Over FY20, the total value of bank frauds in India more than doubled, with an increase of 159 percent by value compared to FY19. The Reserve Bank of India attributed this surge to delays in the detection of fraud, stating that the average lag between the date of occurrence of these frauds and their detection, was 24 months during FY20. This number was even higher for frauds of over Rs 100 crore, at 63 months.
Early Detection of Fraud - A Key Challenge for Banks and Financial Institutions
The main causes for the delay in unearthing fraudulent activity, the apex bank highlighted, were the weak implementation of early warning signals (EWS) by banks and the non-detection of EWS during internal audits. These observations warrant a pressing need for banks to set up robust EWS mechanisms that catch red flags when the fraud is in the nascent stages.
RBI Directives on Red Flags
A red flag is an indicator that serves as a warning of a potential threat to an individual or business loan account. In 2015, RBI mandated a tough stance on systems and processes, which include compliance requirements for EWS. The central bank along with the Department of Financial Services defined a comprehensive framework for an efficient EWS system, listing 42 and 83 signals respectively.
The signals can be broadly classified into categories such as:
Accounting Red Flags: Accounting red flags include missing audited/latest financial statements, excessive or unaccounted for cash transactions, and unreconciled bank account statements, among others
Corporate Governance Red Flags: Disproportionate compensation schemes, flouting of regulatory norms, weak or non-existing internal controls, excessive management turnover are some examples of corporate governance red flags
Entity Red Flags: These may include transactions with undisclosed related parties, the existence of shell entities, history of fraud, etc
General Red Flags: These include red flags in individual accounts such as sudden large purchases, missing KYC documents, piling debt, multiple lines of credit, and so on
Alternative Red Flags: These could include any negative mentions of corporate/leadership on social media or news sources
Post the 2015 directive, there have been further stringent monitoring norms suggested by the Malegam Committee in light of the latest frauds identified, and initiatives such as dedicated market intelligence units and increased use of data analytics were put into place.
Robust AI/ML-based EWS Systems can help Banks and FIs avert fraud
Artificial Intelligence and Machine Learning are playing a key role in devising strong EWS systems, in response to the increasing push for timely fraud detection from governing bodies and banks & FIs alike. An AI/ML-based EWS system can identify these irregularities by skimming through hidden linkages of entities, their compliance and litigation history, their transactions, and sanction & negative lists, apart from alternative data points such as social media chatter, news sources, and so on.
EWS using AI/ML algorithms also have the capability to collate information from databases related to credit ratings, shareholding, statutory registrations, etc. and learn from historical fraud records to predict default. Companies have been constantly working with industry leaders to address the challenges related to EWS, by employing AI/ML-driven algorithms that sift through huge datasets 24/7.
These systems help identify signs of deteriorating economic conditions of borrowers by monitoring their health in real-time and classifying them as Red Flagged Accounts, thereby saving the system crores of rupees that could be lost to fraud and wilful defaults.
The author, Omkar Shirhatti, is CEO and Co-Founder at Karza Technologies. The views expressed are personal
First Published: IST