Cross-Border Payment Infrastructure: SWIFT and Correspondent Banking

The Money Movement Misconception

Most people think "moving money" means something actually travels from point A to point B. Cash in a briefcase, electrons zipping through fiber optic cables, some digital parcel landing in another account. That's not what happens. Not even close. What we call international money movement is really a coordinated series of offsetting debt agreements - a chain of banks quietly agreeing to owe different people different amounts.

Domestically, the plumbing is straightforward enough. Your bank says "we owe this customer less now," the recipient's bank says "we owe this person more," and some intermediary system keeps the ledgers honest. No physical dollars change hands. It's all bookkeeping.

Cross-border works on the same principle, but there's a wrinkle that changes everything: banks largely can't hold money extraterritorially for most practical purposes. So they lean on something called correspondent banking relationships. And that's where it gets interesting.

Correspondent Banking Fundamentals

Banks hold accounts at other banks. This happens constantly - with extreme frequency, in fact - and for a bunch of reasons. But the big international one? Facilitating payments in currencies and jurisdictions that aren't your own.

Here is a practical example I keep coming back to. A banking customer in Japan needs to send U.S. dollar payments to American institutions. The Japanese bank would rather deal in dollars than yen for this transaction, while the recipient on the other end has difficulty accepting international payments. Pretty common situation.

The local Japanese institution carries some dollars on its books. Could be hundreds of millions. But does it physically control more than a tiny fraction of that? No. That small slice is paper currency sitting in branch offices, available for purchase in small quantities at substantial spreads. The vast majority of those "dollar holdings" are really just IOUs from bigger correspondent banks further up the chain.

The Correspondent Chain

So the major Japanese banks supply yen/dollar liquidity domestically, but they don't have direct access to the U.S. banking system themselves. Instead, they maintain accounts at various U.S. banks. Which means the dollars "owned" by Japanese institutions are really owed to them by American correspondents. Nobody in this chain - not the Japanese banks, not their local account holders - has actual custody of the dollars they intend to move. Think of it like a game of telephone, except every person in the chain is simultaneously an accountant.

The correspondent banks with full U.S. financial system access (including FedWire for domestic wire transfers) are the crucial intermediaries. When local banks execute wires, instructions get passed to major correspondents, which pass them to U.S. intermediaries, which effect the actual funds transfer through FedWire running through the Federal Reserve.

What you end up with is this rapid chain of bilateral agreements about who owes whom what. The sending party has less yen. The receiving party has more dollars. And at least five entities along the way collect fees. That's the business model right there.

Correspondent banking has operated in these broad strokes for a very long time. And here is what jumped out at me when I first really dug into this: there's no explicit technological substrate that makes it work. The process could run on TCP/IP, telegraph, or horseback letter delivery. All of those methods have, in fact, been extensively employed in correspondent banking across centuries. The technology is just the wrapping paper.

SWIFT: Encrypted Messaging Infrastructure

SWIFT is both a company and - like Kleenex or Xerox - a metonymic term for its best-known product: FIN, an extremely specialized low-volume encrypted messaging platform. "Low volume" sounds funny here. 5 billion annual messages? That's a lot of messages, sure. But in terms of computing? Inconsequential. We're talking roughly 160 transactions per second. Your average social media platform would call that a rounding error.

The reason SWIFT has become almost synonymous with international wires is simple: it is the primary interoperation method banks have chosen for coordinating wire transfers. Specifically, they send MT 103 messages - slightly longer than tweets - and then each bank goes back to its own internal books and systems to make those encoded requests a reality. Or to fail gracefully, which happens more often than people realize.

Network Effects Dominance

The technical details of MT 103 field allowances are fascinating if you're a payments professional, but they don't actually explain why SWIFT matters. What matters is something much more basic: network effects, layered on top of each other. Join the platform, and suddenly you can reach thousands of counterparties with minimal friction. Regulators are far more comfortable with SWIFT than with alternative solutions. Compliance officers have decades of institutional knowledge about SWIFT's operational quirks and oddities.

These advantages compound. Each marginal financial institution that joins makes SWIFT more attractive for everyone else already on it, which makes the next institution more likely to join too. Powerful lock-in. The kind of moat that, frankly, most tech companies would kill for.

SWIFT Limitations

But SWIFT does not have a monopoly on international money movement. Primarily because - and this is the part people consistently get wrong - it doesn't directly move money at all. Money doesn't travel over SWIFT any more than it travels over napkins, though either one could theoretically contain instructions that a bank might choose to act on.

SWIFT doesn't even monopolize the messaging function people assume it does. Banks keep documented procedures for "moving money when SWIFT is down" that don't involve pausing all economic activity. The major correspondents know the phone numbers of the U.S. institutions holding billions of their dollars. When you have deposits that size, you tend to have a direct line. And you can transact through it.

This is not some academic curiosity. It is the key to understanding how financial sanctions actually work - and where they fall short. Because SWIFT is a messaging layer, not the money itself, cutting a bank off from SWIFT does not destroy or confiscate the funds sitting in its correspondent accounts. Those dollars still exist as accounting entries at U.S. correspondent institutions. What SWIFT exclusion does destroy is the standardized, trusted communication channel through which every other institution in the network processes transactions. And in a system built entirely on network effects and compliance familiarity, that functional inaccessibility is nearly as devastating as the money not existing at all. Nearly. But the limits follow the same logic: correspondent banks outside the SWIFT network, or ones willing to operate through bilateral channels outside standard compliance frameworks, can still intermediate some transactions. Which is precisely why Russian entities found partial workarounds through non-Western banking relationships after the 2022 exclusions. Understanding the plumbing is what separates a surface-level reading of sanctions headlines from a genuine grasp of when the weapon has teeth and when it doesn't.

Banks as Policy Arms

Banks operate as policy arms for the governments that hold jurisdiction over them. In return, they get guaranteed monopolization of some spectacularly lucrative franchises. It is a bargain both sides have honored for centuries, and this pattern shows up everywhere you look in financial infrastructure.

SWIFT is theoretically a Belgian cooperative. The Federal Reserve is theoretically a joint-stock company owned by member banks, not a component of the U.S. government. These are consensual fictions - the kind everyone agrees to maintain because the alternative (admitting the actual power structure) is diplomatically inconvenient. SWIFT even publishes governance documentation that emphasizes "strict neutrality" in its final clause. Another consensual fiction.

In practice, SWIFT functions as a policy arm answering to the EU. It can be - and has been - directed against disfavored individuals, organizations, and entire governments, sometimes indiscriminately. This happened before, and all parties understood it could happen again. But this represents only part of the available policy toolkit.

Commander's Intent Framework

The U.S. military has a concept called "commander's intent." The idea is to balance institutional order-following with the messy reality of ground conditions that no order-writer could have fully anticipated. You tell subordinates what you actually want, so they can intelligently fill in the gaps.

I've noticed commentators routinely confuse the actual effects of severing particular banks from SWIFT with whatever policy goals seem to be motivating the action. But what matters more than either of those is what these actions communicate about commander's intent to the compliance officers and policy arms responsible for enforcement. That's the real signal.

Specifically, big moves like these communicate that money which previously flew under the radar now faces serious scrutiny. The unspoken message goes something like: "Billions in fines will be distributed stochastically over coming years. Institutions that successfully guess who belongs on the Bad Risks list will fare better. Previous tolerance of questionable practices is going to drop sharply. And everybody will conveniently forget they used to look the other way, provided they start being prudent now."

AML Enforcement Complexity

Using sanctions enforcement as a policymaking tool carries deep concerns. I won't pretend otherwise. But it is absolutely pervasive in financial industry regulation, particularly around anti-money laundering (AML). A huge amount of sanctioned activity gets enforced under AML regimes for pretty pragmatic reasons: the laws already exist, there are relatively capable enforcement agents already in place, the rules apply to essentially all human activity that touches money, and you can build process crimes from almost any underlying predicate.

SWIFT dominates today's headlines because, as nearly pervasive infrastructure, it offers one-stop-shopping from any political actor's perspective. But the quieter and arguably more important dynamic? Local compliance officers are assumed to read the news carefully and adjust their practices accordingly. That's where the real enforcement happens - not in Brussels or Washington, but in risk departments at mid-tier banks around the world, recalibrating their tolerance levels in real time.

Unintended Consequences

The global financial system is a blunt instrument. Incredibly blunt. Governments haven't explicitly banned oligarchs from using their wealth before - but in the process of trying, they've already made routine banking severely more difficult for ordinary citizens in affected regions. The collateral damage is real, and it is large.

After 2014, many U.S. banks simply stopped serving customers with certain passports. Nobody sat in a boardroom and decided "Your country got invaded, so you should have less financial access on the other side of the planet." It happened indirectly, through a chain of risk assessments: heightened money laundering concerns, inability to distinguish between different regions within a country, the calculation that AML compliance costs outweigh whatever revenue comes from a small foreign client base, and - at the end of all that math - blanket restrictions. Clean and easy for the bank. Devastating for the people caught in the net.

Here is the part that really bothers me. Nobody - not the regulators, not the compliance departments, not the front-line employees delivering these decisions - considers themselves responsible for what happens to the tens of thousands of individuals affected. There is no accountability loop. And with current conflicts, the number of people hit by these adverse actions against innocents is going to increase. That's not speculation. It's a near-certainty, given everything we know about how this machinery works.