How many comparables does a benchmarking study need?
There is no count that makes a study safe, and chasing one distracts from what is actually examined. A search that starts from a wide database population and narrows through stated screens can end with a handful of genuinely similar companies and be defensible, while a long list assembled by loose screening is not. What carries the weight is the search strategy written down as it was run: the population, each quantitative screen, each qualitative screen, and the reason every candidate was set aside. If the study ends with very few survivors, say so and explain the screens that did the work, rather than relaxing them quietly until the list looks comfortable.
What is a rejection log and why does an auditor ask for it?
The rejection log is the record of every candidate the search threw up and did not keep, with the reason against each one. It is asked for first because it is the only part of a study that shows judgement rather than output. A range can be recomputed by anyone from the surviving set, but whether that set was chosen honestly is visible only in what was discarded. Common reasons are a different functional profile, a controlled ownership position, insufficient financial history, or a business description that does not match the tested activity. A log that says nothing more than unsuitable against each name invites the examiner to rebuild the search from scratch.
Can I reuse last year's benchmarking study?
A search can often stand for more than one year, provided the facts it rests on have not moved and the financial data behind the surviving set is brought up to date. What cannot be reused is a study whose tested activity has changed: a new product line, a shifted risk profile, or a reorganised function means the screens were run against a business that no longer exists. Treat the two halves separately. Refresh the financial data annually, revisit the search itself when the tested activity changes or the database population has plainly moved, and write down which of the two you did and why.
Can I use comparables from another country?
Sometimes, and the answer turns on whether the wider search is a considered choice or a convenience. Where the local market has too few independent companies performing the tested activity, widening to a region can be the only way to build a set at all, but the study then has to address what that widening changes: market size, cost structures, regulation and the accounting basis the figures are drawn on. State the geographic screen, say why it was set where it was, and note any adjustment made for the difference. A regional set adopted silently, with no reasoning recorded, is the version that comes back as a query.
What makes a comparable fail a qualitative screen?
Quantitative screens are mechanical and easy to describe: independence, data availability, persistent losses, activity codes. Qualitative screening is where the study is actually made, and it means reading each surviving company's own description of itself and asking whether it does what the tested activity does. Candidates fall away because they own the intangibles rather than use them, because they carry inventory and market risk the tested activity does not, because a business description reveals a second unrelated segment, or because the accounts consolidate activities that cannot be separated. Each of those readings is a judgement, which is why each one has to be written down at the time it is made.
Does the study have to be finished before I file?
Treat it as work for the year rather than work for the query. Documentation put together after an authority has asked a question no longer satisfies a contemporaneous requirement, and that is not a technicality: a study written afterwards was necessarily built knowing the reported result, which is precisely what a search is supposed to be independent of. Practically, that means the tested activity is defined, the screens are run and the range is computed while the year is being closed, so the reported result can be compared against the range before the return goes in rather than explained once it has.
What is an intercompany agreement, and do we need one?
It is the contract between the related parties — who does what, who bears which risk, what is charged and on what basis. It matters because when there is no agreement, an auditor prices the transaction from the conduct they can observe rather than from the arrangement you intended, and conduct rarely tells the whole story. Signed agreements that match the invoices and the actual functions are the cheapest transfer pricing protection there is. See our transfer pricing work.
What is a totalization agreement and how do I use one?
A social security agreement that stops you contributing to two systems for the same work, and lets periods in both count towards benefit eligibility in either. Which system you stay in depends on the agreement's rules for your situation — a seconded employee usually remains in the home system for a set period, a locally hired one usually joins the host system. You evidence it with a certificate of coverage obtained before or shortly after the assignment starts. See certificates of coverage.
