BG Pattern
BG Pattern
BG Pattern

November 17, 2026

What a group transfer pricing policy should fix before year-end

An open policy booklet, a pen on an unmarked signature line and a stack of study files

A group transfer pricing policy settles, once and in advance, how the group's intercompany prices will be set, tested and signed off. Where it is silent, the year-end close answers those questions under deadline, and the study written afterwards records whatever the numbers turned out to be.

The order matters more than the paperwork. A study evidences a year already gone. A policy should have said, before the first invoice, what shape that year's numbers would take.

Each decision inside it can be tested against a clause of Rule 10D, and the missing clauses surface in December, when the numbers are posted.

  • Rule 10D(1)(k) already requires a record of the assumptions, policies and price negotiations that critically affected the arm's length price.

  • Clause (i) asks for the method chosen and the reason, a decision taken before the transaction and reported after it.

  • A policy that names the tested party and the method for each class stops the study re-arguing the same structural points.

  • HMRC lists policies out of line with the OECD guidance among the causes of arrangements failing scrutiny, along with a drift between documented facts and the ground.

  • The document needs a refresh trigger and a named signatory, or the close becomes the place where policy gets decided.

What a group transfer pricing policy has to fix

A group transfer pricing policy names, for each class of intercompany transaction, which entity is tested, which method applies, how often the comparables are refreshed and who signs the result. It is written before the year begins, and a year that requires it to be rewritten is usually a year in which it was not followed.

The document has to settle a small set of decisions, and each one is a question the study would otherwise reopen.

  1. Which entity is the tested party for each class, and why the measurement sits with that entity.

  2. Which method applies to which class, set out class by class rather than as one preferred method for the group.

  3. When the comparables are refreshed, who runs the search, and what evidence closes it.

  4. Who signs the policy, what the signatory sees before signing, and who may change a price mid-year.

That makes it a governance instrument before it is a tax one. The close is a separate exercise: operational transfer pricing at the close is the quarterly testing of cost base and margin that keeps the posted numbers near the policy, and it cannot repair a policy that never named the tested party. If your copy has gone unopened, contact us, and we will read it against the transactions it governs.

Why a policy is not a study, and a study is not a policy

A transfer pricing study answers a question about the past. It records what the transactions were, what the comparables showed, and why the price sat where it did. A policy answers a question about the future, and the two fail differently.

A group with a study and no policy produces a defensible file each year and a fresh argument each year about the same structural choices: which entity carries the risk, which method fits the services, whether a guarantee has been priced. The file is complete, and the reasoning behind it is re-decided every year.

A group with a policy and no study holds a document nobody can test against a transaction. That is the older failure, and it has become the more expensive one, because the end of file-first, study-later puts the comparables and the margin range into the return itself.

The two also answer to different readers. The study travels with the file to the tax officer. The policy goes to the board, the auditor and the local finance director, which is why checking a transfer pricing study report covers different ground.

What Rule 10D already assumes the policy says

Rule 10D requires information and documents, and it does not ask for a document called a policy. It asks for the content of a policy, clause by clause.

Clause (k) is the one to read first. The rule requires "the assumptions, policies and price negotiations, if any, which have critically affected the determination of the arm's length price", in the text of Rule 10D. A group that has never written those assumptions down has nothing to put in the record, and rebuilding them a year later is a writing exercise with an audience.

Clause (i) asks for "the method selected as the most appropriate method along with explanations as to why such method was so selected". That is a decision about which method fits the facts of each transaction class, and it belongs to the year in which the transactions were priced.

Clause (f) reaches further back: "a record of the economic and market analyses, forecasts, budgets or any other financial estimates prepared by the assessee for the business as a whole". Forecasts are where a policy and a business plan meet, and the clause treats them as evidence of the price.

None of these asks the taxpayer to invent a document. Each asks for a record of decisions that already exist, which is an argument for writing them once and citing them, and that record is the starting point for what the documentation has to prove. The benchmark sits inside the policy too: a fresh search is expected each year, and a rolled-forward set is not treated as compliant, as the Academy of Tax Law's country guide has it.

The five clauses most often missing

The omissions tend to be the same ones, whichever jurisdiction the group files in, which makes them easy to check against your own document.

  1. Tested party, with the reason. Many policies name the entity whose margin is measured and stop there. The clause that prevents trouble says why, tied to the functions, assets and risks that entity carries.

  2. Method by transaction class. One preferred method for the group is the usual defect. Services, royalties and guarantees rarely fit the same method, so the policy should say which applies to each.

  3. Benchmark refresh cycle. Who runs the search, when, and what evidence closes it. A policy that names an annual refresh and an owner survives a busy year; one that says the benchmarks are reviewed periodically does not.

  4. The mid-year change rule. Prices move when volumes, functions or a contract change. The policy should say who may approve a change, on what test, and how it is recorded, so the file can show the reasoning at the time.

  5. Sign-off, with evidence. A signature with no stated basis carries little weight before an auditor. Name the signatory, list what they see before signing, and record the date.

Each clause is quick to draft and expensive to reconstruct. The practical test is whether a finance director in another jurisdiction could apply the policy next month without calling the group tax team to ask what it meant.

What the audit committee should ask before year-end

The policy is a governance document, so the committee's questions are about evidence: what was done, who approved it, and what the file shows.

HMRC's compliance facility lists one recurring cause of arrangements failing scrutiny: the facts the analysis rests on diverge from what happens in the business. That is set out in its guidance, which also notes that businesses change over time so that the functional profile "may become different from that originally assumed or intended".

Inland Revenue New Zealand puts the responsibility where the committee can see it: "It remains the responsibility of local management to ensure a company's transfer prices are in accordance with the arm's length standard." Its documentation guidance adds that where material local facts are omitted or differ from a group position, "the local documentation will be inadequate", and gives a policy to charge royalties at a uniform global rate as its example.

So the questions can be short. Has anything changed in what this entity does? Is the functional profile in this year's file the one the business runs, or the one the policy assumed?

Who signed, on what evidence, and could they produce it now? If those answers are thin, they will be thinner under a notice. Contact us before the numbers are locked, and we will test the document against the year it governs.

A group transfer pricing policy settles, once and in advance, how the group's intercompany prices will be set, tested and signed off. Where it is silent, the year-end close answers those questions under deadline, and the study written afterwards records whatever the numbers turned out to be.

The order matters more than the paperwork. A study evidences a year already gone. A policy should have said, before the first invoice, what shape that year's numbers would take.

Each decision inside it can be tested against a clause of Rule 10D, and the missing clauses surface in December, when the numbers are posted.

  • Rule 10D(1)(k) already requires a record of the assumptions, policies and price negotiations that critically affected the arm's length price.

  • Clause (i) asks for the method chosen and the reason, a decision taken before the transaction and reported after it.

  • A policy that names the tested party and the method for each class stops the study re-arguing the same structural points.

  • HMRC lists policies out of line with the OECD guidance among the causes of arrangements failing scrutiny, along with a drift between documented facts and the ground.

  • The document needs a refresh trigger and a named signatory, or the close becomes the place where policy gets decided.

What a group transfer pricing policy has to fix

A group transfer pricing policy names, for each class of intercompany transaction, which entity is tested, which method applies, how often the comparables are refreshed and who signs the result. It is written before the year begins, and a year that requires it to be rewritten is usually a year in which it was not followed.

The document has to settle a small set of decisions, and each one is a question the study would otherwise reopen.

  1. Which entity is the tested party for each class, and why the measurement sits with that entity.

  2. Which method applies to which class, set out class by class rather than as one preferred method for the group.

  3. When the comparables are refreshed, who runs the search, and what evidence closes it.

  4. Who signs the policy, what the signatory sees before signing, and who may change a price mid-year.

That makes it a governance instrument before it is a tax one. The close is a separate exercise: operational transfer pricing at the close is the quarterly testing of cost base and margin that keeps the posted numbers near the policy, and it cannot repair a policy that never named the tested party. If your copy has gone unopened, contact us, and we will read it against the transactions it governs.

Why a policy is not a study, and a study is not a policy

A transfer pricing study answers a question about the past. It records what the transactions were, what the comparables showed, and why the price sat where it did. A policy answers a question about the future, and the two fail differently.

A group with a study and no policy produces a defensible file each year and a fresh argument each year about the same structural choices: which entity carries the risk, which method fits the services, whether a guarantee has been priced. The file is complete, and the reasoning behind it is re-decided every year.

A group with a policy and no study holds a document nobody can test against a transaction. That is the older failure, and it has become the more expensive one, because the end of file-first, study-later puts the comparables and the margin range into the return itself.

The two also answer to different readers. The study travels with the file to the tax officer. The policy goes to the board, the auditor and the local finance director, which is why checking a transfer pricing study report covers different ground.

What Rule 10D already assumes the policy says

Rule 10D requires information and documents, and it does not ask for a document called a policy. It asks for the content of a policy, clause by clause.

Clause (k) is the one to read first. The rule requires "the assumptions, policies and price negotiations, if any, which have critically affected the determination of the arm's length price", in the text of Rule 10D. A group that has never written those assumptions down has nothing to put in the record, and rebuilding them a year later is a writing exercise with an audience.

Clause (i) asks for "the method selected as the most appropriate method along with explanations as to why such method was so selected". That is a decision about which method fits the facts of each transaction class, and it belongs to the year in which the transactions were priced.

Clause (f) reaches further back: "a record of the economic and market analyses, forecasts, budgets or any other financial estimates prepared by the assessee for the business as a whole". Forecasts are where a policy and a business plan meet, and the clause treats them as evidence of the price.

None of these asks the taxpayer to invent a document. Each asks for a record of decisions that already exist, which is an argument for writing them once and citing them, and that record is the starting point for what the documentation has to prove. The benchmark sits inside the policy too: a fresh search is expected each year, and a rolled-forward set is not treated as compliant, as the Academy of Tax Law's country guide has it.

The five clauses most often missing

The omissions tend to be the same ones, whichever jurisdiction the group files in, which makes them easy to check against your own document.

  1. Tested party, with the reason. Many policies name the entity whose margin is measured and stop there. The clause that prevents trouble says why, tied to the functions, assets and risks that entity carries.

  2. Method by transaction class. One preferred method for the group is the usual defect. Services, royalties and guarantees rarely fit the same method, so the policy should say which applies to each.

  3. Benchmark refresh cycle. Who runs the search, when, and what evidence closes it. A policy that names an annual refresh and an owner survives a busy year; one that says the benchmarks are reviewed periodically does not.

  4. The mid-year change rule. Prices move when volumes, functions or a contract change. The policy should say who may approve a change, on what test, and how it is recorded, so the file can show the reasoning at the time.

  5. Sign-off, with evidence. A signature with no stated basis carries little weight before an auditor. Name the signatory, list what they see before signing, and record the date.

Each clause is quick to draft and expensive to reconstruct. The practical test is whether a finance director in another jurisdiction could apply the policy next month without calling the group tax team to ask what it meant.

What the audit committee should ask before year-end

The policy is a governance document, so the committee's questions are about evidence: what was done, who approved it, and what the file shows.

HMRC's compliance facility lists one recurring cause of arrangements failing scrutiny: the facts the analysis rests on diverge from what happens in the business. That is set out in its guidance, which also notes that businesses change over time so that the functional profile "may become different from that originally assumed or intended".

Inland Revenue New Zealand puts the responsibility where the committee can see it: "It remains the responsibility of local management to ensure a company's transfer prices are in accordance with the arm's length standard." Its documentation guidance adds that where material local facts are omitted or differ from a group position, "the local documentation will be inadequate", and gives a policy to charge royalties at a uniform global rate as its example.

So the questions can be short. Has anything changed in what this entity does? Is the functional profile in this year's file the one the business runs, or the one the policy assumed?

Who signed, on what evidence, and could they produce it now? If those answers are thin, they will be thinner under a notice. Contact us before the numbers are locked, and we will test the document against the year it governs.

Cubic Pattern

Get started today

Let’s talk

If you are evaluating cross-border expansion, restructuring, or strengthening compliance and audit readiness, we can help you plan and execute with clarity.

Cubic Pattern

Get started today

Let’s talk

If you are evaluating cross-border expansion, restructuring, or strengthening compliance and audit readiness, we can help you plan and execute with clarity.

Cubic Pattern

Get started today

Let’s talk

If you are evaluating cross-border expansion, restructuring, or strengthening compliance and audit readiness, we can help you plan and execute with clarity.