There’s an enormous amount of money-laundering (AML) and financing of terrorism. Current regulations are causing significant pain, though clearly not enough to stop the flows, and regulatory scrutiny will increase. So, it’s both good citizenship and good business for firms to invest in better AML process and technology.
Regulatory pressure is on the rise
Regulation and enforcement are already painful, with AML penalties projected to be $1.4 billion in 2020. And regulators may get tougher still. The September leak of FINCEN files brought global attention to the issue. And in the U.S., the new administration is likely to take a more assertive stance on regulation in general and AML in particular. For example, Congress just passed a bill requiring companies to disclose their owner or owners to FinCEN.
It’s also likely that enforcement will be stricter. For instance, there may be less leeway to use deferred prosecution agreements (DPAs) to avoid penalties for large banks as well as greater enforcement of DPAs once they’re signed. There may also be more stringent demands for information about who is being referred to in suspicious activity transaction reports (SARs).
Outside the U.S., we’re likely to see a similar trend around enforcement – and perhaps even tighter regulation – in the UK and Europe.
In parallel with those regulatory and enforcement trends, the industry itself is beginning to feel the pressure to empower compliance officers to act on, rather than ignore, SARs.
The technology exists
Many firms struggling to manage the current scrutiny, much less prepare for even more stringent rules and enforcement. All too often, the status quo is inadequate: reporting is largely manual, alerts aren’t frequent enough, and firms rely on a grab bag of old school, rules-based systems that are scattered across the organization. The result is slow, wasteful and risky. Too many false positive alerts, not enough Suspicious Activity Reports.
Looking at the problem from a data perspective, banks are wrestling with large volumes, disparate sources, time series data, and both structured and unstructured information. From this they need to identify complex and changing patterns and then investigate, explain and report.
Framing AML as a data science problem points to a clear solution: machine learning. And that solution starts by pulling together a coherent view of all of those data sources. Then you develop and train algorithms to detect suspicious patterns and improve detection over time. In addition, you need to make sure that the risk factors and the rationale for alerts are easy to understand and explain to the business and regulators.
There’s a real cost to ripping out old systems and process – and to buying and implementing new ones. But with better processes and smarter tools, banks will be able to find money-laundering activity faster and more accurately, reduce operational costs, and avoid the risk of massively expensive and reputation-destroying penalties.