XTXFX Business Plan — Credit Policy and Risk Management

The underwriting model, affordability assessment, scorecard design and collections policy — where the entire return is actually made.

Credit Policy and Risk Management

Jump to section
On this page

  • 6.1 The underwriting stack
  • 6.2 Collections
  • 6.3 Provisioning, vintages and the loss build

The sensitivity analysis in this plan shows that XTXFX’s viability turns almost entirely on credit performance. The credit function is therefore designed as the core operating capability of the business rather than as a control layer over a marketing operation.

6.1 The underwriting stack

Stage

What it does

Why it matters

1. Identity and fraud

Home Affairs identity verification, device and behavioural fingerprinting, and screening against internal and shared fraud databases before any credit assessment is run

Fraud losses present as credit losses and corrupt the scorecard if not separated

2. Bureau assessment

Full enquiry covering payment profile, adverse listings, judgments, administration orders, debt review status and enquiry velocity

Debt review status is an automatic decline. Enquiry velocity is the earliest signal of distress

3. Verified affordability

Consented retrieval and categorisation of at least three months of bank transaction data to establish observed net income, salary regularity, employer identity, existing debit order load and discretionary income

The evidentiary backbone of section 81 compliance as well as the strongest credit signal in the stack

4. Regulatory affordability calculation

Gross income less statutory deductions, less the prescribed minimum expense allowance, less existing obligations from the bureau, producing the maximum permissible instalment

XTXFX applies an internal buffer below the regulatory maximum. The regulatory maximum is a ceiling, not a target

5. Behavioural scorecard

From Year 2, a proprietary scorecard trained on XTXFX’s own repayment data supplements the bureau score

The asset that compounds — and the reason the Flex product exists

6. Credit committee governance

Policy changes, scorecard cut-off movements and exception rates approved by a credit committee with independent representation, reported to the board monthly

Cut-off authority sits outside the commercial function by design

6.2 Collections

  • DebiCheck as the default. All instalments are collected by authenticated debit order aligned to the customer’s verified salary date, with the mandate authenticated at origination.
  • Salary-date intelligence. Collection dates are set from observed salary deposits in the customer’s transaction history rather than from declared pay dates, and are re-sequenced automatically when the pattern changes.
  • Early-stage contact. Automated WhatsApp and SMS engagement from day one of arrears, with self-service payment arrangement options. The overwhelming majority of recoverable value is recovered in the first thirty days.
  • Restructure before default. A documented forbearance policy allowing term extension or instalment reduction for customers in temporary difficulty, applied consistently and recorded, which protects both recovery rates and the reckless credit position.
  • External escalation. Referral to a panel of external collections agents and attorneys at a defined arrears stage, budgeted at 1.2 per cent of amounts disbursed.

6.3 Provisioning, vintages and the loss build

Credit losses and the vintage effect
Figure 9. Credit losses and the vintage effect.

Year 1

Year 2

Year 3

Year 4

Year 5

Term share of value disbursed

32.0%

42.6%

49.1%

53.5%

56.3%

Mature blended write-off on that mix

12.20%

11.93%

11.77%

11.66%

11.59%

Vintage multiplier

1.35x

1.15x

1.00x

1.00x

1.00x

Credit loss rate on disbursements

16.47%

13.72%

11.77%

11.66%

11.59%

Credit loss charge, R million

5.1

15.1

26.7

41.2

55.6

Expected credit loss provision, R million

0.8

3.2

7.3

12.1

17.0

XTXFX will apply an IFRS 9 expected credit loss model with three stages and forward-looking macroeconomic overlays, calibrated initially on industry benchmarks and migrated to own-book vintage curves from Year 2. Vintage analysis by month of origination, channel and score band is reported to the board monthly.