How to Use Analytics to Detect Money Laundering, Part 2
June 21, 2017
In Part I of the recap of our recent webinar, How to Use Data Analytics to Detect Money Laundering, we looked at how data analytics can help money services businesses (MSBs) detect crimes, fulfill compliance requirements and improve their anti-money laundering (AML) programs. Part II of this series takes a deeper dive into how advanced analytics can detect criminal activities, and will arm companies with strategies to prevent them.
Advanced Analytics for Anomaly Detection
It’s important for MSBs to profile their customer information and transactions by refining and segmenting the data. This will position MSBs to use advanced analytics to detect anomalies. Segmentation will also allow MSBs to break customers into groups and verify that they are behaving similarly. Attributes such as the frequency of transactions, location, time/day of transactions, source and destinations are areas where anomaly detection models can be built to find transactions. With good detection, once a transaction or outlier is found that doesn’t match the normal behavior, the MSB is alerted and can examine the transactions to determine whether the behavior is abnormal.
Strategies for Detection
Some of the models that MSBs can implement to improve their anomaly detection include the following:
- Network Linking – Network linking improves risk model construction. This strategy detects customers that may not have exhibited suspicious behavior but are linked to a high-risk person. As a result of this relationship, that customer should inherit some of their risk, causing their own risk score to increase.
- Predictive Analytics – Predictive analytics use machine learning to train models and predict outcomes. By remembering the patterns or suspicious activity around customers, the model can predict future outcomes about that customer. With a case management system, staff are alerted to look at the pattern and flag it as positive (confirming suspicious activity) or as a false positive, and correct the model to improve future predictions.
- Merging the Models – This approach merges rules-based models, scenario-based models, anomaly detection and predictive analytics, and is then leveraged by the risk-scoring engine. With this model, make sure to look at the customer’s cumulative score and behavior rather than only one analytic on a particular transaction. A customer may have multiple lower-risk transactions that may cumulatively amount to a high score.
The risks and rewards in the MSB business are big. While MSBs can be profitable for retail stores, it’s important to be aware that the space is often targeted by criminals because there is such a large volume of transactions—making it easier to disguise money laundering. Advanced analytics help MSBs get ahead of risk by drawing attention to behavior they did not know they should be watching for. When combined with case management and automated regulatory reporting capabilities, MSBs can fulfill their compliance requirements and be on the path to a lucrative business.
About Anu Sood
Anu Sood (LinkedIn | Twitter) is the Director Marketing at CaseWare RCM and is responsible for the company’s global marketing strategy. She has over 20 years of experience in product development, product management, product marketing, corporate communications, demand generation, content marketing and strategic marketing in high-tech industries.