The correspondent banking retreat from Africa

The correspondent banking system is invisible to almost everyone. It is the inter-bank plumbing that lets a Lagos importer pay a Hamburg exporter, that lets a Nairobi flower farm receive payment from the Aalsmeer auction, that lets a Casablanca importer settle in euros against a French capital-goods supplier. When it works, no one notices. When it does not work, the import order does not ship, the receivable does not get paid, and the corporate treasurer in either jurisdiction is told that the issue is "with the correspondent."

For roughly the last fifteen years, in much of Africa, the issue has been with the correspondent. The retreat of Western banks from African correspondent activity is the single largest reason that trade finance in the corridor has not scaled to match the underlying volume of trade. This note works through what happened, why, and where the architecture sits today.

What "correspondent banking" actually is

A correspondent banking relationship is an account that one bank maintains with another, typically in a different jurisdiction, to enable cross-border payments in the foreign bank's currency. A Kenyan bank that wants to pay a German exporter in EUR needs to either (a) hold its own EUR account with a German or European bank, or (b) use the EUR clearing of a third bank that does. That third bank is the correspondent.

Correspondent banking is the substrate. Trade finance — letters of credit, bills of exchange, documentary collections — runs on top of it. Without the correspondent, the trade finance product cannot settle.

What changed, and why

Between 2011 and 2020, the FSB tracked a roughly 25 per cent decline in the global number of active correspondent banking relationships. Africa saw a sharper contraction than that average. Some of the most visible exits were public — Citibank narrowing its Nigerian correspondent activity in 2021; Standard Chartered consolidating its African correspondent footprint over the same period; the prudential withdrawal of multiple US regional banks from any African counterparty exposure following the FATF re-grey-listing waves of 2018 to 2022.

The underlying drivers were two. The first was the post-2008 enforcement intensification on AML and sanctions screening, which raised the per-relationship compliance cost dramatically. Banks that had historically maintained dozens of African correspondent accounts at modest individual revenue found themselves running each at a loss after compliance allocation. The second was a narrowing of the regulatory tolerance for AML failure: a single missed flagged transaction in a correspondent chain became, post-HSBC, a material enforcement risk that bank boards declined to carry for marginal revenue.

The result was not a series of explicit refusals to bank a particular African country. It was a series of cost-allocation decisions taken inside Western banks that resulted in fewer African correspondent relationships, fewer banks willing to clear USD or EUR for tier-two African counterparties, and longer settlement chains for the relationships that remained.

The decisions were driven by AML and KYC cost, not by African counterparty risk in the underlying commercial sense. Most exited banks were not loss-making; they were unprofitable after compliance allocation.

Who picked up the slack, and where it does not work

Three structures have grown into the gap.

The first is the pan-African bank network. Ecobank, with its presence across thirty-three African markets, operates a hub-and-spoke USD clearing model through its Lomé treasury that allows it to net intra-African flows and reduce the number of external correspondent legs. United Bank for Africa runs a similar architecture out of Lagos. Standard Bank operates a comparable pan-African treasury through Johannesburg. For the volumes that can be intra-mediated through these banks, the correspondent retreat has been substantially mitigated.

The second is the European-subsidiary route. Many African banks of meaningful size operate London, Paris, or Brussels subsidiaries — UBA UK, Ecobank UK, Bank of Africa France — that hold direct membership of SEPA and access to EUR clearing. For EUR-denominated EU-Africa trade, this is increasingly the practical settlement route. It works less well for USD, where direct US correspondent access is harder to maintain for the same compliance reasons.

The third is the stablecoin overlay. The growth of USDT and USDC settlement in African corridors is not a marketing story; it is a substitution for the broken correspondent leg, particularly for SME-tier flows that the pan-African and European-subsidiary routes do not economically serve. Chainalysis data for 2024 to 2025 shows Nigeria at $92 billion in on-chain volume, with stablecoins as the dominant component. The flows are concentrated in import payments, supplier settlement, and treasury management — the activities that correspondent banking should be intermediating.

Where the architecture leaves holes

The structures above mitigate the correspondent retreat for most of the corridors that matter. They do not close it. Three categories of flow remain particularly difficult.

The first is SME flows in markets where the pan-African banks do not have meaningful retail presence. Senegal, Sierra Leone, Liberia, Guinea, Niger — markets where Ecobank or UBA presence exists but is concentrated in capital-city corporate banking — leave SME importers and exporters with the original correspondent problem unmitigated.

The second is USD-denominated flows for jurisdictions on or near FATF grey lists. Even where a pan-African bank can clear, the European or US bank standing behind the clearing chain may decline the underlying transaction if the originating jurisdiction is currently subject to enhanced due diligence. The hassle cost is back-end-loaded to the customer.

The third is flows from non-sanctioned but politically sensitive jurisdictions where Western banks have adopted bilateral risk-off positions. Mali, Burkina Faso, and parts of the Sahel since the political transitions of 2022 to 2024 sit in this category. Trade finance for these markets, even where the underlying commercial transaction is clean, runs into bilateral bank-level reticence that is not formally regulatory.

What an advisor actually does about it

In a corridor mandate, the first practical task is to map the live correspondent state. This is operational work, not strategic work; it means knowing which banks currently clear which currencies for which counterparties, and how long it takes. The published surveys — FSB, BIS, Wolfsberg — are useful for direction but lag the actual state by twelve to eighteen months. The mandates that work in 2026 are run with current information.

The second task is to structure around what does not work. Where direct correspondent clearing fails, route via the European subsidiary of a pan-African bank. Where that also fails, route via a confirmed correspondent of a tier-one EU or UK bank under a credit-insured wrap. Where neither works, use a stablecoin settlement leg with explicit AML and conversion documentation. Each of these is bankable; what matters is matching the route to the corridor.

The third task is to document what was done. A trade finance transaction structured around a non-standard settlement architecture has to be defensible to the auditor of the corporate treasurer, to the credit committee of the funding bank, and to any subsequent enforcement enquiry. That is a documentation discipline; it is also where most mandates fail to close even when the commercial logic is sound.

The correspondent banking architecture for African trade is unlikely to return to the structure of 2005. The work is to operate inside the architecture that actually exists.