Africa's Next Credit Debate Is Not About Formalisation

The usual story says informal African businesses need to become formal before serious investors can fund them. The evidence points elsewhere: the decisive question is whether a company controls the payment, asset, inventory or servicing loop well enough to turn activity into enforceable cash flow.

M-KOPA, the Kenyan company that finances smartphones through small digital instalments, made an unusually blunt admission in evidence considered in a Kenyan tax judgment on October 4, 2024.

When customers stopped paying, the company said recovering the handset was difficult. The devices had little resale value. Physical repossession, the comforting old-fashioned image of a lender taking back an asset, was not doing much work.

What did work was software.

The case was a tax dispute, not a test of whether M-KOPA could collect consumer loans. The research brief does not provide the tribunal’s final ruling or the tax amount at stake, so it would be wrong to invent either. Its importance lies elsewhere: the judgment records M-KOPA’s own explanation of the commercial stakes. A financed handset was not reliably valuable because it could be recovered and sold. It mattered because M-KOPA could suspend its use when an account fell into arrears.

Under M-KOPA’s Kenyan terms, customers pay through M-Pesa or the M-KOPA app rather than cash. If payments fall into arrears, M-KOPA can disable the device. The company can also assign the credit agreement. This was the operating architecture behind a $50 million-equivalent facility from the International Finance Corporation, or IFC, for M-KOPA Kenya, a $15 million Uganda facility in May 2023 and a $51 million commitment announced by DFC, a development-finance provider, in May 2024.

M-KOPA’s financed phones matter because software can suspend their use when instalments stop. Photo: Nicholas Githiri / Pexels, Pexels licence (free commercial use).

The phone was not strong security because somebody could sell it. It was strong because a customer needed it to work every day, and M-KOPA controlled that utility.

That raises a more useful question than the one investors usually ask about Africa’s informal economy.

Not: *How large is the market?*

Ask instead: *Where is the loop?*

Nearly nine in ten workers in Sub-Saharan Africa were informally employed in 2024, according to the International Labour Organization. That is a huge economic fact. It is not, by itself, an investment case. A vast market of cash transactions, irregular records and weak collection rights is not a portfolio. It is a very large claim waiting to disappoint someone.

The myth: registration makes a business investable

The conventional view is understandable. Informal merchants may lack conventional accounts, formal credit histories and assets that can easily be pledged. So the answer appears simple: wait until they register, digitise and resemble conventional small businesses.

But registration is a weak proxy for whether a lender gets repaid.

Imagine two shopkeepers. One is registered but takes most payments in cash, buys from several suppliers and can switch payment providers tomorrow. The other may have a thin formal record, yet sells through a payment rail that records every transaction, receives stock from a controlled distributor and repays automatically from daily sales.

Which one gives an investor more confidence?

The interesting part is that the second merchant may be far more financeable. Not because informality has vanished, but because the transaction has stopped being invisible.

CGAP, a financial-inclusion research organisation, found in March 2024 that transactional data could predict repayment about as well as conventional credit history. Combining the two performed better than either alone. Earlier CGAP research identified direct payment integration, repayment taken “off the top” and inventory disbursement as ways to reduce both uncertainty about a borrower and the risk of having to chase them later.

A spreadsheet full of user numbers cannot do this. Neither can a fleet of dormant payment terminals.

Nigeria’s terminal count was never the point

Moniepoint began in Nigeria by supplying technology to banks. Founded by Tosin Eniolorunda and Felix Ike, it then built merchant payments, accounts, bookkeeping and credit around the daily operating life of businesses.

In October 2025, Moniepoint announced that it had completed a Series C exceeding $200 million. The round was led by Development Partners International, an investment firm, and included LeapFrog, another specialist investment firm, alongside IFC, Visa, Google’s Africa Investment Fund and others. In that same company announcement, Moniepoint claimed more than ten million active business and personal customers, annual payment value above $250 billion and sustained profitability.

Those figures are company claims, not independently corroborated operating metrics in the material provided for this article. They are useful evidence of what investors were asked to back. They are not enough to underwrite a credit book.

Moniepoint’s advantage is active merchant payment flow, not terminals counted at installation. Photo: Ebuka TheArtDairector / Pexels, Pexels licence (free commercial use).

CGAP reported in January 2026 that Moniepoint paid agents for transaction usage rather than simply for signing merchants up, while providing near-real-time settlement and reversals. In plain English: it rewarded behaviour that kept money moving through its system, not hardware sitting on a counter.

That is the difference between a merchant-acquisition story and operating infrastructure.

A terminal sign-up produces a photograph for an investor deck. Active payment throughput produces recurring revenue, transaction histories and a chance to understand how a merchant actually trades. It may later support credit. It may not. But it creates the raw material.

Here is what the numbers do not tell you. Moniepoint’s public disclosures do not reveal loan vintages, default rates or whether repayments are automatically swept from merchant receipts. For an equity investor, its reported scale and profitability are relevant. For a credit investor, those missing answers are material.

This is where most informal-economy investment pitches stop looking. They see millions of users and assume the lending layer follows naturally. It does not.

Equity investors can fund growth before this question is settled because their return depends on the company becoming more valuable. Credit investors lend money that must return on a timetable. Before they fund a repayment loop, they need to know exactly how it captures cash, what stops a customer bypassing it and what happens when the loop breaks.

Data is not collateral

MercadoLibre, the Argentina-founded Latin American marketplace and payments company, offers the clearest warning against that assumption.

It uses marketplace sales, Mercado Pago payment history and internal repayment behaviour in its credit models. Its 2025 annual filing recorded a gross merchant portfolio of $2.009 billion at December 31, 2025, against a $747 million allowance for expected losses. Its broader credit expansion absorbed more than $6.5 billion of cash during 2025, partly financed by approximately $2.4 billion of third-party fintech funding. In Mexico, it also executed a true sale of up to $100 million of loans while retaining collection servicing.

MercadoLibre shows how marketplace and payment data can improve underwriting without removing credit risk. Photo: Fernando Martello / Wikimedia Commons, CC BY-SA 4.0.

The system is powerful because it sees commerce and payment behaviour together, then stays involved in servicing. But even a large ecosystem does not make lending capital-light or loss-free. More lending consumes funding. More lending requires reserves for customers who do not repay. There is no app-shaped escape hatch from that arithmetic.

Stone, the Brazilian payments company that rebuilt its credit operation around registered receivables and financial-institution licences, shows the legal edge of the problem. Its 2025 annual filing recorded a credit portfolio of R$2.84 billion. Loans more than 90 days overdue had increased to 5.21%, from 3.61% in 2024.

Stone’s experience cuts through a common misunderstanding. Processing a merchant’s payments does not automatically give a lender priority over those payments. A merchant can route sales elsewhere. Another claimant may have a better legal right to the receivable. If that happens, the lender has observed cash flow without controlling it.

Observation is not collateral.

For businesses building African merchant payments, that distinction should shape product design before a lending product is launched, not after losses begin to climb.

Indonesia’s lesson: the loop can be human

Not every control loop is digital.

Amartha, founded by Andi Taufan Garuda Putra in Indonesia in 2010, started with Grameen-style groups of rural women. Its model used weekly engagement, joint responsibility and field servicing. During its first five years, Amartha reported a zero NPL rate, meaning no reported loans were more than 90 days overdue. It later became a peer-to-peer platform supervised by OJK, Indonesia’s financial-services regulator.

Amartha’s own platform information says it has disbursed more than Rp35 trillion to over 3.3 million microbusinesses. Those are company-reported figures and should be treated as such. The research brief also refers to an unnamed 2024 regional case study that reported a 98.24% TKB90, a measure of loans settled within 90 days. But the brief provides no publisher, title or URL for that case study. The figure therefore cannot be independently verified from the permitted evidence and is not used here as a basis for an investment conclusion.

Amartha’s field workflow demonstrates that human servicing can be a repayment control mechanism. Photo: Nasrulloh jatim / Wikimedia Commons, CC BY-SA 4.0.

At first glance, this seems like a victory for technology replacing the expensive routines of traditional microfinance. It is almost the opposite. The repeated group meeting and field visit were part of the control system. They created information, social pressure and a practical path to collection.

That system has a cost. And cost matters because a repayment rate does not tell an investor whether a model can scale profitably.

Bank Jatim, an Indonesian bank, recorded a 41% reduction in Amartha-linked accounts after it stopped that particular lending product in 2024. The bank did not specify why it stopped. Still, the episode is a useful restraint on easy conclusions: strong repayment statistics do not guarantee that every funding partnership remains attractive or scalable.

High-touch control can substitute for conventional collateral. It cannot be treated as free.

When a law broke the loop

The sharpest counterexample comes from Andhra Pradesh, India.

Before the October 2010 Andhra Pradesh microfinance ordinance, Indian microfinance institutions relied heavily on weekly group meetings and near-universal repayment. The ordinance restricted field collection and changed repayment frequency.

Collections in the state reportedly fell from about 99% to below 20%.

SKS Microfinance, which had 28% of its portfolio in Andhra Pradesh when the ordinance arrived, recorded state-wide MFI disbursements collapsing from Rs5,035 crore in the first half of fiscal 2011 to Rs8.5 crore in the second. Its provisions and write-offs rose 360% year on year.

The surprise was not that borrowers suddenly became different people. The collection workflow changed, and the old credit performance went with it.

Put Kenya, Nigeria, Indonesia, Latin America and India side by side and a pattern appears that none of the reports states quite this plainly: repayment behaviour is often produced by the system around the borrower. A disabled phone, an active merchant-payment rail, a weekly group meeting and legal priority over receivables all make payment more likely. Remove the system and yesterday’s repayment record can become a misleading comfort.

That is why a historical default rate, on its own, is not enough.

Where the argument breaks

Control loops are not magic. They can be bypassed, regulated away or made unprofitable by the cost of maintaining them.

M-KOPA’s device control depends on a customer continuing to value access to the phone. Amartha’s field model must bear the cost of people and repeated servicing. Stone shows that legal priority matters. SKS shows that a regulator can interrupt a collection practice. MercadoLibre shows that integrated data does not eliminate funding needs or credit losses.

The right conclusion is not that every closed system is investable. It is narrower, and more useful: without a real loop, informal-market credit is usually difficult to underwrite; with one, investors still need to test whether the loop survives competition, law and cost.

What founders should bring before asking for money

A first-time founder asking where to find investors for a Ghanaian, Nigerian, Kenyan or South African business often starts with geography: local angels, regional funds, global venture capital, private equity or development-finance institutions.

Start one step earlier. Match the capital provider to the evidence your model can honestly produce.

If the business is early and has not yet proved repeat transactions, its case is largely an equity case: demonstrate customer behaviour, retention and whether customers will keep using the workflow. If the company wants inventory finance, asset finance or a larger lending facility, it needs a credit case: show where money is intercepted, who has the legal right to it, what happens on default and whether collections remain profitable after acquisition, logistics and servicing costs.

Do not lead with the size of the informal economy. Lead with a cohort of merchants or customers whose transaction frequency, repeat purchasing, repayment and retention can be traced over time. Show the working-capital duration. Show losses and recoveries by vintage, meaning by the group of loans issued in a particular period. Show what happens if customers use cash, another payment provider, another supplier or another SIM.

This is related to the problem explored in why lenders can love your customers and still reject your business. Demand is not control. A good customer base is not automatically a bankable receivable.

GI Network's view: The investable opportunity is not “the informal economy”. It is the company that can prove, contractually and operationally, that value cannot quietly leave its repayment loop.

GI Network would begin by mapping the precise control point in the model: payment routing, asset access, inventory release, a collection mandate or a serviced group workflow. We would test the cash flows and unit economics under bypass and regulatory scenarios, identify the evidence missing from the data room, then structure the capital case around the investors equipped for that risk, whether equity, private credit or blended capital. Before outreach, we would rehearse the investment-committee objection that matters most: “What, exactly, prevents this cash flow from disappearing?” That is often the question behind a stalled raise, as the deal that dies in the IC memo explains.

The five-part loop test

A founder and an investor can use one simple tool before talking about market size: the LOOP test.

L is for location. Where exactly does the company touch the transaction: the payment, the stock, the device, the income stream or the weekly meeting?

O is for observation. Does that touchpoint generate frequent, reliable records of trading and repayment, rather than app downloads or one-off sign-ups?

O is for ownership of rights. Can the company enforce title, assignment, a payment mandate, receivables priority or asset access if the customer stops paying?

P is for profitability. After customer acquisition, logistics, servicing, losses and the cost of local-currency funding, does the loop make money?

P is for pressure test. What breaks if the customer pays cash, changes supplier, switches acquirer, uses another SIM or faces a regulatory change?

If those five answers are clear, informal activity can become financeable cash flow. If they are vague, the business may still be useful, popular and growing. It is simply not yet ready to ask institutional capital to treat that activity as security.