Multi-currency cash forecasting: building a reliable 13-week forecast

There is plenty written about cash forecasting. Almost all of it implicitly assumes the company has one currency. Yet the moment a dollar account, a sterling receipt or a yuan supplier enters the perimeter, three problems appear that none of those methods address: at what rate do you convert the forecast, how do you tell an activity variance from an FX variance, and what do you do with the hedges already in place.

This guide covers the full build, starting from the simple case and adding the currency dimension at each step.

1. Why 13 weeks

The 13-week horizon — a rolling quarter — is not an arbitrary convention. It matches the window over which a finance team has both reasonable visibility and real room to act.

Past 13 weeks, a weekly forecast becomes an extrapolation of the budget: it consumes time without producing a decision. Inside 13 weeks, you do not have the lead time to negotiate a facility or defer an investment.

2. How the table is built

One rule of construction: forecast by nature of flow, never by bank account. The bank account is the result, not the unit of analysis.

3. Handling currencies: the question of the rate

This is where most forecasts lose their usefulness. Three practices coexist, and only one is workable.

The logic is direct: a hedged flow already has its rate, and that rate is known. An unhedged flow will be converted at a future date, and the best unbiased estimate of that future rate is the forward rate for that maturity — not today's spot, and not a market forecast.

4. Keep a forecast per currency AND in euros

The table must exist in two views, produced from the same base.

• A per-currency view: how many dollars come in and go out each week. This drives conversion and hedging decisions, and depends on no rate at all.
• A consolidated euro view: the group's cash position. This drives financing and investment decisions.

5. Analysing variances: the three-way split

A forecast is only worth having if the gap between forecast and actual is explained. Across currencies, a headline variance means nothing until it is broken down.

1. 1
   Volume variance

   The currency flow was larger or smaller than expected. That is an activity variance — it belongs to sales and operations.

2. 2
   Timing variance

   The flow arrived, but not in the expected week. That is a collections or process variance — it belongs to credit control and sales administration.

3. 3
   FX variance

   The flow arrived in the right amount on the right date, but converted at a different rate. That is a financial variance — it belongs to hedging.

Mixed together, these three causes make the forecast unusable and demoralise the team producing it: every variance looks like FX, so nothing gets fixed. Separated, each one points to an owner and an action.

6. Updating: rhythm and discipline

7. The mistakes that ruin a forecast

• Forecasting from revenue rather than from cash receipts: between the two sit payment terms, average delay and disputes. The gap routinely reaches three weeks.
• Forgetting VAT: in an importing business, import VAT and its reverse charge create timing gaps of hundreds of thousands of euros that appear in no operating budget.
• Ignoring margin calls on hedges: they land at the worst moment by construction — when the market moves sharply.
• Converting at today's spot rate: see point 3. This is the error that makes variance analysis impossible.
• Producing a forecast nobody reads: if the executive committee does not take one decision a quarter from this document, it serves no purpose and will eventually be abandoned.