Steps Companies Can Take to Improve Compliance and Reduce Risk
Every year, billions of laundered dollars are moving through financial institutions without detection. Current detection systems aren’t keeping pace with criminals’ ability to outmaneuver them. Financial crime is a complex issue for financial institutions to tackle but AI is offering a clear and effective way forward.
The topics of money laundering and financial crime have been getting plenty of headlines in Europe lately. A recent investigation found that up to $230 billion of ex-Soviet and Russian money had passed through the Estonian branch of Denmark’s largest bank, Danske Bank.
Although it may seem like a European problem at the moment, financial fraud happens everywhere. Billions of laundered dollars are moving through financial institutions without detection. Over the past few years, dirty money has flooded into the U.S. as European countries have enacted laws and regulations to curb this activity. According to the 2015 National Money Laundering Risk Assessment, an estimated $300 billion in illegal proceeds is generated annually in the U.S. The money comes from many sources, including narco-trafficking, international organized crime, foreign corruption, cyrptocurrency-based money laundering and garden-variety fraud.
The bottom line is that financial crime, especially money laundering, remains a complex issue for financial institutions to tackle. All banks have anti-money laundering (AML) systems in place, yet global money laundering transactions are still estimated at 2 to 5 percent of global GDP.
It’s no wonder that the current approach to AML transaction monitoring is being questioned by many organizations and in the majority of cases, it is deemed to be ineffective at managing the risk or the cost of compliance. Despite the fact that regulators are holding financial institutions responsible for the real-life consequences of AML failures, these firms are still struggling to manage the financial and regulatory burden and, despite their best efforts, are still unintentionally facilitating money laundering.
Why? Three main reasons:
1. Outdated Systems
Banks across the sector installed their current AML systems as a quick, short-term reaction to increasing regulatory pressure. Subsequently, most financial institutions have attempted to repurpose retail bank AML or market abuse system models from a different era that aren’t sophisticated enough for current requirements. These models offer a limited understanding of risk exposures and don’t have the capacity to identify and act on suspicious behavior, causing laundered money to become increasingly unrecognized.
2. Money Laundering Can be Hard to Detect
Money laundering typically starts with lower dollar amounts that can go easily undetected before it is moved around the world in large volumes. And it involves a complex web of companies, individuals, trades, settlements and payments organized by inconspicuous individuals. With traditional methods, basic detection occurs at a transaction or account level in isolation, without an understanding of the wider context surrounding individual activity. This prevents many money launderers from being identified. Banks and financial institutions must see the wider picture of individual activity in order to reduce their vulnerability to illegal activity. The increase in connected devices only makes the money laundering web more complex. Experts predict that there will be more than 50 billion connected devices across the world by 2020, giving criminals more communication options and a better chance to hide their activities.
3. False Positives
The staggering number of false positives flagged by current AML systems continues to haunt banks in their efforts to combat money laundering. Layered company structures, usually across international networks, make it extremely difficult for them to define certain AML transaction monitoring systems’ (TMS) requirements that identify risk at an acceptable level of false positives. In an attempt to avoid missing any potential criminal activity, current TMS flag tenuous links that aren’t comprehensively connected, including two people living at the same address, attending the same school or sharing the same name. On top of that, these investigations are labor- and cost-intensive, keeping banks in a vulnerable position as they continue to waste time investigating false positives and making it more difficult to spot cases of true illegal activity.
Today, 90 to 95 percent of alerts are false positives, yet analysts are legally obliged to investigate, regardless of legitimacy, due to the fear of enormous fines and the fact that there is now an increased drive to hold compliance officers, senior executives and board members personally liable for failing to have an adequate AML programs in place.
A New Approach
In order to address these vulnerabilities, institutions need to rethink how they combat fraudulent activity. It is clear that banks are treating launderers as individual transactions, rather than a web of individuals.
The good news is that new practices are becoming more accessible. Contextual monitoring uses entity and network analysis techniques, in combination with AI, to detect anomalous and suspect activity. Understanding the network and its wider context is the first step to reducing false positives and becoming more efficient and effective in the fight against criminal activity. However, organisations have to use care when considering how to deploy AI, as models should be encouraged to “learn” behavior of large AML operational teams that may not be consistently reviewing alerts or following best practices. Furthermore, organisations do need to be able to get any models though model governance and demonstrate to regulators how their models work. The right approach is to combine statistical techniques with context, coupled with the insight and guidance of the very best AML investigators – AI with an expert systems approach.
Money laundering continues to remain a large-scale issue for banks and financial institutions alike. As the criminals get smarter, current stubborn AML systems remain in the dark ages. To make use of the infinite data now made accessible to banks, they need to adopt new compliance technologies and understand criminal networks as an entity, rather than a single transaction. Only then will banks and financial institutions alike reduce their vulnerability to money laundering.
Imam Hoque is COO & Head of Product at Quantexa. He has spent 29 years in the IT services and software industry and now specializes in large data-driven AI solutions with a focus on financial crime and network analytics. Formerly the CTO, Founder and Board Director for Detica NetReveal, he grew this international team to more than 300 specialists and went on to acquire and integrate Norkom AML. Prior to Quantexa, Imam led SAS’ Financial Crime practice for EMEA/AP and was the CTO at an internet startup. In his career, he has been a developer, enterprise architect, strategist, business analyst and project manager.
