The premise
Subbarao is right about the easy part
In a Times of India editorial titled “Small price, smart call, smooth going”, former RBI Governor Duvvuri Subbarao argues that a modest charge on higher-value UPI transactions can help finance the payment infrastructure’s upkeep while preserving UPI’s wider benefits.
As far as infrastructure financing goes, that’s a reasonable position. A payment system operating at UPI’s scale cannot remain economically invisible forever. The infrastructure, security, fraud management, technology and banking ecosystem behind it all carry real, ongoing cost. Someone eventually has to fund that — free was never going to be permanent.
But I think there’s a larger question sitting underneath the financing question, and it’s the one the MDR debate keeps skipping past: who should benefit from the economic value UPI has created?
The distinction
UPI was never a card network with different branding
Visa and Mastercard evolved around a commercial payment and credit ecosystem in which the network, issuer, acquirer and merchant all had identifiable economic relationships — interchange existed because everyone in that chain had a defined commercial role from the start.
UPI was built differently: as interoperable public payment infrastructure around bank accounts, not around a commercial network of issuers and acquirers. Its extraordinary adoption happened precisely because the transaction itself was largely free to the consumer and, historically, largely free to the merchant.
That zero-cost model created enormous economic value. It also did something less obvious: it obscured the question of how that value should be distributed. When nobody is paying for something, nobody has to argue about who deserves the upside either.
The reframe
MDR isn’t a price. It’s a new economic layer
Introducing MDR on UPI shouldn’t be understood as simply putting a price on a free service. It’s effectively the creation of a new economic layer around one of the world’s largest payment networks.
If the new MDR creates a ₹16,000–20,000 crore annual pool, as various brokerage estimates suggest, the important question isn’t simply whether that’s enough to keep UPI running. It’s how that pool gets divided between banks, payment applications, payment aggregators, technology providers and, ultimately, merchants. A financing question and a distribution question look identical in a headline number and completely different in who ends up better off.
The gap
Payment infrastructure isn’t financial infrastructure
Here’s where I think the bigger opportunity sits. Every merchant payment through UPI carries information about turnover, frequency, ticket size, seasonality and cash flow. Yet the payment itself doesn’t automatically create a financial identity or a credit relationship for that merchant. The transaction happens, and the information mostly just sits there.
UPI answers one question well:
Can money move between two bank accounts efficiently?
The next question — the one nobody’s pricing, financing, or even really debating — is a different one entirely:
Can that economic activity be converted into financial identity, better underwriting, working capital and other financial services?
That gap — between moving money and understanding the economic activity behind the movement — is where the real economics are so much larger than MDR alone.
The flywheel
From payment to financial opportunity
A merchant generating ₹25 lakh a month through UPI might create only a relatively small payment margin for the ecosystem. But that same merchant represents a much larger banking relationship waiting to be built: a current account, deposits, reconciliation, accounting services, insurance, working capital and other financial products.
So the real opportunity isn’t simply to charge merchants for UPI. It’s to use the payment relationship as the foundation for a merchant financial operating system — UPI as the interoperable payment rail, the bank as the regulated banking relationship, the payment app as the interface, and a financial-intelligence layer turning transaction history and consented financial data into a usable financial identity and credit profile.
Figure 1 · the economic flywheel
That’s fundamentally different from simply monetising a payment transaction. A transaction fee is collected once and forgotten. A financial identity compounds — it gets more valuable to the merchant, the lender and the ecosystem every time the loop runs again.
The scorecard
Two very different ways to measure UPI’s success
Sustainability and value-capture get measured on completely different scales. One asks how much revenue a fee generates. The other asks how much of India’s informal economic activity became visible, bankable and financeable because it happened over UPI in the first place.
Figure 2 · illustrative comparison, not a revenue forecast
The real success of UPI shouldn’t be measured by how much MDR it can collect. It should be measured by how effectively the payment network converts economic activity into financial opportunity.
The build
This isn’t hypothetical — it’s what LumexPay already runs
The flywheel above isn’t a proposal for someone else to build. It’s the architecture behind OpenCredit, the platform LumexPay operates — live today at credit.lumexpay.com, not on a roadmap.
Each stage in the loop maps to a coordinated agentic layer already running: a Credibility Agent turns consented UPI, GST and Account Aggregator data into a Trust Confidence Profile; an Underwriting Agent translates that profile into lending-ready, lender-type-specific risk parameters; a Matching Agent routes each borrower to the right cooperative bank, NBFC or merchant bank; and an Offer Orchestration Agent manages transparent, borrower-facing offer comparison and delivery.
Every score behind that loop comes from a deterministic, rules-based engine — not a black box — and every explanation reaches the borrower in their own language, not just English. The merchant doesn’t need a new account with a new fee attached. The same UPI activity that already moved the payment is what builds the financial identity on top of it.
That’s the actual test of the thesis above: once a merchant has credit and working capital, the same UPI activity keeps compounding into a stronger financial identity the next time around — which is exactly the loop the flywheel describes, running in production rather than on a whiteboard.
The principle
Sustainability shouldn’t be confused with merely recovering infrastructure cost. The larger question is whether India uses UPI’s network to make financial services reachable for merchants who’ve always had transactions but never had a financial identity.
I agree with Subbarao’s underlying proposition that UPI needs a sustainable economic model — a payment system at this scale cannot stay economically invisible forever. But a fee schedule answers a financing question. It doesn’t answer the distribution question, and it doesn’t come close to answering the opportunity question.
UPI didn’t just prove that India could move money instantly and for free. It proved that India could generate an enormous, continuous stream of verified economic activity — turnover, frequency, cash flow, seasonality — from millions of businesses that formal finance had never been able to see clearly before. Pricing that stream is a much smaller ambition than building on it.
The payment was never the product—the financial identity it can build was.