India's payment rail goes looking for borrowers

A payments network that charges nothing has to make its money somewhere else. India built the largest such network in the world and is now finding out where.

UPI transactions reached 314 trillion rupees, about $3.2 trillion, with PhonePe and Google Pay together accounting for more than seventy per cent of them, according to the National Payments Corporation of India. Paytm is third; WhatsApp Pay ranked eighth and Amazon Pay twelfth on July figures.

Those shares describe a settled market. When two players hold seventy per cent of a zero-fee rail, the competition stops being about payments and starts being about what can be sold to the people using it.

Credit is the answer everyone reached. At the Global Fintech Festival on Thursday, the Tiger Global-backed BharatPe launched BharatPe Flex, a pre-approved credit line of up to 60,000 rupees usable for UPI payments at online and offline merchants.

The company's chief executive, Nalin Negi, said at the launch that there is a significant opportunity to make formal credit more accessible to consumers. The framing is not marketing so much as a description of the gap.

The gap is large and measurable. A Deloitte report in June put the share of Indian adults with access to formal credit at fifteen per cent, against a global average of twenty-four, and found that more than eighty-five per cent of micro, small and medium enterprises relied on informal and usurious financing.

That second figure is the one that should hold attention. Informal lending is not an absence of credit; it is credit at a price set without competition, and a formal product does not have to be cheap to beat it.

What makes the product possible now is not appetite but data. Digital payments generate a record of how a borrower actually behaves, and a lender assessing a merchant's daily settlement flow is looking at something more current than any filed account.

Vipin V., a managing director and partner at BCG, told CNBC that demand for small, unsecured loans is rising and that fintechs have cracked the model through reach and convenience. Both halves of that are distribution arguments rather than credit ones.

Which is the risk worth stating plainly. Underwriting on transaction data has been tested through a long expansion; it has not been tested through a downturn, when the behaviour the model learned from stops predicting the behaviour that follows.

The structure being built is nonetheless the logical end of the design. A free rail moves the money, the data describes the borrower, and the lending sits on top. India is simply the first place where the rail got large enough for the rest to follow.