Hedging strategies in forex: calculating the net cost of carry

A forex hedge priced at 42 forward points is not a 42-point hedge cost. The economic result also includes the bid-ask spread, execution markup, cross-currency basis, collateral funding, maturity mismatch and, for rolling instruments, the path of future rollover charges. The interest-rate differential is only one input.

This distinction matters most when hedge notionals are large and maturities are short. A 10-basis-point pricing error on a $50 million equivalent exposure is not analytical noise. It is a direct variance in hedge P&L. Hedging strategies in forex must therefore be measured through executable forward pricing, not through policy-rate comparisons.

The hedge cost is the price of converting an uncertain future spot rate into a known forward rate, plus every friction required to maintain that conversion.

Beyond interest-rate differentials: the carry stack

The market convention starts with covered interest parity. Ignoring transaction costs, the forward-to-spot relationship is:

\(F/S = (1 + i^*) / (1 + i)\)

Where:

• \(S\) is the spot exchange rate.
• \(F\) is the forward exchange rate for the same settlement horizon.
• \(i^*\) is the gross return in one currency.
• \(i\) is the gross return in the other currency.
• The interpretation depends on the quotation convention.

This is a parity relationship, not a forecast. A forward premium does not establish that the spot rate will rise. A forward discount does not establish that it will fall. It establishes the rate at which the market will exchange two currencies at a future date, conditional on current funding conditions and balance-sheet constraints.

For a foreign-currency payable, quoted as domestic currency per one unit of foreign currency, the basic forward-versus-spot difference is:

\(Carry_{forward} = N \times (F - S)\)

Where \(N\) is the foreign-currency notional.

If \(F > S\), locking the payable at the forward rate produces a positive cash difference relative to the current spot reference. If \(F < S\), the same formula produces a benefit. The sign reverses when the exposure reverses. It also reverses when the quotation convention changes. This is where many retail hedge calculations fail: the trader applies the same "positive swap equals cost" rule to EUR/USD, USD/JPY and an inverted internal accounting quote.

A complete net-carry estimate requires at least five components:

1. Embedded forward-point carry. The difference between the executable forward rate and spot rate, adjusted for hedge direction and notional.

2. Cross-currency basis. The premium or discount embedded in the swap curve beyond the simple risk-free rate differential.

3. Bid-ask cost. The difference between the rate available to buy protection and the mid-market rate used in preliminary models.

4. Commissions and dealer markup. Explicit commissions, ticket fees, prime-broker charges, or compensation embedded in a wider quote.

5. Rollover and collateral effects. Relevant for rolling spot, CFDs, non-deliverable structures and collateralised derivatives. Not automatically applicable to listed FX futures.

The market does not charge a single universal "forex hedge rate." It produces instrument-specific pricing.

Decoding forward points without creating a sign error

Forward points are defined as:

\(Forward\ Points = F - S\)

They are the observable carry input. They are not, by themselves, an annualised percentage and not a complete P&L estimate.

Consider a hypothetical importer with a EUR 12 million payable in 90 days. The accounting currency is USD. The market quote is USD per EUR.

The gross forward carry cost relative to spot is USD 90,000:

\(12,000,000 \times (1.0875 - 1.0800) = 90,000\)

That number is not yet the all-in hedge cost. It excludes the rate at which the importer can actually buy EUR forward. A dealer's executable offer may be above the displayed midpoint. It excludes bank fees. It excludes any internal funding charge attached to posting collateral. It also excludes the result on the underlying payable, which is the reason the hedge exists.

A hedge should not be evaluated against a hypothetical unhedged spot outcome. That outcome is random. The operational comparison is between:

• a known locked conversion rate;
• the budget or accounting reference rate;
• the current spot rate;
• the distribution of possible future spot outcomes;
• the cost of residual mismatch after the hedge is placed.

The forward is a certainty instrument. The unhedged position is a volatility position.

Match the quote before calculating anything

The calculation must first identify the numerator and denominator currencies.

For EUR/USD at 1.0800, the quotation means USD 1.0800 per EUR 1. A EUR payable is multiplied by the rate to obtain the USD obligation. A higher EUR/USD forward therefore raises the locked USD cost.

For USD/JPY at 150.00, the quotation means JPY 150 per USD 1. A USD payable is multiplied by the rate to obtain the JPY obligation. A higher USD/JPY forward increases the locked JPY cost. But an analyst reporting P&L in USD must convert the resulting JPY amount using a separate convention. The economic direction cannot be inferred from "points are positive" alone.

The minimum control set is mechanical:

• State the exposure currency and the reporting currency.
• State whether the exposure is a payable, receivable, asset, liability or forecast transaction.
• State the market quotation convention.
• Use the executable bid or ask, not the screen midpoint.
• Match the forward maturity to the underlying cash-flow date.
• Separate the hedge's standalone carry from the combined hedge-plus-underlying outcome.

This is not administrative detail. A reversed quote or wrong-side quote converts a hedge estimate into a directional error.

Forward points are observable. Net carry is constructed. The difference is execution, basis and instrument design.

Cross-currency basis: the spread between parity theory and funding reality

Covered interest parity is the baseline. It is not a guarantee that the traded forward curve equals the simple ratio of two benchmark interest rates.

FX swap rates can include an additional premium related to the supply and demand for a currency's funding, particularly US dollars. This premium is commonly expressed through the cross-currency basis. It reflects the cost of obtaining balance-sheet capacity and funding through swap markets rather than through a frictionless cash-and-bond transaction.

The basis cannot be assumed to be zero. Persistent deviations from covered interest parity have been linked to strong forward hedging demand and the balance-sheet costs carried by arbitrage intermediaries. The practical implication is direct: a hedge may be more expensive than the policy-rate differential implies even when the exposure, maturity and counterparty credit profile appear unchanged.

The 2022 USD funding episode provides a useful stress reference. During February through December of that year, temporary US-dollar FX-swap borrowing premia were reported in a range of roughly 80 to 150 basis points across maturities. That was not a permanent level and not a current quote. It demonstrates the scale at which basis can dominate a simplified carry model.

A treasury desk holding foreign assets often sees this effect most clearly. The asset may earn a higher local nominal yield than the domestic market. Once the currency is fully hedged, the forward curve and basis can remove most or all of that nominal advantage. In some cases, the hedged yield is below the domestic alternative.

The relevant calculation is not:

\(Foreign\ yield - Domestic\ yield\)

It is closer to:

\(Hedged\ return = Foreign\ asset\ return - FX\ hedge\ cost - execution\ cost - residual\ basis\ risk\)

The final term remains even after a forward is executed if the underlying asset's cash flows do not align exactly with the hedge maturity, currency denomination, valuation convention or hedge ratio.

Basis risk is not only cross-currency basis

Cross-currency basis is a market pricing component. Basis risk is broader. It includes every mismatch between the exposure and the hedge.

A portfolio manager hedging a CAD-denominated equity portfolio with USD/CAD forwards may still retain equity-linked currency exposure if the portfolio value changes materially before the forward matures. A fixed notional hedge becomes under-hedged after an equity rally and over-hedged after an equity decline.

Cross-hedging techniques add another layer. A manager may hedge an illiquid currency exposure with a correlated liquid currency. The hedge ratio cannot then be set at 100% merely because the two currencies have historically moved in the same direction.

The minimum-variance hedge ratio is commonly expressed as:

\(h^* = \rho \times \sigma_E / \sigma_H\)

Where:

• \(h^*\) is the minimum-variance hedge ratio.
• \(\rho\) is the currency correlation coefficient between the exposure and hedge instrument.
• \(\sigma_E\) is the standard deviation of changes in the unhedged exposure.
• \(\sigma_H\) is the standard deviation of changes in the hedge instrument.

A correlation coefficient of 0.70 is not a 70% hedge. The volatility ratio matters. Correlation also changes across market regimes. A 12-month sample can understate stress-period correlation breaks, particularly where commodity prices, capital controls or local liquidity dominate the exposure currency.

Quantifying spreads, commissions and rollover charges

The net cost of carry should be evaluated at executable prices. Mid-market pricing is useful for attribution. It is insufficient for trade approval.

For a forward hedge, the pricing sequence is straightforward:

1. Identify the spot bid or ask appropriate to the transaction direction.

2. Identify the forward points bid or ask for the same maturity.

3. Construct or obtain the outright executable forward rate.

4. Calculate the cash difference against the chosen reference rate.

5. Add explicit commissions, settlement fees and credit charges.

6. Add estimated rollover costs only if the hedge will be rolled rather than held to settlement.

7. Report the result in both currency units and basis points of notional.

The bid-ask spread is a direct cost. A commission-free retail FX transaction is not a cost-free transaction. Dealer compensation can be embedded in the spread. This is especially relevant where a retail platform quotes a narrow headline spread during liquid hours but applies wider pricing around roll, data releases or thin regional sessions.

For rolling spot or CFD-style positions, daily financing must be treated separately from forward points. Providers may post debits or credits linked to rate differentials, but the effective rate can diverge from an interbank reference. Liquidity access, provider funding, risk management and internal markup all affect the final charge.

A practical all-in framework is:

\(Net\ hedge\ cost = Forward\ point\ effect + spread + commissions + markup + projected\ rollover + collateral\ funding\)

Each variable must be converted into the same reporting currency and horizon. A spread quoted in pips, a commission quoted per million and a rollover quoted in annualised percentage terms cannot be added until they are normalised.

A rolling-hedge example

Assume a USD-based investor owns a foreign asset and maintains a series of one-month currency hedges rather than a single annual forward. The annual hedge cost cannot be inferred by multiplying one month's forward points by 12.

The realised annual result depends on:

• the forward curve on each roll date;
• spot changes that alter the hedge notional relative to the asset value;
• the bid-ask paid at each roll;
• holiday and settlement-calendar effects;
• changing cross-currency basis;
• any cash collateral funding;
• the degree to which the asset distribution dates match the hedge schedule.

The forward curve is repriced each month. A hedge programme is therefore a sequence of executed transactions, not a single static interest-rate calculation.

A rolling hedge is best understood as a calendar problem with balance-sheet consequences: each roll date is a fresh market event, not an automatic extension of the prior leg. Settled assumptions about recurring forward points break down the moment the curve, basis or funding window shifts between rolls. The same scheduling discipline applies to FX settlement dates — they are fixed at execution, and missing a date through a calendar error is a costly operational failure.

Comparing hedge vehicles: forwards, FX swaps, CFDs and futures

The correct instrument depends on the cash-flow structure. A corporate payable, an institutional portfolio hedge and a retail directional position do not carry the same financing mechanics.

An FX swap consists of an initial exchange of currency principals and a reverse exchange at a rate agreed when the transaction is entered. It is frequently used for funding and liquidity management. The swap points on the far leg represent the embedded carry. The position may be rolled, but each roll is a new market transaction.

In exchange-traded FX futures, the interest-rate differential is incorporated into the futures curve at entry. Adding a retail-platform daily rollover charge on top of the futures carry would double-count financing. Futures do create margin variation and contract-roll requirements, but those are not the same as a CFD overnight swap debit.

Retail CFDs are operationally simpler for small positions but require the most scrutiny of published financing terms. The effective overnight rate may be based on a benchmark plus or minus a provider adjustment. The adjustment can be more material than the nominal interest differential for short-duration hedges.

A forward can be economically cleaner for a known receivable or payable because the notional and settlement date are fixed. It becomes less precise when the exposure is uncertain, such as a fluctuating foreign equity allocation or a projected revenue stream. In that case, staggered maturities and dynamic hedge ratios reduce concentration at a single roll date.

Why the hedge ratio matters as much as the instrument

The choice of vehicle sets the carry structure. The hedge ratio determines how much of that carry is actually deployed.

A 100% nominal hedge ratio assumes a one-to-one sensitivity between the exposure and the hedge instrument. That assumption fails as soon as the underlying exposure moves, the chosen hedge currency drifts away from the exposure currency, or the hedge notional is held constant while the exposure fluctuates.

Dynamic rebalancing – resetting the hedge notional at defined intervals or tolerance bands – restores the ratio but introduces transaction costs on every adjustment. Those costs sit inside the net carry calculation and must be amortised over the planned hedge horizon rather than ignored.

A static hedge ratio is acceptable when the exposure is contractually fixed. It is rarely acceptable for mark-to-market portfolios, projected cash flows or balance-sheet items that respond to operating or equity-market movements.

A disciplined net-carry model for hedge approval

A hedge model should produce three outputs: locked cash flow, all-in cost and residual risk.

The locked cash flow is the contractual conversion outcome. The all-in cost is the difference from the selected reference after spreads, commissions and funding. Residual risk measures what remains unhedged due to timing, notional, correlation or instrument mismatch.

For a foreign-currency payable, the core worksheet can be structured as follows:

Each line is a control point. A missing input is a missing cost.

The model should also be tested against three sensitivities:

• Curve sensitivity. Move the forward rate by one standard deviation of its recent 30-day distribution and recompute. If the result is identical to the base case, the inputs are not independent and the model is masking volatility.
• Basis sensitivity. Add and subtract a representative basis shock (for example, 25 basis points for a major pair, more for an emerging-market currency) and recompute. A hedge that looks cheap under a flat basis can become expensive under a stressed basis.
• Rollover sensitivity. Re-run the calculation using the highest and lowest daily financing rate observed in the prior quarter. The midpoint gives a false sense of precision.

What the model should not do

It should not produce a single "hedge cost" without a defined reporting currency, a defined reference rate and a defined hedge direction. It should not combine the standalone carry of the hedge with the P&L of the underlying exposure into a single number, because the hedge is a certainty instrument and the underlying is a volatility position. The two are reported separately and then combined only at the reporting layer.

It should not treat a screen midpoint as an executable price. It should not treat one provider's published rollover as a market reference. It should not extrapolate a short-dated forward points quote to a longer horizon without checking the full curve.

Most importantly, it should not absorb a unidirectional view on the future spot rate. Hedging strategies in forex are valuation exercises, not directional bets. If the model begins to talk about "expected spot" inside the carry calculation, it has crossed into a different problem.

Hedging strategies in forex are not priced by interest-rate differentials. They are priced by what the market will actually deliver at settlement, after every friction between the model and the trade ticket.

The discipline is unglamorous. It is also the difference between a hedge that protects a margin and a hedge that quietly costs the same margin it was supposed to protect.