The Wise model: cross-border without correspondent banking
Money never crosses borders. The accounting just makes it look like it did.
By Solomon Ajayi · Free to read, no signup
Wise (and now most cross-border-payments fintechs) sidestep correspondent banking entirely. The trick: hold real local bank accounts in every currency you support. When a user in Nigeria sends ₦1,650,000 to a recipient in Singapore for $1,000, Wise doesn't actually move money across borders. The NGN goes into Wise's Nigerian account; an equivalent $1,000 worth of SGD goes out of Wise's Singaporean account to the recipient. Money never leaves either country. The two-sided accounting illusion is what makes the model work, and why Wise can charge ~0.5% instead of ~5%. This lesson shows the four-line journal entry that makes one cross-border 'wire' actually two same-country movements with one FX revaluation in the middle.
Correspondent banking is how money traditionally crossed borders: a chain of banks holding accounts with each other, passing a SWIFT message and a few dollars of fees down the line at every hop. It is slow, opaque, and expensive, and a single international transfer can touch four or five intermediaries each taking a cut. Wise looked at that chain and asked a simpler question. What if the money never actually crosses the border at all?
The answer is to pre-fund a real local bank account in every currency you serve. A Nigerian sender's naira lands in your Lagos NGN account, and your Singapore SGD account pays out an equivalent amount to the recipient through local Singaporean rails. Two domestic movements, one for each leg, joined only by an internal FX reclassification on your own books. The user experiences one wire; your ledger records two same-country transfers and the spread you kept between them.
The accounting is genuinely simpler than a correspondent chain, but the operational burden moves onto you. Your local accounts can run dry if outflows in one currency outpace inflows, so you must continuously rebalance positions, fund accounts daily, and run an FX desk that prices spread against the mid-market rate. The FX Spread Revenue line in this lesson is the whole business model: it is what you earn for warehousing currency risk so the user does not have to.
Worked example, step by step
User in Lagos sends ₦1,650,000 for a SGD payout
User pays ₦1,650,000 into Wise's NGN account in Lagos. Of that, ₦1,500,000 covers the SGD payout at mid-market, ₦150,000 is the FX spread Wise keeps. Note: the SGD hasn't moved yet. We've only seen the NGN side.
| Account | Debit | Credit |
|---|---|---|
| User Wallet (2000) | ₦1,650,000.00 | |
| NGN Account (Lagos) (1310) | ₦1,500,000.00 | |
| FX Spread Revenue (4400) | ₦150,000.00 |
User Wallet DOWN ₦1,650,000 (debit). NGN Account UP ₦1,500,000 (asset, debit). FX Spread Revenue UP ₦150,000 (income, credit). Three lines, balanced. The NGN side is fully booked. No money has left Nigeria.
Payout in Singapore from SGD account: $1,000 (book at 1500 NGN/SGD for simplicity)
Wise's Singaporean account pays $1,000 SGD to the recipient via local Singaporean rails (FAST, instant). The SGD account is DENOMINATED IN SGD; on your NGN reporting books, that $1,000 SGD outflow translates to ₦1,500,000 at the SGD book rate. The SGD account goes DOWN.
| Account | Debit | Credit |
|---|---|---|
| NGN Account (Lagos) (1310) | ₦1,500,000.00 | |
| SGD Account (Singapore) (1330) | SGD 1,500,000.00 |
SGD Account DOWN $1,000 SGD = ₦1,500,000 at book rate (asset, credit). NGN Account DOWN ₦1,500,000 (asset, credit). Wait, that double-credits. The pattern is: the SGD outflow drops the SGD asset; on your NGN reporting books, the OFFSET is the NGN you 'consumed' to fund the SGD position. Re-think the entry as: SGD Account DOWN $1,000 (credit), with the COUNTER-debit being the NGN Account being topped up via the FX desk's settlement. Simplified two-line entry: SGD Account DOWN ₦1,500,000 (credit), NGN Account UP ₦1,500,000 (debit), internal reclassification, no NEW money in or out. The recipient sees $1,000 SGD land in their account. The NGN balance gains an offsetting position via end-of-day FX settlement.
Takeaway
The Wise / Revolut / Sendwave model collapses cross-border money movement into two same-country movements joined by FX settlement at the operator's level. The user perceives ONE 'wire'; the books show two. No correspondent chain, no Nostro/Vostro fees, no SWIFT messages, no $35-per-wire intermediary cuts. The cost-of-service collapses by 10x. The price you pay: you must continuously balance your local-currency positions, fund them daily, and run an in-house FX desk that prices spread against mid-market. The accounting is simpler than correspondent banking; the OPERATIONS is harder. Worth it at scale.
Practice this on a real ledger
Reading is half of it. Open this lesson in the lab to post the entries yourself against a real Postgres-backed double-entry ledger, with the validation on. Free, your sandbox is yours.