Risk and hedging transactions: what software can run and what still needs a treasurer

Software can aggregate exposures, calculate hedge positions, match confirmations, post hedge accounting entries and reconcile to the ledger. People still set the strategy, decide when and how much to hedge, and approve trades. None of it works until exposure data is complete and trusted.

Where software already does the work

The mechanical end of hedging is well served. A treasury system can pull balances and forecast flows by currency and entity, then net them into a single exposure view. It can value open positions against a market data feed every day without anyone rekeying a rate.

Once a trade is booked, the system can match the bank's confirmation against the deal ticket and flag any field that disagrees. Settlement instructions can be generated from standing data. Foreign exchange conversions on settlement follow the same rules every time, which is exactly what software is good at.

Hedge accounting is the other strong candidate. Revaluation entries, reclassifications out of reserves and the journals that follow a hedge relationship through its life can be produced by rule. Posting them to the general ledger and reconciling the hedge subledger back to ledger balances is routine work. Each balance should trace to a trade, and each trade to its confirmation.

AI tools add something narrower. They are useful for reading free text confirmations, spotting unusual exposure movements in a forecast, and drafting hedge documentation or commentary for a treasurer to edit. They do not replace the controls around a trade.

Where a person has to stay in the loop

The hedging strategy is a judgment about risk appetite. Someone has to decide which exposures matter, what hedge ratio the business can live with, and which instruments are allowed. That belongs to the treasurer and whoever approves the policy.

Timing is human too. Interest rate and currency calls depend on market conditions, funding plans and conversations a model never hears. A system can propose a trade to bring a position back inside policy limits. Approving it, choosing the counterparty and executing within delegated authority stays with a dealer and an approver who are not the same person.

Hedge designation needs care. Deciding that a derivative hedges a specific forecast transaction, and documenting why, involves accounting judgment. When effectiveness testing fails or a forecast flow no longer looks likely, someone has to decide what happens to the relationship and the reserve balance. Exceptions in general, from a mismatched confirmation to a counterparty limit breach, need a named owner.

What has to be true about the data first

The weak point is almost always the exposure forecast. Business units submit flows late, in different formats, with currencies and entities coded inconsistently. Automating on top of that just produces wrong hedges faster.

Before switching anything on, entity, currency and counterparty codes need to be identical across the ERP, the treasury system and the bank feeds. Every trade should be captured in one place, not partly in a spreadsheet. The market data source used for valuation must be agreed with the accounting team so revaluations match what auditors will test.

Hedge relationships have to be linked to their underlying items in the system itself, not described in a separate memo. Without that link, hedge accounting cannot be generated by rule and the reconciliation will never close cleanly.

Standing settlement instructions deserve their own check. A wrong account number held in master data pays the wrong party every time.

Questions to ask the people who run it

• Where does the exposure number actually come from on the day a hedge is placed, and who adjusts it by hand?
• Which trades are agreed by phone or chat before they appear in the system?
• How are bank confirmations checked today, and what happens when one is late?
• Who can change settlement instructions, and is that change reviewed by anyone else?
• When a forecast flow is cancelled, how does the hedge accounting team find out?
• Which spreadsheets would break the month end if they disappeared?
• How often does the hedge position get overridden outside the policy, and who signs that off?
• What do the auditors ask for every year that takes longest to produce?

The answers usually reveal a side process the documentation never mentions. That side process is where automation either needs to absorb the work or will quietly fail.

Where automation tends to go wrong

Teams often automate the trade booking first because it is visible. The bigger gain sits upstream in exposure collection and downstream in accounting. Starting in the middle leaves manual effort at both ends.

Segregation of duties can also erode. When one system can propose, approve and settle, roles must be configured so no single user holds every permission. Review the access setup before go live and again after the first change request.

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.