Use of AI will prevent formation of stressed assets: IMF's Subramanian
Mumbai, Feb 23, 2024
Subramanian said India should follow the global practice where the bourses and the regulator need to be informed when there is a violation of a technical covenant
The use of data and Artificial Intelligence (AI) by banks will help to avert moral hazard and adverse selection, and appropriate disclosure of default will help in the prevention of stressed assets, said Krishnamurthy Subramanian, Executive Director, International Monetary Fund.
Speaking at the 7th National Summit for Stressed Assets, Subramanian said, “The disclosure requirement is very important, which is the first. The second is the onus actually needs to be on banks. Our banks need to get far more intensive on data, artificial intelligence, and machine learning in addressing both the adverse selection and the moral hazard problems.”
Subramanian said India should follow the global practice where the bourses and the regulator need to be informed when there is a violation of a technical covenant, irrespective of whether the company has sold equity or debt securities.
Subramanian also added that banks need to get more intensive on data and artificial intelligence, and machine learning to address problems related to adverse selection and moral hazard.
“So adverse selection is the right kind of borrowers not being given loans and the moral hazard is after having taken the loan, the borrower actually does not pay properly. So I think continuous monitoring, using data, artificial intelligence, and machine learning are incredibly important. Now, there is enough research that can enable banks to assess using data or objective measures, not just the ability of the borrower to repay, or even their willingness to repay,” Subramanian said.
He also noted that post the formation of the Insolvency and Bankruptcy Code (IBC), the phenomenon of super equity where the promoter remained in control has changed. The threat of losing the company has made a big difference to the promoters.
[The Business Standard]