TBML’s unsolvable problem?

Occasionally the opaque origin of trade becomes suddenly and jarringly public. In 2019 Tesla’s use of cobalt from mines that employed child labor was one example. Another was 2020’s alleged fraud from Singapore-based commodities firm Agritrade that left banks with more than $600 million in liabilities.

Why doesn’t trade-based money laundering (TBML) ever seem to stop, and only get worse each year?

There are lots of contributing factors, but one in particular pervades them all. Even as we deal largely in digitized information today, the real world remains analogue.

Regulators have made some important inroads on this seemingly unsolvable problem. Much of the focus has been on data collection. These efforts have yielded important gains as both data sharing and analysis have improved. And where data cannot be shared freely, collaborative consultation models improve insight.

Existing technology failed the Covid stress-test

When the Covid-19 pandemic hit, the financial sector had to adjust on the fly. In a March 2020 publication, R3 along with the ICC and others took stock of some of the ad hoc measures that banks were implementing just to keep trade moving. This included taking pictures of signatures and emailing copies of documents that would normally need to be in original formats. Banks just didn’t have the right technology to scale digital solutions quickly.

Scammers took advantage of all of this. They exploited the confluence of an explosion in demand for medical equipment with the difficulty of tracking cross-border transactions. The result was some fancy new schemes for money laundering and fraud. FATF has gone into great detail about these new types of TBML techniques. They range from abuse of economic stimulus measures to counterfeiting and non-delivery of PPE.

Even though 2020’s fines continued to climb higher than previous years, and the technological tools were better than ever, the estimated detection rate remained only around 0.1%.

The human limits of comprehension (or, too much data too quickly)

The reason TBML technology hasn’t made more progress is because the technologies we use today were built to solve a very specific problem: how to cope with lots of data.

Data is a tricky problem in TBML. Specifically, there is both never enough and at the same time, far too much. Since the advent of big data in the mid-2000s, the volume of information has increased dramatically. But more data requires technology to sort, clean and analyze it.

This problem is by no means specific to financial crime. In studies of how physicians cope with information overload, research has shown that technology is a critical tool. It enables doctors to absorb the vast and constant flood of information about new medications and novel uses of existing medications.

The technologies available to combat TBML today are certainly at the cutting edge of data analysis — Artificial Intelligence (AI), Optical Character Recognition (OCR) and Machine Learning. And they have all made important inroads into fighting TBML. OCR scans through unstructured data, adding to the volume and depth of big data, while AI and Machine Learning analyze it. What all of these have in common is that these tools execute only once data is gathered and combined.

These examples illustrate both the advances and the limitations of technology in a digital world. They have improved the processing of data that has already come into the system. The limitation is that in order to gain access to more data, analysts have to shift to efforts focused on changes in policy and external requests for assistance.

Blockchain has changed trade finance, can it also change TBML?

This is where blockchain offers a unique approach to innovation. Rather than improving existing data, it enables connectivity across a greater variety of reporting sources. This is a basic, but dramatic difference in data collection. Also, in combination with existing technologies, it can solve some of the problems that continue to dog the industry.

There are already applications in the AML blockchain space that take advantage of this to move collaboration forward technologically with blockchain. MonetaGo’s fraud mitigation network is aimed at reducing fraud in receivables by verifying the authenticity of waybills and invoices. NICE Actimize talked about their the financial crime compliance revolution during a recent R3 webinar. Both have rolled out solutions that complement policy collaboration by facilitating instant and automatic data sharing.

We’ve been thinking a lot about blockchain’s role in making existing TBML technologies ever more effective. Check out the recent paper we published on Collaborative Combat. It highlights how blockchain, together with existing technological developments, can link the data available from all different parts of the trade process.