XTXFX Business Plan — Risk Management

The principal risks facing a regulated micro-lender, from credit deterioration and funding withdrawal to regulatory change, with controls.

Risk Management

Jump to section

On this page

14.1 The risks that matter
14.2 Risks sized against the plan
14.3 Trigger points and management response
14.4 Concentration limits and portfolio controls

14.1 The risks that matter

Credit losses exceeding assumptions is the risk that decides this investment. A two-point deterioration removes R20.8 million of Year 5 profit and break-even sits only 3.7 points above the base case. It is managed by conservative initial cut-offs, small first-vintage exposure through the Flex product, monthly vintage review by channel and score band, pre-agreed policy tightening triggers, and growth taps that can be closed within one origination cycle.

Regulatory non-compliance is low in likelihood and catastrophic in impact, and it has two distinct forms. Originating while unregistered voids the agreement under section 89(2)(d); pricing above the section 103 ceiling is unlawful conduct that the Tribunal can treat the same way. Both are managed by registration before origination, a dedicated compliance officer from month one, caps enforced in code rather than by policy, full decision logging, and an annual independent compliance review.

The nominal freeze in prescribed fees is high in likelihood and cumulative in impact. Fee income is roughly half of revenue and erodes in real terms every year, with no relief assumed over the plan period. It is managed by growing average loan size within affordability limits and by driving cost to serve down faster than fee value erodes, fixed cost per loan falls from R1 679 to R400 across the plan for exactly this reason.

Funding withdrawal is moderate in likelihood and high in impact. It is managed by diversified facility providers, covenant headroom maintained well inside limits, and the structural protection that the book self-liquidates within fifteen months, so an origination freeze converts it to cash quickly.

Interest rate movement is a partial hedge rather than a pure risk. The NCA ceiling rises with the repo rate, so pricing headroom widens as funding costs rise and the two partially offset. The exposure that remains is timing: the facility reprices immediately while the existing book is written at the old rate.

14.2 Risks sized against the plan

Risk	Assessment	Effect on Year 5 profit	Mitigation
Credit losses exceed assumptions	High likelihood in early vintages; high impact	(R20.8m) for a 2.0 point deterioration	Conservative initial cut-offs; small first-vintage exposure via Flex; monthly vintage review; pre-agreed tightening triggers; growth taps closable within one origination cycle
Regulatory non-compliance	Low likelihood, catastrophic impact	Voids the agreement under section 89(2)(d)	Registration before origination; dedicated compliance officer from month one; caps enforced in code; full decision logging; annual independent compliance review
Prescribed fees remain frozen	High likelihood, moderate and cumulative impact	Roughly half of revenue erodes in real terms each year	Grow average loan size within affordability limits; drive cost to serve down faster than fee value erodes
Origination volume shortfall	Moderate likelihood, high impact	(R10.1m) for a 10% revenue shortfall	Diversified channel mix; employer partnerships from Year 2; repeat lending reaching 72% of originations
Funding withdrawal	Moderate likelihood, high impact	Growth halts; the book runs off	Diversified facility providers; covenant headroom well inside limits; the book self-liquidates within fifteen months
Interest rate increases	Moderate likelihood, moderate impact	(R2.2m) for 300 basis points	The NCA cap rises with repo, so pricing headroom widens as funding costs rise; the two partially offset
Macroeconomic deterioration	Moderate likelihood, high impact	Presents as credit loss	Target segment is formally employed; employer concentration limits applied; sector concentration monitored monthly
Fraud and identity theft	Moderate likelihood, moderate impact	Presents as credit loss and corrupts the scorecard	Device fingerprinting, Home Affairs verification, bank account ownership validation, shared fraud databases, first-payment-default monitoring by channel
Key person dependency	Moderate likelihood, moderate impact	Loss of credit judgement	Documented credit policy, dual-approval governance, key person insurance, no single individual able to override the decision engine

14.3 Trigger points and management response

Trigger	Measured	Response
First payment default above 3.0% on any channel	Monthly by channel	Suspend the channel pending investigation. First payment default is fraud or scorecard failure, not credit deterioration
Vintage loss curve above the modelled band at month six	Monthly by origination cohort	Tighten cut-offs within one origination cycle. Do not wait for the vintage to mature
Blended loss rate above 13.0% of disbursements	Quarterly	Halt growth; hold volume flat while policy is re-cut on observed data
Blended loss rate above 14.5%	Quarterly	Close new-customer origination; lend only to proven repeat customers while the book runs down
Cost per funded loan above R450 on any channel	Monthly	Reduce or close the channel. Aggregator channels are adverse-selection prone and capped by policy
Cash below R5.0 million	Continuous	Draw the facility; approach funders before a covenant breach rather than after
Any affordability assessment without complete documentary evidence	Continuous	Stop origination on that journey until the evidence gap is closed. This is the reckless credit exposure
Two consecutive vintages outside the modelled loss range	Quarterly	Board-level review of whether to recapitalise, halt growth or wind the book down. This is a governance failure, not a market failure

14.4 Concentration limits and portfolio controls

Control	Limit	Rationale
Single employer exposure	No more than 5% of the gross book	The employer channel offers verified employment and lower losses, but a retrenchment event at one employer becomes a correlated default
Single sector exposure	No more than 20% of the gross book	Retail, security, logistics and hospitality all respond to the same consumer cycle
Aggregator channel share	Capped at 15% of originations	Purchased leads are adverse-selection prone; the applicant is shopping every lender simultaneously
New-customer share of Term originations	No more than 30%	Term is the profit engine and should be lent predominantly to known payers
Exception rate against scorecard cut-off	Below 3% of approvals, reported monthly	Exceptions are how cut-offs erode without a decision being taken
Maximum instalment against affordability	An internal buffer below the regulatory maximum	The section 81 maximum is a legal ceiling, not a credit target
Loans per customer outstanding	One Flex and one Term concurrently at most	Multiple concurrent exposures to the same customer defeats the affordability assessment

Every limit above is enforced in the decision engine rather than by policy, and breaches are reported to the credit committee monthly. The exception rate is the most instructive of the seven: a lender with a 1 per cent exception rate is running its scorecard, and a lender with a 12 per cent exception rate has replaced its scorecard with individual judgement without ever deciding to.

Previous section13. Sensitivity and Scenario Analysis

Next section15. Implementation Roadmap