Chapter 19 · BNPL and Credit on UPI
Chapter 18 built the underwriting engine; this chapter tears down the products it powers. Buy-now-pay-later's two structures — and the 2022 crackdown that killed the first — the merchant subsidy that funds "free" credit, the three verified routes by which credit now rides UPI, and the classification question every fintech lawyer bills by the hour to answer: is this a payment product or a loan?
🎯 Learning Objectives
- Anatomise BNPL: the checkout moment, the zero-to-customer price, and the merchant who pays
- Distinguish BNPL's two structures and explain why the RBI's 2022 PPI circular killed the first
- State the three routes credit rides UPI — RuPay cards, PPI interoperability, pre-sanctioned credit lines — with their dates and instruments
- Apply the classification question: payment product or loan, and why the answer decides everything
- Compute the merchant-subsidy arithmetic and judge its sustainability
- Teardown any credit-at-checkout product on four dimensions: who lends, who pays, what discloses, what collects
1. What BNPL Actually Is
Strip the marketing and buy-now-pay-later is a short-duration, small-ticket instalment loan offered at the moment of purchase — pre-approved upstream, invisible downstream. The customer experience is the pitch: no card number, no branch, a two-tap approval at checkout, and — the defining feature — a price of zero to the customer if she repays on schedule. Whether it is a "product innovation" or just a revolving credit line wearing a friendly interface is exactly the fight this chapter referees.
Structurally, Indian BNPL has lived two lives. BNPL 1.0 — the wallet version: the fintech issued a prepaid wallet, loaded it with a partner's credit, and the customer spent from the wallet. The credit sat off any regulated entity's loan book, disclosure was whatever the app showed, and recovery leaned on the app's own "engagement" tools. This structure died by regulation in June 2022, when the RBI barred loading credit lines onto PPIs — a single circular that ended the model overnight and forced the industry's rebuild. BNPL 2.0 — the balance-sheet version: the loan now sits on a bank's or NBFC's book (the RE of Chapter 17), the fintech fronts the journey as an LSP, and the 2025 Directions' full apparatus — KFS, APR, direct disbursal, recovery conduct — applies. Same checkout experience; completely different legal animal.
2. Who Pays: The Merchant Subsidy
If the customer pays zero, who pays? The merchant — through an MDR-style fee on every BNPL transaction, typically a few percentage points, often above card rates. The economics in one sentence: the merchant buys higher conversion and larger baskets by funding the customer's credit. A shopper who will not — or cannot — pay ₹9,000 today pays ₹1,500 six times; the sale happens; the merchant happily pays the fee out of margin.
Compare the two cross-subsidies you now know. Cards: revolvers fund transactors (Chapter 12), with merchants paying 1.5–2.5 per cent. BNPL: merchants fund everyone, often at higher rates, betting on basket growth. Both hide the credit's true cost from the person deciding to borrow — which is a conduct problem, not a pricing one, and Chapter 20's opening scene.
3. Credit Rides UPI: Three Routes
While BNPL rebuilt itself, UPI — built for "pay now" from deposits — learned to carry credit. Three routes, each with its own instrument and date.
Route one: RuPay credit cards on UPI (2022). A RuPay credit card links to a UPI app; scanning a QR spends the card's sanctioned limit. The significance is strategic as much as functional: it broke the international networks' assumption that QR-code India and credit-card India were separate worlds, and gave RuPay's credit issuance a payments-side reason to exist (Chapter 12's story, cashed).
Route two: PPI interoperability through UPI (April 2023). Full-KYC wallets — which, recall, are prepaid instruments, spending loaded money, never credit — became usable through UPI, letting wallet issuers ride the public rail instead of closed loops. Note the boundary the RBI drew and kept: interoperability is for prepaid balances; loading credit lines into wallets remains barred (Section 1's 2022 rule).
Route three: pre-sanctioned credit lines through UPI. Announced on 6 April 2023 and operationalised by the circular of 4 September 2023 (under the PSS Act, 2007): a scheduled commercial bank's customer, with prior consent, may link a pre-sanctioned credit line as a funding account for UPI payments — the limit, tenure and rate set by the bank's board-approved policy. The facility was extended to small finance banks in December 2024. And in June 2026, the RBI settled the prudential question with a new chapter in the Credit Facilities Directions: the regulatory treatment of any credit facility linked to a payment instrument is determined solely by the nature of the underlying credit facility — the channel, the instrument, the technology change nothing. A credit line on UPI is a loan wearing a QR code, and it is supervised like one.
Sources: RBI Statement on Developmental and Regulatory Policies (6 April 2023); circular CO.DPSS.POLC.No.S-567/02-23-001 (4 September 2023, updated February 2025); Credit Facilities Fourth Amendment Directions (23 June 2026).
4. The Classification Question
Every product in this chapter lives or dies on one question: is this a payment product or a credit product? Payment products are licensed under the PSS Act and earn fees; credit products live on a regulated entity's balance sheet and earn spreads under the full weight of Chapters 16-17's rules. The industry's favourite trick was to build a credit product and call it a payment product — earning payment-side lightness for loan-side risk.
The regulators' answer, in three moves. The June 2022 PPI bar killed the wallet structure that kept credit off the books. The Digital Lending Directions (2022, consolidated 2025) made the RE's book non-negotiable: no LSP lends, every loan sits on an RE, and the KFS-APR apparatus attaches. And the June 2026 prudential clarification closed the last creative door: whatever the channel, the underlying facility's nature governs — classification follows substance, not packaging. The examiner's discipline (and IE2's) falls out directly: when you meet a credit-at-checkout product, do not ask what it calls itself; ask whose balance sheet carries the default, and you have your classification.
5. The Four-Dimension Teardown
The session's working tool. Any credit-at-checkout product — BNPL, UPI credit line, pay-later at a food-delivery app — tears down on four questions.
| Dimension | The question | What the answer decides |
|---|---|---|
| Who lends | Whose balance sheet carries the default? | Classification — RE loan (Ch 17 applies) or barred structure |
| Who pays | Where does the revenue come from — merchant fee, customer interest, late fees? | Business model and its conduct pressure points |
| What discloses | Is there a KFS with an all-inclusive APR before the tap? | Compliance; the hidden-APR dark pattern (Ch 20) |
| What collects | E-mandate autopay, or "engagement" nudges? | Recovery conduct; who calls when the free turns fee-charging |
Notice how the four dimensions collapse the marketing: a "frictionless pay-later" that discloses nothing, collects through guilt-trip notifications and lends from an unidentifiable book is not a payment innovation — it is Chapter 16's journey with every failure-mode box ticked. The teardown is the immunity.
6. Where This Goes
Credit at the checkout is now plumbing — which means the differentiation has moved to the questions around it: pricing honesty, collections conduct, over-indebtedness at the small ticket. Those are Chapter 20's subjects, and this chapter's products supply its case studies. Beyond conduct, the same four routes point forward: pre-sanctioned lines and RuPay credit on UPI are, in effect, the e-rupee debate's retail context — by the time Part IV asks whether India needs a new form of digital money, credit will already be flowing through the existing one at scale. Hold that thought for Chapter 22.
🛠 Hands-On: The Product Teardown
Two products to tear down, then a verdict to defend.
Product A — "PayL8r": checkout pay-later, six instalments, "always 0% to you"; a ₹199 "convenience fee" per instalment appears in the app's terms, page 9; the lender is "PayL8r Finance Pvt Ltd" (a non-bank, unregistered); collection is by app notifications referencing "your family contacts' guarantee".
Product B — "QuickLine": a scheduled commercial bank's pre-sanctioned credit line linked to UPI; KFS with an all-inclusive APR shown before the first tap; limit ₹20,000 at 18 per cent; repayment by UPI Autopay mandate set at sanction.
Steps
- Run both products through the four-dimension teardown table.
- Classify each: payment product, RE loan, or barred structure — naming the instrument that decides.
- Compute Product A's true cost: a ₹9,000 purchase over six instalments with ₹199 fees per instalment. What is the effective total cost, and roughly what APR does a ₹9,000 six-month loan at "0%" with those fees imply?
- Compute the merchant's side: ₹10 lakh of monthly PayL8r volume at a 4 per cent fee, with a conversion lift worth ₹2 lakh of extra sales at 25 per cent margin. Net win or loss?
- In 120 words: which product survives your teardown, and what single change would fix the other?
View Solution / Walkthrough
1-2. PayL8r: lends from an unregistered non-bank — barred structure, full stop (no RE, no licence); revenue from hidden fees plus merchant charges; discloses on page 9 of terms — no KFS before contract; collects by contact-referencing harassment — the Chapter 20 catalogue. QuickLine: an SCB credit line via the 4 September 2023 circular — an RE loan; revenue is transparent interest at 18 per cent; KFS with APR upfront; collects by an e-mandate the customer signed at sanction. QuickLine is precisely the product the June 2026 prudential clarification had in mind: a loan, supervised as a loan, whatever rail it rides.
3. Fees: 6 × ₹199 = ₹1,194 on a ₹9,000 purchase — 13.3 per cent of principal for six months of credit. As an annualised rate, roughly 26-27 per cent APR — and materially higher in true effective terms, because instalments shrink the outstanding balance each month while fees stay flat. "0%" is a marketing sentence; ₹1,194 is a number.
4. BNPL fee: 4% × ₹10,00,000 = ₹40,000. Conversion lift: ₹2,00,000 extra sales × 25% margin = ₹50,000 contribution. Net win: ₹10,000 — the subsidy pays, but only while the lift is real and measured. (The worked example of Section 2 shows the same arithmetic failing at half the lift.)
5. Strong verdicts: QuickLine survives on every dimension; PayL8r fails four of four — and the fatal failure is the lender's identity, because no disclosure or collection fix can cure "no RE". The single change that would fix PayL8r: move the loan onto a registered RE's book (a marketplace LSP model, Chapter 16), with the ₹1,194 of fees inside a KFS APR — at which point the "0%" claim must also change. Half-measures (better fonts on page 9) earn no marks.
Key Takeaways
BNPL is checkout credit at zero visible customer price. Structure decides everything: wallet-BNPL died in June 2022; balance-sheet BNPL lives under the full 2025 Directions.
The merchant funds the "free": the subsidy pays only while the conversion lift is real — measure it, don't assume it.
Credit rides UPI on three routes — RuPay credit cards (2022), PPI interoperability (April 2023), pre-sanctioned credit lines (announced April 2023, circular 4 September 2023, SFBs December 2024).
June 2026 settled classification: the underlying facility's nature governs, whatever the channel. Ask whose balance sheet carries the default.
Teardown on four dimensions — who lends, who pays, what discloses, what collects — and no marketing survives the table.
Test Your Understanding
Choose one best answer per question. A calculator is allowed for the arithmetic items; the teardown products of the hands-on are the data. Aim for 24/30 before moving to Chapter 20.
1. Stripped of marketing, BNPL is:
2. The defining feature of BNPL 1.0, the wallet structure, was that credit: (Hand arithmetic allowed.)
3. The regulation that killed BNPL 1.0 was the RBI's 2022 bar on:
4. In BNPL 2.0, the loan sits on: (Hand arithmetic allowed.)
5. In BNPL's standard economics, the credit is funded by:
6. A merchant's BNPL: ₹10 lakh monthly volume at a 4 per cent fee; conversion lift worth ₹2 lakh of extra sales at 25 per cent margin. Net result: (Calculator allowed.)
7. At half the conversion lift (₹1 lakh extra sales, same 4 per cent fee on ₹10 lakh), the merchant: (Calculator allowed.)
8. The first route by which credit rode UPI was:
9. The boundary the RBI kept around PPI interoperability on UPI is that wallets carry: (Hand arithmetic allowed.)
10. Pre-sanctioned credit lines through UPI were announced and operationalised respectively on:
11. Under the September 2023 circular, a UPI-linked credit line's terms are set by:
12. UPI-linked credit lines, initially for scheduled commercial banks, were extended to small finance banks in:
13. The June 2026 prudential clarification established that a credit facility linked to any payment instrument is treated according to:
14. The single question that classifies any credit-at-checkout product:
15. The June 2022 PPI bar is this book's purest specimen of:
16. The cross-subsidy comparison: cards versus BNPL. (Hand arithmetic allowed.)
17. The first dimension of the four-dimension teardown asks: (Hand arithmetic allowed.)
18. PayL8r's fatal teardown failure is: (Hand arithmetic allowed.)
19. PayL8r's true cost on a ₹9,000 six-instalment purchase: (Calculator allowed.)
20. QuickLine — the SCB's UPI-linked credit line with a KFS and autopay — is:
21. The single change that could fix PayL8r:
22. The strategic significance of RuPay credit cards on UPI is that it:
23. The disclosure dimension of the teardown fails a product whose APR appears: (Hand arithmetic allowed.)
24. On the collections dimension, the clean design is:
25. The forward pointer this chapter plants for Part IV:
26. The chapter's closing discipline on "0%" claims:
27. BNPL 2.0's fintech role is, in Chapter 16's vocabulary:
28. The distinctive credit risk of small-ticket checkout credit is:
29. The chapter's one-line verdict on the classification question:
30. The next chapter takes up: