South Africa's BNPL market processed R815.1 million in transactions in 2025. A meaningful share of that consumer credit exposure was extended without an affordability assessment. Not illegally, at least not clearly. The regulatory framework that governs BNPL in SA has a structural gap that BNPL providers have been operating in for years, and the IFWG (Intergovernmental Fintech Working Group) has acknowledged as a "regulatory void." That void is closing. The question for compliance and legal teams at BNPL providers, and for every other credit provider who must account for BNPL obligations in their own assessments, is what to do before the rules land.
How BNPL works and why the regulatory classification matters
The BNPL model splits a purchase into instalments (typically three or four equal payments) collected over weeks or months. The retailer or platform pays the merchant upfront and collects from the consumer. Some platforms charge a flat fee for the service; others charge no explicit cost to the consumer and generate revenue from merchant discount rates.
That distinction, fee or no fee, matters under the NCA. The Act's credit classification framework turns on whether a transaction constitutes a credit agreement, an incidental credit arrangement, or something outside the Act's scope entirely. If a BNPL product is a credit agreement, the provider must register with the NCR, conduct affordability assessments per Regulation 23A, and comply with the full consumer credit framework. If it is not, none of those obligations apply.
For a R600 buy-now-pay-later arrangement, the difference between "credit agreement" and "not credit" is the difference between a detailed regulatory compliance programme and essentially no compliance obligation at all.
The incidental credit loophole: how BNPL providers stay outside the NCA today
NCA Section 4(6)(b) defines incidental credit as arising when a fee, charge, or interest is levied in relation to goods or services 20 or more business days after the amount became due and payable. A provider that charges no interest and collects instalments within 20 business days of each due date can argue it has never triggered the incidental credit provision at all, and therefore never became a credit provider under the Act.
This is not a loophole created by regulatory sloppiness. It reflects a 2005 legislative architecture that did not anticipate modern BNPL mechanics. The NCA's drafters were thinking about retailer accounts and deferred payment plans, not real-time instalment payment rails embedded in e-commerce checkouts. BNPL platforms operating under the 20-business-day window are, on the face of the statute, structurally outside the NCA.
The Regulation 3 threshold exemptions add a further layer. Small-value credit agreements below the thresholds prescribed by regulation are exempt from various NCA requirements regardless of structure. Platforms that keep individual transaction values below these thresholds can argue a double exemption.
Webber Wentzel's June 2025 legal analysis of BNPL in South Africa describes the current position plainly: the regulatory framework does not clearly capture most BNPL products, and no single regulator has claimed jurisdiction. The NCR regulates credit; the FSCA regulates financial products and market conduct; neither has formally declared BNPL within its remit. De Rebus has published commentary calling explicitly for legislative reform, describing the current position as a consumer protection gap.
The R815 million question: why this gap is a systemic risk
R815.1 million in BNPL transactions in 2025, projected to reach R1.3 billion by 2030, represents a significant and growing volume of consumer credit commitments that sit outside the affordability assessment requirement. Consumers are accumulating BNPL repayment obligations that reduce their real disposable income, but no credit provider (including the BNPL provider itself) is required to verify whether those obligations are serviceable before the credit is extended.
The systemic risk is not hypothetical. A consumer with three active BNPL arrangements totalling R1,200 per month in repayments is functionally over-extended if those obligations are not visible in their affordability calculation. They are not on bureau; most BNPL providers in SA are not registered credit providers and do not report to credit bureaus. They are not declared; consumers routinely under-report small recurring debit commitments in self-declaration forms. But they appear in the bank statement. Every instalment collection runs as a debit order or card transaction that leaves a traceable record in the transaction history.
Media coverage in SA has been tracking the regulatory debate actively. Mail & Guardian and IOL both ran BNPL regulatory coverage in May–June 2025, citing the IFWG's acknowledgment of the gap and the proposed direction of reform. The direction is toward NCA coverage, not away from it.
What the 2025–2026 reform proposals would change
The reform trajectory, as documented by the IFWG and in academic commentary, points in a consistent direction: mandatory NCA coverage for BNPL regardless of fee structure; threshold-independent registration requirements; and mandatory affordability checks at origination. The specific legislative vehicle and timeline remain uncertain as of mid-2026. Promulgation has not occurred, but the direction is clear.
A mandatory affordability check requirement for BNPL would mean that every split-payment transaction requires the provider to verify, at minimum, gross income and existing obligations, and document the assessment in a retrievable record. For high-volume, low-value BNPL at checkout, that requirement would need to be met in seconds, which pushes toward automated statement analysis rather than manual review.
BNPL providers who have built their operations on the assumption of no compliance obligation will face a transition period if and when the regulation lands. Providers who have invested in affordability infrastructure ahead of that transition will have a competitive and operational advantage. The same pattern played out when Regulation 23A tightened the affordability assessment standard for conventional credit providers. Firms that had already built structured affordability processes adapted quickly; those who had not scrambled.
Why credit providers should count BNPL obligations now
Even where BNPL sits outside the NCA, the repayment obligation is real. A consumer with two BNPL arrangements in active repayment is committed to those monthly debits regardless of whether the BNPL provider reports them to a bureau. Regulation 23A requires assessment of "all monthly debt repayment obligations of the consumer," not "all monthly debt repayment obligations reported to a registered credit bureau."
The word "all" is load-bearing. A credit provider who conducts an affordability assessment against bureau data alone and misses two active BNPL debit orders totalling R800 per month has not, on a plain reading of the regulation, assessed all monthly obligations. The NCR's enforcement posture toward incomplete affordability assessments has become materially more aggressive since the Tsoelli 2023 registration cancellation, and "the BNPL provider is not registered with us" is not a defence for a lender who obtained the consumer's bank statement and could have seen the BNPL debits.
BNPL debit orders are identifiable in bank statement transaction data. They appear as recurring card charges or debit orders from platform names (Payflex, PayJustNow, Mobicred, and others) with consistent reference patterns. A structured statement analysis that categorises transactions identifies them alongside conventional instalment obligations. For credit providers who already obtain three months of bank statements (as Reg 23A requires for income verification), the BNPL detection comes at no additional data cost. It requires only that the categorisation layer be comprehensive enough to classify BNPL transactions as obligations rather than as general spending.
Practical implications: what to do before the regulation lands
The practical response differs by role.
For BNPL providers: Voluntary affordability screening is the responsible position. You do not need to wait for mandatory compliance to check whether a consumer has the income and disposable cash to service a new BNPL obligation. Running a lightweight statement check at approval (income present, existing obligations identified, approximate disposable income calculated) reduces default rates, creates a compliance-ready audit trail, and demonstrates responsible lending conduct to regulators who will eventually claim jurisdiction. The cost of lightweight screening is small relative to the cost of the default rate on unchecked approvals.
For credit providers assessing BNPL consumers: Review your categorisation logic. If your bank statement analysis engine currently classifies Payflex or PayJustNow debits as "discretionary spending" rather than "financial obligations," your affordability calculations are systematically understating obligations for a meaningful segment of applicants. The fix is a taxonomy update, not a process redesign.
AffyScore's statement analysis detects BNPL debit orders in transaction data and includes them in the Reg 23A affordability obligation count, regardless of the NCA registration status of the BNPL provider. For BNPL providers themselves, AffyScore provides the lightweight pre-approval affordability screening that can be embedded at checkout without adding friction to the consumer experience. The API response time is under three seconds, fast enough for a real-time purchase decision. For a more detailed comparison with device finance credit obligations, which are fully NCA-captured, the structural difference is worth understanding.
