UPI’s Next Test Is Credit, Not Payments

A QR code at an Indian shop once offered a fairly simple proposition: scan it, authorise a transfer and spend money already held in a bank account. That proposition is changing. With RuPay credit cards and pre-sanctioned bank credit lines enabled on the Unified Payments Interface, the same familiar checkout can increasingly become an entrance to borrowing. The visible act remains almost identical. The financial consequences do not.

This is a consequential development in India’s digital public infrastructure. UPI has made electronic payments accessible to merchants for whom conventional card acceptance was often unattractive. Connecting credit to those rails could extend the usefulness of that network. It could also import the complexity of consumer lending into a product whose appeal rests on simplicity. India’s task is not to resist that transition, but to make its new boundaries unmistakable.

From moving money to allocating it

UPI, operated by the National Payments Corporation of India, allows participating applications and banks to exchange payment instructions through a common system. Its central achievement is interoperability: customers and merchants need not use the same bank or application. That has reduced the coordination problem that makes payment networks difficult to build. A widely recognised QR code can connect a neighbourhood seller to customers using different services.

Credit changes what this network does economically. In 2022, the Reserve Bank of India permitted RuPay credit cards to be linked to UPI. In 2023, it allowed transfers from pre-sanctioned bank credit lines through UPI. These are distinct products, with different eligibility conditions and repayment structures. Neither means that every UPI user automatically receives credit. Both, however, separate the payment experience from the immediate availability of a bank balance.

The distinction matters because payments and lending solve different problems. A payment system must move an authorised amount reliably. A lender must decide whether an advance can be repaid, price uncertainty and manage distress. Excellent payment infrastructure cannot answer those questions by itself. When both services share an interface, users may reasonably assume a continuity of simplicity that the underlying contracts do not provide.

The public-policy opportunity is therefore narrower, and more credible, than a promise of universal credit. Existing acceptance infrastructure can lower the friction of using approved borrowing facilities. Whether it lowers the cost of borrowing, or reaches people previously excluded, depends on competition, underwriting and product design—not on the QR code alone.

Acceptance is not inclusion

For a small merchant, accepting a credit-funded UPI payment may make a sale possible without installing a conventional card terminal. For a customer facing a short gap between an essential purchase and incoming wages, a clearly priced credit line could be useful. These are genuine gains. Informal borrowing can be costly and opaque; a regulated alternative deserves consideration rather than automatic suspicion.

But easier spending should not be confused with broader access to suitable finance. Banks may initially prefer customers whose income, repayment histories and account relationships are already well documented. Expanding the places where such customers can borrow is commercially sensible. It is not the same as bringing an excluded household into formal credit. India needs evidence about new borrowers and loan terms, not merely another rising transaction count.

There is also an important distinction on the merchant’s side. Receiving a customer’s credit-funded payment does not itself give the shopkeeper working capital. A retailer who needs to replenish inventory faces a separate underwriting decision. Better payment records may help establish business activity, but sales receipts are not profits. Rent, supplier obligations, household withdrawals and seasonal fluctuations determine how much debt a business can carry.

For households with irregular earnings, even a modest facility can be poorly matched to cash flow. A fixed due date may work for a salaried borrower and fail a worker paid unpredictably. The relevant innovation is not simply a smaller loan delivered faster. It is a repayment structure that reflects income patterns without disguising a higher effective price.

The subsidy question returns

UPI’s success has accustomed customers and many merchants to payments with little or no visible charge. Ordinary bank-account UPI transactions operate under a zero merchant discount rate policy. Credit-linked transactions have different commercial arrangements and applicable rules. That distinction is easy to lose when both arrive through the same checkout experience. It should not be lost in the policy debate.

Someone must pay for infrastructure, fraud controls, customer service and capital. Credit adds funding costs and the risk of non-payment. If merchant fees are constrained, providers may look elsewhere for revenue: interest, penalties, distribution income or sales of additional products. None is necessarily illegitimate. The danger arises when the headline promise of frictionless payment obscures the actual source of returns.

Merchant economics also deserve scrutiny. A shop may welcome additional sales while resisting acceptance costs on thin-margin goods. Clear onboarding and statements should explain which transactions attract charges and why. Otherwise, merchants may respond with selective refusal or surcharges, weakening the predictability that made UPI useful. Acceptance cannot be sustained by assuming that every participant benefits equally from every transaction.

Public support should consequently distinguish the shared payment utility from competitive lending businesses built upon it. There is a case for sustaining reliable, broadly accessible payment rails. There is a weaker case for subsidising the acquisition of profitable credit customers without demonstrable public benefit. Where incentives are offered, their objectives and beneficiaries should be explicit. Transaction volume alone is an inadequate measure of value.

Consent must survive the checkout

The most immediate safeguard is a visible distinction between spending deposits and drawing debt. Before authorisation, the application should identify the funding source in plain language. For credit, users should be able to see the applicable interest terms, relevant fees and repayment date without navigating a chain of documents. A familiar PIN should authorise an informed choice, not substitute for one.

India already has regulatory requirements around lending disclosures, including key facts statements for covered loan products. The challenge is to translate contractual disclosure into usable comprehension. A document presented during enrolment may be forgotten weeks later at a crowded counter. Short, transaction-relevant information should complement the full agreement. Local-language explanations need testing for understanding, not merely confirmation that a translation exists.

Defaults matter as much as disclosures. An application should not quietly prioritise a credit facility when a user expects to pay from a deposit account. Limit increases and promotional offers should require meaningful choice. Notifications should distinguish an outstanding borrowing balance from ordinary payment activity. The design objective should be to prevent surprise, rather than secure a technically defensible record of consent.

Data deserves similar restraint. A rich payment history can help a lender assess a borrower, but it can also reveal sensitive patterns of daily life. Access should be purpose-specific, proportionate and governed by applicable privacy and financial rules. More data does not automatically produce fairer decisions. Lenders should test whether their models systematically misread irregular incomes or mistake low digital activity for weak creditworthiness.

Judge the system after the sale

Responsibility becomes harder to follow when an application, a payment network and a lending bank participate in one transaction. A disputed purchase, an unauthorised payment and an incorrectly calculated interest charge are different complaints. Customers should not have to diagnose the institutional problem before obtaining help. A clear front door for complaints, with accountable routing and published timelines, is essential.

Supervision should also look beyond origination. Useful indicators include repayment stress, repeat borrowing, complaints, fraud losses and the share of customers new to formal credit. Aggregate results should be published in forms that protect privacy while allowing independent scrutiny. Rapid growth accompanied by deteriorating outcomes is not evidence of inclusion. It may indicate that distribution has improved faster than lending discipline.

The same principle applies to competition. Interoperable payment rails do not guarantee a competitive market for the credit offered on top of them. Distribution power can concentrate in a few applications, influencing which products users encounter. Policy should favour transparent offers, avoid unnecessary barriers to switching and examine whether commercial incentives distort presentation. Borrowers need meaningful alternatives, not simply several routes to the same opaque proposition.

UPI has demonstrated the value of making payments easier. Its credit phase must demonstrate something harder: that convenience can coexist with intelligible obligations and responsible allocation of risk. India should preserve speed at checkout while insisting on clarity before borrowing and accountability afterwards. The next achievement will not be making debt feel exactly like money in the bank. It will be ensuring that nobody mistakes one for the other.