The potential for data analytics to transform the fight against money laundering and terrorism financing lies not only in increasing adoption, but in the interconnectivity of data sources and analytics applications.
This was one of the points highlighted by Singapore’s Data Analytics Working Group in a paper that was published last week to encourage financial institutions (FIs) to adopt analytics in a bigger way and to guide the use of analytics in AML/CFT efforts.
The working group, led by DBS Bank, was set up was set up to leverage the collective experience of The Anti-Money Laundering and Countering the Financing of Terrorism Industry Partnership (ACIP) members in using AML/CFT data analytics to better detect suspicious client profiles and activities.
ACIP was set up in April 2017. It is a private public partnership that is co-chaired by Singapore’s Commercial Affairs Department (CAD) and MAS, and brings together the financial sector, regulators, law enforcement agencies and other government entities to collaboratively identify, assess and mitigate key and emerging money laundering and terrorism financing risks facing Singapore.
The paper discussed how the adoption of analytics can be driven through the use of simple and widely available analytics tools. It noted that much of the value to be unlocked lies not in cutting edge technology but in the interlinking and integration of existing data and capabilities.
Currently, there are private sector projects in Singapore which aim to build platforms that leverage and interlink analytics capabilities to enable data aggregation. This would provide a single view of customers and cross border coordination across financial intelligence units of multiple jurisdictions.
Given that money laundering, sanctions evasion and terrorism finance are by their nature complex and cross border, and that current data and operational silos hinder effective AML/CFT measures as well as result in significant inefficiency, these projects propose to leverage existing initiatives and technology to eliminate such silos as far as possible, reduce unnecessary de-risking, and improve efficiency across the AML/CFT framework, noted the paper.
“While such projects are likely to be extremely challenging, if successful, they could point to a future state in which the global AML/CFT framework is much more tightly linked and responsive to financial crime threats,” it said.
Challenges inherent in the nature of such projects include governance and ownership issues, given the wide range of stakeholders involved, and funding, privacy, regulatory and other technology requirements, among others.
The ACIP previously found that use of data analytics techniques has helped to effectively enhance the battle against financial crime, by bringing about 40% reduction in false positives and 5% increase in true positives, delivered by one bank’s proof-of-concept conducted on a machine learning solution for transaction monitoring.
Closer industry and private-public cooperation could yield significant benefits, noted the paper. These include dedicated career paths and skills development for AML/CFT analytics professionals, as well as workshops for financial institutions. Meanwhile, regulatory agencies like the MAS and CAD could collectively address key policy and operational issues in AML/CFT analytics such as model governance and using DA to target high-risk areas.
MAS assistant managing director (banking and insurance) Ho Hern Shin, also the ACIP co-chair, said, “MAS strongly encourages the use of data analytics in AML/CFT, which has the potential for bringing about transformative change in our approach to combating financial crime. The strong showing by analytics solutions providers during the recent Singapore Fintech Festival shows the growing opportunities to adopt such techniques. We are heartened that the ACIP banks are willing to share their experiences with other financial institutions looking to embark on such projects. MAS also looks forward to working with the industry on areas of collaboration in AML/CFT analytics.”
Recognising that the field is nascent, the paper also provides practical suggestions, drawn from the experiences of major banks, to address key governance and implementation issues. These include important considerations on the validation, audit and explainability of DA models to gain assurance that the models built can reliably improve the detection of illicit activities.
The paper can be accessed here.