Currency exposure and hedging: what to automate and what to keep human

Software can gather open foreign currency balances, revalue them, match hedges to exposures and draft the accounting entries. People must still set the hedge ratio, judge whether forecasts are believable and approve trades. None of it works until every exposure carries a clean currency code, a reliable date and an owning entity.

How the work actually flows

Exposure comes from two places. Some of it is booked: receivables, payables, intercompany loans and cash held in a currency other than the entity's functional one. The rest is forecast: orders placed but not yet invoiced, expected sales, planned capital spend. Booked exposure is a fact sitting in the ledger. Forecast exposure is an opinion, and it changes.

Treasury nets these by currency and by entity, compares the result against policy, and decides what to cover. Trades are placed with banks, confirmed, and recorded. At each close, the hedges and the underlying items are revalued, effectiveness is tested where hedge accounting applies, and a foreign currency report goes to leadership. That report must tie back to general ledger balances, or nobody trusts it.

Steps software can run now

Collecting booked exposure is the clearest win. A treasury system or a well built data feed can pull open items from receivables, payables and intercompany ledgers, group them by currency pair and maturity, and present a net position without anyone touching a spreadsheet. Aging foreign currency receivables belongs here too, since a stale receivable is still an open exposure until it is collected or written off.

Revaluation is mechanical. Given trusted rates, a system applies them consistently and posts the gain or loss.

Trade confirmation matching is another strong candidate. Bank confirmations arrive in structured formats, and comparing them against what was booked in the treasury system is rule based work. Mismatches get flagged; matches get filed.

AI tools are useful at the edges. They can read free text on purchase orders to spot a currency clause that was never coded, or compare this period's forecast against last period's and highlight which business units keep overstating. They should suggest, not post.

Drafting the foreign currency report and checking each figure against ledger account balances can also be automated, as long as the mapping between treasury categories and ledger accounts is maintained.

Steps that still need a person

Deciding how much of a forecast to hedge is a judgment call. It depends on how often that business unit's forecasts have held up, on market conditions and on how much earnings volatility the board will tolerate. A model can propose a ratio. Someone accountable has to accept it.

Approving and executing trades needs a human with authority, and policy usually requires a second one. Counterparty choice involves credit limits, relationships and pricing that shift.

Hedge accounting designation is another. Documenting the relationship at inception and defending it to auditors is professional judgment, and errors here move results in ways that are hard to unwind.

Exceptions also stay human. When an exposure suddenly appears or vanishes, someone has to call the business and find out whether a deal slipped, was cancelled or was simply miscoded.

What the data has to look like first

Every open item needs a transaction currency distinct from the entity's functional currency, held in a field and not buried in a description. Due dates must be real, not defaulted. Intercompany balances must agree on both sides, or netting will show exposure that does not exist.

Forecasts need an owner and a version. Without both, nobody can tell whether a change in exposure reflects the business or just a new spreadsheet.

Rates must come from one agreed source with a fixed capture point. Mixing sources between revaluation and reporting creates differences that look like hedge ineffectiveness and are not.

Orders and commitments raised in foreign currency should be visible before they become invoices. If procurement records commitments only in the functional currency, the earliest signal of exposure is lost, and hedging always starts late.

Finally, the mapping from treasury positions to ledger accounts has to be documented and reconciled. Automation built on an unreconciled mapping produces fast, confident and wrong reports.

Questions to ask the people who run it

• Where does the exposure number really come from today, and who adjusts it before it reaches treasury?
• Which business units' forecasts get discounted informally, and on what basis?
• When a bank confirmation does not match, what happens next and who is told?
• Are there exposures that never appear in the system, such as contracts priced in foreign currency but invoiced locally?
• Which intercompany balances routinely disagree, and how is the difference cleared?
• What gets changed in the foreign currency report after it is generated, and why?
• Who has placed a trade outside policy, and what made it seem necessary?
• Which rate source does each team use, and has anyone checked that they agree?
• What would break first if the person who maintains the spreadsheet left?

APQC's Process Classification Framework® (PCF) is an open standard developed by APQC, a nonprofit that promotes benchmarking and best practices worldwide. To download the full PCF or to view definitions and measures, please visit www.apqc.org/pcf.