You send the money on a Tuesday. You get a confirmation number, a small fee deducted, a notification that the transfer is processing. By Thursday it still hasn't arrived. You try again Friday. The family in Mogadishu, or Port-au-Prince, or Caracas, keeps waiting. No regulator held a press conference. No announcement came. A corridor simply closed, quietly, because of decisions made inside institutions you have never heard of and will never deal with directly.
That is how remittance routes die. Not with a ban, usually. With a chain.
The plumbing nobody sees
When you send money internationally through a bank, you almost certainly believe two banks are involved: yours, and the recipient's. In practice, the transaction may pass through three, four, or five institutions before it arrives. Each link in that chain is a correspondent banking relationship: a formal agreement in which one bank holds an account on behalf of another, called a nostro account, and agrees to settle transactions on its behalf in a currency or jurisdiction where the first bank has no direct presence.
The structure matters enormously. A bank in the Pacific Islands that wants to move US dollars must route those dollars through a correspondent that holds a dollar account at the Federal Reserve. It cannot simply decide to stop doing that and go direct. The architecture is not optional.
Think of it less like a road network and more like a series of locks on a canal. Every vessel passes through the same gates, in sequence, and if one lock-keeper decides to stop operating, the whole canal goes still above that point. The image is not dramatic. It is precise.
The correspondent at the top of the chain, typically a large American, European, or Japanese bank, is the one that sets the true terms of access. When that institution decides a particular downstream relationship carries too much compliance risk, it doesn't need to refuse individual transactions. It terminates the relationship with the smaller bank. That smaller bank, now without a correspondent, cannot process dollar payments at all. Every corridor it was serving goes dark simultaneously.
How risk accounting inside one institution ripples across continents
The decision that closes a corridor is almost never made by anyone who thinks of it in those terms. It is made by a compliance officer, or a risk committee, running what the industry calls a profitability-adjusted risk review. The question is not whether a given downstream bank is doing anything illegal. The question is whether the expected revenue from the relationship justifies the compliance overhead required to monitor it adequately.
For a Tier-1 correspondent, a relationship with a small Somali remittance bank might generate fees in the low tens of thousands of dollars annually. The compliance cost of properly monitoring that relationship, given the regulatory scrutiny applied to Somalia as a jurisdiction, can run into six figures. The math is straightforward and the outcome is predictable.
Consider two people who both wanted to send money from the United Kingdom to Somalia in the same month. Call them Amina and David. Amina used a specialist Somali remittance operator that had maintained a correspondent relationship with a mid-tier European bank for eleven years. David used a digital transfer app that had built its own direct relationship with a larger institution. When the mid-tier European bank quietly terminated its correspondent agreements with high-risk-jurisdiction operators, Amina's operator lost its dollar and euro clearing access within sixty days. The corridor didn't degrade. It stopped. David's app, routed through a different chain entirely, kept working. The difference between them was not the destination, not the amount, not the compliance record of either operator. It was the internal structure of the chain each one depended on.
This is what development economists mean when they call de-risking a structural problem rather than a behavioral one. The operators being cut off are frequently the ones with the strongest compliance records in their tier, cut loose not because they are doing anything wrong but because their jurisdiction is simply too expensive to monitor. That distinction matters, and the industry has been too slow to make it plainly.
What people get wrong about why corridors close
The most common misconception is that remittance corridors close because of sanctions, or because money laundering is actually occurring. Sometimes that is true. More often, it is not.
The Financial Stability Board has documented a broad decline in correspondent banking relationships across the Pacific, Caribbean, and sub-Saharan Africa that is only loosely correlated with actual enforcement actions. The corridors closing fastest are frequently in jurisdictions with no active sanctions exposure and no recent major enforcement cases. They close because the compliance cost structure of Tier-1 banking makes any relationship with a small institution in a poor country financially irrational, regardless of that institution's actual behavior.
A second misconception: that fintech and mobile money solve the problem. They can route around correspondent chains for small domestic transfers. The moment a cross-border payment involves US dollars or euros, it must touch the correspondent system somewhere. There is no workaround for that.
And a third, perhaps the most persistent: that the receiving country can simply build its own financial relationships. Correspondent banking is not reciprocal in the way people assume. A Haitian bank cannot simply approach a major institution and ask to open a nostro account. Access is granted at the discretion of the Tier-1 institution, which is running the same profitability-adjusted risk calculation on every application. The door opens from one side only.
The corridor as a thing that can be designed, or abandoned
The corridors that survive are, almost without exception, the ones where the chain has been deliberately engineered to reduce the number of links and the compliance burden at each one. The UK-to-Nigeria corridor remains relatively robust not by accident but because several large operators invested in direct relationships with Nigerian banks that meet international compliance standards, and because the corridor's volume makes the economics work for Tier-1 correspondents.
Volume is the variable that almost nobody sending a remittance thinks about, but it is the variable that determines whether their corridor exists at all. A corridor processing two billion dollars a year will retain correspondent relationships almost automatically. A corridor processing forty million dollars a year, scattered across a dozen small operators, is permanently at risk. The risk is not the forty million dollars. The risk is that no single institution in the chain has enough revenue from that corridor to justify carrying its compliance costs.
Ask yourself: who, exactly, is supposed to fix that? The small operators lack the capital. The receiving-country banks lack the access. The Tier-1 correspondents have no commercial incentive to act. The regulators who created the compliance burden have been slow to acknowledge that their rules produce this outcome as a side effect. The history of financial infrastructure suggests that gaps of this kind persist for a long time before anyone with the power to close them decides it is worth the trouble.
The people who lose access to a remittance corridor rarely know why it happened. They know the transfer failed. They know the operator is telling them to try a different service. They do not know that a risk committee in Frankfurt or New York looked at a spreadsheet some months earlier and decided the numbers didn't work. They do not know that their corridor was never a permanent fixture, that it existed only because someone, somewhere up the chain, decided the math was acceptable for long enough.
When the math changes, the canal goes still. The family waits. The compliance officer moves on to the next line item, and the corridor's closure does not appear in any ledger as a cost to anyone who mattered in that room.