Samuel M

Kenya’s National Payment System Bill prices the multi-rail model directly. Section 10(6): full minimum capital for your highest licence category, plus 50% of the minimum for every additional one. Card acquiring plus wallet plus remittance is KSh 90m before you touch e-money. Section 36(5): trust balances must never fall below what’s owed to customers. Not at the close. Never. You can’t evidence that without knowing daily what you owe and what you hold, by currency, across every provider. Fourth Schedule caps any one bank at KSh 500m or 25% of trust monies. So trust balances split across banks, and the daily customer money position has to be assembled across all of them. One year to comply.

A Montreal-adjacent story keeps repeating itself across African fintech announcements: a company launches with a big coverage map, and the map is real, but the depth behind it is concentrated in the two or three currencies that already had liquidity before the company existed. One live example right now: a stablecoin settlement platform recently announced African infrastructure running on Aptos, through a partnership with HashKey MENA, positioning it as the African node in a network that already connects Hong Kong, the Philippines, Vietnam, and the UAE. The real story underneath the announcement isn't "African stablecoin infrastructure gets built." It's which existing Asia-Middle East rail network wins the plug-in slot for Africa first, and what happens to pricing once two or three networks are competing for the same node. For anyone evaluating settlement partners in this space, the coverage map on day one tells you almost nothing. What matters is whether the currencies that were shallow before the partnership are still shallow twelve months later.

WAEMU's PI-SPI interbank scheme carries a mandatory connection deadline of 30 September, already extended once. Eighty institutions across eight countries are connected. Thirty million users are enrolled. Individual transfers are free by design. Those numbers describe adoption. They say nothing about whether connecting is actually earning its cost for any single institution, or being carried as a compliance expense with no revenue path attached. Free transfers mean the business case has to come from somewhere else entirely, float income, cross-subsidy, a longer positioning bet. None of that shows up in the public rollout figures. It shows up in each institution's own books, which nobody publishes. The institutions in the best position after 30 September won't be the ones that connected fastest. They'll be the ones that connected with an actual answer to where the revenue comes from. If your institution is facing this deadline and that answer isn't written down yet, that's the conversation worth having before the date forces the decision for you.

Banca d'Italia sent real $200 USDC transfers across ten corridors to test the stablecoin cost claim properly. The blockchain leg cost 0.4% and settled in minutes. Total costs still ranged up to 9%, because everything before and after the chain, buying in, converting out, still runs through the same banking rails it always did. The token isn't the bottleneck. It never was.

West Africa solved the wrong half of cross-border money movement, and this thread from Cotonou names exactly which half. A Beninese trader can move XOF into naira and back in seconds through mobile money. What that same trader cannot do is take savings and buy a Nigerian treasury bill, or hold a reconciled position on the Nigerian exchange, or access anything resembling a cross-border investment account. Payments got solved. Capital did not. The instinct is to point at PAPSS or the African Exchanges Linkage Project and say the infrastructure already exists. It's worth being precise about what AELP actually is, because the precision matters here. AELP has been live since 2022, connecting brokers across BRVM, NGX, JSE, NSE, EGX and several other exchanges. But by its own design, the sponsoring broker clears and settles every trade in the host market's local currency, under that market's own rules. That means AELP assumes the investor already has a way to convert into naira or CFA before the trade even happens. It was built to let brokers see and execute across each other's order books. It was never built to solve currency convertibility for a retail saver in Cotonou. Four years live, and the exact gap this essay describes hasn't moved, because AELP was never the layer designed to close it. That's not a criticism of the project, it's a precise description of what problem it was actually built to solve, and what problem is still sitting untouched next to it. The retail cross-border capital rail this essay is asking for, a Beninese saver holding a reconciled Nigerian T-bill position, needs something structurally different: an identity layer that both regulators recognize, a settlement account that doesn't require choosing one country's currency over the other, and custody that survives a transaction crossing a currency union border. None of the infrastructure currently celebrated as "Africa's cross-border payments success story" was built with that in mind. It's worth reading the full piece. The frictions named here, CFA-naira FX regimes, OHADA versus common law, no ECOWAS-wide identity equivalent to Nigeria's BVN, are the actual list anyone building this next layer needs to solve, in order.

Uber suspended Visa in Kenya for eight months. Mastercard, facing the same stated cost pressures, never stopped working. The difference wasn't the network, it was the settlement path. Visa ran through an offshore merchant-of-record structure with FX spread and interchange stacked on top. Same market, same riders, different routing, different outcome. The rail isn't the variable. The settlement architecture behind it is.

Ghana's credit bureaus have operated under a 2007 act, supervised by the central bank, since before mobile money existed at scale. Tax administration runs under a completely separate 2009 act, under a different regulator entirely. No statutory bridge connects the two. The practical result: years of reliable, verifiable payment history sitting inside a country's own tax and mobile-money records never reaches the underwriting system that could turn it into credit access. A market woman who's been paid reliably through mobile money for a decade still has no file a bank, a landlord, or a lender will recognise. This isn't unique to Ghana. It's a pattern worth checking anywhere an institution is building an alternative-data lending strategy on mobile-money history, presumptive-tax records, or utility payments. Data existing and being collected reliably doesn't mean it's legally reachable by whoever wants to underwrite against it. Worth confirming which side of that gap your own data strategy actually sits on.

Zambia's Currency Directives, in force since December, require domestic transactions to settle in Kwacha. Most coverage stops at the headline rule. Fewer people have read far enough to find the personal liability clause: directors and senior management who authorise a breach face up to two years imprisonment, alongside institutional fines. Then there's the mining exemption, which sounds like a simple carve-out until you read the condition attached to it. A mining transaction only qualifies for foreign-currency settlement if the payer remits taxes to the Zambia Revenue Authority through selling that same foreign currency to the Bank of Zambia first. The exemption only holds if the dollars flow through the central bank first. Get that sequencing wrong, and a transaction that looked exempt on paper becomes a personal criminal exposure question. If your institution has Zambian counterparties, suppliers, or subsidiaries, the operative question isn't whether the directives apply to you. It's whether every foreign-currency transaction in your current contract book has actually been tested against that exemption schedule, and whether whoever signs off on payment knows which exemptions are conditional.

Ghana’s fintech stack is getting denser: passporting, open banking, dedicated EMI licensing, and digital trade/payment credentials all advanced this week. The strategic question is no longer only “are we licensed?” But “can permission, data, settlement and evidence operate as one controlled model?”

Ghana's central bank ran a gold-buying operation for years that quietly lost money, 9 billion cedis in the most recent reported year alone. The IMF eventually flagged it as a threat to central bank independence, a quasi-fiscal activity dressed up as monetary operations, and pushed for the function to move to a separate state entity, GoldBod, earlier this year. What's happening now is the more interesting part. GoldBod is actively trying to convert gold to dollars without routing through

Standard Chartered has confirmed correspondent banking exits across several African markets over the past two years, Botswana, Cameroon, Gambia, Sierra Leone, Zimbabwe. In the same window, SC Ventures, the bank's own venture arm, led a funding round into licensed African stablecoin settlement infrastructure. One arm of the same institution closing a relationship it no longer wants to carry on its own balance sheet. The other arm quietly capitalizing the rail built to take its place. That's not an exit from these markets. It's a bet on a different architecture replacing the one being wound down. For any institution still relying on the correspondent relationships currently being closed, the useful diligence question isn't whether access is at risk. It's whether the alternative rail being built to replace it is one you've actually evaluated, or one you'll be forced onto without ever having tested it.

There's a real number behind the phrase "prefunding is a silent tax." A payments company moving money across Nigeria, Ghana, and Kenya can't send it the moment a sender initiates. It has to already be holding local currency in each market, waiting, so the payout is instant on the receiving end. Multiply that across every corridor, and a growing company ends up running twenty-plus local accounts, each one holding capital that earns nothing. Just-in-time stablecoin credit has genuinely solved half of this. A company can borrow dollars for a moment instead of holding them idle for months. But that only fixes the sender's side. Someone still needs cedi sitting in Accra, ready, at four o'clock on a Friday. Stablecoin credit doesn't conjure that into existence. What actually happens when half the problem gets solved is that the tax doesn't disappear. It moves. From the currency everyone already has deep markets for, to the currencies where liquidity is thinnest. If your growth plan assumes prefunding is a solved problem, worth checking which half of your corridor map that assumption is actually true for.

In August, a Nigerian trading firm's currency swap fell apart in the way these things usually do quietly. Eight million dollars agreed, matched, and moving. Then one side didn't fully deliver. $4.2 million stayed outstanding. Sixty three accounts got frozen while the dispute worked itself out. The platform that matched the trade wasn't the problem. Matching two willing counterparties at an agreed rate is the easy part of any FX marketplace. What failed was everything after the match, the assumption that both sides would deliver what they'd promised, with nothing forcing that to be true. A CLOB prices a trade. It doesn't guarantee it. That guarantee needs a default fund, prefunded margin, or a counterparty willing to stand behind both legs. Most African FX platforms selling "instant settlement" right now don't have that layer built. If your treasury team is routing meaningful size through one of them, the honest question isn't whether the platform can match your trade. It's what happens on the day the other side doesn't deliver.

Kenya's US remittance share fell below 50% for the first time since CBK started publishing corridor-level data in 2019, down to 48.24% from 53.51% a year earlier, as US inflows dropped 12.6% while the UAE, UK, and Australia partially absorbed the shortfall. CBK has cut its full-year growth forecast to 0.7%, which would be the weakest expansion in over 15 years. Worth adding a detail most coverage of this story leaves out. A separate KNBS household survey found that total diaspora inflows, including informal channels, were 43% larger than CBK's formal figure over the same period. If that gap holds, the entire "US remittances are declining, growth has stalled" narrative describes a shrinking slice of a considerably bigger picture than the official data captures, not necessarily a shrinking total. For any institution pricing Kenya's FX outlook, remittance-backed lending, or diaspora-focused product strategy off the CBK print alone, The formal number may be measuring corridor-mix shift, US senders moving toward informal channels as formal costs or friction rise, more than an actual contraction in diaspora support. Those are very different planning assumptions, and only one of them shows up in the headline data.

A Dutch foundation just won the right to collect a debt in Kenya. Winning is not the same as collecting. In October 2016, a Netherlands-incorporated entity called Stichting Rabobank Foundation agreed to provide financial support of roughly $180,000 to a company called AVA Chem Limited, operating in Kenya. As is standard in cross-border lending of this size, the arrangement came with a personal guarantee. A director of AVA Chem, a man named Christopher Irungu Mwangi, signed a deed of suretyship the following year, putting his own name behind the company's obligation to repay. It was, on paper, an unremarkable piece of trade and development finance, the kind of transaction that happens hundreds of times a year between European lenders and African borrowers, usually without anyone outside the two parties ever hearing about it. AVA Chem defaulted. Mwangi, by the Foundation's account, acknowledged the debt and agreed he would honour the guarantee if the company couldn't pay. When the money still didn't come, the Foundation did what any lender in its position would do: in September 2022, it filed suit in the High Court in Nairobi, seeking recovery of just over $230,000, the original amount plus accrued interest and costs. What happened next is where this stops being an ordinary debt collection story and becomes something worth every cross-border lender's attention. AVA Chem's defense wasn't that the debt didn't exist. Nobody disputed that money had changed hands, that a guarantee had been signed, that repayment hadn't happened. Instead, the company's lawyers reached for a different argument entirely, a procedural one, buried in Kenya's Companies Act. Section 974 requires any foreign company "carrying on business" in Kenya to register locally. Stichting Rabobank Foundation, being a Dutch entity that had never registered in Kenya, hadn't done that. AVA Chem's argument was elegant in its bluntness: if the Foundation had never properly registered, it had no legal standing to sue in a Kenyan court at all. Not that it should lose the case. That it shouldn't be allowed to bring the case in the first place. In July 2024, the High Court agreed. The judge accepted that extending a cross-border loan to a Kenyan company amounted to "carrying on business" under the statute, found that the Foundation lacked the standing to sue, and struck out the entire claim. A legitimate, undisputed debt of $230,000 evaporated, not because the borrower proved it didn't owe the money, but because the lender had never filled out the right paperwork with the Kenyan registrar of companies. For any foreign institution that had ever lent money into Kenya without registering there first, and that describes a great many development finance institutions, trade financiers, and cross-border lenders, this ruling was the kind of thing that keeps a general counsel up at night. The Foundation appealed. And on the 31st of July this year, nearly a decade after the original loan agreement was signed, the Court of Appeal reversed the High Court entirely. The appellate judges drew a sharp distinction the lower court had missed: registering as a foreign company under section 974 is a regulatory requirement, not a condition of legal existence. A company incorporated in the Netherlands does not cease to be a legal person the moment it crosses into Kenya without paperwork. Registration and the right to sue, the court held, are two separate questions, and conflating them had been an error. The case was reinstated. Stichting Rabobank Foundation can now proceed to actually litigate whether AVA Chem and its guarantor owe the money. This is, genuinely, an important ruling, and the reaction from Kenya's legal and finance community has treated it as such. It removes a specific and previously real risk that any unregistered foreign lender could find their entire claim thrown out on a technicality, regardless of how clear-cut the underlying debt was. For development finance institutions, export credit agencies, and trade-finance desks structuring exposure into Kenya, this is a meaningful reduction in legal uncertainty. Advocates are right to be discussing it. Lenders are right to feel some relief. But there's a second half to this story that almost nobody discussing the ruling is asking about, and it's the half that actually determines whether the money comes back. Winning the right to sue is not the same as collecting a debt. Even now, with the Court of Appeal squarely on the Foundation's side, the practical journey from a favourable Kenyan court judgment to Dutch euros sitting back in the Foundation's own account runs through a completely different set of obstacles, ones that have nothing to do with company law and everything to do with how money actually moves across a border. Say the Foundation wins at trial. Say a Kenyan court orders AVA Chem, or Mwangi personally, to pay the outstanding amount. What happens then is not a wire transfer. It's a foreign-currency surrender process, because Kenya, like most African markets, has rules governing how and when funds get converted and released for cross-border payout. It's a question of which Kenyan bank account the recovered funds land in, and how quickly that institution is willing and able to move foreign currency out again, because local banks manage their own dollar and euro liquidity carefully and a large outbound payment doesn't clear instantly just because a court has ordered it. And it's a question of correspondent banking, because a Kenyan bank sending funds back to a Dutch foundation is routing that payment through the same web of correspondent relationships that has been getting slower, pricier, and more heavily scrutinized across African corridors for years, the exact friction that has pushed so much of the continent's payments innovation toward local rails and stablecoin alternatives in the first place. None of that shows up in a Court of Appeal judgment. None of it gets discussed in the legal commentary celebrating this ruling as a win for foreign lenders. And all of it can turn a legally airtight, court-confirmed debt into a recovery that takes months longer and costs more than anyone structuring the original guarantee expected. This is the part worth asking any DFI, trade-finance desk, or corporate treasury team financing exposure into Kenya right now: has the legal team celebrating a ruling like this actually mapped what happens after judgment. Sue in Nairobi, certainly, that battle just got considerably easier to win. But collect through what. A Kenyan correspondent bank with its own liquidity constraints and its own queue. A mobile-money payout rail, if the amount and structure even allow for one. A local account that has to hold the recovered funds before FX surrender rules let any of it leave the country. The legal risk in cross-border Kenyan lending just got meaningfully smaller. The settlement risk didn't move at all. And the institutions that will actually collect efficiently on the next decade of Kenya-facing debt won't just be the ones with the strongest legal position. They'll be the ones who structured their guarantees with an actual answer to the question nobody in this week's legal commentary is asking: once you've won, how does the money get home?

Kenya's US remittance share just fell below 50% for the first time in seven years of CBK data. What most coverage skips: a KNBS household survey found total diaspora inflows, including informal channels, run 43% higher than the official CBK figure. The formal number may be capturing a mix shift toward informal channels, not an actual decline in diaspora support

There is a particular kind of confidence that shows up in founder replies during a crisis in someone else's business, and it's worth paying close attention to, because it usually reveals more about a product's actual architecture than any pitch deck does. In mid-August, when the details of a broken $8 million currency swap started circulating, a swap where $4.2 million remained outstanding and sixty three accounts sat frozen while the dispute worked itself out, Michael Anyi's response to the story was two words. His platform solves this. No elaboration, no mechanism, just the confidence that a matching engine was the missing piece in a story that was actually about something else entirely. It's worth walking through what actually happened, because the gap between "solves this" and what would genuinely have prevented it is the whole argument. A currency swap of that size involves two parties agreeing to exchange one currency for another, at an agreed rate, with delivery expected on both sides. The trade itself, the agreement on price and size, is the easy part. Any reasonably built matching system can find two parties willing to trade $8 million at a given rate and record that they've agreed to do so. What that matching system cannot do, by itself, is guarantee that both parties will actually deliver what they promised. That guarantee requires something structurally different: a mechanism that holds both sides accountable simultaneously, so that currency A only changes hands at the exact moment currency B does, with no window where one party has delivered and the other hasn't. In financial infrastructure, this is called payment versus payment, and it typically requires either a trusted intermediary holding both legs in escrow, or a central counterparty willing to step in and guarantee settlement if one side fails, backed by a default fund built for exactly this scenario. None of that is what a central limit order book does. A CLOB is a screen. It shows prices, matches willing buyers and sellers, and produces a trade. What happens after the match, whether the naira actually lands and the dollars actually move, whether one side can walk away with $4.2 million and leave the other holding nothing, is a separate question entirely, and it's the question that actually determined the outcome of the story Anyi was replying to. This distinction matters more in African cross-border FX than in almost any other market, precisely because the infrastructure that G10 currencies take for granted, decades-old settlement systems with built-in guarantees, simply doesn't exist yet for most African currency pairs. When something goes wrong in a naira to dollar swap outside a properly guaranteed settlement system, there's no automatic backstop. There's a dispute, frozen accounts, months of negotiation, and in the worst cases, money that simply doesn't come back. The Plaude and Zacuten story wasn't a matching failure. It was a settlement failure, the exact category of risk that "we match trades" does nothing to address. So the honest version of "solves this" would sound different. It would specify whether the platform actually novates the trade, meaning it inserts itself as the legal counterparty to both sides so that each party is trading with the platform rather than with each other, removing the risk of one side defaulting on the other directly. It would specify whether there's a prefunded margin requirement before a trade of that size is even allowed to match, so that both parties have already put up collateral before the trade executes. And it would specify what happens, mechanically, on the day one side fails to deliver, whether there's a default waterfall, a fund built specifically to absorb exactly this kind of loss, or whether the platform is simply hoping this particular failure mode never happens to a trade it matched. That's not a hostile question. It's the question every serious institutional counterparty asks before routing real size through a new FX venue, and it's the question that separates infrastructure a corridor can actually trust from a well-designed screen that happens to work until the day it doesn't. A $4.2 million hole is not a hypothetical. It already happened once, in public, in the exact market this platform is trying to serve. The next one is a question of when, not if, and the platforms that survive that moment will be the ones that answered the settlement question before they needed to, not the ones still finding out what "solves this" actually meant when it mattered.

There is a particular kind of confidence that shows up in founder replies during a crisis in someone else's business, and it's worth paying close attention to, because it usually reveals more about a product's actual architecture than any pitch deck does. In mid-August, when the details of a broken $8 million currency swap started circulating, a swap where $4.2 million remained outstanding and sixty three accounts sat frozen while the dispute worked itself out, Michael Anyi's response to the story was two words. His platform solves this. No elaboration, no mechanism, just the confidence that a matching engine was the missing piece in a story that was actually about something else entirely. It's worth walking through what actually happened, because the gap between "solves this" and what would genuinely have prevented it is the whole argument. A currency swap of that size involves two parties agreeing to exchange one currency for another, at an agreed rate, with delivery expected on both sides. The trade itself, the agreement on price and size, is the easy part. Any reasonably built matching system can find two parties willing to trade $8 million at a given rate and record that they've agreed to do so. What that matching system cannot do, by itself, is guarantee that both parties will actually deliver what they promised. That guarantee requires something structurally different: a mechanism that holds both sides accountable simultaneously, so that currency A only changes hands at the exact moment currency B does, with no window where one party has delivered and the other hasn't. In financial infrastructure, this is called payment versus payment, and it typically requires either a trusted intermediary holding both legs in escrow, or a central counterparty willing to step in and guarantee settlement if one side fails, backed by a default fund built for exactly this scenario. None of that is what a central limit order book does. A CLOB is a screen. It shows prices, matches willing buyers and sellers, and produces a trade. What happens after the match, whether the naira actually lands and the dollars actually move, whether one side can walk away with $4.2 million and leave the other holding nothing, is a separate question entirely, and it's the question that actually determined the outcome of the story Anyi was replying to. This distinction matters more in African cross-border FX than in almost any other market, precisely because the infrastructure that G10 currencies take for granted, decades-old settlement systems with built-in guarantees, simply doesn't exist yet for most African currency pairs. When something goes wrong in a naira to dollar swap outside a properly guaranteed settlement system, there's no automatic backstop. There's a dispute, frozen accounts, months of negotiation, and in the worst cases, money that simply doesn't come back. The Plaude and Zacuten story wasn't a matching failure. It was a settlement failure, the exact category of risk that "we match trades" does nothing to address. So the honest version of "solves this" would sound different. It would specify whether the platform actually novates the trade, meaning it inserts itself as the legal counterparty to both sides so that each party is trading with the platform rather than with each other, removing the risk of one side defaulting on the other directly. It would specify whether there's a prefunded margin requirement before a trade of that size is even allowed to match, so that both parties have already put up collateral before the trade executes. And it would specify what happens, mechanically, on the day one side fails to deliver, whether there's a default waterfall, a fund built specifically to absorb exactly this kind of loss, or whether the platform is simply hoping this particular failure mode never happens to a trade it matched. That's not a hostile question. It's the question every serious institutional counterparty asks before routing real size through a new FX venue, and it's the question that separates infrastructure a corridor can actually trust from a well-designed screen that happens to work until the day it doesn't. A $4.2 million hole is not a hypothetical. It already happened once, in public, in the exact market this platform is trying to serve. The next one is a question of when, not if, and the platforms that survive that moment will be the ones that answered the settlement question before they needed to, not the ones still finding out what "solves this" actually meant when it mattered.

There's a phrase making its way through African fintech circles that deserves to stick around longer than most: prefunding as a silent tax. It's a good phrase because it's precisely accurate. Every business moving money across African borders today is, in effect, paying an invisible levy, not to a government, not to a bank in a line item anyone can point to, but to the structure of the system itself. That levy is capital. Specifically, it's capital sitting idle in accounts around the world, waiting to be used, earning nothing, existing purely because the alternative, moving money without it, isn't currently possible. Here's the mechanism, stated plainly. A payments company operating across, say, Nigeria, Ghana, and Kenya cannot simply move money from a Nigerian sender to a Kenyan recipient the moment the sender initiates the transfer. Instead, that company typically has to hold a working balance of Kenyan shillings in a Kenyan account in advance, so that when the Nigerian sender's money arrives, the company can immediately pay out shillings from its own pre-existing local balance, and separately settle the naira side later. Multiply that structure across every corridor a company operates in, and you get a business that is, in effect, running twenty or more local bank accounts around the world, each one holding capital that exists purely to make instant payouts possible, capital that could otherwise be lent, invested, or deployed toward growth. That's the tax. It's real, it's large, and it's the actual reason so many African cross-border payments companies struggle to scale despite strong transaction volume, because growth means opening more accounts and parking more idle capital, not less. This is precisely the problem a company like MANSA has built its business around solving, at least for one half of the equation. The pitch is elegant: instead of a payments company pre-funding a foreign account with its own capital, it borrows stablecoin credit just in time, uses it to cover the immediate payout, and repays once the underlying transaction settles on the other end. In principle, this removes the need to keep large dollar balances sitting idle, freeing up capital that would otherwise be locked in nostro accounts around the world. Sanoh's framing of the underlying problem, that Africa's payments bottleneck isn't a missing API but idle capital sitting in twenty different nostro accounts, is correct, and worth repeating more often than it currently is. But here's where the essay, and the broader excitement around solving prefunding, tends to stop half a step too early. Just-in-time credit on the dollar leg solves the sender's side of the problem. It does nothing, by itself, for the recipient's side. When a Nigerian sender's money needs to become Ghanaian cedi in a recipient's hands by four o'clock on a Friday afternoon, someone still has to have cedi sitting in a Ghanaian account, ready to disburse, at that exact moment. Stablecoin credit can eliminate the need to hold dollars idle waiting for that transaction. It cannot conjure cedi into existence in Accra in real time. That local payout leg still requires either a prefunded local account, held by someone, or an instantaneous local liquidity source that, in most African markets, simply doesn't exist yet at the depth required. What actually happens, in practice, when a company solves the dollar side of prefunding without solving the local side, is not that the tax disappears. It's that the tax changes currency. The idle capital that used to sit in a dollar account now needs to sit in a cedi account, or a shilling account, or a franc account, wherever the payout actually happens. The total amount of dead capital in the system may shrink somewhat, because just-in-time dollar credit is genuinely more efficient than a company holding its own idle dollar reserves. But the structural need for someone, somewhere, to hold local currency ready for instant disbursement doesn't go away. It just moves downstream, from the sender's side of the transaction to the recipient's side, from the currency everyone already has deep, liquid markets for, to the currencies where liquidity is thinnest and hardest to source. The genuinely interesting question, the one worth asking any company solving half of this problem, is what happens to that second half. Who is holding GHS, KES, or XOF ready for four o'clock on a Friday, and what does that capital cost them. Until that question has as clean an answer as the dollar-side pitch does, prefunding hasn't been solved. It's been relocated, to the exact corridors where solving it is hardest, and where the businesses trying to operate there are least equipped to absorb the cost of doing so themselves.

Zambia's Currency Directives, effective since December, carry personal criminal liability for directors and senior management who authorise a breach, up to two years imprisonment, alongside institutional fines. Most coverage of the directives focuses on the headline rule: domestic transactions settle in Kwacha. Fewer institutions have mapped the exemption schedule closely enough to know where their own transactions actually sit. The mining exemption is a useful example of how granular this gets. A mining-sector transaction only qualifies for foreign-currency settlement if the paying party remits taxes to the Zambia Revenue Authority through a sale of that same foreign currency to the Bank of Zambia. The exemption isn't unconditional, it's structurally tied to routing through the central bank. Get that sequencing wrong and a transaction that looked exempt on paper becomes a compliance breach with personal exposure attached. For any institution with Zambian counterparties, suppliers, or subsidiaries, the operative question isn't whether the directives apply. It's whether every foreign-currency-denominated transaction in the current contract book has been individually tested against the exemption schedule, and whether the people signing off on payment know which exemptions are conditional versus absolute.