State Bank of India's Managing Director advocates for leveraging alternative data sources such as power consumption and telecom records to enhance MSME credit assessment, addressing the limitations of GST data for a significant portion of unregistered businesses.
Key Points
- Bankers struggle with MSME lending using only GST data due to many unregistered businesses.
- SBI suggests using alternative data like power consumption, telecom usage, and digital footprints for credit assessment.
- Routing alternative customer data through credit bureaus could create a more comprehensive borrower profile.
- Current integration of diverse data sources is complex and costly for lenders, hindering broader adoption.
- Automated credit models using alternative data are in early stages due to insufficient historical data for development.
Bankers are finding it challenging to do MSME lending only with GST data, a top official from SBI said on Tuesday, pitching for the use of alternate sources such as power consumption, telecom and other digital records to assess the creditworthiness of an entity.
Ashwini Kumar Tewari, Managing Director at State Bank of India, said that while GST and banking data can help lenders assess a significant portion of the MSME sector, a large number of businesses remain outside the formal data ecosystem. "Only 1 crore (of) total 9 crore are GST registered. So, everybody is trying to actually mine that section. That is also a large number. But, there are so many, which are not GST registered. And their data is not available," Tewari said.
Challenges With Current Data Ecosystem
He suggested that alternative customer data, including electricity consumption, telecom usage and other digital footprints, could be routed through credit bureaus to build a more comprehensive picture of borrowers. "It is a good suggestion that all the data of coordinates, which is numerical, that is power consumption, data ownership, mobile phone ownership, all of them, if that can be routed through the bureau itself, that would be eligible data," he said while addressing the annual Fibac event here.
At present, lenders seeking to use such information have to integrate data from multiple platforms and service providers, making the process difficult and costly, he said. "Today, if somebody wants to use all this data, it has to be integrated to all the separate platforms. There are three or four telecom providers. Some of them have to integrate with all of them. We all charge money. So, it becomes very, very difficult," he said.
Developing Automated Credit Models
Tewari also stressed the need to develop automated models using such alternative data, although he acknowledged that these models are still at an early stage because of the lack of sufficient historical data. "The time has come to use the automated data through the bureau mechanism and then develop those models. Those models are still at a very early stage. There is no historical data that can be used in these models," he said.
