The transfer that lost value at every border
The number on the closing document said $480,000. The number that arrived in the seller’s account, three business days later and four time zones away, was $468,240. No one stole anything. No error was made. No one even had to lie. Eleven thousand, seven hundred and sixty dollars simply ceased to exist somewhere between a bank in Miami and a private account in Zurich, consumed by a system so layered, so architecturally accustomed to helping itself, that the people who designed it no longer think of it as a cost. They think of it as gravity.
This piece is not about why international wires are slow. That story has been told. This is a forensic trace of where value actually goes — the precise anatomy of erosion, hop by hop, spread by spread, from the moment funds leave one account to the moment something diminished lands in another. The professionals who orchestrate these closings — the brokers, the advisors, the attorneys coordinating disbursement — send these wires constantly, sometimes dozens in a month. They absorb the shortfall, explain it away, or build it into the deal. What they rarely do is track the money through every gate it passes.
It’s worth tracking it. The math is instructive and, once seen, impossible to unsee.
The architecture of the correspondent chain
To understand where the money goes, you have to understand why it travels the way it does.
SWIFT sends payment instructions between banks but never touches the money. Settlement runs through correspondent accounts, not through SWIFT. This distinction — between message and money — is fundamental to everything that follows. When a broker’s client wires funds from a US bank to a counterparty overseas, they are not sending money through a pipe. They are sending an instruction through a network of trusted relationships, each of which has to be honoured by a separate institution, on a separate ledger, under a separate set of compliance rules and cost structures.
SWIFT settlement runs on bilateral banking relationships managed through nostro and vostro accounts. The terms come from Latin (“ours” and “yours”) and describe the same account from opposite perspectives. A nostro account is a foreign-currency account your bank keeps at an overseas correspondent bank. That same account shows up as a vostro account on the correspondent’s books. Each bank in the chain holds pre-funded pools of currency in these accounts, and when a transfer passes through, the ledger entries move rather than the physical dollars. The money doesn’t so much travel as get reassigned, institution by institution, until the final bank in the chain credits the recipient.
For funds to move physically from your bank to their destination, the sending and receiving banks must have a relationship with each other. This isn’t always the case, so when making an international transfer via the SWIFT network, your money will often pass through one or several correspondent or intermediary banks before reaching its final destination.
Here is what determines the length of that chain: relationship geography. Every bank in the world does not have a direct relationship with every other bank. The more remote the corridor — geographically, linguistically, or by currency — the more intermediaries have to be recruited to bridge the gap. If your bank and the recipient’s bank don’t have a direct relationship, the SWIFT network routes the payment through one to three intermediary banks, each adding time and fees. A transfer from a small US bank to a regional bank in Southeast Asia, for example, might pass through two or three intermediaries.
In some cases, a common transfer involves four banks working together through the SWIFT network. Sometimes, up to five banks are required to process international transfers, particularly when trading exotic currencies. The more banks that are involved, the higher the correspondent fees are likely to be, and the longer the transfer will take.
Five banks. Five separate institutions, each operating under its own commercial logic, each permitted — by convention and by the mechanics of the system — to help themselves to a slice before passing the remainder along.
Now run the numbers.
The toll booth at the originating bank
The wire begins with what the sender sees: a flat fee, the sending bank’s charge for initiating the transfer. A typical outgoing domestic wire runs $25 to $35, an outgoing international wire $35 to $50. On a large deal — say, $480,000 moving from a US account — this fee feels negligible. It’s visible, explicit, and priced at a level that doesn’t invite scrutiny.
Ask a CFO what it costs to send an international wire, and they will usually quote you a number between $25 and $50. That is the fee their bank charges to initiate the transfer. It is also a fraction of what the transfer actually costs. The rest is buried in places that do not show up on any single line item: exchange rate markups baked into the conversion, correspondent bank fees deducted from the transfer mid-flight, and the working capital locked up while the payment crawls through a multi-day settlement chain.
That flat fee is the most honest cost in the entire process, which is a damning thing to say about a $45 line item. Everything that follows is less transparent, less predictable, and — aggregated across professional deal volumes — far more consequential.
Call this Step One. The transfer leaves the originating bank, $480,000 minus a $45 wire fee. $479,955 enters the SWIFT network. The sender receives a confirmation number and, often, a false sense of completion.
The FX spread: the largest fee no one sees
The transfer must convert from US dollars. The destination account, in our illustrative scenario, is denominated in Swiss francs. This conversion will happen exactly once, at exactly one institution, at exactly one moment — and that institution will not give the sender the real rate.
The most expensive part of an international wire transfer is rarely the flat international wire transfer fee printed on your bank’s schedule. It is the exchange rate markup your bank applies when converting your money from one currency to another. This markup is built invisibly into the exchange rate your bank offers, making it effectively invisible unless you actively compare it to the real mid-market rate.
The mid-market rate, also called the interbank rate, is the midpoint between the buy and sell prices for any currency pair at a given moment. It is the rate shown on Google, Reuters, or xe.com. Banks do not offer this rate to retail customers. Instead, they offer a rate 2 to 4 percent worse, and keep the difference as profit.
On our $479,955 transfer, a 2.5% FX markup — squarely within the typical range for a major US bank on a USD/CHF conversion — costs $11,999. Not as a fee. Not as a line item. As a rate. The sender sees an exchange rate. They may even find it reasonable. What they do not see is what the rate would have been if they had received what the market actually offers.
Banks apply a markup typically ranging from 1.5% to 5%, depending on the currency pair, the bank, and the client relationship. On a $100,000 transfer, a 2% FX markup costs $2,000. That dwarfs the $40 wire fee.
FX markup is the 2–4% spread that banks add to the interbank exchange rate on wire transfers and currency conversions. Unlike foreign transaction fees, FX markup remains hidden because banks embed it in the exchange rate rather than listing it separately.
The mechanics of this concealment matter. When a bank quotes an exchange rate on an international wire, it presents a single number. That number is the mid-market rate plus its margin, blended together so seamlessly that the sender receives no indication of where market reality ends and bank profit begins. Traditional bank wires bundle the margin into the rate, which is why two “no-fee” international wires can carry very different true costs.
After the FX conversion, $479,955 has become approximately $467,956 in equivalent value delivered — before a single correspondent bank has touched it.
The sender’s confirmation still reads $480,000 sent. The bank is not wrong. It is, however, incomplete.
The first correspondent cut: the clearing bank
The converted funds now need to move through the network. The first stop is typically a major international clearing institution — a global bank large enough to have direct relationships in both currency zones, acting as the primary correspondent for the originating bank in this corridor.
SWIFT payments often hop through one or more correspondent banks before reaching the destination, and each can deduct a fee from the principal as it passes through. You sent $10,000; your supplier sees $9,930; the missing $70 went to banks you never chose and can’t see on your statement.
Correspondent banks are permitted to detract a fee of any amount directly from funds being transferred using the SWIFT network, all without your knowledge or permission.
This is worth pausing on. The correspondent bank does not need to notify the sender. It does not need to disclose its fee schedule in advance. It simply reads the payment instruction, processes the transfer, deducts its charge, and forwards the remainder. If you end up being charged a correspondent bank fee, the amount can vary significantly. What’s more, the amount will not be made transparent to you — not even in your bank’s fee schedule. This is because it’s the correspondent bank, rather than your own bank, that subtracts the fee from your transfer.
Intermediary and correspondent banks on SWIFT routes can deduct “lifting fees” of roughly $15 to $50 each, so your supplier receives less than you sent.
Call this first correspondent’s fee $35 — the lower-to-mid range for a major international clearing institution on a well-worn USD/CHF corridor. The remainder: approximately $467,921 in equivalent value, continuing toward Zurich.
The sender, if they check the status of their wire at this point, will likely see it “in process” or “pending.” The system has no obligation to tell them that $35 just left.
The second correspondent cut: the regional bridge
The first clearing bank does not have a direct nostro relationship with the recipient bank in Zurich. This is more common than it sounds. The more common instance where an intermediary bank is required is where a relationship does not exist between the issuing bank’s correspondent bank and the receiving bank’s correspondent bank. To facilitate the transfer, an intermediary bank that holds a relationship with these two parties needs to be engaged, so that the funds can pass through a trail of connected banks.
A regional European bank steps into the chain. Its job is to bridge the clearing institution’s network to the recipient bank’s network. This is not a clerical function; it is a commercial one. The regional bank is a separate legal entity with its own compliance infrastructure, its own risk management, its own cost base — and its own incentive to extract value from every message that passes through its accounts.
Each institution involved in a cross-border transaction might charge its own processing or currency conversion fees. These costs can accumulate and make it difficult for senders and recipients to know the exact cost of a transaction in advance.
Under the correspondent banking model, each bank involved in the cross-border payment processing chain can apply a fee. This also extends to the beneficiary bank, which can also apply a fee. These additional correspondent banking and beneficiary banking fees can be greater than the initial charge made by the originating bank.
The regional bridge bank charges $30. The running total: approximately $467,891 in equivalent value, forwarded to the final institution.
Two correspondent fees. Neither disclosed upfront. Neither visible in any confirmation the sender will ever receive unless they request a detailed SWIFT trace — a process that itself typically costs money and takes time.
The receiving bank’s incoming fee
The funds reach the recipient’s bank in Zurich. The transaction is almost complete. Almost.
Even after your money arrives at the destination bank, the recipient’s institution may charge a fee to process and credit the incoming wire. This is called the receiving fee or incoming wire fee. Receiving fees vary widely by country and institution. In some markets, they are negligible. In others, they can reach $15 to $30 per transfer.
Swiss private banking institutions typically sit toward the upper end of that range. The recipient bank charges a $25 incoming wire fee, deducted from the amount before credit.
Regardless of correspondent bank charges, any recipient bank fees will still be owed by the beneficiary and likely deducted from their receiving total, as you’re rarely able to cover these fees as a sender. Recipient fees vary by country. Some banks charge a flat receiving fee, while others charge a variable fee depending on where the funds are being transferred from.
The final credited amount: approximately $467,866 in CHF equivalent — translated back to USD at the same mid-market rate, for the purposes of this accounting.
The full ledger
Here is the complete anatomy, laid out without editorial:
Sent: $480,000
Originating bank wire fee: –$45
FX spread (2.5% on USD/CHF conversion): –$11,999
First correspondent bank lifting fee: –$35
Second correspondent bank lifting fee: –$30
Recipient bank incoming fee: –$25
Total deducted: $12,134
Delivered: approximately $467,866
The leakage is 2.53% of the original transfer. On a transaction of this size, in the context of a professional deal, that number represents real money — not a rounding error, not an acceptable overhead, but a structural extraction built into every payment of this type, every time.
Total actual cost on a comparable large international transfer can run roughly 2% of the transfer amount. The bank’s quoted wire fee can represent less than 1% of the real cost. In this anatomy, the sender paid a $45 fee and received confirmation. The system collected $12,134.
Scale that to a company making 50 similar payments per month and the annual cost approaches $2.5 million. Most of it invisible on any single invoice or bank statement. This is a structural feature of the correspondent banking model. Each bank in the chain is a separate business, operating under its own regulatory regime, maintaining its own compliance infrastructure, and pricing its services to cover its own costs and margin.
The correspondent banking model is, in this sense, honest about what it is. It is not designed to deliver money cheaply. It is designed to deliver money reliably, through a chain of trusted institutions, each of which expects to be compensated. The cost of that reliability is the spread between what was sent and what arrived.
What makes this hard to fix
The professionals who work in this environment — the closing attorneys, the international brokers, the advisors coordinating disbursements across multiple jurisdictions — know all of this, intuitively if not forensically. They have watched clients blink at settlement statements. They have fielded calls from counterparties asking why the amount was short. They have learned to over-wire, to build cushion into instructions, to pre-negotiate who will cover the shortfall.
None of that is satisfying. It is adaptation, not resolution.
The deeper structural problem is that the correspondent chain is opaque by nature. It is very hard to know how many intermediary banks are going to be in the transaction and how much each will charge, since it depends on the various currencies involved, the countries, and the banking relationships. Before a wire is sent, no institution in the chain is obligated to publish its lifting fee for that specific route, that specific currency, that specific day. The actual cost of the transfer is unknowable in advance — it can only be measured after the fact, by comparing what was sent to what arrived.
This matters enormously in professional contexts. A dealmaker coordinating the disbursement of proceeds across multiple parties — a selling broker, an advisor, a referral partner in another country — cannot build a precise financial model around a number that can only be known retrospectively. They can estimate. They can use historical ranges. They cannot know.
The frustrating part is that these fees are often poorly disclosed upfront, making it difficult to predict the total cost of your transfer. In a deal context, that unpredictability is not merely inconvenient. It affects the economics of every disbursement made across a border.
There is also the matter of float. In some cases, a cross-border payment can still take 24 hours to reach the beneficiary bank account. This means one correspondent bank in the cross-border payment chain has effectively taken one day’s float in addition to its processing fees — this is significant given interest rates remain relatively high. On a large transfer, a single day’s float on a six-figure sum is real money, silently earned by an institution the sender never chose and may never be able to identify.
The operational work involved and lack of visibility on available funds resulting in overfunding the account represents a significant portion of the cost of making cross-border payments. This is a major pain-point for many banks — and, downstream, for every professional whose clients sit on the other end of these transactions, waiting for funds that have theoretically been sent.
The timing problem compounds everything
The erosion of value is bad enough in isolation. In a deal context, it interacts with something equally corrosive: timing.
SWIFT transfers typically take one to five business days and are more expensive, since several banks may be involved. In a cross-border closing — commercial real estate, a yacht changing ownership in a different currency zone, a structured business sale between parties on different continents — the closing date is not aspirational. It is contractual. Sellers have made plans. Advisors have coordinated. The entire machinery of a deal is built around a specific date, and that date depends on funds clearing in time.
International wire transfers are one of the biggest reasons foreign national mortgage closings miss their original closing date. In most cases, the delay is not caused by underwriting or missing loan documents. The loan may already be fully approved while the transaction remains stuck waiting for the borrower’s wire transfer to clear intermediary banks, AML review, or final confirmation.
Exchange-rate fluctuations can create small but important shortfalls between the expected cash-to-close amount and the final balance. If the receiving party gets less than required, the transaction may require a supplemental wire before recording can proceed.
A supplemental wire, of course, moves through the same correspondent chain, pays the same fees, converts at the same spread, and arrives on the same uncertain timeline. The problem is self-reinforcing.
International payments often pass through one or more correspondent banks, so settlement times can range from 24 hours to several days. On top of that, each intermediary might apply its own compliance checks, cut-off times, and processing rules, further increasing delays.
Bank cut-off times are their own quiet source of disruption. One of the most common reasons for a delay in wire transfers is bank cut-off times. Banks often have a specific time of day after which wire transfers will not be processed until the next business day. If your closing is scheduled later in the afternoon, you may miss the cut-off window, causing a delay.
A Friday closing with a European counterparty, initiated after 2:00 PM Eastern time, does not close Friday. It closes Monday — if the correspondent chain is efficient, if nothing flags in compliance, if the receiving bank’s cutoff has not also been missed. The professionals involved know this. They negotiate around it. But that negotiation is itself a cost: of time, of attention, of the contingency planning that shouldn’t be necessary.
The split problem, specific to professional deals
For professionals managing multi-party disbursements — a broker splitting a commission with a co-broker overseas, an advisor dividing proceeds among beneficiaries in different countries, a dealmaker coordinating payments to multiple advisors simultaneously — the correspondent chain problem is not merely linear. It multiplies.
Each payment is a separate wire. Each wire pays the originating bank’s fee. Each wire moves through its own correspondent chain, accumulating its own lifting fees, converting at its own FX spread. If three parties are owed money in three different countries, the sending party initiates three wires, each of which loses value independently, and the professional orchestrating the disbursement has no way of knowing, in advance, exactly how much will arrive in any of the three accounts.
In these scenarios, the value erosion is not a rounding error. It is a systemic overhead that the professionals involved have normalized because they have had no alternative. The broker based in Miami who routinely co-brokers with a partner in London has probably never sat down and calculated what the correspondent chain has cost their partnership over the course of a year. The math is uncomfortable.
Traditional bank wire total costs typically range from 4% to 8% of the transfer amount, combining sending fees, exchange rate markups, and potential intermediary charges. On a $100,000 commission split, 4% is $4,000. It does not show up as a line item. It simply never arrives.
The point of no return
There is a moment in every wire transfer — it comes quickly, usually within seconds of initiation — when the transaction becomes irreversible. Wire transfers cannot be reversed once initiated, so accuracy is critical. This finality is, in most respects, a feature. It is why closing attorneys and title companies prefer wires over checks. Wire transfers are preferred over checks for real estate closings because they provide speed, verification, and certainty. Checks can bounce or face delays, but wire transfers are immediate and permanent once processed.
But that same finality means there is no mechanism to recover fees extracted mid-flight. Once the correspondent bank deducts its lifting fee and forwards the remainder, the deduction is done. The sender can request a SWIFT trace. They can open a dispute. They can ask their bank to investigate. What they cannot do is recover the $35 that a bank they never heard of took from a transaction they initiated in good faith.
The shortfall arrives at the recipient’s account. The recipient flags it. The professional coordinating the disbursement is now in the position of explaining to their client or counterparty that the amount is short — not through any error or bad faith, but because that’s what the system does. It is a professionally awkward conversation that happens every day, in every deal corridor, at every commission split and proceeds disbursement that crosses a border.
What changes when the chain shrinks
The anatomy above describes what happens when money moves through the traditional correspondent network in full. The length of the chain is not fixed. It is a function of the relationship density between the sending and receiving institutions — and that relationship density varies considerably by corridor, by institution, and, increasingly, by payment rail.
Fewer intermediaries in the chain means fewer toll booths. A payment model that compresses the correspondent chain — or addresses it for the cross-border leg — structurally reduces the fee stack. Each hop removed eliminates not just the lifting fee but the compliance lag, the float, the timing uncertainty, and the irreducible unpredictability of a deduction that will only be known after the fact.
For professionals coordinating multi-party disbursements, the more structurally important question is not “how do I send wires more cheaply” but “how do I ensure that everyone who is owed money receives exactly the right amount, with certainty, at the moment the deal closes.” Those are different problems, and the second one has a different shape to its solution.
This is where tools like Shaka become legible. When a broker sets up a payment link with recipient wallets and split percentages configured in advance — before the deal closes — the disbursement is executed as a single transaction, onchain, with funds moving directly to each wallet without passing through a correspondent chain. There are no lifting fees deducted mid-flight, because there is no mid-flight. There is no FX spread embedded in a rate, because the conversion — where it exists — is made once, explicitly, at a visible price. The money lands, split as specified, simultaneously, with the finality that a wire offers and the transparency that a wire never provides.
The professional’s role doesn’t change. They still coordinate the deal, establish the relationships, negotiate the terms, and determine who is owed what. What changes is the payment layer — the mechanism by which the closing translates into settled balances in the right accounts. Instead of initiating three wires and hoping each one clears within the window and that the recipient doesn’t call asking why the amount was short, the disbursement happens in one gesture, to every party at once, for exactly the amount specified.
That is a different experience of closing.
The invisible tax on professional deals
Return to the original number. $12,134, vanished from a $480,000 transfer. Not stolen. Not misapplied. Simply collected — by a sending bank, by two correspondent institutions, by a receiving bank — as the operational cost of moving money through a chain of trust relationships built over decades.
The professionals reading this understand that cost structurally. What they may not have done is run it across their own volume. Take a broker who closes twelve international deals a year, each with proceeds disbursement in the $400,000–$600,000 range. At 2.5% leakage, the system is extracting between $120,000 and $180,000 annually from transactions that professional touches. None of it appears on any statement. None of it is recoverable. All of it is, in the strictest sense, avoidable — but only by someone who understands the anatomy well enough to choose differently.
Most businesses remain unaware of their actual FX markup costs because the markup stays invisible on wire transfer confirmations. The same is true for most professionals in deal-intensive fields. The confirmation says sent. The amount that arrives says something else. The difference, if they have never run the math, is simply assumed to be the normal cost of doing business internationally.
It is the cost of doing business internationally a particular way. The correspondent chain is not gravity. It is infrastructure — legacy infrastructure, built for a world without programmable payment rails, without smart contracts, without the ability to specify in advance exactly how proceeds should be distributed and to execute that distribution atomically. That world no longer exists. The infrastructure, however, persists.
The transfer that lost value at every border will keep losing value at every border, on every deal, for every professional who hasn’t stopped to measure what the chain actually costs — and decided that measurement itself is the beginning of doing something about it.
