Forensic Payroll Analytics for IPPIS: A Hybrid Anomaly-Detection Framework to Expose Payroll Fraud, Improve Data Governance, and Protect Employee Rights

The Integrated Payroll and Personnel Information System (IPPIS) is one of the largest government digitization reforms in Nigeria. The purpose is to ensure payroll accuracy, fight corruption, and weed out ghost workers. While many papers have dealt with the evident successes and challenges, very few have actually ventured to propose forensic analytic frameworks that are deployable in anomaly detection within IPPIS payroll data. This paper presents a forensic hybrid model that generates synthetic data to be fed to unsupervised anomaly detection algorithms and subsequently uses a set of rule-based forensic checks to flag irregularities in payroll patterns. Having been developed to operate under possible data-access restrictions, this framework uses metadata and pseudonymized identifiers in order to circumvent most privacy and governance concerns. The study essentially contributes in three ways: (i) to describing a methodological approach to building synthetic training datasets for payroll anomaly research; (ii) combining several detection techniques into a composite risk-scoring model; and (iii) exhibiting governance controls for ethically applying the research. The results also seem to prove that such a framework renders the presence of stronger payroll integrity, lessening the burden on investigations and increasing transparency.