Working with Cross-Border Payments When You're Not a Bank
I spent about four years handling payment operations for a mid-size fintech, and the thing nobody tells you about international money movement is that it is almost entirely a problem of exceptions. The happy path works fine. Swift messages hit the right node, the intermediary bank takes its cut, the beneficiary gets paid. Then something goes wrong and you are spending six hours on a conference call trying to figure out why a payment intended for Lagos got routed through Luxembourg and landed in a holding account at a correspondents correspondent. If you are just getting into this space and have come across resources like International Money And Finance Melvin or similar academic frameworks, those give you the theory. Good. Now let me tell you what actually happens when you try to move real money across borders.
Why Your Payment Is Stuck in Correspondent Limbo
Here is the core structure you need to understand before anything else. Most international payments do not go directly from your bank to the recipient bank. They travel through a chain of correspondent accounts. Your bank has a dollar account at a US bank, which may have another account at another US bank, and the receiving bank in the other country has an account at one of those same institutions. Each hop is a separate SWIFT message, and each institution can hold, reject, or reverse the payment at will. I had a case once where a €12,000 transfer from a German corporate client to a supplier in Vietnam was stuck for eleven days. The Vietnamese bank had flagged it for enhanced due diligence because the sender's company name contained a character that looked similar to a term on a sanctions screening list. Not the same term. A character that resembled it. The German bank refused to release the funds without explicit clearance, and the Vietnamese bank would not release them either. Both banks had the same transaction reference but were communicating through three different intermediary hops, so neither side could see the complete picture. The workaround was to have the originator provide a certified letter of explanation in English, fax it to the compliance team at the US correspondent bank that sat between both sides, and get a manual override on the transaction. That process took eight business days and required my direct intervention. The sender's payment system showed the funds as "in transit" the entire time, which is the banking industry's polite way of saying nobody knows where it is.
Setting Up a Reliable Cross-Border Payment Flow
The first decision you need to make is which currency corridor you are working with. Some corridors are trivial. USD to CAD through major banks is usually next-day with minimal friction. USD to USD for a US bank sending to another US bank with a foreign beneficiary is routine. The problems show up in what we call emerging market corridors, and they get worse from there. Structure your account relationships around your primary corridors, not around what looks cheapest on paper. A bank that charges lower fees but lacks a direct correspondent relationship in your target country will end up costing you more in delays, reversals, and manual handling. I learned this the hard way when we switched our EUR corridor from Deutsche Bank to a smaller regional lender to save 0.15 percent in fees. The regional lender had no direct euro clearing in Frankfurt, so every payment got routed through a third bank that added an extra intermediary fee of €45 per transaction and introduced a 24-to-48-hour delay. We switched back within three months. The math was simple once you stopped looking at headline fees.
Get the Full Details

SWIFT vs. Alternative Routing Networks
The SWIFT network handles the messaging. It does not handle the money. Understanding that distinction saves you a lot of headaches. SWIFT sends instructions. The actual settlement happens through correspondent accounts, central bank systems, or increasingly through bilateral arrangements and newer platforms. For specific corridors, alternative routing can be significantly faster. The SEPA system in Europe handles euro payments efficiently within the eurozone. CHAPS handles sterling in the UK. For Asia-Pacific, you have options like the Hong Kong FPS, Singpaore's PayNow cross-border links, and various central bank real-time gross settlement systems. None of these are universally better than SWIFT. They are corridor-specific. If you are moving money into Nigeria, SEPA will not help you. If you are moving money within the eurozone, SWIFT is slower and more expensive than SEPA credit transfers. I used a setup where we maintained direct correspondent accounts at three major US banks for dollar clearing, two European banks for euro, and a dedicated Asian correspondent for regional payments. The overhead was real, but the reliability was dramatically better than trying to route everything through a single provider. It also gave you leverage. When one bank started holding payments for routine screening, you could shift volume to another provider while the issue was being resolved.
Compliance Screening: Where Everything Goes Wrong
Sanctions screening is the single biggest source of delays in international payments. OFAC, EU sanctions, UN lists, HM Treasury, local central bank lists. Every jurisdiction has its own. Every bank interprets the lists slightly differently. A payment can be blocked at origin, at an intermediary, or at destination, and each block point has different release requirements. The counter-intuitive part is that stricter screening is not always better. I worked with a payments team that implemented an automated screening rule that blocked any transaction containing a name with more than eight characters that did not match a known beneficiary. This sounds reasonable until you realize that Vietnamese names, Arabic names, and many African names commonly exceed eight characters. We lost approximately 34 percent of our Vietnam-bound payments to false positives in the first two weeks after rolling that rule out. The false positive rate dropped to about 2 percent after we adjusted the rule to account for name length by country of origin and beneficiary region. The lesson is that screening rules need to be calibrated to your actual payment flow, not just set to maximum strictness. Here is what most beginners miss about compliance: the risk is not just in the destination country. It is in the origin, the intermediary hops, and the beneficiary structure. A payment from France to Kenya that passes through a US correspondent bank is subject to US sanctions screening regardless of where the money originates or where it ends up. US correspondent banks will block transactions involving Iran, Syria, Crimea, and several other jurisdictions even if neither the sender nor the receiver is connected to those places. This is extraterritorial application of US sanctions, and it is a permanent feature of the dollar clearing system.
Handling Reconciliation When Things Break
The worst part of international finance is reconciliation. A payment leaves your account on Monday. It shows as sent in your system. Three days later, the beneficiary says they have not received it. You check your system, and the payment status is "in progress." You message the receiving bank. They say they have no record of it. You check the SWIFT message trail. It is sitting at an intermediary bank that does not provide real-time status updates. The workaround I found most effective was maintaining a parallel tracking system. Every outbound payment got a unique internal reference code, and we logged the SWIFT message type, the expected intermediate banks, the value date, and the correspondent account details in a spreadsheet that we updated daily. When a payment went missing, we could trace it through the chain instead of starting from zero. This reduced our average resolution time from about four business days to roughly eight hours for most cases. The remaining cases were the truly difficult ones involving sanctions blocks or regulatory holds, and those required the manual intervention I described earlier.

Fee Structures and Hidden Costs
When you are comparing international payment providers, the fee schedule is almost never the full cost. Look at the OTF, OUR, and SHA options on SWIFT messages. In the SHA (shared) arrangement, which is the default for most retail and small business transfers, each intermediary bank in the chain deducts its fee from the principal amount. A €5,000 transfer might arrive as €4,940 because three intermediary banks each took €20. The beneficiary complains about missing funds. You have to explain that the fees were deducted en route. Nobody is happy. The OUR option means the sender pays all fees upfront. This is more expensive on the surface but eliminates the mystery deductions and the complaints from beneficiaries. For business payments where the invoice amount is fixed, paying extra to ensure the full amount arrives is usually worth it. I calculated it once for our operation: switching from SHA to OUR for our primary corridors reduced complaint volume by about 60 percent and cut reconciliation time by roughly 40 percent because we stopped spending hours tracking down missing intermediary fees.
When International Money And Finance Melvin Comes Up in Study or Practice
People studying this area often encounter references to Melvin in the context of international monetary economics. The textbooks and papers use certain models around exchange rate determination, balance of payments adjustment, and the constraints that countries face when they try to manage capital flows. Those models are useful for understanding why the system behaves the way it does, but they do not prepare you for the operational reality of moving money through a chain of banks that each have their own compliance teams, their own risk appetites, and their own interpretations of regulatory guidance. The gap between the theory and the practice is where most people stumble. You can understand the Mundell-Fleming model perfectly and still have no idea why a payment to a particular country is being held for three weeks because the correspondent bank's compliance officer decided to request additional documentation. The operational side is less about elegant models and more about relationships, documentation, patience, and knowing which phone number to call when the automated systems fail.
Practical Rules That Actually Matter
Verify the beneficiary bank's SWIFT BIC before sending. Not just the country code. The full eight or eleven character code. A wrong digit sends money to the wrong bank entirely, and recovering it from the unintended recipient bank involves a whole separate process that is slower and less guaranteed than getting it right the first time. Build in buffer time for intermediate jurisdictions. A payment that should take one business day often takes three when it involves a country with heavy capital controls or a correspondent chain that requires manual review. I started quoting three-business-day timelines for most emerging market corridors and two-day timelines for developed market corridors. The estimates were conservative, which meant we rarely missed them, which meant fewer anxious calls from clients. Keep records of every interaction. When a payment goes wrong, the bank will ask for documentation. If you have a clear paper trail with timestamps, reference numbers, and correspondence, you resolve issues faster. If you do not, you are starting from scratch while the payment sits in limbo.

The systems work most of the time. The times they do not work are when you need them to, and being prepared for those moments is what separates people who can operate in this space from people who spend all their time putting out fires.