Part II — Payments

Chapter 11 · Competition in Payments

The state built a free, open, neutral rail — and two apps came to carry four of every five transactions on it. NPCI's answer was a 30 per cent market-share cap, proposed in 2020, deferred twice, and now due for enforcement on 31 December 2026. This chapter gives you the cap's real machinery, the concentration it targets, and the debate this session exists for: is the cap sound competition policy — and should it be enforced?

📖 Chapter 11 of 35 ⏱ ~60 min read 🎓 Session 14 · CILO 3 Core 60% Stable 20% Volatile 20% 📝 Quiz · 30 MCQs

🎯 Learning Objectives

1. The Question Behind the Cap

Two kinds of rules govern a payment system, and confusing them wrecks the debate. Prudential regulation protects the system's safety: capital, settlement guarantees, fraud controls — Chapter 6's territory. Competition policy protects the market's structure: who may grow how large, so that entry stays possible and no gatekeeper can tax the flow. A safety rule asks "can this fail?"; a competition rule asks "can anyone else win?"

The TPAP market-share cap looks like the first kind but is mostly the second. Yes, it has a safety story — an app carrying half a nation's payments is a single point of failure, as Chapter 9's March 2020 PSP outage taught at smaller scale. But its core purpose, stated in NPCI's own circular of 5 November 2020, was "to address the risks and protect the UPI ecosystem" — with competition named as the object: no single third-party app should control the rail's traffic. That makes this chapter's question, and this session's debate, a question of competition policy. And it is a genuinely hard one, because the context is unusual: a rail that is free, open and interoperable still concentrated.

📝
Hold the paradox in view: interoperability means any app can reach any account — the classic barrier to entry is gone. Users can switch apps in minutes at zero cost. And still, two apps won ~80 per cent of the traffic. Any theory of this market must explain that fact before it prescribes anything.

2. How a Free Rail Concentrated

Interoperability removed the technical barriers. Four forces rebuilt concentration on top of it.

Defaults and distribution. Payments ride phones, and phones ship with defaults: app-store prominence, pre-installed applications, placement decisions made by operating-system owners who also own one of the leading payment apps. The merchant side compounds it — a stall displays one QR code prominently, and that QR belongs to whichever app acquired the merchant first. Habit and word of mouth. Payments are a repeated, low-attention act; users settle early on the app their circle uses and stop comparing. The absence of price competition. Zero MDR made every payment free, so apps could not compete on price — only on experience, distribution and marketing, where scale begets scale. Data flywheels. Every transaction trains fraud models, reliability engineering and merchant tools; the biggest apps learn fastest.

The result, sustained for years: PhonePe and Google Pay together carrying roughly four of every five UPI transactions — about 85 per cent at the peak era of the deferrals, still 77.8 per cent of listed-app volume in September 2026. Paytm, the durable third, held about 8 per cent after its 2024 payments-bank shock. Banks' own apps — the actual PSPs, exempt from the cap — hold single-digit shares collectively: an instructive fact, to which the debate returns.

📋 Stable content — Reviewed: October 2026 · Next review: April 2027

3. The Cap and Its Machinery

State the rule exactly, because most commentary gets it wrong in one of three ways. The cap, from NPCI's circular of 5 November 2020:

⚠
Warning: The enforcement consequence is the chapter's most-missed fact. An app over the cap does not lose users or volume directly — it freezes new users. Think through what that does and does not achieve when the incumbent already has hundreds of millions of them: the mechanism manages the future, not the installed base.
⚠ Volatile content — Reviewed: October 2026 · Next review: January 2027

4. The State of Play, September 2026

From NPCI's app-level data, September 2026: PhonePe 45.6 per cent and Google Pay 32.2 per cent of listed-app volume — 77.8 per cent combined, down from 81.9 per cent a year earlier. Paytm third at 8.1, Navi fourth, and a tail of smaller apps gaining slowly. In May 2026 the leaders' combined share dipped below 80 per cent for the first time since NPCI began publishing app-level data — a real, slow decline, and one that headlines over-read: the cap binds each provider separately, so share trickling from the leaders to small apps does not move PhonePe's own number anywhere near 30. On the rolling three-month measure, both leaders remain far over the cap as the 31 December 2026 deadline arrives.

Two further facts matter for the debate. In December 2024, alongside the second extension, NPCI lifted WhatsApp Pay's user onboarding limits, freeing it to serve its full Indian base — a supply-side attempt to seed a challenger, which had moved its share barely. And in April 2026, a group of challengers — Amazon, Meta, CRED, MobiKwik and Flipkart's Super.money — met NPCI to press for enforcement of the cap: the smaller apps want the rule applied, a useful corrective to the assumption that "industry" uniformly opposes it.

Source: NPCI UPI applications data as reported for September 2026; shares of listed-app volume (23.93 Bn), not the full 24.07 Bn network volume — NPCI publishes no listed-app total, a denominator caution Chapter 10 trained you to note.

5. The Debate: Enforce, Extend, Replace?

The session's debate motion, argued here at full strength on both sides: This House believes NPCI should enforce the 30 per cent cap as scheduled on 31 December 2026.

5.1 The case for enforcement

Contestability is the rail's bargain. The public-rail thesis of Chapter 5 promised competition at the app layer in exchange for the state building the infrastructure. An 80 per cent duopoly on a free public rail is that bargain broken — private gatekeepers capturing the strategic position (users, merchants, data) that public money created, without even the discipline of price competition to constrain them. Concentration is a live risk. One app failing, being hacked, or suffering a bad update would take out nearly half of national retail payments in an afternoon — Chapter 9's Yes Bank lesson scaled to the app layer. The mechanism is gentler than claimed. Breach freezes onboarding, not users; nobody's payment stops. And challengers themselves demand it — the April 2026 lobby shows the rule's intended beneficiaries still want it, five years into deferral.

5.2 The case against enforcement

The mechanism cannot deliver the outcome. Freezing new users while incumbents hold installed bases of hundreds of millions reduces share only if rivals grow — and rivals struggle to, because zero MDR starves them of acquisition funding: with no payment revenue, smaller apps cannot raise or spend what share-building requires. Enforcing the cap without an MDR decision, industry voices argue, penalises the leaders without helping anyone catch them. The cap punishes merit. The duopoly was won on reliability and experience under rules open to all; freezing a superior product's growth is an odd competition policy. It binds future challengers too: the same 30 per cent ceiling would stop the very third app the policy hopes to nurture. And the asymmetry undermines it: banks' apps are exempt despite years of small shares — evidence that rules, not products, are not what is holding entrants back.

5.3 What evidence would decide it

Three observable tests, worth naming in any written case. The trajectory test: if the leaders' shares keep sliding a point or two a year without enforcement, the "market will resolve it" argument gains; if they plateau near 80, it loses. The MDR test: a price side to payments (Chapter 12) would let challengers fund acquisition — the cap-with-MDR combination is coherent in a way cap- alone is not. The substitute test: if NPCI instead attacks the actual bottlenecks — defaults, app-store placement, QR exclusivity — would contestability improve without freezing anyone? State in advance what would change your mind; that is what separates a position from a preference.

🌎
Real World: Notice who sits on each side. Foreign-owned leaders benefit from deferral; Indian challengers lobby for enforcement; banks are exempt and quiet; the payments industry's own council split publicly. When you read any commentary on this debate, first find the author's app — then weigh the argument.

6. Competition Law's Other Lens

India also has a competition authority with different tools. The Competition Commission of India works largely ex post — it investigates conduct after the fact under abuse-of-dominance law, and in October 2022 fined Google ₹1,133 crore over Play Store billing practices, in a cluster of cases about exactly the layer this chapter keeps pointing to: mobile distribution. The TPAP cap is the opposite instrument — ex ante structural regulation, limiting size before any misconduct is shown, closer in spirit to spectrum caps in telecom than to antitrust fines.

The two lenses together expose the debate's deepest point. The payment rail is public, open and neutral — the gatekeeping power in this market does not sit on the rail. It sits in the operating systems and app stores that decide defaults and placement, the merchant QR real estate, and the funding economics that zero MDR shapes. A volume cap on apps manages a symptom of those layers; the CCI's cases and the RBI's Payments Vision 2028 — with its own competition measures — work nearer the causes. Expect this chapter's question to be re-fought there, and expect Chapter 34's embedded-finance world to reopen it: when payments hide inside every commerce app, "share of a payments app" starts to mean less.

One thread carries straight onward. Every serious position in this debate — enforce, extend, replace — now runs through a single prior decision: should anything be priced on this rail again? That is the MDR question, and it is Chapter 12 in full: cards, RuPay, and the fight over who pays for a free utility.

🛠 Hands-On: The Market Share Debate

This is the session's assessed-format debate, made self-paced. The motion: This House believes NPCI should enforce the 30 per cent cap as scheduled on 31 December 2026. You will prepare both sides, then adjudicate — because the exercise's point, as the course outline says, is holding a contested position you may not agree with.

Steps

  1. Build the case FOR enforcement — three arguments, each with a dated number attached (shares, the rail's structure, the April 2026 challenger lobby).
  2. Build the case AGAINST — three arguments, including the enforcement-mechanism critique and the zero-MDR funding problem.
  3. Rebuttal matrix. For each of your six arguments, write the strongest opposing rebuttal in one sentence — the version an intelligent opponent would actually use.
  4. The third positions. Draft two alternatives in three sentences each: "extend again with conditions" and "replace with a defaults-and-distribution rule". Name which evidence would make each preferable to both enforcement and abandonment.
  5. Adjudicate. In 200 words, deliver a verdict using the September 2026 numbers. You must engage the strongest argument on the losing side, and state one piece of future evidence that would reverse your verdict.
View Solution / Walkthrough

FOR, with numbers: (1) The bargain — a public rail (Chapters 5, 9) now carries 77.8 per cent of listed-app volume through two private apps (September 2026), both far above the 30 per cent cap on the rolling measure. (2) Risk — PhonePe alone at 45.6 per cent means one app's bad afternoon is a national payments outage; March 2020 showed the shape of that failure at PSP scale. (3) Demand — the intended beneficiaries (Amazon, Meta, CRED, MobiKwik, Super.money, April 2026) still ask for enforcement; deferral has a constituency, but so does the rule.

AGAINST, with numbers: (1) The mechanism — breach freezes new-user onboarding while incumbents hold hundreds of millions of users; nothing in the rule reduces existing share. (2) The funding asymmetry — with zero MDR there is no payment revenue to fund a challenger's acquisition; enforcing the cap without pricing the rail caps the leaders and starves the pursuers together. (3) The asymmetry — banks' apps, exempt since 2020, still hold single-digit shares collectively, suggesting the bottleneck is product and distribution, not the leaders' size.

Third positions: Extend with conditions — tie the extension to observable share-erosion milestones (say, leaders below 40/28 by end-2027) and to a defaults code; preferable if the trajectory test is passing. Replace — prohibit OS-level payment defaults and exclusive QR placement, and let growth compete; preferable if the CCI cases show the bottleneck truly sits in distribution.

Model adjudication (extract): The enforcement case wins on the rail's own logic — a public utility's strategic layer cannot sensibly be a private duopoly — but its mechanism cannot deliver its goal, which is why the honest verdict conditions enforcement on an MDR decision (Chapter 12) that lets challengers fund growth. The strongest losing argument — that freezing merit punishes the apps users chose — would reverse this verdict if leaders' shares were collapsing on their own, which 77.8 per cent and falling roughly a point a year is not.

Key Takeaways

1

The cap is competition policy wearing a safety jacket: its object is market structure, not solvency.

2

A free, interoperable rail still concentrated — defaults, habit, data flywheels and the absence of price competition rebuilt scale advantages on top of openness.

3

The machinery: 30 per cent of volume per provider, rolling three months, banks' apps exempt — and breach freezes new onboarding, never existing users.

4

Three deadlines, one pending: December 2022, December 2024, 31 December 2026 — with leaders at 45.6 and 32.2 per cent (September 2026), the cap's promise and its mechanism remain untested.

5

Both sides are arguable at full strength — and every position now runs through one prior question: should the rail be priced again? Chapter 12 takes it.

Test Your Understanding

Choose one best answer per question. Select Check Answers to score yourself. An explanation appears after each question once checked. Aim for 24/30 before moving to Chapter 12.

1. The TPAP cap limits any single third-party app to:

2. The 30 per cent cap was first proposed in:

3. Which category is exempt from the market-share cap?

4. What actually happens to a TPAP that exceeds the cap?

5. The deadline history of the cap runs:

6. In September 2026, the individual listed-app volume shares were roughly:

7. The milestone of May 2026 was that the two leaders' combined share:

8. "The duopoly is cracking" headlines, based on the combined share slipping, over-read because:

9. Which mechanism most directly removed price competition among UPI apps?

10. The distribution-side force the chapter holds most responsible for concentration is:

11. The systemic-risk argument for the cap echoes which episode from Chapter 9?

12. The zero-MDR critique of cap enforcement says:

13. The mechanical weakness of the enforcement consequence is that:

14. How would the cap treat a hypothetical third app that scaled successfully to 35 per cent?

15. The "bargain" argument for enforcement states that:

16. The April 2026 meeting at NPCI is notable because challengers:

17. The December 2024 decision about WhatsApp Pay was to:

18. Opponents of enforcement cite the banks' apps because those apps:

19. The Competition Commission's October 2022 Google Play fine (₹1,133 crore) illustrates:

20. The TPAP cap, as an instrument, is best classified as:

21. The chapter's claim about where this market's gatekeeping actually sits:

22. The "trajectory test" in the debate asks whether:

23. The "MDR test" holds that cap enforcement becomes coherent when:

24. The competition concern about on-us transactions, from Chapter 9, is that:

25. The RBI document containing competition measures for payments is:

26. Why can no single month's app-share table prove a cap breach?

27. Under NPCI's operating ladder, alerts to an app begin when its volume share reaches:

28. In the assessed debate, attacking only the weakest opposing argument:

29. The chapter's one-line frame of the whole question:

30. Every position in this debate now runs through one prior question, taken up next in: