Test anfragen
That distinction decides most arguments about benchmarking scope — how wide the screen, how many sources, how often the study is refreshed. Those arguments are usually conducted as if the Guidelines demanded maximum effort. They say the opposite, in four consecutive paragraphs that are rarely quoted.
Para. 3.80 — cost is a legitimate concern. Chapter III of the OECD Transfer Pricing Guidelines opens the topic by asking "the extent of the burden and costs that should be borne by a taxpayer to identify possible comparables", and answers that the cost of information can be a real concern, "especially for small to medium sized operations, but also for those MNEs that deal with a very large number of controlled transactions in many countries". Cost is not something the taxpayer smuggles in. The Guidelines put it on the table.
Para. 3.81 — no exhaustive search. "When undertaking a comparability analysis, there is no requirement for an exhaustive search of all possible relevant sources of information. Taxpayers and tax administrations should exercise judgment to determine whether particular comparables are reliable." The test is the reliability of what you used, not coverage of everything that exists.
Para. 3.82 — prudent business management. Transfer prices should be established, monitored and reviewed "taking into account the size of the transactions, their complexity, level of risk involved, and whether they are performed in a stable or changing environment". It is therefore reasonable "to devote relatively less effort to finding information on comparables supporting less significant or less material controlled transactions" — and for simple transactions in a stable environment whose characteristics stay the same, "a detailed comparability (including functional) analysis may not be needed every year".
Para. 3.83 — SMEs. The arm's length principle applies to everyone, but "pragmatic solutions may be appropriate in order to make it possible to find a reasonable response to each transfer pricing case".
Chapter V repeats the point for documentation: taxpayers "should not be expected to incur disproportionately high costs and burdens in producing documentation" (para. 5.28); not every related-party transaction is material enough for full local-file treatment (para. 5.32); SMEs should not be asked for the documentation volume expected of large groups (para. 5.33).
Proportionality is not one rule but a scaling factor. The table maps the transaction profile to the effort the Guidelines support.
Two consequences follow for practice.
The first is the refresh rhythm: document the search strategy once, update the financial data of the accepted comparables annually, re-run the full search every three years unless something changes. Some jurisdictions have written that cycle into law — under art. 11r of the Polish CIT Act a benchmarking analysis must be updated at least every three years, unless a change in the economic environment materially affecting it forces an earlier update. Para. 3.82 carries the same rhythm elsewhere. Nothing requires a new screen every twelve months for a transaction that has not moved.
The second is that "unchanged" has to be true. A stable environment is a factual claim in your documentation. If the functional profile shifts, a business is restructured, margins swing or the group enters a new market, the relief from the annual re-run lapses with it.
Para. 3.81 limits the scope of the search. It says nothing about the quality of the source, and taxpayers who read it as a licence for weak data run into trouble.
What auditors reject is rarely a search that was too narrow. It is a search that cannot be reproduced: undated extracts, no record of the screening steps, rejections with no stated reason, ratios that cannot be traced back to a company's own financial statements. A defensible benchmarking study documents its search strategy, its accept and reject criteria, and its data lineage. That is the bar the effort has to clear: a recognised data source at a reasonable cost, never a weaker source at any cost.
Local law is where that bar acquires teeth, and it varies. Some jurisdictions specify what the file must contain and what happens when it falls short — unusable records treated as never created, estimation powers, fixed penalties. Others leave more to audit practice. Read the Guidelines for the principle and the domestic rules for the consequence; where the two are read together, proportionality holds up. Our companion piece sets out how German law resolves the same question, provision by provision: Angemessenheit in der Verrechnungspreisdokumentation.
Several reliefs are in force and underused.
Low value-adding intra-group services. Chapter VII offers an elective simplified approach: the cost pool plus a 5 % mark-up, with no benchmarking study for the mark-up. It covers supportive services outside the core business that involve no unique intangibles and no significant risk, and it is elective per jurisdiction, so confirm both sides accept it. The OECD's consultation document revising Chapter VII of 1 June 2026 carries the simplified approach forward largely unchanged.
Materiality in the local file. Para. 5.32 invites jurisdictions to set materiality thresholds, in relative or absolute terms, so immaterial transactions do not attract full documentation. Most have. Check the threshold before commissioning a study, not after.
De minimis and SME regimes. Many jurisdictions exempt smaller taxpayers from formal documentation below a turnover or transaction-volume threshold, replacing it with an obligation to answer questions and produce existing records. The exemption is from the file, not from the arm's length principle.
Aggregation. Where transactions are economically comparable in functions and risks, they can usually be documented as a group rather than one by one — provided the grouping follows rules set in advance and stated in the file.
Narrower filing obligations. The trend in recent reform is to reduce what must be filed unprompted and to keep the rest available on request. The mandatory perimeter is narrowing, not widening.
Taken seriously, proportionality inverts the selection question. A taxpayer does not have to justify why a data source costs less than the most expensive option on the market. The burden runs the other way: any budget committed has to be proportionate to the size, complexity and risk of the transaction it documents.
Two requirements hold at once.
Where a provider sits on price is a commercial question. Whether its data is recognised and traceable is a documentation question, and only the second is decided by the Guidelines. smartZebra's transfer pricing database was built for that second test: 500,000+ company comparables drawn from published financial statements, built on 50,000+ company profiles, each result carrying a calculation log back to its source. The entity and methodology facts are published in full.
Proportionality is a defence you document before you need it. To see what a reproducible benchmarking search looks like end to end, the transfer pricing module shows the TNMM workflow, the screening steps and the calculation log on live data.
No. OECD TPG para. 3.81 states there is no requirement for an exhaustive search of all possible relevant sources of information; taxpayers and tax administrations exercise judgment on whether the comparables used are reliable. The obligation is to document the search strategy applied and why the accepted comparables are reliable, not to prove that no other source exists.
Not necessarily. Under para. 3.82, a detailed comparability analysis may not be needed every year for simple transactions carried out in a stable environment whose characteristics remain the same. Established practice is to refresh the financial data of the accepted comparables annually and re-run the full search every three years — or immediately when the functional profile, the terms or the market change. Some jurisdictions, such as Poland, have codified that three-year cycle.
The arm's length principle applies equally; the effort does not have to. Paras. 3.80 and 3.83 recognise the cost of information as a real concern for smaller operations and endorse pragmatic solutions, and para. 5.33 advises against expecting SMEs to produce the documentation volume of large enterprises. Most jurisdictions add de minimis thresholds below which no formal file is required.
No tax authority prescribes a specific database. What is expected is a recognised source whose results can be reproduced: identifiable companies, disclosed accounting basis and period, a documented search and rejection strategy, and figures traceable to published financial statements. A study a reviewer cannot re-walk is the problem, not the price of the licence.
The same scaling applies, but financing transactions are rarely routine: terms move with the market, so rate and credit data have to be current at the transaction date. Chapter X analyses sit at the higher end of the effort scale — see our note on arm's-length interest rates for intercompany loans.
%20Loans%20in%20Private%20Credit.png)
