When can I use the CUP method instead of a margin method?
When a genuinely comparable transaction exists. That means an uncontrolled deal in the same product or right, on broadly the same terms, at a broadly similar point in the chain — a commodity with a quoted price, or a licence whose terms are public. Where such a comparable is available it is the most reliable method there is, because it compares the price itself rather than a margin derived from it. The difficulty is availability: most intercompany dealings have no close equivalent, which is why the most direct method is also the one least often usable in practice.
Is a published commodity price a valid comparable for intercompany sales?
It can be a strong starting point, because the product is standardised and the price is observable. What decides it is the distance between the quotation and your transaction. Quality or grade differences, the delivery point, who pays freight and insurance, contract volume, the date the price was struck, and payment terms all move the number, and each difference you bridge is an adjustment that has to be quantified and explained. Where those adjustments stay few and small, the comparison holds up well. Where they accumulate, the price comparison stops being the most reliable evidence available.
We licence our brand to a subsidiary — can we benchmark the royalty rate?
Sometimes. Licence agreements do reach the public record, and where you can find licences of comparable rights on comparable terms, a direct comparison of rates is possible. The care goes into whether the rights really match: the territory, exclusivity, the field of use, what support comes with the licence, who funds the marketing, and the term. A rate is only meaningful alongside those terms. Where the available agreements differ from yours in ways you cannot adjust for with any confidence, the analysis is better anchored on the returns the parties actually earn.
Do I have to test the CUP method first before using another one?
You do have to consider it, and say why you did not use it if you did not. Method selection is part of the documentation, so a file that goes straight to a margin method without addressing the availability of a price comparison has left out a step a reviewer expects to see. The record can be short: what you looked for, where you looked, and what you found or did not. That note is worth keeping even when the conclusion is negative, because the absence of a comparable is itself a finding about the transaction.
How many adjustments can I make before a comparable is no longer comparable?
There is no fixed limit; reliability is what you are watching. Each adjustment adds an assumption, and once the adjustments carry more weight than the observed price the comparison is really an estimate dressed as a comparison. The practical test is whether you could defend each adjustment separately and show where its inputs came from. When the answer starts to be no, reliability has shifted to a margin-based method, and choosing that method deliberately, with the CUP attempt documented, is a stronger position than a heavily adjusted price comparison.
Can a sale to a third-party customer support my intercompany price?
An internal comparable of that kind is often strong evidence, and it is routinely overlooked. If the group sells the same product or licences the same right to an unrelated party, the terms of that deal are directly on point and the supporting records already exist. Check the differences carefully: volume, market, stage in the chain, the customer's function, and the date. Where the third-party sale is at a different point in the chain — to an end user rather than to a reseller — the comparison needs an adjustment, and whether that adjustment is reliable decides whether the method holds.
What is the CUP method?
Comparable uncontrolled price. You find the price charged in a comparable transaction between unrelated parties and test your intercompany price against it. It is the most direct of the methods and the most persuasive when it fits, because it compares like with like at the transaction level. Its limit is data: close comparables exist for commodities and standard products, rarely for bespoke services or unique intangibles, which is where the margin-based methods take over. See our transfer pricing work.
Do we need transfer pricing documentation for a small group?
The obligation follows the existence of cross-border transactions with related parties, not the size of the group — which surprises founders with one foreign subsidiary and a management fee. Size affects which report is required: a local file, a master file, a country-by-country report. In Canada the practical trigger is timing, because documentation prepared by the filing due date is what stands between an adjustment and a penalty on top of it. See contemporaneous documentation in Canada.
