When it comes to assessing the success of anti-money laundering (AML) measures, banks are reliant on metrics, which often provide an incomplete and inaccurate picture. Flawed metrics make it hard to understand both the scale of the problem and – critically – the impact of measures designed to address it.
With the rise of the European Union’s 6th AML directive (6AMLD) putting the focus squarely on 22 predicate offences, do we now have an opportunity to re-evaluate how we measure AML detection success and effectively tackle organised crime?
Measuring the number of alerts
Every day, millions of alerts are generated in response to unusual or suspicious transactional behaviours. This is the first line of defence, and the net is cast widely. These alerts could pick up a highly sophisticated criminal attempting to move their money offshore, but they could also pick up you or I innocently sending money to a relative in Spain. With more than 95% of alerts proving to be false positives, there’s a lot of noise and limited insight into AML effectiveness.
Number of open cases or ongoing investigations
In come the army of analysts, dedicated to trawl through all of those alerts and figure out which ones are people innocently sending money to their relatives abroad, and which are criminal activity. Most cases are closed without further investigation due to lack of information or seemingly innocent activities. As above, there is a lot of noise here and no gauge on AML effectiveness.
Number of cases filed to law enforcement
For those cases that do seem suspicious enough, a report is filed to law enforcement. But, sadly due to a large proportion of defensive filing, a lack of law enforcement resources, and limited communication between law enforcement and banks, only ~10% of suspicious transaction reports received by national financial intelligence units in the EU are further investigated. So again, not a useful measure of success as the pool of reported suspicious behaviour isn’t effective in aiding investigations.
Can we change this? What are the solutions?
A potential solution could lie in the re-aligning of AML detection around the 22 predicate offences named in the 6th AMLD. However, measuring the success of detecting these crimes would require input and commitment from the whole industry – regulators, financial institutions and law enforcement.
The reality is that all the predicate crimes are connected by money laundering, for which banks have primary responsibility over. However, law enforcement is often frustrated by banks thinking of money laundering strictly as placement (moving money into), layering (moving about) and integration (bringing it out).
Up until now, banks have had a narrow focus and pre-conception of what their responsibility looks like. The intelligence gathered by banks and sent to law enforcement sometimes doesn’t contain the contextualised information needed around the activities that lead to money laundering. In turn, this has created friction between law enforcement and banks –one potential contributor to the introduction of 6AMLD.
Crucial to solving this problem is closing the feedback loop and developing better ways for law enforcement to feedback information about cases to financial institutions. This must be done in a way that is secure, sharable and actionable by financial institutions, and which focusses on the intelligence law enforcement need to prosecute.
Perhaps, with the introduction of 6AMLD, we now have the opportunity to prompt financial institutions to commit to implementing predicate offence-led controls, on the advice of law enforcement. At the same time, financial institutions need to be afforded the opportunity by the regulator to turn off legacy generic controls and focus on crime detection.
We know that almost all crimes have some sort of financial footprint, but we also know that banks are struggling to leverage financial intelligence to disrupt crime, provide evidence and seize assets. If the industry doesn’t align around this, there is a danger that we won’t understand whether a wave of new detection tools are helping or hindering progress. Before we get excited about next-gen analytics, we need to think more about what good looks like.
Complex… but possible?
In this blog, Harriet Shaw questions the measures we use today and examines the possibilities that 6AMLD presents
