Picture the moment. A wire transfer lands on a compliance officer's queue somewhere in a large clearing bank. She can see the amount: $340,000. She can see it came from a bank she knows, a counterparty her institution vetted two years ago and approved. What she cannot see is the trading company that originated the payment, the city it came from, the number of institutions it has already passed through, or whether the bank two hops upstream employs a compliance team that amounts to more than a filing cabinet and good intentions. The money is real. The paper trail is a hallway of closed doors.
This is not a software glitch, not a regulatory gap that someone forgot to patch. It is the structural consequence of how correspondent banking chains are built, and understanding the architecture explains why financial crime can move through the global payments system with a confidence that would embarrass a shoplifter.
A chain that was never designed to be transparent
Correspondent banking is the plumbing of cross-border finance. When a small bank in Nairobi needs to send dollars to a client's supplier in Houston, it almost certainly does not hold a dollar account at a Federal Reserve member bank. So it maintains a relationship with a larger bank that does, paying fees in exchange for access to that bank's accounts and network. That larger bank may itself rely on a third institution for certain currencies or jurisdictions. Stack enough of these relationships together and you have a correspondent chain: a series of banks, each holding accounts for the one below it, passing payments along like a relay race where no runner ever sees the full course.
The critical structural feature is that each link in the chain sees only its immediate neighbors. Bank A instructs Bank B. Bank B instructs Bank C. Bank C credits the final beneficiary. When Bank B processes its leg of the journey, its systems record a payment from Bank A to Bank C. The originating customer's name, address, and account number may or may not travel with the message, depending on the payment format used, the instructions Bank A chose to include, and whether any intermediary stripped or truncated the data along the way.
This is sometimes called the nested correspondent problem, and it is worth sitting with the geometry for a moment. A bank at the center of a large correspondent network might process payments on behalf of dozens of respondent banks, each of which has its own downstream clients. The center bank's compliance team is not reviewing the end customers of those respondent banks. It is reviewing the respondent banks themselves, as institutions, usually at onboarding and periodically thereafter. The individual transactions flow through as bulk traffic. Think of it less like a border checkpoint and more like a toll booth on a motorway: the booth records that a vehicle passed, not who was sitting in the back seat.
Where the visibility actually breaks down
Consider how this plays out in practice. A trading company in a mid-sized city opens an account at a regional bank, call it Meridian Bank, operating in a jurisdiction with functional but lightly staffed financial intelligence infrastructure. Meridian holds a correspondent account at a larger regional hub, Continental Bank, which in turn holds a nostro account at a major clearing bank in New York, Global Clearing Co.
The trading company instructs Meridian to send $340,000 to a supplier account. Meridian sends a SWIFT MT103 message to Continental. Depending on how Meridian formats that message, the originator fields may be complete, abbreviated, or populated with Meridian's own identifiers rather than the underlying customer's details. When Continental forwards the payment to Global Clearing Co, it may reformat the message entirely, substituting its own account reference. By the time Global Clearing Co's compliance system runs its screening, the transaction looks like an interbank transfer from Continental, a bank Global Clearing Co has already vetted and approved. The $340,000 moves. Nobody at Global Clearing Co ever ran the trading company's name through a sanctions list.
This is not a hypothetical edge case. It is the documented mechanics behind enforcement actions against major institutions across multiple continents over the past two decades. The SWIFT messaging standards themselves were updated, via rules requiring full originator and beneficiary data in cross-border wires, precisely because regulators recognized that data was being lost or suppressed in transit. Compliance gaps remain, though, because the rules govern what banks are supposed to include, not what every counterparty actually checks on receipt. There is a meaningful difference between a standard and an enforcement, and the distance between them is exactly where the problem lives.
What people get wrong about this
The popular assumption is that the big clearing bank at the top of the chain is the last line of defense, the institution with the most sophisticated screening tools and the most to lose. That is true as far as it goes. But it rests on a premise that frequently fails: that the big bank has the data it needs to screen against.
Screening software, however advanced, can only work on the fields it receives. A sanctions-screening system that sees "Continental Bank / ref 0042817" cannot tell you whether the underlying customer is a designated individual. It will pass the payment cleanly. The compliance officer's screen shows green. She has, by any reasonable measure, done her job correctly given what she was given, and blaming her for the outcome is the kind of analysis that sounds satisfying and explains nothing.
The other common misreading is that nested correspondent relationships are inherently suspicious or somehow unusual. They are neither. They are the ordinary mechanism by which a bank in a smaller market accesses global payment rails. The vast majority of transactions flowing through these chains are entirely legitimate commercial payments, the unglamorous movement of trade finance, payroll, and supplier settlements that keeps the world's supply chains turning. The structural opacity is a side effect of a system optimized for speed and interoperability. That distinction matters for policy, even if it provides cold comfort to the compliance officer who cannot see past the first door.
The consequence that compounds
Two colleagues, Marcus and Yemi, joined the same correspondent banking compliance team on the same day. Marcus was assigned to review new respondent bank relationships: the due diligence files, the ownership structures, the AML policy documents. Yemi was assigned to transaction monitoring, the daily flow of actual payments. Within a year, Yemi had noticed something Marcus's work could not catch. A respondent bank with a perfectly acceptable compliance policy on paper was sending payments with unusual consistency in round numbers, always just below reporting thresholds, always to a narrow cluster of beneficiaries. The policy documents said the right things. The transaction pattern said something else entirely.
Yemi's observation points to the only real mitigation available within the current architecture: behavioral analysis of correspondent flows over time, rather than relying on the presence of correct data fields in any single message. Regulators in several jurisdictions now expect large correspondent banks to monitor not just individual transactions but aggregate patterns from their respondent banks, looking for the kind of clustering and structuring that individual-transaction screening will always miss.
It helps. It is not the same as being able to see the original customer.
So ask yourself what it would actually take to fix this. Not patch it, not add another layer of guidance that well-intentioned banks will follow and others will note with interest before filing away: actually fix it. The answer requires rethinking the incentive structure of correspondent relationships at a level that no single jurisdiction can impose unilaterally, which is precisely why the problem has outlasted so many rounds of regulatory tightening. The architecture of the correspondent chain was built to move money efficiently across borders that would otherwise be impenetrable to smaller institutions, and it succeeded at that, thoroughly. The visibility problem is not a flaw waiting to be engineered away so much as a structural trade-off that was made before anyone was fully counting the cost. Compliance officers working inside these chains are not failing at their jobs. They are doing their jobs inside a system that was never designed to give them a clear view of what they are being asked to watch. The distinction is important, because confusing the two tends to produce enforcement that punishes individuals for institutional architecture, which is a reliable way to generate fines without generating change.