When the currency is not freely convertible

Most conversations about foreign exchange are conversations about the spread. How many basis points, on what reference, against what benchmark. In a freely traded currency pair, that is a reasonable place to focus. In a constrained corridor, it is close to irrelevant — and a company that negotiates hard on the rate while leaving the structure untouched will usually keep paying, just less visibly.

What exchange control actually asks of you

A currency that is not freely convertible is not simply a currency that trades badly. It is a currency whose movements are supervised, where conversions and cross-border transfers require an approved justification, and where the approval is a document produced by someone, checked by someone else, and dated.

In practice, this means a payment file rather than a payment instruction: the underlying contract, the invoice, a purpose-of-payment code that matches both, sometimes evidence that local obligations have been settled first, occasionally a prior authorisation with a validity period of its own.

The delay a company experiences is almost never the settlement. It is the file. And the reason the file takes weeks is usually that no one involved has assembled that particular file before, so each element is discovered in sequence rather than prepared in parallel.

The round trip nobody decided to take

The second pattern is subtler, and we find it most often in groups that consider their FX exposure to be under control.

A company’s cash sits in one currency. Part of its cost base is in another. Each time a payment falls due in the second currency, the treasury converts at whatever the market offers that day. Months later, receivables in that same second currency arrive from clients, land on an account with no immediate use, and are converted straight back.

The same underlying flow crosses the market twice, in opposite directions. Two spreads are paid. And between the two conversions sits a period of directional exposure that nobody deliberately took — the company is, in effect, holding a currency view it never formulated, for as long as the gap between the two legs lasts.

A round trip is not a bad rate. It is a structure that requires you to keep buying what you already receive.

The fix is rarely a better rate. It is an account structure that lets the company hold each currency as it receives it, pay from it, and only convert the genuine surplus. That is an architecture question, and it is usually settled long before anyone talks to a dealer.

Collections deserve the same attention as payments

Companies collecting online are exposed to a variation on the same problem. A payment gateway domiciled in one jurisdiction may convert incoming card payments automatically into the local currency of the merchant entity, at terms that are not negotiable and not always visible on the settlement report.

For a business collecting in several currencies from several regions, that conversion applies to every transaction, every day, in one direction only. There is no rate to negotiate here either. There is a gateway and account structure that either allows the company to hold what it collects, or does not.

Making an unpredictable corridor repeatable

Where a corridor is genuinely constrained and cannot be avoided, the objective is not speed. It is predictability — and predictability is achievable even when speed is not.

Three things make the difference in practice. A documented file structure, assembled once and reused every month rather than rebuilt each time. A banking partner that has executed this specific corridor before, and can therefore tell you in advance what will be asked. And a funding calendar built around the real settlement window rather than the theoretical one, so the local entity is funded ahead of its obligations instead of chasing them.

None of this removes the constraint. It converts it from a monthly uncertainty into a known process — which, for a subsidiary that has to meet payroll on a fixed date, is most of what was actually needed.

Where to look first

If a company suspects it is paying more than it should on currency, the spread is the last place to look, not the first. Ahead of it come three questions with larger answers: does the group hold each currency it regularly receives, or convert it on arrival? Does any flow cross the market twice in opposite directions within the same year? And in constrained corridors, does a documented, repeatable process exist, or does each transfer start from a blank page?

The rate matters. It just tends to be the smallest number in the equation.