What it means
Forecasts come in different time horizons. Short-term forecasts, measured in minutes to days, are mainly the concern of traders, while medium and long-term forecasts of months or years feed into budgets, investment appraisals and financial plans.
Forecasting software can use several approaches. Some programs apply fundamental models based on interest rates, inflation and trade flows, others scan price charts for patterns, and newer products use statistical learning to find relationships in large data sets.
Good software typically includes data feeds, charting tools, back-testing (running a method on past data to see how it would have performed), alerts and reports. A treasury team may also value features such as scenario analysis, which shows the effect of different rates on profit.
No forecast is reliable. Exchange rates respond to news, politics and market mood, and research on currency prediction repeatedly finds that beating a simple random benchmark is very difficult.
For that reason, sensible businesses treat forecasts as a range with probabilities, and avoid basing the entire business plan on a single number. They often set budget rates conservatively, hedge a proportion of exposures, and track their forecast accuracy over time.
When choosing software, ask what data it uses, how its accuracy is measured, and whether it can be tested on past periods. Be wary of any product that promises guaranteed results, since that claim alone is a warning sign.
In practice
Real-world examples.
Example
A US toy importer buys from Europe and uses forecasting software to set its budget exchange rate for next year. It sets the rate at the midpoint of the software's range and hedges half of its exposure. The other half is left open so the company still benefits if the forecast turns out too pessimistic.
Example
A retail trader subscribes to a forecasting service that issues daily direction signals. She tests the service on a demo account for two months and tracks how often its calls turn out correct. She also notes the size of the losses on wrong calls, not only how often the service is right.
Example
A bank's economists publish a quarterly forecast for the pound and dollar, based on expected interest rate changes. Corporate customers use the forecast as one input to their hedging decisions. Smart treasurers compare several banks' forecasts and treat the spread between them as a measure of uncertainty.
Formula
Calculation
Forecast error (%) = (forecast rate - actual rate) / actual rate x 100
Dollar impact = foreign currency amount x (forecast rate - actual rate)
Suppose a US company forecast EUR/USD at 1.2240 for the end of the quarter, when it plans to buy 500,000 euros. The actual rate turns out to be 1.2000.
The forecast error is (1.2240 - 1.2000) / 1.2000 x 100 = 0.0240 / 1.2000 x 100 = 2.0%. The dollar impact is 500,000 x 0.0240 = $12,000, so the company budgeted $12,000 more than it needed to spend.
Tracking this error over many quarters shows whether the forecasting method is adding value or only adding noise.Case study
Seen in the real world.
Greenfield Logistics is an illustrative, fictional US freight business that pays several suppliers in Mexican pesos. Its finance team bought forecasting software and set budget rates from the software's predictions.
After a year, the controller compared forecast and actual rates and found that the software had been right on direction about half the time, no better than a coin toss. The average miss was large enough to move quarterly profit by several per cent.
In this illustrative case, the team kept the software for data and alerts but stopped using it as the sole basis for the budget. It adopted a rule to hedge a fixed proportion of exposure regardless of the forecast, so the budget no longer relied on a prediction. The controller now reports forecast error to the board each quarter as a standing item.
Watch out
Common mistakes.
- Treating a forecast as a fact, when it is an estimate with a wide range of possible outcomes.
- Paying for software without testing it, when back-testing and a trial period show if it adds value.
- Using a single forecast for the whole budget, when scenarios at higher and lower rates give a better view of risk.
Questions
People also ask.
Can forecasting software predict exchange rates accurately?
Not consistently, because rates react to unexpected events, so it is better viewed as a decision aid than as a prediction machine.
How should a business choose a budget rate?
Many use a forward rate or a conservative estimate, then hedge part of the exposure, so results are less dependent on guesses. A forward rate is a market price, which is a more neutral starting point than an internal opinion.
What does back-testing mean?
It means running a forecasting method on past data to see how it would have performed, although good past results do not guarantee future ones. Beware of methods tuned so finely to history that they only fit the past.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%Related