Why Testing is Key to Improving Banks’ AML and Human Trafficking Detection Tools
- Harriet Shaw, Venture Lead at BAE Systems Applied Intelligence
- 15.12.2020 02:30 pm AML , Banking
Financial crimes like money laundering can often seem far removed from the reality of daily life and the humdrum world of international business. But the truth is there’s a huge human cost to the criminally obtained funds flowing through the global financial system—and perhaps nowhere is this greater than in the area of human trafficking. It’s a cost increasingly weighing on the minds of banking customers and compliance professionals alike.
The good news for financial institutions is that, despite significant industry and technological challenges, there are cloud-based tools out there capable of more accurately rooting out serious anti-money laundering (AML) predicate offences like trafficking. The first step towards success is to accurately test your existing systems.
A lesson in human tragedy
There are 22 predicate offences associated with money laundering that will be recognised under the EU’s Sixth Money Laundering Directive (6AMLD) when it’s introduced in June 2021. These include things like corruption, fraud, cybercrime, insider trading and environmental crime.[i] Few on the list, aside perhaps from murder, terrorism and sexual exploitation, are quite as visceral as human trafficking.
The UN currently estimates that 30 million people live in conditions of human trafficking today, with many sold into sexual exploitation, forced labour, organ removal, forced marriage and other hardships. [ii] This is human tragedy writ large. It’s estimated, for example, that 79% of human trafficking victims end up being exploited sexually, while 20% of victims are children.[iii]
Unfortunately, the economic ravages of COVID-19 may be exposing even more vulnerable individuals to human traffickers, as spikes in unemployment and falling incomes make it hard for many to make ends meet. For those already being confined by traffickers, government lockdowns could “drastically” reduce their chances of being identified and rescued, the UN warns.[iv]
Why does it matter?
It’s heartening, therefore, to see that half of global banking compliance professionals BAE Systems polled recently for an AML report cited human trafficking or sexual exploitation as the money laundering-related crimes that concern them most.[v]
Finding ways to improve detection of such crimes is certainly important for compliance with AML regulations, as well as for wider financial reasons. Human trafficking came third behind fraud and corruption as the predicate offences that caused the largest monetary losses to banks. However, the need to improve detection of perpetrators through financial transactions is increasingly also being viewed through a moral lens. We found banking customers in large numbers worried, angered and appalled at the link between money laundering and serious offences like human trafficking. More than half (51%) said they believed banks were already able to identify cases by analysing financial transactions.
Unfortunately, the reality is somewhat different. Nearly two-thirds (61%) of compliance professionals said that doing so is “somewhat” (35%) or “extremely” difficult (26%). This figure rose to 71% in the US. In the UK, half (47%) of financial institutions said they have to investigate criminal financial activity linked to human trafficking, but just a fifth (20%) are confident they can spot and stop such transactions or behaviour.
A tech challenge
It doesn’t help that many AML compliance professionals feel they are being held back by external factors. Large numbers complained that money laundering typologies are difficult to distinguish from one another, and even that the metrics they’re forced by regulators to work with are misaligned with the practical task of catching criminals. On that note, many also complained that when they do raise suspicious activity reports (SARs), subjects often fail to face justice. It’s a challenge that was laid bare by the recent FinCEN leaks, which revealed how no action was taken against some individuals despite multiple SARs being raised about them over the years.[vi]
There may be little AML compliance professionals can do about this, but they can take steps towards improving their own detection capabilities. Finding patterns in financial data indicative of human trafficking can be incredibly challenging. This is why we advocate machine learning-based tools which can baseline what “normal” looks like so they’re better able to flag potentially suspicious activity. Network analytics are particularly useful, in rapidly revealing real-world connections across large volumes of apparently unconnected financial data.
It’s all about finding the needle in the haystack. By producing fewer, higher fidelity alerts, such systems can maximise the productivity of your in-house teams so they’re focused on higher value tasks. Meanwhile, automation and analytics work in the background to help with the heavy lifting of processing huge amounts of data.
Testing first
But how do you know if your current systems are up to par? BAE Systems’ Futures team has just finished a highly successful trial of our FinCrime Testing Service with two banks. Crucially, we’re now able to feed into the service data insights gathered from up-to-date, jurisdiction-specific financial crime threat intelligence, indicative of human trafficking. The result is that we can run highly realistic simulations of criminal and victim behaviour to test the effectiveness of your AML detection systems. The trial was specific to human trafficking, but the plan is to extend this to all 22 predicate offences.
It’s a long overdue step which we think will help AML compliance teams to gain valuable insight into the effectiveness of their current detection systems, and useful evidence to build a business case for investments in more intelligent, cloud-based solutions. With COVID-19 set to hit budgets, this kind of data could be invaluable in helping to green light those much-needed investment in nuanced detection approaches.