Transfer Pricing Documentation Isn't a One-Off

What goes in it, when it's required, and why the version you commissioned stops being true

For CFOs

Copy link

Most transfer pricing documentation gets read twice. Once by the person who wrote it, and once, a couple of years later, by someone checking whether it still matches the company.

This documentation is the written evidence that the pricing method(s) your group of companies charge each other are akin to how independent parties would interact. In most countries it follows the OECD's format:

local file - the detailed picture of one entity's intercompany transactions and why the pricing holds up
master file - the group-level view of your structure, where the IP sits, and how the group prices things
country-by-country report (CbCR) - for groups with revenue over €750 million

You almost never file any of it with a return. You keep it, and hand it over when a tax authority, an auditor or a diligence team asks. That's what makes it an odd deliverable; its whole job is to avoid the sentence: “Yes, technically that was the structure at the time.”

So what actually goes in it?

However thick the document gets, the local file is really answering four questions:

What does this entity do? The functional analysis: who does what, who carries the risk, who owns the assets.
What does it charge its sister companies, and for what? The charges it makes, and the agreements sitting behind them.
How was each price set? The method, in plain terms: cost plus a margin, a percentage royalty, a distributor's margin.
Why is that price fair? The benchmarking: evidence that independent companies doing similar work earn similar margins.

The master file zooms out to the group: the structure, where the valuable stuff sits, the policy that ties it together.

None of this is hard if the design underneath is sound. If the decisions were made reactively, one entity at a time, the document describes a company that only exists on paper.

When do you actually need it?

It depends on the country.

In the UK, the prescribed master file and local file format only kicks in past that same €750 million mark, for accounting periods from April 2023, and has to reach HMRC within 30 days of a request. But that's a rule about the format, not about whether the evidence needs to exist. Every transaction has two ends, and the country on the other end sets its own bar.

In the US, nobody forces you to prepare documentation at all. It's simply what protects you if your pricing is ever questioned. It has to exist by the time the return is filed, and be on the IRS's desk within 30 days of them asking.

Most countries sit somewhere between those two, with their own thresholds and their own clocks. Which local file requirements apply to you is a per-country question.

In practice, though, the deadline that arrives first for a scaling company is rarely a tax authority's. It's a diligence request list with "transfer pricing documentation" on it.

The passport photo problem

Documentation captures your group at a single moment: entities, team, flows, margins.

Then you open a new entity, move engineering to Romania, and add a licence fee that didn't exist last year. Now the document is about as useful as an old passport photo.

Worse than going stale, it can start contradicting your books. A report saying the UK entity charges a 10% mark-up for services, sitting next to the books showing nothing was ever charged, raises more questions.

So the useful question is always "does the document actually describe the company?" Worth asking every year, and whenever something material changes.

Consultant, spreadsheet, or software?

Companies produce this three ways.

A one-off consultant report is a decent snapshot and an expensive one to refresh. The numbers start ageing the day it lands.
Spreadsheets, a template, and a dream are fine before you’ve hit local file thresholds; but you shouldn’t be wasting time on this if you don’t have the expertise.
Documentation software automates the writing, which helps, but the output is only as good as the pricing design and agreements underneath it; it can make managing the documents easier. It cannot polish the turd underneath.

CONCLUSION

The Caribou version

As you can probably tell by now, the order of operations is critical. Caribou sets the overall design first, so by the time the local file needs to be written, the intercompany agreements, pricing and accounting entries already line up. The local file comes out of that live setup rather than being reconstructed from memory. And it moves when the group does. So when someone does ask, the answer is already audit-ready. If your documentation hasn't been opened since the year it was written, book a demo.

Know someone who needs to read this?

Copy link