Picture Ahmed on a Tuesday morning, sitting across from his trade finance officer at a small Nairobi bank, hearing for the first time that the confirmation he has relied on for a decade is simply no longer available. His shipment is already on the water. The season for selling winter textiles runs roughly eight weeks. He has maybe three of them left by the time the goods could clear customs, if they clear at all. Nobody called him. Nobody warned him. The line just closed.
Not dramatically. No single bank called a press conference. The European correspondent that had been backing his bank's instruments quietly hit its sub-Saharan Africa limit and stopped confirming. Ahmed, who knows every customs officer in Mombasa by name and has never missed a payment, found that his bank could no longer get a major institution to back the instrument. The shipment sat. The season was lost. The reason was structural, not personal.
A letter of credit chain is, at its core, a sequence of promises. The importer's bank issues the LC. A larger correspondent bank confirms it. Sometimes a second correspondent adds its own confirmation for the beneficiary's benefit in a third country. Each institution in that chain stakes its own creditworthiness on the one below it. When the chain is short and every link is well-capitalised, the whole thing flows. When it stretches across three or four tiers, the arithmetic of counterparty risk compounds in a way most importers never see until it bites them.
The weight travels downward, not upward
Think of the chain the way you would think about load-bearing in a suspension bridge. The anchor cables carry the weight of everything hanging beneath them. The issuing bank at the top of a wholesale LC chain, typically a mid-tier regional institution in a developing market, carries the credit risk of the importer directly. But every confirming bank above it bears the risk of the issuing bank itself. They are not assessing the importer's balance sheet. They are assessing the issuing bank's.
That is the structural fact that determines who gets cut off first.
When a correspondent bank tightens its exposure, it does so by setting country limits, bank-specific limits, and aggregate limits across its correspondent network. Those limits are allocated from the top of its internal priority list downward. The largest-volume relationships, the ones generating the most fee income and the longest history, get their limits preserved. The smaller issuing banks, representing smaller importers in smaller markets, sit at the bottom of that allocation. Last to be onboarded, first to be squeezed.
Consider Ahmed alongside Maria, who buys similar goods through a mid-sized Nairobi bank processing roughly 40 million dollars in LC volume annually. Ahmed's bank manages maybe 4 million. When a major European correspondent tightens its sub-Saharan Africa limits by twenty percent, Maria's bank likely absorbs a proportional reduction and adjusts its pricing. Ahmed's bank may find its confirmation line suspended entirely. The correspondent's credit committee is not being cruel. It is optimising its own capital allocation, and a 4-million-dollar relationship does not justify the compliance overhead once limits get scarce. That 36-million-dollar gap in annual flow is the entire explanation.
The importer has no visibility into any of this. Ahmed never knew there was a line. He just knew his bank could confirm LCs, until it couldn't.
When the chain has three links instead of two
The problem intensifies when a so-called super-correspondent sits in the middle, a structure that is not the exception but the norm across most of Africa, South Asia, and parts of Southeast Asia. This happens in trade corridors where the issuing bank has no direct relationship with a major Western institution and routes through a regional hub bank instead. The regional hub has its own correspondent line. The major bank's limit now covers the hub, which sub-allocates to the issuing bank, which serves the importer. Three links, each applying its own risk margin, each carrying its own compliance threshold.
When any single link decides to reduce exposure, the effect is amplified downward. A ten percent tightening at the top can translate to a full suspension two tiers below, because the middle tier, now working within tighter limits itself, preserves its own best-performing issuing bank relationships and drops the marginal ones. The mathematics here are unforgiving, and no amount of creditworthiness on Ahmed's part changes them.
Ask yourself honestly: how many steps separate your bank from the institution whose risk appetite actually governs your access to trade finance? Most small importers have never thought to count.
Small importers at the end of long chains do not lose trade finance because they are bad credit risks in any direct sense. They lose it because they are invisible to the institutions whose decisions actually govern their access. The correspondent bank that ends the chain has probably never heard their name. That invisibility is the real exposure, and it is a design feature of the correspondent banking model, not a bug that regulators have simply missed.
The practical upshot is not cheerful, but it is actionable. The length of your LC chain is not fixed. Smaller regional development banks, multilateral-backed trade finance programmes, and fintech-enabled open account structures exist precisely because the correspondent banking model has this flaw built into its foundations. A shorter chain costs more in fees upfront; a suspended confirmation costs a season. Run those numbers against each other. Understanding that your vulnerability is positional, a function of where you sit in a hierarchy of credit relationships rather than your own financial standing, points directly toward the remedy: find a shorter chain, or find a different instrument. The problem is not that you are too small to trade. It is that you are too far from the anchor cable, and distance, unlike creditworthiness, is something you can actually change.