Author(s)
JONATHAN GIBRILL KARGBO, Maurice Sesay, Habibu Rasin Bundu, Precious Satta Bockarie
- ISSN (P): 3139-8464
- Manuscript ID: 140863
- Volume: 2
- Issue: 10
- Pages: 407–423
Subject Area: Engineering
Abstract
Mobile money services have become the primary financial infrastructure for millions of Sierra Leoneans, yet agent-led networks remain acutely vulnerable to diverse and evolving fraud typologies. This paper presents findings from a cross-sectional mixed-methods study involving 600 respondents drawn systematically from seven districts — Freetown (Western Area), Bo, Kenema, Makeni (Bombali), Moyamba, Port Loko, and Kono — conducted between October 2024 and February 2025. Survey instruments, in-depth interviews (n=42), and mystery-shopping audits of 80 registered mobile money agent outlets were employed. Results reveal that 61.4% of respondents reported at least one fraud victimization experience in the preceding 12 months, with Bo (48%) and Kenema (44%) recording the highest district incidence rates. Eight distinct fraud typologies were identified, with agent PIN-disclosure fraud (34.2%), SIM-swap/account takeover (27.8%), and social engineering (24.4%) dominating. Binary logistic regression identifies informal agent network use (OR=2.350, p<.001), low financial literacy (OR=2.077, p<.001), and rural location (OR=1.844, p<.001) as the strongest victimization predictors. Machine-learning benchmarks on a 500,000-record MoMTSim Sierra Leone-calibrated simulation demonstrate that XGBoost achieves the highest Matthews Correlation Coefficient (MCC = 0.874) and AUC-ROC (0.961), reducing estimated false positives to 240 per 100,000 transactions versus 1,140 for logistic regression. SHAP analysis identifies customer–agent interaction frequency and transaction velocity as the dominant fraud-discriminating features. The study maps critical regulatory gaps including the unresolved NaTCA–Bank of Sierra Leone jurisdictional overlap and proposes an evidence-based eight-pillar governance framework. Findings carry direct implications for regulators, mobile network operators, and policymakers across agent-dependent mobile money ecosystems in Sub-Saharan Africa.