Multi-Currency Accounting for Exporters

Three currencies, not one

Most confusion about foreign exchange in accounting comes from not separating three distinct roles a currency can play.

Your ledger is kept in the functional currency. A USD invoice therefore has to be recorded twice over: at its face value in USD, and at its converted value in PKR. Both are stored, and the relationship between them is where every gain and loss comes from.

Two dates, two rates

A foreign-currency receivable touches the rate at least twice, and usually three times.

1. Transaction date: the invoice is recorded at the rate ruling that day.
2. Reporting date: if the invoice is still open at period end, it is revalued at the closing rate.
3. Settlement date: when payment arrives, the actual cash received is converted at that day's rate.

The customer owes a fixed number of dollars throughout. What changes is what those dollars are worth to you, and the ledger has to record that movement somewhere.

Realised and unrealised

The distinction is simply whether the money has actually moved.

An unrealised gain or loss arises from revaluing something still outstanding. It is an accounting measurement, not cash, and it reverses if the rate moves back.

A realised gain or loss arises on settlement. The cash is in the bank, the amount is final, and nothing reverses it.

A worked example

A Pakistani exporter invoices a US customer 50,000 USD on 10 March. Functional currency is PKR.

At invoicing, receivables are debited 13,950,000 and revenue credited the same. Revenue is fixed at that moment and never moves again — a point frequently got wrong.

At 31 March the receivable is still open, so it is revalued to 14,175,000. The 225,000 increase is an unrealised exchange gain: debit receivables, credit unrealised FX gain.

On 22 April the cash arrives, worth 14,060,000. The receivable is carried at 14,175,000, so there is a 115,000 shortfall against the carrying amount — a realised exchange loss. Debit bank 14,060,000, debit realised FX loss 115,000, credit receivables 14,175,000.

Net across both periods: a gain of 225,000 then a loss of 115,000, a net 110,000 gain, which is exactly the difference between the 279.00 invoice rate and the 281.20 settlement rate on 50,000 dollars. The arithmetic closes, and the split across periods is what revaluation exists to achieve.

What gets revalued, and what does not

The rule is monetary versus non-monetary. Monetary items — a fixed number of currency units receivable or payable — are revalued at each reporting date. Non-monetary items are not.

• Revalued: foreign-currency receivables, payables, bank accounts, loans.
• Not revalued: inventory bought in foreign currency, fixed assets, prepayments. These stay at the rate on the date of the transaction.

Inventory is where this is most often got wrong. Stock bought for 10,000 USD is recorded at the PKR value on the purchase date and stays there. It does not move with the rate, because you own goods rather than a claim to dollars.

Where the rates come from

Every figure above depends on a rate, and rates are a policy decision rather than a fact you look up. Two businesses can book the same invoice at different values, both correctly.

The decisions to make explicitly:

• Which source. A central bank reference rate, a commercial provider, or your own bank's published rate. Pick one and document it; mixing sources makes variances unexplainable.
• Which rate type. Spot, mid-market, or the buy and sell rates your bank actually gives you. Mid-market is conventional for the ledger, but it is never what you receive, so expect a permanent small loss on settlement.
• When it is captured. Daily rates loaded overnight are standard. A monthly average rate is acceptable for high-volume low-value transactions and misleading for a large one in a volatile month.
• Who may override. Occasionally a transaction has a contractually fixed rate. That should be enterable, restricted, and reportable, because an overridden rate is also the easiest way to hide a problem.

A practical control is a tolerance check on manual entry: if someone keys a rate more than a set percentage from the loaded rate for that day, it needs a second approval. Transposition errors in rates are easy to make and produce balances large enough to distort a period.

For a business in a country with a managed or rapidly moving currency, the frequency question becomes the important one. Where the rate moves several per cent in a month, a monthly average quietly misstates every transaction in the period, and daily rates stop being a refinement and become a requirement.

Why the ledger and the bank disagree

A recurring question with a straightforward answer. A USD bank account has two true balances: the dollars actually in it, and its PKR carrying value.

The dollar balance is reconciled against the bank statement. The PKR balance is what appears in the trial balance and moves every time the account is revalued, with no corresponding movement in dollars.

So reconcile a foreign-currency account in its own currency, always. Attempting to reconcile the converted figure against a statement is a guaranteed waste of an afternoon — the two are measuring different things, and only one of them is what the bank holds.

Errors worth checking for

• Revenue restated when the rate moves. Revenue is fixed at the transaction date; only the receivable moves.
• A single rate used for a whole month. Defensible for low-value items, misleading for large ones in a volatile period.
• Revaluing inventory or fixed assets. Non-monetary items stay at historic rate.
• Realised and unrealised posted to one account, so nobody can tell cash results from measurement noise.
• Revaluation run after the statements rather than before, so every reviewed figure moves afterwards.