Reporting for Duty

Searchspace monitors transactions at Nedbank and reports on money laundering activity.

The Financial Intelligence Centre Act of 2001 (FICA) has put immense pressure on South African financial institutions to identify money laundering and report suspicious transactions -- or face heavy penalties. Though Sandton, Johannesburg-based Nedbank began searching for an automated solution almost immediately, the bank was forced to rely on employee reactions to client interactions, or monitoring of exception reports and transaction activity to comply. Branches reported suspicions of possible money laundering to Nedbank's Group Regulatory Risk Services (GRRS).

But, "It was impractical to manually monitor the high volume of transactions that are coming through different product systems and profiling the clients as specified in the regulations," explains Andre Wentzel, head of GRRS. "A detection software solution was necessary, which would enhance and supplement the unusual-transaction reporting process."

Something for Everyone

With 450 locations in South Africa, as well as branches in London and other African countries, Nedbank ($56 billion in assets) needed a detection solution that encompassed case management and workflow functionality in order to identify and report on suspicious activity, manage rules-based detection, investigate high-risk transactions and minimize the risks of damaging its relationships with correspondent banks. After examining solutions from 10 vendors, in early 2004, Nedbank narrowed its search down to SAS (Cary, N.C.), Mantas (Herndon, Va.) and Searchspace (London).

Attracted to its automated money laundering identification, reporting, monitoring and client profiling capabilities, Nedbank purchased Searchspace's Anti-Money Laundering Solution, which included IBM (Armonk, N.Y.) as an implementation partner. In April 2005, the AML software was deployed onto the bank's Wintel/Unix platform.

Searchspace configured the AML solution's back-load software to the bank's environmental and architectural requirements, and a minimum of three months of transactional and reference data was back loaded into the AML database from Nedbank's Oracle (Redwood Shores, Calif.) database, which runs on an IBM DB2 server, to establish client profiles and transactional behavior.

Searchspace trained bank employees on system usage, alert interpretation, troubleshooting and other functions. Nedbank conducted user tests and integrated acceptance testing of all programs prior to piloting. The pilot took place between November 2005 and mid-December 2005, and the detection software went live before the end of the year.

The AML system monitors every transaction on private individual and business accounts at the bank, according to Wentzel. It can detect both known and previously unknown patterns of money laundering behavior while minimizing costs through the reduction of false positives. "We have the capability ... to do money laundering client profiling in order to detect suspicious transactions and assist with management decisions with regards to these clients," says Wentzel, noting that manually reported cases of possible money laundering also can be incorporated into the system.

Analysts in Nedbank's GRRS department are immediately notified of any unusual activity so they can investigate further. "At present, an average of every 100 alerts analyzed results in one matter registered for further investigation. We trust that this ratio will continually improve as our rules are refined and data matures," says Wentzel.

Nedbank continues to optimize the system. At the end of March, based on additional accumulated data, the bank plans to conduct an alert tuning exercise with Searchspace and IBM consultants.

Fraud Detection

***Institution: Nedbank (Sandton, Johannesburg, South Africa).

** Assets: $56 billion.

** Business Challenge: Automate monitoring and reporting of money laundering activities.