The leg that earns nothing would be worth more than the industry that runs it

The answer

This report has never once said how much money is at stake. It says the merchant leg carries 63.9% of transactions and earns nothing, that the state pays 7.05 basis points to hold it there, and that distribution earns wider margins than transactions. All true, and none of it is a number a budget can hold.

Here it is. Priced at the 30 basis points NPCI itself permits on merchant UPI, the leg would generate Rs 15,450 crore a year on FY2023-24’s national merchant value of Rs 51.5 lakh crore. That is 1.9 times the entire revenue of One97 (Paytm), India’s largest listed payments company, across all of its business lines. The state replaces it with Rs 3,631 crore, which is 23.5% of it.

The foregone pool is not a rounding error against the industry. It is larger than the industry.

Three supporting arguments

1. The gap between the permitted rate and the charged rate is the whole market. NPCI permits 30bps and statute sets it to zero, so the distance between them is not a pricing decision any operator can make. At 10bps the leg would still produce Rs 5,150 crore, comfortably more than One97 (Paytm) earns today. No operator is failing to capture this. It is not available to capture.

2. What the state pays is not a substitute for the price, and was never sized as one. Rs 3,631 crore against a 30bps pool of Rs 15,450 crore means the subsidy stands in for about 24% of the foregone rate. Sub-module B shows that share falling every year without a decision. An operator modelling the incentive as revenue is modelling 24% of a price that does not exist, on an appropriation that is shrinking.

3. The state’s own stated target has a price, and nobody has printed it. Sub-module B recommends indexing the outlay to merchant volume so the effective rate stops falling by arithmetic. Holding 7.05bps in FY2024-25 costs between Rs 4,181 crore and Rs 5,017 crore, against Rs 1,500 crore appropriated: a fiscal ask of roughly Rs 2,681 crore to Rs 3,517 crore. It is a range rather than a figure because the published merchant base for that year covers 10 months, and a ten-month base annualised is an assumption, not a measurement.

So what

• For an operator: the number that matters is not the pool, it is that the pool is unavailable. Size the distribution opportunity against Rs 8,192 crore, what the largest listed operator actually earns, not against Rs 15,450 crore, which is what the transaction would earn in a country that priced it.
• For an investor: any model whose upside depends on MDR returning is underwriting a Rs 15,450 crore policy reversal. State that as the size of the bet.
• For a policymaker: the choice has a price and now it has a figure. Zero MDR moves roughly Rs 15,450 crore a year from acquirers to merchants and consumers, and the state buys back a fraction of it.

Method and its limits

Every figure above is a published rate multiplied by a published base. National merchant value and the incentive paid are the PIB tables sub-module B already uses; the operator revenue is the filed account sub-module J already uses. Nothing is modelled and no new source is introduced.

What this is not. It is not a revenue forecast: it prices a leg at rates that are not charged, to show the size of what the statute forgoes, and an MDR actually introduced would change behaviour, shift volume and not simply multiply out. It is not a take rate: no operator’s revenue is divided by national volume anywhere here, because those perimeters do not line up. And the fiscal ask is a range rather than a figure, because the merchant base published for that year covers ten months and annualising it would be an assumption presented as a measurement.