Africa’s missing dollars are stuck in Europe’s pipes.

Every working day, importers in Nairobi, Lagos, and Accra pay for goods made in Guangzhou, Ningbo, and Shenzhen. Trade between China and Africa reached a record $348bn in 2025, up 17.7 percent in a year, according to Chinese customs data. For a decade the IMF, the rating agencies, and Africa’s own central banks have explained the friction in that trade with a single diagnosis: a chronic shortage of dollars.

The continent’s accounts do not fit the diagnosis. African Export-Import Bank (Afreximbank) puts Africa’s exports at $682bn in 2024 and its imports at $719bn, a deficit of $37bn, or 2.6 percent of total trade, down from $80bn the year before. Africa sold the European Union €23.3bn more than it bought from it in 2024, by Eurostat’s count. Against China it ran a deficit of $102bn in 2025. The continent earns in Europe and spends in Asia, and its money has no fast road between the two.

The export proceeds exist. They sit in correspondent accounts in Frankfurt, London, and Paris, split across dozens of African banks, each with its own relationships and none with a pool to draw on. Much of what passes for a dollar shortage is a plumbing failure. The money is there, scattered across too many accounts and too slow to move, while a supplier in Guangzhou waits for it to be found, screened, and sent.

How chief executives, banks and regulators fix that plumbing will decide whether African enterprise stays in low-velocity, high-margin survival or competes at industrial scale.

Part I · The architecture of fragmentation

China’s trade with the 10 members of ASEAN (Association of Southeast Asian Nations) passed $1tn for the first time in 2025, reaching $1.05tn, according to China’s commerce ministry. In 2010 that trade was worth $293bn, while China’s trade with Africa was about $127bn. In 15 years the gap between the two corridors has grown from about $166bn to about $700bn. South-east Asia’s buyers and suppliers share time zones, banks, and a clearing center with China. Africa’s corridor shares none of the three.

The imbalance with China is widening fast. Chinese exports to Africa rose 25.8 percent in 2025 to $225bn, while imports from Africa rose 5.4 percent to $123bn, and the deficit grew by 64.5 percent from about $62bn in 2024. Solar panels, transformers, construction machinery and cars carried much of the surge. Every one of those invoices has to be paid, and most are paid in dollars.

Hong Kong shows what a pool looks like. Since 2000 its US dollar clearing system, USD CHATS, has settled dollar payments between banks in the city in real time during the Hong Kong business day, with links to the Hong Kong dollar, euro, and renminbi systems and to the settlement systems of Malaysia, Indonesia and Thailand. A dollar received by one bank in Hong Kong can pay a supplier who banks at another the same afternoon, and payments running in opposite directions offset inside one system. Africa has built the Pan-African Payment and Settlement System, PAPSS, for its own currencies. It has no equivalent pool for the dollars it earns.

Take a bank in Nairobi or Lagos holding $100m from European buyers in a correspondent account in Frankfurt. A manufacturing client asks it to pay $20m to a supplier in Guangzhou. Unless both banks sit inside an offshore system such as Hong Kong’s, dollars between banks settle in New York, so the instruction leaves Nairobi for Frankfurt, crosses to the correspondent’s dollar account in New York, moves to the Chinese bank’s correspondent, often through Hong Kong, and reaches Guangzhou several time zones and compliance desks later. Each hop screens the payment, and each can deduct a fee. A compliance officer in Frankfurt sees a Kenyan payer, a Chinese payee, and an invoice for machinery, and asks for the bill of lading, the end user and the beneficial owner. The payment waits while the answers travel back up the same chain.

The same trade can clear in Hong Kong and be held in Frankfurt. The difference lies in legibility. Banks in Asia see Africa’s trade with China every day and know what a normal invoice looks like. Compliance teams in Europe see little of it, and they treat what they cannot read as risk. Fewer banks are willing to read it at all. Africa lost 35.6 percent of its active US dollar correspondent relationships between 2011 and 2020, according to the Committee on Payments and Market Infrastructures at the Bank for International Settlements. The Financial Stability Board’s 2025 progress report found sub-Saharan Africa the slowest region in the world for receiving wholesale cross-border payments.

The continent earns in Europe and spends in Asia, and its money has no fast road between the two.

Part II · Velocity over margin.

Chinese manufacturers and traders commonly work on margins in single digits and turn their working capital every few weeks. Many African importers price far higher, because their money moves slowly and every slow month has to be paid for.

Profit on a pool of working capital equals the margin earned each time the money goes round, multiplied by the number of times it goes round in 365 days. Take two illustrative businesses, each with $1m. A Nairobi importer prices at a 35 percent margin and turns its capital twice a year. A Guangzhou trader prices at 5 percent and turns it monthly. The importer earns 70 percent on its capital and the trader 60 percent. If the importer’s money costs 15 percent a year and the trader’s 3 percent, both finish close to 55 percent.

The returns are similar. The prices are 30 percentage points apart. Velocity buys a company the right to be cheap, and the cheaper seller takes the volume, the shelf space and in time the market. A slow importer has to be expensive to survive.

Velocity buys a company the right to be cheap.

Payment friction sets much of that speed. Suppose the physical cycle of an import, from deposit through production, the voyage to Mombasa or Lagos, port clearance, sale and collection, takes 90 days. Capital then turns about four times a year. If dollar sourcing, letters of credit and compliance holds add 30 days, it turns three times, and a quarter of the year’s profit disappears. If they add 90 days, turns halve to two. At a 5 percent margin, four turns earn 20 percent and two turns earn 10 percent. To get back to 20 percent on two turns the importer must charge 10 percent, and every month of friction ends up in the shelf price.

Faster payment lifts turns only when the rest of the cycle can use the time saved. A trader who pays a supplier in minutes still waits weeks for the ship. The payment leg is the part of the cycle a treasurer controls, and in many African import businesses it is among the longest.

The second drag sits on the balance sheet. Payout businesses, remittance operators and airlines that must deliver money on weekends and public holidays keep local currency waiting in every market where they promise to pay. That pre-funded float is capital sitting in someone else’s bank account, earning nothing. A business paying out $5m a day in each of six markets, which must cover Friday evening to Monday morning, keeps at least $90m idle before any buffer for public holidays. At a 10 percent cost of capital that float costs $9m a year, and it carries the exchange-rate risk of six currencies.

Part III · The new settlement stack.

Dollar stablecoins, tokens backed one for one by cash and short-dated US Treasury bills, have moved from speculation into trade settlement because their cost has collapsed. A few years ago, converting shillings or naira into a stablecoin and back out into another currency cost about 2 percent at each end, or 4 percent for the round trip, too much for trade on thin margins. One operator serving institutional clients in Kenya now puts the cost at 20 to 30 basis points a leg, about 50 basis points end to end. On a $20m invoice the round trip has fallen from about $800,000 to about $100,000. Chainalysis estimates that sub-Saharan Africa received more than $205bn in on-chain value in the year to June 2025, up about 52 percent, driven largely by stablecoins and bitcoin used for payments and savings.

The first product built on that cost curve is known in the industry as the stablecoin sandwich. A company pays shillings or naira into a local account through a web-based treasury portal. The provider converts the money into a dollar stablecoin, moves it on-chain in minutes, and a partner at the far end converts it into renminbi, dollars or euros and pays the supplier’s bank account, with the know-your-business file, invoice, and shipping documents attached. The finance director never holds a private key or a crypto wallet. She sees shillings leave one account and renminbi arrive in another. Conduit, one of the larger providers, which counts the pan-African network Onafriq among its clients, expects Africa to become its largest market by the end of 2026.

The second product attacks the float. A payout business that can top up its balance in stablecoins at any hour, with the money usable within minutes, no longer has to park days of local currency in every market before the weekend. The furthest step, funding each payout at the moment it is triggered, removes most of the float altogether. Eversend , a payout platform operating across East and West Africa, has taken its own treasury off SWIFT and is moving in that direction. The same rails give exporters, tour operators and freelancers dollar, sterling and euro account details for collecting from customers abroad, settled into one balance.

The third product bridges African currencies to each other. PAPSS lets a bank in one African country pay a bank in another in local currencies, with Afreximbank as settlement agent, and payment networks such as Onafriq have begun connecting mobile money wallets and bank accounts to it. The constraint is liquidity. “Technology is ahead of liquidity,” Raj Kamal, chief executive of the payments firm TransFi, told Daily Maverick in October. A Nigerian payment to Tanzania can settle in real time only if someone holds the Tanzanian shillings to pay it out. Where a currency pair has no market of its own, a stablecoin can stand in the middle.

Part IV · The tier-2 banking unlock.

The banks that hurt most by de-risking are the ones closest to Africa’s traders. Mid-sized and small lenders finance the truck fleets, discount the invoices and know which distributor pays on time. When their client needs to pay Guangzhou, many must route the payment through nested correspondent relationships or through a larger domestic rival, which then sees their client’s trade. The client learns that the bank that knows it best cannot pay its suppliers quickly, and moves its trade business to a bank that can.

The new settlement stack offers these banks a way back. Infrastructure firms now sell banks the whole engine, conversion, on-chain settlement, custody and the compliance data package, under the bank’s own brand, so a tier-2 lender can offer the same speed to Shenzhen as its largest competitor. The price of entry is institutional. Banks expect regulated custody, SOC 2 or ISO 27001 security audits and rigorous checks on every business counterparty before they connect, and compliance sets the ceiling on how fast any provider can grow.

Banks of every size have begun to move. Credit Bank, one of Kenya’s smaller lenders, agreed in April to explore embedding a dollar stablecoin in its services, subject to engagement with the Central Bank of Kenya. Access Bank South Africa began exploring stablecoin settlement for African businesses trading with Asia in August. At the top of the market, Absa launched institutional digital-asset custody for South African clients at the end of September, the first African bank to do so, and it has said that extending the service to other African markets is now mainly a matter of licences. Custody at the large banks and settlement partners for the smaller ones, taken together, could give an exporter in Eldoret the same speed to Shenzhen that a multinational enjoys.

Part V · The sovereign frame.

Central banks tried to answer the speed problem with their own digital money, and most of the attempts have stalled. In any given week 98.5 percent of Nigeria’s eNaira wallets went unused, the IMF reported in May 2023. Brazil’s central bank decided in August 2025 to launch the first phase of its Drex project without blockchain. China turned the e-CNY into an interest-bearing deposit from January 2026 to lift adoption. The BIS left Project mBridge, the multi-central-bank platform it had helped to build, in October 2024. The Central Bank of Kenya concluded in 2023 that a CBDC “may not be a compelling priority in the short to medium term.”

Privately issued, regulated money has moved faster. The United States passed the GENIUS Act in July 2025, giving payment stablecoins a federal framework built on full reserves. J.P. Morgan put its deposit token on a public blockchain for institutional clients in November 2025. In April the Hong Kong Monetary Authority granted its first stablecoin issuer licences, to HSBC and to a joint venture led by Standard Chartered .

That speed carries a risk for African banks. A dollar stablecoin’s reserves sit in US Treasury bills and US bank deposits. If African companies and households move their savings from local deposits into dollar tokens, the deposits that fund local-currency lending shrink, and the savings end up financing Washington. A continent short of domestic credit cannot afford to export its deposits by default.

The counter-move is a regulated local-currency stablecoin. Nigeria’s cNGN, a naira token issued by a consortium under the Securities and Exchange Commission’s regulatory incubation program, began trading in February 2025. Kenya’s Virtual Asset Service Providers Regulations, gazetted in July 2026, make stablecoin issuance a licensed activity with reserve, redemption-at-par, audit requirements, and paid-up capital of KSh 300m. Kenya’s e-money issuers already hold customer funds in trust at banks, so a shilling token would be M-Pesa’s float written on a different ledger. Dollar tokens can carry trade with the rest of the world, while shilling and naira tokens keep savings, and the credit they fund, at home.

Part VI · Four reforms.

The market has built most of the rails. Four rules stand between them and the average African company.

Accounting. IFRS has no standard written for stablecoins. Guidance from 2019 places cryptocurrencies that carry no contractual right to cash under inventory or intangible-asset accounting, which leaves a fiat-backed stablecoin with a direct redemption right in an awkward space. America’s Financial Accounting Standards Board proposed in August 2026 to show when certain stablecoins can count as cash equivalents. A conservative board will not move treasury onto rails its auditors cannot classify. A practice note from ICPAK and its regional peers on classification, measurement and audit evidence would let a reserve attestation and a wallet confirmation stand beside a bank statement.

Banking access. A Central Bank of Kenya circular from 2015 warned banks against dealing in virtual currencies and limited the sector’s access to formal banking. Kenya’s new regime requires licence applicants to keep a local banking arrangement, which a bank still guided by the 2015 circular may hesitate to provide. Firms that cannot open accounts push their flows towards peer-to-peer channels that supervisors see less clearly. A joint circular from the CBK and the Capital Markets Authority setting risk-based account-opening standards for licensed providers, with written reasons for refusals, would bring that money back into view.

E-money parity. A licensed e-money issuer in Kenya already ring-fences customer funds in trust at strong-rated banks and may not lend or invest them. Those are the protections a stablecoin regime asks for. A fast track for licensed payment providers to issue shilling tokens, with their existing trust arrangements recognized, would put local-currency stablecoins in the hands of firms that already serve millions.

Two-way flow. Nigeria’s money transfer rules of January 2024 limit licensed operators to inbound transfers and bar outbound ones. An operator that can only receive cannot net a diaspora inflow against an importer’s outflow, so every outbound dollar has to be bought in the market and spreads stay wide. Licences that allow money to move both ways are the cheapest liquidity a regulator can create.

Part VII · The verdict.

Some companies will spend the next decade asking central bank governors for preferential allocations of foreign exchange. Some banks will defend correspondent fees with the vocabulary of compliance. The executives who win will count settlement days with the care they give margin points, because the days decide the price they can charge.

On Monday morning that work has three parts. A finance director can trace the last 10 supplier payments from invoice to the supplier’s receipt, and count the days, the intermediary deductions and the spread on each, corridor by corridor, in basis points.

She can ask the company’s bank and payment partners for an on-chain settlement option that carries the full compliance pack, and run one corridor through it for a quarter. And the board can back the push for accounting guidance and local-currency stablecoin rules, because no single company can write those rules alone.

The importers in Nairobi, Lagos and Accra already price every week of delay into their goods. Their suppliers in Guangzhou price in days. The gap between those two clocks is the cost of Africa’s plumbing. The dollars were never missing. They have been waiting in Frankfurt for a faster road to Guangzhou.