries are linked together?” 

Although identifying such networks is possible manually, technology is now able to consume entire corporate registries and link entities automatically, taking into account factors that may be missed such as differences in naming conventions. 

Patrick Craig, EY’s financial crime technology lead for Europe, the Middle East, India and Africa, says the “more advanced” financial institutions are already deploying such tools to ascertain whether seemingly unrelated counterparties may in fact be working together. 

“There are technology solutions out there that connect datasets, use entity analytics and network analytics to truly understand the counterparties that are transacting, and whether there is counterparty risk,” he told GTR at Sibos. 

“[Understanding] networks brings in all sorts of views around the context of transactional activity, and you could look at collusive activities.” 

Mapping potential collusion cannot prevent all types of trade-based financial crime, however. Banks are also using technology to identify potential anomalies in transactions, where activity undertaken does not fit an entity’s historic trading patterns. 

Sometimes, unusual activity will be obvious, for example if a trader starts handling goods that are outside of its typical profile – a steel exporter suddenly selling sugar, for example – says David Cooperman, executive director for business consulting, financial services, at EY. 

But often, changes in activity will be more subtle – for example whether goods move through higher risk ports than usual, or where quantities and values are outside established patterns. 

“There’s no way the human mind can detect that,” Cooperman tells GTR. “We can catch the big things, but what about if there’s a subtle pattern shift? A machine can look at every single transaction that a bank has done between two parties, and see patterns that a human being will never see.” 

Identifying broader patterns of behaviour can also help banks keep up with changing techniques used by fraudsters and money launderers. 

“The agility of criminals – how quickly new criminal techniques emerge – puts banks in a reactive cat-and-mouse game with a counterpart that has almost endless possibilities available,” says Sjoerd Slot, chief executive of Netherlands-based technology provider Sygno. “Criminals will also return to previous techniques just as easily, based on whatever works.” 

Slot tells GTR that technology such as machine learning “can detect suspicious activity without knowing the actual modus operandi, just by modelling around what behaviour would be expected in any situation”.