ICICI Bank official has cautioned against excessive use of computer algorithms in lending and has called for applying "human judgement".
Mumbai: Stating that people can "game" credit bureau scores and doubt the authenticity of tax returns, a top ICICI Bank official has cautioned against excessive use of computer algorithms in lending and has called for applying "human judgement".
The warning by one of the largest private sector lenders comes even as banks, including ICICI Bank itself, and new-age fintechs are increasingly using credit scores issued by companies like Cibil and other online tools to assess creditworthiness of a borrower for quicker turnaround of credit proposals.
"Straight processing using single input of only a credit bureau is a lag indicator I think we should be cautious and should not throw caution to the wind," its executive director Anup Bagchi said speaking at a Bankbazar conference late Tuesday. Earlier, Bankbazar co-founder and chief executive Adhil Shetty said the financial products distribution company is seeing a huge spurt in paperless lending which depends on online tools.
Bagchi said it is possible to "game" the Cibil data on credit scores and also other inputs, including social media profiles and liability profiles, which may be used by the algorithms for understanding a potential borrower. Commenting on the tax returns, which can be another set of data that can be used by the algorithms, Bagchi made clear his uneasiness at the veracity of such datapoints.
"Lots of income tax returns are filed. Lots of those are low income. I don't know the quality of those assessees," he said. It can be noted that banks are offering credit almost instantaneously using online tools and advancement in data availability.
Bagchi seemed to suggest that we can depend more on the systems for lending in cases where the securities are in demat form and more liquid assets like securities and mutual funds but "human judgement" is important in the rest. He also appreciated the efforts taken by various states to increase digitisation of land records, which can help in loan underwriting.