Arm's Length Principle

The arm's length principle is the test every transfer price is measured against. A controlled transaction meets it when the result is consistent with the result that would have been realized if uncontrolled taxpayers had engaged in the same transaction under the same circumstances. In the US regulations the phrase is the arm's length standard; in the OECD Guidelines and in most treaties it is the arm's length principle, and Article 9 of the model treaty is where it lives. The formulations differ in wording and in some of their machinery. They agree on the idea.

The word doing the work is "comparable." A result is only as reliable as the uncontrolled transactions it is measured against, and the regulations set out the factors that determine whether two transactions are comparable enough to support a conclusion: the functions performed by each party, the contractual terms allocating responsibility and benefit, the risks each party bears, the economic conditions of the relevant markets, and the property or services transferred. Where material differences exist, adjustments can sometimes bridge them, and where they cannot be reliably adjusted for, the comparable should be rejected. The rejection reasoning is itself part of a defensible study.

Two features of the standard get misunderstood in practice.

The first is that it produces a range rather than a number. Applying a method to a set of comparables gives a set of results, and any point in the arm's length range satisfies the standard. Where the comparables are of sufficiently high quality, the full range applies. Where comparability is imperfect, which describes almost every real analysis, the range is narrowed to the interquartile range, and an adjustment is generally made to the median rather than to the nearest edge of the range. This is why a taxpayer sitting just inside the range is in a materially better position than one sitting just outside it.

The second is that risk allocation is tested against conduct, not just contracts. A written agreement assigning a risk to one party carries weight only where that party has the financial capacity to bear the risk and actually exercises control over it. An entity contractually bearing inventory risk while another entity makes every purchasing decision will not be respected as the risk-bearer, and the return attributed to it will not survive.

The principle applies transaction by transaction rather than to a group's overall profitability. Aggregation is permitted where transactions are so interrelated that evaluating them together provides the most reliable measure, but a global operating margin that looks reasonable does not establish that a particular royalty or service charge was arm's length.

Related terms: Transfer Pricing, Section 482, Form 5472, Economic Nexus

Go deeper: Transfer Pricing Documentation: A Practitioner's Guide to Scoping and Defending It

Comparability work is evidence assembly: functional facts, contract terms, and the segmented financials to test them against. Adopt's agents build that record from the source systems and carry it forward year over year. Sign up free.