ADVERTISEMENT

Fintech Lending To Higher Risk Customers: White Space Or Red Flag?

Are digital lenders increasing systemic risk by focusing on lower credit score customers.

Payment businesses from global names like Google and WhatsApp, to the country’s banks, its postal service, e-commerce firms, Fintechs  and telecom companies are all  competing for customers in the digital age. However, all of these entities rely on the common network and payment infrastructure designed and created by the RBI and NPCI. (Photographer: Samyukta Lakshmi/Bloomberg)
Payment businesses from global names like Google and WhatsApp, to the country’s banks, its postal service, e-commerce firms, Fintechs and telecom companies are all competing for customers in the digital age. However, all of these entities rely on the common network and payment infrastructure designed and created by the RBI and NPCI. (Photographer: Samyukta Lakshmi/Bloomberg)

It was the early 2000s and a pick-up in credit card spending was all the rage. A number of private banks were falling over themselves to issue credit cards. Few believed that this form of unsecured credit would prove to be a headache in later years.

Then came the global financial crisis and a slowdown in the Indian economy. Many lenders were left holding a pile of bad credit card debt.

Fast-forward to 2018 and the rage is now about lending to customers with lower credit scores. This time its not the banks but fintech firms which are arguing that improved credit assessment tools will minimize the higher risk associated with such borrowers.

They may very well be right but some analysts are keeping a wary eye out for any build-up in risk.

“Risk models work until they don't,” wrote Gautam Chhugani, analyst at Bernstein in a March 11 note titled ‘Are fintech lenders increasing systemic risk?’

“Risk is priced to earn a good return on capital but that doesn't guarantee return of capital. Growth aspirations lead to underwriting of the edge cases,” Chhugani cautioned. Bernstein, a believer in the power of fintech, said that not all payment firms will be in a position to successfully pivot towards a lending model, which many are attempting.

White Space Or Red Flag?

The concern, while nascent, is emerging from the increased enthusiasm to lend to customers with lower credit scores.

Consumers with a score of more than 750 assigned by credit bureau Transunion CIBIL are considered to be prime, while those with a score about 800 are considered super-prime. At the other end of the spectrum, customers with a score of 650-750 are considered ‘near prime’ and those below 650 are considered ‘sub prime’.

Banks are most comfortable lending to those with a CIBIL score of 750 and above. Fintech firms see an opportunity in lending just below that segment.

There is a huge white space in the consumer lending landscape in India—credit score of 550-750 and loan size of below Rs 1 lakh with a one-year tenure, according to EasySalary, whose chief executive officer Akshay Mehrotra spoke at a recent conference organised by Kotak Institutional Equities. The comments were published in a report by the brokerage house.

This segment, of customers with a score below 750, makes up nearly a third of the personal loan segment, which, in turn, is the second largest segment in retail credit after housing loans, said Satyam Kumar, chief executive officer of LoanTap. Kumar told BloombergQuint that it makes sense for fintech firms to target this segment since it includes, in large part, salaried millennials who are ideal customers for digital lending.

The trend of lending to those in the ‘near prime’ credit range has picked up in the last 1.5-to-2 years, said Hrushikesh Mehta, country manager at credit data firm ClearScore. “Preliminary studies suggests that the model is holding quite well though it is still too early to reach a conclusion,” said Mehta.

Data included by Bernstein in their report showed that a large part of the incremental lending is going to the 650-750 credit score bucket. The share of incremental lending to ‘super-prime’ customers was just at 5 percent in the 12-months till September 2018, while 60 percent of the new loans went to ‘near-prime’ customers, showed the data.

The data was based on an analysis of banks and non-bank financial companies, including digital lenders.

So Where’s The Problem?

Fintech firms don’t see this as a problem. A few that BloombergQuint spoke to cited shortcomings of credit bureau data, particularly in the case of younger customers. Others felt that newer credit appraisal tools give them enough confidence to lend to this segment.

Every segment has customers with good and bad profiles, said Rajan Bajaj, chief executive officer of SlicePay. You can split the segment in a way that you do business with the best customers in that segment, he added.

According to Raman Kumar, founder of CASHe, a pure-play digital lender, credit scores work for customers with a substantial credit history. For others, data-based algorithms may be more effective. “One reason why fintechs build algorithms to track their borrowers is because credit bureau scores are not perfect,” Kumar said.

Mehrotra of EarlySalary also believes that credit scores tend to be biased against young and new-to-credit customers.

Ajit Kumar, chief executive officer of P2P lender RupeeCircle shared that view. Credit bureau scores merely measure the intention of the borrower to repay their loans and not the borrowers’ ability to repay, Kumar said.

And so, fintech firms are relying increasingly on customer data beyond credit scores. This data ranges from assessing a customer’s cashflows, to their spending patterns, their family commitments and even their social media habits.

Satyam Kumar, chief executive officer of LoanTap said that defaults depend on specific demographic characteristics of the borrower. For instance, people in low-income categories typically tend to have repayment issues only when there are a large number of dependents in the family, he said. Kumar claims that his borrowers ‘only face job risks and not macro-economic risks’ and adds that higher defaults are likely for customers with a salary of Rs 10,000 to Rs 30,000. LoanTap builds in these learnings from its data analysis into lending and interest rate decisions.

Assessing Default Rates

Credit bureau Transunion CIBIL declined to share data on default rates across different credit score segments.

Crif Highmark, another credit bureau, also did not share any data. A spokesperson for Crif Highmark said that a borrower with a score of 700+ is likely to have all loans repaid in time and have good levels of utilisation of credit limits. Whereas, someone with a credit score of 600 may have minor lapses in repayments or high average utilization of available credit.

Delinquency rates published by some of these lenders on their websites show a wide divergence. The rate of default ranges from 0.5 percent at LoanTap to 4 percent at SlicePay. Since industry-wide data for this segment is not available, benchmarking these default rates is difficult.

Chhugani of Bernstein is cautious in accepting the efficacy of alternative credit-scoring models.

Hoping to tap into the social and mobile footprint of customers to assess credit quality is experimental at best, Chhugani wrote in his note. There is a good chance that lending would exceed the confidence levels in these data models, he added.

Data is trying to replace loan collateral, resulting in growth of unsecured lending... Whoever has access to customer transaction data believes they can make an underwriting model out of it. However, not everyone will have the right to win.
Gautam Chhugani, Analyst, Bernstein