BFSI bets on blockchain and ML to sort the bad debt mess

n October last year, the newly appointed SBI Chairman, Rajnish Kumar stated that “tackling bad debts will be the top priority for State Bank of India.” Three months later, the country’s largest bank revealed it had accrued a massive amount of Rs 23,239 crore in bad loans.

And it wasn’t just SBI. Behemoths like Punjab National Bank, HDFC Bank, Axis Bank and YES BANK also reported humongous amounts in Non-Performing Assets (NPAs).

CIO India speaks to IT heads from the BFSI sector to understand how technology can be leveraged to flag NPAs earlier in the loan lifecycle.

Munish Mittal, Group Head – IT and CIO at HDFC Bank believes that technology is the key ingredient to automate underwriting and make it more intelligent. “At the end of the day, it is a policy administration automation, which is decided by the credit team. It’s ultimately the outcome of credit policies and market conditions,” he says.

While the RBI has taken a tough stance in regulating and cracking down on defaulters, its blanket rule to classify legitimate performance-related loan issues and wilful defaulters under the same umbrella leaves well-intentioned MSMEs short-changed.

Economist Sriram Balasubramanian, in his opinion piece, points out that the top 10 corporates in India owe more than Rs 5.73 trillion to state banks across the country.

Mithilesh Singh, Director-Technology Audit at IDFC Bank says: “RBI, as a regulator, has done a fabulous job in cracking down on bad debts. The problem with NPAs is that it not only impacts the bank’s finance, but also the overall economy of the country.”

Singh points out that there are solutions that help banks flag defaulters, for instance, the Early Warning Signals (EWS). These signals in a loan account immediately put the bank on alert regarding a weakness or wrongdoing. The RBI mandates that a bank cannot afford to ignore such EWS, but must instead use them as a trigger to launch a detailed investigation into a Red Flagged Account (RFA).

Blockchain and machine learning to the rescue

Seema Gaur, executive VP and head-IT at Iffco-Tokio General Insurance opines: “With respect to NPAs, I feel blockchain could be deployed amongst banks, to have better visibility and transparency on assets under debt.”

She goes on to explain that in this way, NPAs could be isolated, pinned down and the banks alerted, right at the infancy stage of such assets.

Singh seconds Gaur’s view on banks leveraging emerging technologies like machine learning and blockchain to help prevent bad debts. He believes that these technologies can help banks come up with a technology-driven, statistics-based approach to deal with NPAs.

Mittal opines that more than AI, it is machine learning that holds a lot of potential. “Machine learning can bring a lot of value to the whole process of credit automation. It makes it more robust and tighter,” he explains.

Is it time for the RBI to crack down on institutions running into bad debts?And it’s not just transparency and foresight the industry is gunning for. Automation of the loan processing cycle brings in the much-needed efficiency and resilience in the process.

Given the recent spate of huge amounts of money being written off as bad loans, and the disclosure of Rs 11,000 crore loan fraud at the Punjab National Bank, Singh of IDFC Bank believes that RBI needs to take stringent action against the banks involved.

“Bad debts are unhealthy not only for the bank, but it affects the economy to a large extent. The impact also trickles down to the citizens of the country, because government initiatives that could otherwise be utilized for development are being written off by banks,” he says.

He adds that banks should also cooperate in bringing clients with humongous NPA amounts to the court of law and take legal course of action. “RBI and the ministry of finance should come up with very strict and harsh measures to curtail these kind of activities,” he concludes.