India should pay for new UPI acceptance before it pays for more volume

India’s early case for subsidising low-value merchant payments was the case for building a network. A customer values a payments app when shops accept it. A shop values the system when customers use it. A subsidy can get both sides over that threshold. As the network becomes routine, the same subsidy increasingly pays for transactions that would have occurred anyway. A new round of support should start by identifying merchants who would lose reliable acceptance without it. DFS, incentive scheme evaluation release · Cabinet, FY 2024–25 UPI incentive rules

The denominator changes the answer

The FY 2024–25 Cabinet scheme had an estimated outlay of ₹1,500 crore. For qualifying payments up to ₹2,000 to small merchants, it offered an incentive of 0.15% of value under zero merchant discount rate. The February evaluation release describes growth in digital transactions and UPI’s rising share. That history is relevant to the scale the policy has helped build. Volume grew alongside wider smartphone use, other changes to the payments ecosystem and shifts in customer habits. Counting transactions gives the size of the rail. It supplies limited evidence on the extra volume attributable to a subsidy. Cabinet, FY 2024–25 UPI incentive rules · DFS, incentive scheme evaluation release

A ₹500 qualifying payment generates 75 paise at the stated rate. Imagine two groups each generating one million such payments. In the first, 20% of payments would have been made in cash without the incentive. In the second, that share is 2%. Total incentive cost in each is ₹7.5 lakh. The cost per additional digital payment is ₹3.75 in the first group and ₹37.50 in the second. These shares are assumptions to illustrate the arithmetic. The gap would be larger if the second group’s additional transactions merely switched from another digital instrument.

Who gets the margin?

A no-fee QR sign can be genuinely valuable to a vendor. It can reduce cash handling, widen the pool of customers and leave a transaction record. The infrastructure still has operating costs. Banks, app providers and merchant acquirers bear costs for systems, dispute handling, settlement and support. A payment linked to transaction value favours the flow a provider can measure easily. It may have little connection to the fixed cost of training a merchant, replacing a broken device or providing reliable support in a district with thin coverage.

A targeted scheme could pay for a merchant becoming active and remaining active through several months, with extra weight for locations where sustained acceptance is low. That would create a direct incentive to support adoption. It would also create a risk of duplicate or nominal registrations, so retention should be measured using activity from different customers and over time. Paying for service availability or timely dispute resolution could make providers compete on features that matter once acceptance is widespread.

A credible case for paying the providers

The strongest counterargument is that a transaction incentive helps cover recurring costs on a rail expected to operate continuously. If provider revenues fall below reasonable operating costs, the first service cuts may be in low-value segments. That risk warrants a serious cost comparison. It is a case for identifying which functions need support and how much they cost. Processing a transaction, maintaining redundancy and resolving a disputed payment each have different cost drivers. A transparent procurement or service payment could fund the last two more directly than a flat share of purchase value.

An evaluation should count substitution across payment modes too. A customer who switches from a card to UPI may reduce some acceptance costs for a shop without expanding participation. The shop that accepts its first digital payment can change which customers it serves. Those are distinct benefits and can justify different funding.

Policymakers should compare matched cohorts of merchants with similar pre-policy activity and local conditions. Measures should include new active outlets, retention, failed payments and complaints, with results split by geography and merchant size. A phased redesign could make that evidence stronger. If acceptance or reliability fell sharply after a carefully measured reduction in a mature cohort, the case for recurring volume support would improve. If both held while uptake rose in previously underserved markets, the fiscal case would favour shifting the marginal rupee towards entry and service quality.