Proposed amendments to South Africa’s anti-money laundering (AML) law – the Financial Intelligence Centre Act (FICA) – will require South African accountable institutions to adopt a risk-based approach to manage money laundering and terrorist financing risks. This follows the Reserve Bank’s imposition of AML penalties on many South African banks.
A recent post by management consulting firm, McKinsey, highlights the importance of quality data in delivering customer risk-rating models.
“Poor data quality is the single biggest contributor to the poor performance of customer risk-rating models. Incorrect know-your-customer (KYC) information, missing information on company suppliers, and erroneous business descriptions impair the effectiveness of screening tools and needlessly raise the workload of investigation teams. In many institutions, over half the cases reviewed have been labeled high risk simply due to poor data quality.” McKinsey
In cracking down on money laundering, South Africa is following a global trend. Delivering risk-based models for AML requires financial institutions to consolidate customer and transactional data and identify suspicious transactions as they occur.
Enhancing data quality is critical to reducing the number of false positives – transactions or customer erroneously identified as suspicious – but is also useful to link related parties and gain a complete view of customer transactions.
For one multinational banking group, traditional relational database approaches to AML have outlived their usefulness. The bank turned to big data technologies, adopting Trillium for Big Data, running natively in Hadoop, to cleanse and match related records and deliver accurate AML alerts.
The banking group needed to scale and enhance existing capabilities to address two key challenges:
- Identification of money moving “silently” between joint account holders
- Managing exploits related to poorly formatted SWIFT messages
Identify risk networks
A common exploit used by fraudsters is to move money silently between organisations using joint account holders. As a simple example, a father may move money into a joint account with his wife, who in turn moves money into a joint account with her daughter. In most cases, banks struggle to identify the relationship between the father and the daughter.
Using Trillium for Big Data’s sophisticated matching capabilities, banks are able to identify related parties (such as joint account holders) to ensure that compliance teams can determine if a transaction belongs to an existing customer, a known external party, or whether a ‘new party’ should be created.
Historical transactions can be stored in a reference hub and used when screening future transactions, ensuring that on-going profiles of linked transactions can be built and the system becomes more effective and resilient over time. This matching happens directly within the big data platform allowing the bank to build risk profile at the individual level and also into his network of contacts.
Prepare and assure SWIFT transactions
SWIFT is the de facto standard for foreign exchange transactions. Yet, SWIFT messages are highly prone to errors as they vary in style and content depending on bank system, operator skill and country of origin.
Experienced money launderers can easily vary information such as names, addresses and even account number layouts to dupe automated checks.
The bank uses Trillium for Big Data to prepare and validate SWIFT messages – ensuring that each record is broken into its key elements which are each standardised and validated. Each transaction is then checked against external reference sources such as international anti-terrorism lists.
By enhancing their big data platform with in built data quality, the bank is now able to process and validate hundreds of millions of transactions daily, significantly reducing their AML risk.
For more information you can download the AML case study.