Now Reading: Insider threat biggest bank fraud risk, Rediff warns
-
01
Insider threat biggest bank fraud risk, Rediff warns
Insider fraud, though only 8% of total banking fraud cases, accounts for a disproportionately high share of losses because employees with legitimate system access can exploit core banking solutions, value dating, and…
Insider fraud, though only 8% of total banking fraud cases, accounts for a disproportionately high share of losses because employees with legitimate system access can exploit core banking solutions, value dating, and suspense accounts to quietly siphon funds. Fraudsters backdate entries, activate dormant accounts, and park money in inter-branch or suspense accounts, often squaring off entries before end-of-day cycles close, making detection difficult through routine exception reports.
While the number of staff fraud cases has dropped from 2,624 in FY21 to 1,935 in FY25, and 400 in the first half of FY26, the value of such frauds remains significant. Banks are urged to shift from annual audits to a real-time 'data + behaviour + controls' framework with automated red-flag rules. Key measures include biometric logins, enforcing the four-eye principle, mandatory job rotation, and user behaviour analytics. AI and machine learning, combined with industry-wide data pooling, could help move from reactive to proactive fraud detection.
Indian Opinion Analysis
The real risk in banking fraud is not just the employee but the lag in detection. Exception reports are reviewed periodically, giving insiders months to exploit gaps. Value dating abuse is particularly hard to catch because each debit is matched by a same-day credit reversal. The RBI has long pushed for stronger internal controls, but implementation varies across banks. The next step is for banks to adopt real-time behavioural monitoring, as mandated under the Reserve Bank of India's 2023 guidelines on internal fraud risk management. Watch for the RBI's upcoming review of compliance with these rules.
Source: rediff.com
This story was synthesised by AI from the source linked above.
