By — Muskkan Talreja
Abstract
This article examines how India’s cashless payment revolution is shifting market power from payment infrastructure to the data and predictive intelligence generated above it. It argues that transaction data reveals actual behaviour, enabling dominant platforms to refine credit models, profile consumers and merchants, and reinforce network effects. The article thereby suggests that the coordination of governance for payment- related data should include purpose- based retention, independent audits of inference models, disclosures of the resulting inferences, portability and antitrust regulation of “inferential concentration.” This can help safeguard privacy, fairness and effective competition.
Introduction
Every time you use your mobile phone to buy coffee, the transaction captures part of your biography: where you are, how frequently you visit and what your purchases reveal about your income, routine and lifestyle. When multiplied across billions of transactions each month, a payment system transcends its role as mere infrastructure and becomes a comprehensive map of an entire population’s economic activity.
Regulatory scrutiny of digital payment markets has traditionally examined visible indicators of competition, such as market share, transaction volumes, pricing and control over payment infrastructure. These indicators remain important, but they do not fully capture a less visible source of market power: the accumulation and analysis of transaction data. The real point of concentration of power is an invisible level above the rails. It is the transaction graph that gives the platform with millions of merchant- consumer interactions and inference power, behavioural profiling, credit signals and merchant intelligence.
Behind Every Payment: The Hidden Chain of Data Collection
A single digital transaction involves several participants. The gateway verifies the user’s identity, the acquiring bank sends the request to the issuing bank and the banks settle the transaction. There are four players on this chain with various interests in the transaction data: consumers who create patterns from their actions, merchants who analyze the patterns of purchases for loyalty programs, traditional banking institutions using history for detecting fraud and processors collecting data across multiple merchants and consumers. This demonstrates a chain no individual bank would be able to mimic.
Payments data has a unique attribute: unlike other forms of digital data, it captures actual actions of people, not their self- reported interests. This is the reason why transaction data can provide more relevant signals to predict economic activity compared to survey data or sentiment analysis, because it is collected at a higher frequency, with minimal delay.
Such precision is key when combining transactions with merchant information, location data and identities: a simple list of products purchased gets translated into a behavioral profile and from there a prediction into the future actions of a person. This is precisely how tech platforms gain competitive advantage.
Combining payments with e-commerce or search data enables companies to build predictive models that may, in certain contexts, outperform traditional credit scoring models. This is because such data captures real time indicators, including cash flow patterns, sales consistency and transaction frequency, that credit histories may not reflect. BIS research found that Mercado Libre’s e-commerce data-based model predicted defaults more accurately than models relying on credit-bureau scores and conventional borrower information.
The Payment- Data Flywheel: When Privacy Becomes the Price
Digital platforms tend to become more valuable to existing users as more people join, adding to the classic network effect. Payment platforms layer a second, quieter flywheel on top. The more transactions a platform processes, the better its behavioural models become and the better its models become, the more attractive it is to merchants seeking fraud detection or targeted offers, which draws in more consumers. India’s competition regulator has recognised, generally in the context of platform markets, that network effects combined with large accumulated datasets can produce ‘winner-takes-all conditions’ smaller rivals cannot realistically overcome. Its research into telecommunications came up with the same conclusions about how privileged access to data is a barrier to entry independent of price or market share.
This raises an uncomfortable question for competition law. If two payment apps both charge nothing, but one extracts and combines substantially more personal data, are they really offering consumers the same price? Another emerging line of thinking suggests that any loss of privacy itself can be considered non- monetary pricing or reduced service quality, thereby placing the issue under competition law jurisdiction and not under the purview of data protection authorities alone. Consent deserves scrutiny too: if bundled privacy policies form part of the price paid for accessing an essential service, consumers need to know if they have consented to all uses of their personal data.
When Payment Data Becomes Power: Discrimination and Merchant Dependence
Once payment histories feed into credit and risk models, uncomfortable questions follow. Could frequent payments to a hospital quietly lower a person’s credit score? Could spending patterns correlated with gender, caste, religion or occupation act as proxies for traits that should never influence a lending decision?
The danger may not come from incorrect data at all but from perfectly accurate data used for an unfair purpose. Another form of dependence exists on the merchant side as well. A payment platform may track the merchant’s turnover, peak time periods and supplier rotation cycles, hence allowing platforms that offer loan services or operate its own marketplace to use this information to determine which merchants it should ally with or enter into which category of its own. This creates a clear conflict of interests, as the platform becomes an infrastructure provider, data gatherer, lender and competitor to the merchant.
The Interoperability Paradox: Open Payments, Concentrated Data
India’s UPI is often held up as a model of interoperability: a customer on one app can pay a merchant on an entirely different app, which appears to reduce lock- in. The scale is remarkable: UPI processed a record 23.2 billion transactionsworth close to Rs 29.9 trillion in May 2026 alone. In the same month, the two prominent apps, PhonePe and Google Pay, saw their combined share fall below 80 percent for the first time.
This occurred years after the National Payments Corporation of India (NPCI), introduced a rule limiting each third- party UPI app, not the two companies collectively, to a maximum market share of 30%. In 2024, however, the deadline for complying with this cap was extended until December 2026.
The implementation was postponed because PhonePe and Google Pay continued to process an overwhelming majority of UPI transactions, making an immediate enforcement potentially disruptive to the payments ecosystem. The market share cap may limit the number of future transactions processed by dominant platforms but it cannot remove the behavioural intelligence and sophisticated fraud- risk models they have already developed from years of accumulated data. Thus, UPI may be interoperable at the level of payments, but genuine competition cannot be fully realised unless users can also carry their transaction history and the value derived from it across platforms.
Conclusion: India’s Next Payment Challenge Is Data, Not Access
An efficient solution would be for India to now develop a regulatory framework for the intelligence generated from payment data. Payment data governance must complement payment system regulation, with the Reserve Bank of India (RBI), National Payments Corporation of India (NPCI), Competition Commission of India (CCI) and Data Protection Board of India (DPBI) coordinating their respective approaches. Platforms should disclose not only the data they collect and retain, but also the behavioural profiles, credit signals and merchant insights derived from that data, as well as whether these inferences are used for lending, advertising or competing with merchants. The DPDP Act already mandates purpose- specific notice, security safeguards and erasure in prescribed circumstances while CCI research recognises that privileged access to data can become a barrier to entry.
Three additional measures are recommended. First, any use of transaction data in determining credit worthiness or any other major decision should be subject to independent auditing of the inferences drawn from their models, including whether ostensibly neutral spending patterns act as proxies for caste, religion, gender or health. Second, there should be a linkage of retention periods to a specific purpose, ensuring data gathered for the processing or security of payments is not indefinitely stored for commercial profiling on unrelated bases. Third, competition assessments should consider the competitive advantage created by access to large and varied datasets. A platform may use this data to generate insights about consumers and merchants that smaller rivals cannot easily replicate, even if its existing market share does not appear dominant.
It is essential that any portability measures safeguard personal financial information and maintain anti-fraud mechanisms. The battle thus, is not just about who processes the payment but who gains the ability to understand and shape the payer.
About the Author
Muskkan Talreja is a fourth- year B.Com. LL.B. (Hons.) student at O.P. Jindal Global University and a member of the Economics and Finance Cluster of Nickeled & Dimed. She is interested in the intersection of law and economics, with a particular focus on infrastructure development, public policy and their impact on economic growth and governance.
Image Source: 12 Digital Payments To Consider Accepting For Your Business