Part III — Credit & Wealth

Chapter 20 · Conduct Risk in Digital Credit

Every chapter since 16 has flagged the failure column of the loan journey; this chapter walks straight down it. The dark patterns that steer borrowing, the coercive recovery that shamed an industry into regulation, the hidden APRs that make "free" a lie — and the audit that recognises manipulation when it sees it, because that recognition is CILO 3's core skill in its purest form.

📖 Chapter 20 of 35 ⏱ ~60 min read 🎓 Session 26 · CILO 3 Core 75% Stable 25% Volatile 0% 📝 Quiz · 30 MCQs

🎯 Learning Objectives

1. What Conduct Risk Is

Conduct risk is the risk that a firm's behaviour — not its solvency, not its systems — harms the customer: mis-selling, hidden charges, manipulated choices, abusive collections. Prudential regulation asks "can the firm fail?"; conduct regulation asks "is the firm cheating?" Digital credit multiplies the exposure on three axes. Scale: a branch officer mis-sells one loan; a dark pattern mis-sells a million. Automation: the harm is applied consistently and without a human to appeal to. Invisibility: the interface is the sales process — and interfaces can be engineered to steer.

The RBI's own preamble to the Digital Lending Directions (Chapter 17) listed the conduct failures that built the framework: "mis-selling, breach of data privacy, unfair business conduct, charging of exorbitant interest rates, and unethical recovery practices." This chapter takes those six charges — plus the dark patterns that enable them — and gives you the recognition vocabulary, because you cannot audit what you cannot name.

📝
The course's position: conduct risk is not a compliance afterthought bolted onto an otherwise sound product. It is a design choice — someone chose the pre-checked box, someone wrote the "act now!" countdown, someone imported the contact list. The audit asks who, and why.

2. Dark Patterns: The Taxonomy

A dark pattern is an interface element deliberately designed to trick users into decisions they would not freely make. The term comes from user-experience design — the same craft that makes good products usable, inverted. Six patterns cover almost everything Indian digital lending has produced.

PatternWhat it looks likeJourney stage corrupted
Drip pricingHeadline "0%" advertised; fees revealed one at a time through the funnelDiscovery → disclosure
Confirm-shaming"No thanks, I like paying more" as the decline button's labelSanction; add-on sales
Pre-checked boxesInsurance, "protection plans", auto-upgrades ticked by defaultSanction; agreement
False urgency"Offer expires in 04:59" timers that reset on refreshDiscovery; look-up period
Obscured cancellationCancel is five screens deep; "Borrow More" is the home buttonLook-up period; servicing
NaggingFull-screen "Increase Your Limit!" pop-ups after every paymentServicing; recovery

Two disciplines for the auditor. First, map every pattern to its stage — a pattern at the disclosure stage (drip pricing) attacks the KFS's legal function; a pattern at the look-up stage (obscured cancellation, false urgency) attacks the borrower's exit right; a pattern at servicing (nagging) attacks the collection relationship. The stage tells you which clause is violated, not just that something feels wrong. Second, distinguish the pattern from the preference: an upsell offer presented clearly at a fair price is commerce; the same offer behind a confirm-shaming label is a dark pattern. The design intent, visible in the interface, is the finding.

3. Coercive Recovery

No part of Indian fintech has caused more harm, and no part forced the regulatory response more directly. The 2020-21 unlicensed lending-app economy (Chapter 16's origin story) industrialised recovery abuse: contact-list harvesting — the app uploading every phone contact at "KYC" — followed by calls and messages to the borrower's family, employer and friends, public shaming ("XYZ is a defaulter" forwarded to relatives), threats of legal action no court had seen, morphed photographs in the worst cases. The borrower's social world was seized as collateral without her consent and without any law permitting it.

The regulatory answer now lives in the Digital Lending Directions and the Fair Practices Code beneath them. Recovery runs through the RE and its agents only, within decency; the grievance redress officer is named in the KFS, with escalation to the RBI Integrated Ombudsman after thirty days; and the conduct travels the accountability chain — an LSP's recovery misconduct is the RE's enforcement problem (Chapter 17's rule). Alongside sits the data-protection wall: DPDP consent is specific and purpose-limited (Chapter 31 opens it fully), so an app that uploaded contacts for "KYC" and used them for recovery violated both the lending rules and the data law.

⚠
Warning: Do not soften this section's history into abstraction. Real borrowers — disproportionately poor, first-generation credit users — were humiliated in front of their families over ₹5,000 loans. The rules exist because the harm was real, documented, and repeated at scale. The audit's tone should match.

4. Hidden APRs

The third family is a pricing lie. Chapter 17 defined the all-inclusive APR and the KFS that must carry it; Chapter 19's PayL8r showed the trick — "0% interest" with ₹199-per-instalment fees that compute to 26-27 per cent. The hidden-APR family is every mechanism that widens the gap between the number the borrower thinks she is paying and the number she pays: fees outside the KFS (now barred outright), fees inside but displayed at page 9 while "0%" is at checkout, penal charges that arrive only in default (legitimate in the APR definition as contingent — but must still be disclosed in the KFS), and the refinancing trap where one "0%" loan is repaid by the next, stacking fees invisibly.

📝
The auditor's arithmetic: every hidden-APR finding ends in a computed number. "The fees feel high" is not a finding. "₹1,194 of fees on ₹9,000 over six months is 13.3 per cent of principal for six months' credit — roughly 26-27 per cent annualised — while the checkout screen says 0%" is a finding. Carry a calculator; the calculator is the audit tool.

The KFS-APR regime is the structural antidote, and its power is pre-emption: if the true APR must appear on one page, before the contract, in a format the borrower can compare, then the pricing lie cannot survive the paperwork — it has to be designed out, or the KFS has to be hidden, and hiding the KFS is itself the violation. One honest page beats a hundred conduct guidelines.

📋 Stable content — Reviewed: November 2026 · Next review: May 2027

5. The Conduct Provisions

Assemble the rulebook the audit runs against — every provision already met in earlier chapters, now drawn into one place.

ProvisionSourceWhat it forbids / requires
KFS before executionDL Directions 2025Hiding or delaying the KFS; any charge outside it, for the loan's whole life
All-inclusive APRDL Directions + KFS circular (Apr 2024)Headline rates that exclude fees; drip-revealed pricing
Look-up periodDL Directions 2025Obstructed exits; penalties inside the window
Direct disbursalDL Directions 2025Pass-through intermediaries controlling the money
Recovery conductDL Directions + Fair Practices CodeHarassment, contact-list misuse, third-party shaming; RE's agents only
Grievance redress officerDL Directions 2025Anonymous complaint routes; unacknowledged complaints
LSP accountabilityDL Directions 2025"The app did it" as a defence; RE answers for its LSP
Data minimisationDPDP Act 2023 + DL DirectionsContact-list harvesting; data collected beyond the lending purpose

6. The Dark Pattern Audit

The session's tool, assembled from everything above. The audit walks one product through five steps.

  1. Capture the journey — screenshots of every screen from install to first repayment; the reviewer's own trace, not the marketing screenshots.
  2. Name every pattern — against the Section 2 taxonomy; if it does not fit a name, describe it precisely and coin a label.
  3. Map to clause — which provision of Section 5 does each pattern violate? A pattern with no violated clause is a commercial choice; say so.
  4. Compute the pricing gap — headline versus true APR, in numbers.
  5. Write the finding — pattern, clause, evidence, severity (harm to a vulnerable borrower outranks harm to a sophisticated one), remedy.

The audit's discipline is the same as IE2's, sharpened: a finding names a pattern, cites a clause, shows a screenshot, and computes a number. "The app feels manipulative" is a feeling. "The cancellation button is absent from the look-up-period screen, and the timer resets on refresh, contrary to the look-up provision" is a finding.

Part III's credit story ends here — the value chain (16), the rulebook (17), the engine (18), the products (19), and now the dark side policed (20). One chapter remains: wealth, where the same digital rails carry savings and investment, and the conduct questions change shape because the customer's own money is at stake. Chapter 21.

🛠 Hands-On: The Dark Pattern Audit

Audit "InstaCash", a personal-loan app. The evidence below is the audit's capture; your job is steps 2-5.

The captured journey:

  1. Install screen: "Get ₹50,000 in 5 minutes! 0% interest for new users!" — with a countdown timer showing "04:59 to unlock".
  2. Permissions: Contacts, SMS, photos and camera all requested at once, "for verification". Decline is grey text; Accept is a large green button.
  3. Loan offer screen: "Your approved limit: ₹50,000" — with a ₹499 "processing fee" shown in 8-point grey text below the Accept button.
  4. KFS: Available behind "View Details" — a link at the very bottom, in the same grey as the background.
  5. Look-up exit: The "Cancel Loan" button appears only after a full-screen "Are you sure? Your credit score may be affected!" warning with two buttons: "Keep My Loan" (green, large) and "Continue with Cancellation" (grey, small, bottom-left).
  6. Servicing: After the first EMI, a full-screen popup: "Congratulations! You qualify for ₹1,00,000! Tap to claim" — with no visible close button; closing requires tapping a 10×10-pixel × in the top corner.
  7. Recovery (after a missed EMI): An SMS to the borrower: "We will inform your contacts about your DEFAULT unless you pay today."

Steps

  1. Name every dark pattern in the capture, using Section 2's taxonomy (or coining a label where needed).
  2. Map each to the clause it violates from Section 5's table. If any captured element violates no clause, classify it as a commercial choice and say so.
  3. Compute the pricing gap: the ₹499 processing fee on a ₹50,000, 12-month loan at a headline 18 per cent. What is the true APR, roughly?
  4. Rank the findings by severity of harm. Justify the ranking in two sentences.
  5. Write the three most severe findings in the audit's full format: pattern, clause, evidence, severity, remedy.
View Solution / Walkthrough

1-2. Pattern map: Screen 1: false urgency (the timer) + drip pricing ("0%" headline). Screen 2: over-collection of data beyond the lending purpose — contacts, SMS, photos requested en masse; also a visual dark pattern (asymmetric buttons). Screen 3: drip pricing (the ₹499 in 8-point grey). Screen 4: obscured KFS — violates the before-execution, intelligible-display requirement. Screen 5: confirm-shaming ("credit score may be affected!") plus obscured cancellation — violates the look-up provision (penalty-free exit). Screen 6: nagging plus obscured dismissal (the invisible ×) — a commercial choice at the servicing stage if the offer is honest, but the invisible close button is a dark pattern regardless. Screen 7: coercive recovery via contact-list threat — violates the recovery conduct provisions, DPDP purpose limitation, and the Fair Practices Code; this is enforcement territory, not a guideline.

3. ₹499 ≈ 1 per cent of principal. On a 12-month amortising loan, a one-time fee of 1 per cent of principal adds roughly 2 percentage points to the effective annual rate (the fee is paid upfront but the principal is used across the year). Headline 18 per cent → true APR ≈ 20 per cent. The gap is 2 points — smaller than PayL8r's, but hidden identically.

4. Severity ranking: (1) the recovery threat — real, targeted harm to a vulnerable person in distress; (2) the obscured KFS and cancellation — traps a borrower in a contract she could not read or exit; (3) the data over-collection — a privacy harm feeding the recovery threat; the pricing drip ranks below these because ₹499 is money, but screens 4, 5 and 7 are traps.

5. Model findings: (1) Coercive recovery — SMS threatening to inform the borrower's contacts; violates recovery-conduct provisions and DPDP; evidence: the SMS text; severity: highest — targeted psychological harm to a distressed borrower; remedy: immediate cessation, contact-list deletion, RE enforcement. (2) Obscured KFS — link at bottom, grey-on-grey; violates the KFS display requirement; severity: high — disables all pricing honesty downstream; remedy: KFS on the offer screen, before Accept. (3) Confirm-shaming + obscured cancellation — the look-up exit behind a threat screen with asymmetric buttons; violates the penalty-free look-up exit; severity: high; remedy: equal-weight exit, no score warnings inside the window.

Key Takeaways

1

Conduct risk is design choice at scale: someone chose the pre-checked box, the timer, the contact harvest. The audit asks who, and why.

2

Six dark patterns cover the field: drip pricing, confirm-shaming, pre-checked boxes, false urgency, obscured cancellation, nagging — each mapped to the journey stage it corrupts.

3

Coercive recovery — contact-list misuse, family shaming, threats — is the harm that forced the Digital Lending Directions into existence. Real borrowers, real humiliation, at scale.

4

Hidden APRs are a pricing lie measured in numbers: the audit's finding ends in a computed gap between headline and truth.

5

A finding names a pattern, cites a clause, shows the evidence, computes the number — and the one-page KFS-APR remains the structural antidote to all of it.

Test Your Understanding

Choose one best answer per question. A calculator is allowed for the arithmetic items; the InstaCash capture of the hands-on is the data. Aim for 24/30 before moving to Chapter 21.

1. Conduct risk asks:

2. The three axes on which digital credit multiplies conduct risk:

3. A dark pattern, defined:

4. InstaCash's "0% interest" headline with the ₹499 fee in grey 8-point text is:

5. The pattern where the decline button reads "No thanks, I like paying more":

6. InstaCash's "04:59 to unlock" timer that resets on refresh is:

7. Cancel being five screens deep while "Borrow More" is the home button is:

8. Mapping dark patterns to stages: a timer at the look-up period attacks:

9. The 2020-21 scandal's signature — uploading every phone contact at "KYC" — violated:

10. Under the DL Directions, the grievance redress officer is named:

11. When an LSP's recovery agent harasses a borrower, accountability lands on:

12. The conduct failures listed in the DL Directions' own preamble include all EXCEPT:

13. "The fees feel high" versus "₹1,194 on ₹9,000 over six months is 26-27% APR while the screen says 0%": only the second is:

14. The structural power of the one-page KFS-APR is that it:

15. The KFS link at the bottom of the offer screen in grey-on-grey:

16. The look-up exit screen's warning with a large green "Keep" and a tiny grey "Cancel" is:

17. InstaCash's pricing gap: ₹499 fee on ₹50,000 at 18% headline, 12 months:

18. The severity principle in the audit's ranking:

19. InstaCash's "We will inform your contacts" SMS ranks first in severity because it is:

20. The full-screen "Increase Your Limit!" popup with the 10×10-pixel close button:

21. The dark pattern audit's five steps, in order:

22. When a captured element violates no clause, the honest audit:

23. Penal charges in the APR framework:

24. The refinancing trap in hidden-APR lending is:

25. Under the recovery conduct provisions, collection must run through:

26. The permissions screen's large green Accept against grey decline text is:

27. Under DPDP, an app's use of "KYC" contacts for recovery violates:

28. The correct remedy for InstaCash's recovery SMS finding:

29. The audit's format rule, in one line:

30. Part III's final chapter takes up: