When the invoice date moves: extending, pre-delivering and closing out forward contracts

Every forward contract carries a date, and every date is a guess. You book USD 2 million for value on 30 September because that is when the supplier said the container would land and the invoice would fall due. Then the vessel is held up in Singapore. Or the supplier ships early and wants paying early. Or the order is cut in half after the customer walks. The hedge is now attached to a cash flow that no longer exists in the shape you hedged.

This is not an edge case. Across a book of a dozen or more forwards, a mid-market importer will have to move, split, bring forward or unwind several of them every year. How you handle those adjustments matters more to your realised hedging cost than the rate you originally dealt at, and it is where most of the quiet losses in an SME hedging programme actually live. It is also where the least attention gets paid, because the adjustment feels like admin. It isn't. Each one is a new trade with a new price, and a couple of them, done badly, will undo a year of disciplined execution.

What you actually hold at maturity

A vanilla forward is an obligation to exchange two currencies on a fixed date at a fixed rate. On the value date you deliver AUD and receive USD, or the reverse. There is no option to walk away and no automatic flexibility on timing. The rate you agreed is spot at the time of dealing plus or minus forward points, and the points are nothing more than the interest rate differential between the two currencies for the period.

That last part is why date changes cost money, or occasionally make money. When you move the delivery date, the interest differential for the changed period has to be settled one way or the other. With AUD and USD cash rates sitting within a few dozen basis points of each other, as they have through most of the past year, the points on a one or two month adjustment are small: a handful of pips on a rate around 0.66. That is not where the cost hides. The cost hides in the spot leg, and in how the provider chooses to show it to you.

Every adjustment to a forward is, mechanically, one of three things. You bring it forward, you push it back, or you close it. Each has an honest version and a version designed to make the P&L look better than it is.

Pre-delivery: taking the currency early

Pre-delivery, sometimes called an early drawdown, is the simplest case. The supplier invoices two weeks ahead of schedule and you need the USD now rather than on the original date.

The provider will let you take some or all of the notional early. The rate you receive is the original forward rate adjusted by the forward points for the period between the new date and the old one. If the points for that period are positive from your side, the rate moves marginally in your favour. If negative, marginally against. With rates roughly at parity between Australia and the United States the adjustment is often under ten pips, and on a USD 500,000 partial drawdown that is a few hundred Australian dollars either way.

Partial pre-delivery is worth knowing about because supplier invoices rarely arrive as one clean amount on one clean date. A USD 2 million forward can be drawn in four tranches over six weeks as the invoices land, with the residue delivered on the original date. Most bank and non-bank providers allow this without fuss for wholesale clients. Some retail-facing providers restrict partial drawdowns or charge a flat administration fee per event.

The point to hold onto: pre-delivery does not reopen the spot rate. You dealt at 0.6650, you still get roughly 0.6650. Nothing about the market move since dealing enters the calculation. That is what makes it the benign adjustment. The other two are not benign.

Extension: pushing the date out

Extension is where the trouble starts, and it is the adjustment you will need most often, because shipments slip far more than they arrive early.

There are two ways to extend a forward. The correct one is a close-out and re-dealing, sometimes written up as a swap. The provider closes the original contract at the current market rate for the original date, crystallising a gain or loss in AUD, and opens a new forward for the new date at the current market rate plus points. You settle the gain or loss on the original value date, or it is netted into the new contract as an explicit, visible cash adjustment. Either way, the number is on the confirmation and in your ledger.

The second way is the historic rate rollover, usually abbreviated to HRR. Here the provider rolls the contract to the new date at, or near, the original rate. The loss on the original contract is not settled. It is folded into the rate of the extended contract, adjusted for the provider's funding cost, so the new deal shows a rate close to what you first booked and the P&L hit disappears from view.

HRRs are attractive for exactly one reason: they hide the loss. And that is the reason to be wary of them. When a provider rolls your out-of-the-money forward at the historic rate, they are lending you the mark-to-market loss until the new maturity, unsecured, and charging you for it inside a rate that you cannot easily decompose. The Australian Prudential Regulation Authority requires banks to treat an HRR as an extension of credit, approve it through the credit function and hold evidence of the client's commercial reason for it. Most of the major banks now refuse them outright for SME clients, or approve them only on a case-by-case basis with credit sign-off. Some non-bank providers still offer them freely. That should tell you something about who benefits.

This is the wrong default. If a shipment slips and the forward is under water, the honest position is that your hedge locked in a rate the market has since moved away from. That is what a hedge does. Take the loss into the P&L where it belongs, offset it against the lower AUD cost of the underlying purchase when it finally settles, and book a clean new forward for the new date. The economics are identical to the HRR, minus the funding charge and minus the opacity. The only thing you give up is the ability to avoid an awkward line in the management accounts, and that is not a treasury objective.

Close-out: when the exposure disappears

The third case is a forward with nothing left to hedge. The customer cancelled, the supplier lost the order, the project was shelved. You hold USD 2 million forward for a payment that will never be made.

A forward cannot be torn up. It can be closed by dealing the opposite contract for the same date, which leaves you with a net AUD amount to pay or receive on the value date and no currency changing hands. If the AUD has weakened since you dealt, you receive cash. If it has strengthened, you pay. That cash is the mark-to-market of the position, crystallised.

Two things catch finance teams here. The first is timing. Close it out promptly, the day you learn the exposure is gone. An unhedged forward is a speculative currency position, and every day you leave it open you are running a directional view with the company's money whether you intend to or not. The discomfort of realising a loss today is not a reason to hope it shrinks by Friday.

The second is credit. A forward book runs against a credit line or a margin facility with your provider, sized on potential future exposure. An out-of-the-money position consumes that line. Close-outs release it, but a close-out that leaves you owing a six-figure AUD amount on a date two months away is still a credit exposure until it settles, and some providers will call for cash margin against it in the interim. Know your facility terms. A margin call on a hedge you no longer need, arriving in the same week the underlying order was lost, is the kind of thing that gets treasury policies rewritten in a hurry.

A worked example

Take an importer with a USD 2,000,000 supplier payment due 30 September. In March, with spot at 0.6640 and six-month forward points of plus ten, the finance manager books a forward to buy USD 2,000,000 at 0.6650. The AUD cost is locked at 3,007,519.

In early August the supplier advises the goods will not ship until late October and the invoice will be payable on 30 November. Spot has moved to 0.6900. The AUD has strengthened, so the forward is out of the money: USD is now cheaper than the rate the importer locked.

The honest extension works like this. The provider closes the original contract by selling USD 2,000,000 for 30 September at the current forward rate for that date, roughly 0.6905. That leg realises AUD 2,896,451. Against the AUD 3,007,519 the importer committed to pay, the loss is AUD 111,068, settled on 30 September or netted into the new deal as a visible adjustment. The provider then opens a new forward to buy USD 2,000,000 for 30 November at around 0.6910, an AUD cost of 2,894,356.

Add the two together and the all-in AUD cost of the USD 2,000,000 is 3,005,424, a few thousand dollars better than the original 3,007,519 because of the forward points earned over the extension. The importer is exactly where it would have been had it dealt for November in the first place. The hedge did its job. The loss on the original contract is real, but it is matched by a purchase that is now AUD 111,000 cheaper in spot terms than it was in March.

Now the HRR version. The provider rolls the contract to 30 November at a rate near 0.6650 plus a funding adjustment. Say it prints at 0.6642. The new confirmation shows USD 2,000,000 at 0.6642, an AUD cost of 3,011,141. No loss appears anywhere. The management accounts for September show no FX line. The board sees a hedge rate that looks like the original one and asks no questions.

But the importer has paid AUD 5,717 more than it would have under the honest extension, that difference being the provider's funding charge on AUD 111,068 for two months at a rate you were never quoted and could not have benchmarked. The loss has been deferred, not avoided, and it has cost money to defer. On one contract that is a rounding error. Across a book that rolls a dozen contracts a year with an average extension of six weeks, it is a material, invisible leakage, paid to the provider for the privilege of not having to explain a number to the board.

Window forwards as the pre-emptive answer

A lot of extension activity is avoidable. If you know the settlement date is uncertain at the time of dealing, and for anything involving sea freight you do, hedge with an instrument that tolerates the uncertainty.

A window forward, sometimes called a flexible or time-option forward, lets you draw down the currency at any point during a specified period rather than on a single date. You might book USD 2 million with a window from 15 September to 15 November. Draw it in one lot or several, on any business day in that range, at the agreed rate.

The price of that flexibility is that the provider assumes you will draw down at the worst point for them within the window, so the rate reflects the forward points for the least favourable date. With the AUD/USD differential close to zero the cost is modest, typically ten to twenty-five pips of give-up relative to a fixed-date forward for the middle of the window. For a two-month window on USD 2 million that is somewhere between AUD 4,500 and 11,000. Set that against the funding charge in the HRR example and the window often pays for itself.

Windows are not the answer to everything. If your payment terms are fixed and your supplier is reliable, a plain forward is cheaper. Use windows for the exposures where the date is a range rather than a point, which for most importers is the majority of the sea-freighted book and very little of the air-freighted one.

Governance: who decides, and on what basis

Adjustments to live forwards need the same authority and documentation as the original trade, and in practice they get less. The original forward is booked under the hedging policy with a designation memo and an approval. The extension three months later is done over the phone by whoever picks up when the dealer calls to say the contract is maturing tomorrow and there is no drawdown instruction on file.

Fix this with three things in the policy.

First, a rule that every adjustment is a new trade requiring the same approval tier as the original, and that historic rate rollovers are prohibited or require CFO approval with a written commercial reason. If your provider offers them by default, tell them in writing not to.

Second, a monthly maturity review. Two to three weeks before each forward matures, someone confirms the underlying cash flow is still expected, on what date, and in what amount. Mismatches get dealt with then, calmly, with time to price alternatives, rather than on maturity day.

Third, a reason code on every adjustment: shipment delay, early invoice, order reduction, order cancellation. Over a year those codes tell you something about your forecasting. If a third of your forwards are being extended for shipment delays, your hedging tenors are systematically too short and you should be dealing for the expected date plus a fortnight, or moving to windows.

The accounting follows the governance. Under AASB 9, rolling a hedging instrument into a new one as part of a documented hedging strategy is not treated as expiry or termination of the hedge, so hedge accounting can continue without the relationship being discontinued. That only holds if the rollover was contemplated in the hedge documentation from the outset. If your designation memo describes a fixed-date forward against a fixed-date payment and says nothing about extension, the auditor is entitled to ask why the hedge relationship survived a change that the documentation did not anticipate. Write the rollover provision into the template now.

Cancellation is the harder accounting case. If the forecast transaction is no longer expected to occur, any cumulative gain or loss sitting in the cash flow hedge reserve in other comprehensive income has to be reclassified to profit or loss immediately. Not when the forward matures. Not when you get round to closing it. When you conclude the transaction will not happen. The auditor will want to see the date you reached that conclusion and the evidence for it.

Common mistakes

Extending to avoid a loss. This is the big one. A shipment slips a fortnight and the forward is out of the money, so the finance manager extends by two months rather than two weeks, reasoning that the AUD might come back. It is a directional bet dressed as an operational adjustment. If the extension period does not match the revised payment date, it is speculation, and it should be treated as such in the policy.

Letting the provider run the maturity process. Providers will call the day before maturity and ask what you want to do. By then your options are whatever they can price in the next hour. Run your own diary.

Ignoring the points. Because the AUD/USD differential has been small, teams have stopped checking the forward points on adjustments. Differentials change. A widening of fifty basis points between the RBA and Fed cash rates turns a two-month roll from a non-event into a meaningful cost, and if you are not checking you will not notice until the year-end reconciliation.

Rolling the same contract repeatedly. A forward that has been extended three times is a forward against an exposure that was never properly forecast. After the second extension, stop and ask whether the underlying transaction is real.

Leaving cancelled exposures open. An order is cancelled on a Tuesday and the forward is closed the following Thursday week because nobody told treasury. The nine business days in between are pure currency speculation, and nobody approved it.

Not knowing the facility. Margin terms, close-out mechanics, whether partial drawdowns are permitted, whether HRRs are offered by default: these are in the terms of business you signed when the facility was opened. Read them before the first adjustment, not during it.

Where this leaves you

The rate you deal at is the part of hedging everyone watches. The adjustments afterwards are the part that determines whether the programme actually delivered the rate you thought you had locked. Pre-delivery is cheap and harmless. Close-out is a loss or gain you already have and should recognise the day the exposure disappears. Extension is where the money leaks, and it leaks through historic rate rollovers that turn a hedging loss into an unpriced loan from your counterparty.

Take losses when they occur, in the ledger, where the offsetting benefit on the underlying purchase can be seen alongside them. Use windows for exposures with uncertain dates. Give every adjustment the same authority as the original trade. Run your own maturity diary rather than the provider's. None of this is difficult. It is just the part of the job that happens after the trade is done, when attention has moved on, and it is precisely because attention has moved on that it is worth writing down.

---

David Dowling • 23 April 2026