Fintechs leverage new-age tech to minimize loan default risks

India’s formal banking sector has long been plagued with bad debts. According to the RBI, public sector banks lost INR 1.07 lakh crore in bad debts in the previous fiscal year.

Central bank data reviewed by Reuters revealed that Indian banks wrote off over USD 30 billion worth of bad debt in FY 2018-19. While these write-offs eased India’s bad debt pile, the underlying problem was unresolved. Weak balance sheets coupled with the ambiguity around the non-performing assets (NPA) only added to the woes.

The Yes Bank crisis further suggests that the bad-debt problem isn’t restricted to public banks. Instead, it can have a far-reaching impact on the overall economy. Private sector banks, NFBCs, peer-to-peer lending platforms are also bearing the brunt of loan defaults.

This scenario draws attention to the inadequacy of traditional credit rating and underwriting methods. The need of the hour is a robust, tech-driven credit rating system that helps lenders mitigate credit risks and weed out bad cases early on. Fintechs can play a critical role here by collaborating with banks and other financial institutions to create transparent, auditable risk measurement frameworks.

Banks have historically collected large volumes of customer data. However, this data is often unstructured and adds no value to the decision-making process. This is where artificial intelligence (AI) comes into the picture.

Fintech companies use AI-powered automated tools to turn vast data sets from multiple, disparate sources into valuable insights. In addition to credit bureau data, this reservoir also includes alternative data like the prospective borrower’s educational background, employment history, daily transactions, utility and recurring payments, and social media activities.

With direct access to actionable insights, they can build a loan distribution model that increases trust, reduces risk and in turn, the acquisition costs. Interestingly, the more data is fed into the system, the more accurate these insights become.

With feedback systems, continuous learning algorithms and deep learning, the data labelling and data cleaning processes keep improving by the day. This leads to more accurate borrower profiles based on their true creditworthiness. For example, high-value loan disbursals by banks using tech-enabled methods like credit scoring and risk analytic engines such as Crediwatch, have led to more prudent credit disbursement systems.

Human bias majorly influences decision-making, which resulted in loan frauds and NPAs. AI can streamline the credit underwriting process with little to no human intervention. By running the acquired data against a set of rules designed to determine acceptability, it helps lenders take an unbiased decision eliminating the scope for any anomalies.

A predictive analytics-driven vertical approach enables lenders to analyze quantitative and qualitative risk factors to create a comprehensive borrower profile for assessment. Additionally, AI allows credit underwriters to focus on complex aspects like looking into other contingencies that the data may not reveal. Hence, the final decision ultimately lies with the lender, but AI-based technologies facilitate more accurate decision making in a faster manner.