×

Evaluating The Effectiveness Of Predictive Data Analytics In Detecting And Combating Money Laundering Layering Schemes: A Case Study Of A Local Bank, Zimbabwe

Author : Abel Mushipe, Faith Nyakudya Journa Name: International Journal of Science, Engineering and Technology Volume: 14 issue: 6 Year: Volume-14-issue-6 Views : 4
Abstract:
Money laundering layering schemes that move funds rapidly across bank accounts, interbank switches and mobile money wallets are difficult to detect with static, rule-based transaction monitoring, particularly in Zimbabwe’s volatile multi-currency environment, where fixed thresholds decay quickly and alert queues are dominated by false positives. This study evaluates the effectiveness of Predictive Data Analytics (PDA), implemented through the Vneuron anti-money laundering platform at the local Bank in Zimbabwe, in de-tecting layering schemes, reducing false positives and supporting regulatory reporting to the Financial Intelligence Unit (FIU). An embedded mixed-methods single-case design was adopted. Questionnaires were completed by 44 compliance, risk, IT forensic and digital banking professionals (88.0% response rate) and eight senior specialists were interviewed, complemented by document analysis and researcher observation. Data were analysed using descriptive statistics, Wilcoxon signed-rank tests and thematic analysis. Respondents rated rigid thresholds, unsustainable false-positive volumes and linear, account-by-account analysis as substantial weaknesses of the legacy system (means 4.16–4.20). Agreement that PDA improved layering detection, network visibility, alert prioritisation and regulatory reporting was high (means 4.02–4.32; 81.8%–90.9% agreeing or strongly agreeing), and Wilcoxon tests showed significant post-implementation improvements in perceived layering detection (Z = ?5.42), hidden-relationship iden-tification (Z = ?5.38) and false-positive control and investigation efficiency (Z = ?5.41), all p < 0.001. Inter-viewees attributed these gains to unsupervised anomaly detection based on statistical distance, graph-based link analysis and automated narrative generation. Interpreted through agency, game and systems theories, the findings indicate that PDA removes the launderer’s advantage of predictability and converts raw alerts into higher-quality evidence for the FIU. Because effectiveness was assessed through practition-er perceptions in a single institution, validation against alert-level performance data is recommended.

Related Indexing Platform

Indexed

Leave Your Comment

Related Reviewers

Chat with Expert