A workable intercompany reconciliation template holds one row per counterparty pair per account category per period, with both sides' amounts in transaction and group currency, the difference, a difference category, an owner and an ageing date. The template matters less than the rules attached to it: a documented matching key, a tolerance policy, a difference taxonomy and a hard cut-off before consolidation.
Intercompany differences are rarely fraud and rarely large individually. They accumulate: a timing difference here, an FX rate difference there, a rebilling booked to the wrong account. By year-end the elimination column carries an unexplained residual, and the auditor tests it at transaction level.
The fix is procedural, not technological. Groups running this well use a plain template with strict rules and a fixed monthly rhythm — long before they invest in a matching engine.
Anything less than this and you will not be able to explain a difference two months later. Anything more and it stops being filled in.
Categorising differences is what turns the reconciliation from a chore into a diagnostic. Five categories cover almost everything.
Intercompany balances are also transfer pricing data. Service recharges, loans and cost allocations that sit unreconciled become documentation gaps in the local file, and interest and financing flows feed Pillar Two computations. A clean pair-level reconciliation with a category and a rate reference gives all three processes the same underlying numbers — which is the whole point of reconciling at category rather than aggregate level.
Counterparty pair, entity-side balances in transaction and group currency, matched amount, difference, difference category, owner, age and resolution status — one row per pair, per period.
A spreadsheet is acceptable as a starting structure, but the reconciliation must be reproducible from the ledgers. Where the spreadsheet holds the only version of the truth, the evidence trail breaks at audit.