The arm's-length principle is the rule at the centre of transfer pricing: a transaction between related parties must be priced as if it had occurred between unrelated parties dealing independently, each looking after its own interest. Everything else in transfer pricing — documentation, method selection, comparables — exists to test and support that one idea. This guide explains the principle itself, what drives it, and how the recognised pricing methods relate to it. For how the principle is applied in Canada specifically, see our transfer pricing primer and documentation guide.
The principle itself
When two parties are related — common ownership, control, or another relationship that removes the independence ordinary trading partners have — there is no natural market tension forcing the price toward a fair outcome. A parent company selling to its own subsidiary has no independent counterparty pushing back on price the way an unrelated buyer would. The arm's-length principle fills that gap: it asks what price, or what terms, unrelated parties acting independently would have reached for the same transaction under the same circumstances, and requires the related parties to use that outcome instead.
What makes parties non-arm's-length
Non-arm's-length status usually flows from control: a parent and its subsidiary, sister companies under common ownership, or a company and a shareholder who controls it are the clearest examples. The concept can also extend to relationships where the facts show the parties are not genuinely dealing with each other independently, even without a formal control relationship — the test looks at whether the parties' interests are actually aligned in a way that removes independent bargaining, not just at the org chart.
Why comparability drives everything
The arm's-length price is only meaningful in reference to what comparable independent parties actually did. That makes comparability the foundation of the entire analysis:
- Product and market comparability. The same or a similar product, sold into a similar market, under similar economic conditions.
- Functional comparability. The comparable transaction or company should perform similar functions, use similar assets and bear similar risks to the related party being tested — a distributor that carries inventory risk is not a good comparable for one that does not.
- Contractual comparability. Similar terms — payment terms, volume, exclusivity, warranties — because these terms materially affect price.
Weak comparables produce an unreliable conclusion regardless of how carefully the arithmetic is done afterward. Most of the analytical work in a transfer-pricing study goes into finding and adjusting comparable data, not into the final calculation.
The recognised pricing method families
A handful of method families are widely recognised for testing or setting an arm's-length price. Each fits a different fact pattern, and none is mechanically correct for every transaction:
| Method | How it works | Best suited to |
|---|---|---|
| Comparable uncontrolled price (CUP) | Compares the price charged in the related-party transaction directly to the price charged in a comparable transaction between unrelated parties | Transactions with close, reliable direct market comparables — commodities, standardized products |
| Resale price method | Works back from the price the related buyer resells the goods for, less an appropriate resale margin | Distribution arrangements where the reseller adds limited value before resale |
| Cost-plus method | Adds an appropriate markup to the costs incurred by the party supplying goods or services | Manufacturing or service arrangements where cost is a reliable base and value-add is limited |
| Transactional profit split | Splits the combined profit from the transaction between the related parties based on their relative contributions | Highly integrated transactions where both parties contribute unique, valuable functions |
| Transactional net margin method (TNMM) | Compares the net profit margin earned on the transaction to margins earned by comparable independent parties | Cases where reliable gross-margin comparables are not available but net-margin comparables are |
Method selection depends on the functional profile of the parties and the quality of comparable data available for each candidate method — the same transaction can sometimes be reasonably tested more than one way, and the choice is a matter of which approach the facts and evidence support best, not a fixed national ranking that applies regardless of circumstances.
How this connects to documentation
The arm's-length principle is the standard being met; documentation is the proof that it was met. A taxpayer's contemporaneous file records the functional analysis, the method chosen and why, and the comparable data relied on — the full chain from "these parties are related" to "here is why this price is arm's-length." See our transfer pricing documentation guide for what that record should contain.
How RN Canada helps
RN Canada helps Alberta and BC businesses with related-party cross-border dealings understand where the arm's-length principle applies, work through the functional analysis that drives method selection, and connect that analysis to the documentation the rules expect. For the structural picture behind cross-border groups, see corporate finance and capital restructuring; the transfer-pricing analysis itself often warrants specialist tax-law support alongside our advisory.
This page is general information, not personalized tax, accounting, or legal advice. Speak with RN Canada about your specific situation.
Frequently asked questions
It is the rule that a transaction between related parties must be priced as if it had occurred between unrelated parties dealing independently, each acting in its own interest. In Canada it underpins the section 247 transfer-pricing rules, which let the CRA adjust prices and reallocate income where a related-party transaction departs from arm's-length terms.
Common ownership or control is the usual trigger — a parent and subsidiary, sister companies under common ownership, or a company and a shareholder who controls it. Parties can also be considered non-arm's-length where the facts show they do not deal with each other independently, even without formal common control.
The arm's-length price is defined entirely by reference to what comparable independent parties would have agreed to. If the comparable data is not genuinely similar in product, market, function and risk, the resulting price conclusion is unreliable no matter how carefully the calculation itself is done. Comparability is the foundation every method is built on.
The commonly recognised method families are the comparable uncontrolled price (CUP) method, the resale price method, the cost-plus method, the transactional profit split method, and the transactional net margin method (TNMM). Each suits a different fact pattern, and choosing between them depends on which one the available comparable data supports most reliably.
The methods are not applied in a rigid, one-size-fits-all hierarchy. The right method for a given transaction depends on the facts — the functions, assets and risks involved, and the quality of comparable data available for each method — rather than defaulting mechanically to the same method regardless of circumstances.