Transfer pricing is not a filing you complete once a year. It is a record you are assumed to have been keeping all along — and the assessment that tests it may arrive years after the transactions it questions.
The obligation starts earlier than most groups expect
The moment an Indian entity transacts with a related company abroad, it has entered the transfer pricing regime. Not when the transaction becomes large. Not at the end of the first profitable year. At the first invoice.
This catches groups out because the first intercompany transactions rarely look like transactions at all. The parent pays for a software licence the Indian team uses. Head office allocates a share of global marketing cost. A regional director spends part of the year supporting the India business and the cost sits centrally. None of these feel like trade, and none of them arrive with a purchase order — but each is an international transaction between associated enterprises, and each needs a defensible basis for the price attached to it.
The regime asks one question of every such transaction: would two unrelated parties, negotiating independently, have agreed this price? That is the arm’s length principle, and everything else — the methods, the benchmarking, the documentation — exists to answer it with evidence rather than assertion.
The transactions that need support
| Transaction type | Typical form in a subsidiary | What tends to go wrong |
|---|---|---|
| Services rendered to the parent | Engineering, support, back-office or R&D work performed in India | Cost-plus markup applied without a benchmarking study to support the margin chosen |
| Management and head-office charges | A share of global leadership, HR or IT cost allocated to India | No evidence the Indian entity actually received a benefit, and no basis for the allocation key |
| Royalties and IP licensing | Use of group brand, software or technology | Rate set by group policy with no independent evidence it is arm’s length in India |
| Intra-group financing | Parent loans, cash pooling, extended intercompany credit terms | Interest rate not benchmarked; long-unpaid receivables treated as if they carried no financing element |
| Cost recharges | Shared licences, travel, seconded staff | Treated as pass-through bookkeeping rather than as reportable transactions |
The pattern across all five is the same. The commercial logic is usually sound — these are real costs for real value. What is missing is the contemporaneous record showing how the number was arrived at, and why an independent party would have accepted it.
Contemporaneous is the word that carries the weight
Documentation prepared at the time is evidence. Documentation assembled after an assessment notice arrives is an argument.
The distinction matters more than the volume of paper. An assessing officer reviewing a benchmarking study can generally tell whether the analysis informed the pricing or was reverse-engineered to justify it — the comparables chosen, the period they cover, the tidiness with which the conclusion lands on the number already booked. A study that genuinely preceded the pricing decision reads differently from one that followed it.
This is the single most consequential habit to establish early, and it costs very little while the business is small. A short annual file — the intercompany agreements, the method chosen and why, the benchmarking analysis, the actual numbers transacted — created each year while the facts are current is worth substantially more than a comprehensive study produced three years later under time pressure.
What a defensible file contains
- The intercompany agreements themselves. Written, signed, and describing what is actually happening rather than what was intended when the template was drafted. An agreement that does not match the substance of the arrangement is worse than no agreement.
- A functional analysis. Which entity performs which functions, holds which assets and bears which risks. This is the foundation of everything that follows — the pricing method and the margin both depend on it, and it is the part most often skipped.
- The method chosen, and why the alternatives were rejected. A method selected without a stated reason invites the assessing officer to propose a different one.
- The benchmarking study. Comparable companies or transactions, the screening criteria applied, and the arm’s length range they produce. Refreshed periodically — a study is not a permanent artefact.
- The actual transacted numbers, reconciled to the financial statements. Documentation that cannot be tied back to the audited accounts is a gap that will be found.
- Form 3CEB, certified by a practising Chartered Accountant. Filed annually, reporting the transactions and the accountant’s view on arm’s length consistency.
Where groups with a global policy still get caught
A well-run multinational usually arrives with a group transfer pricing policy already written, often by advisers in the parent jurisdiction. This is genuinely useful — it establishes consistency, and consistency is itself a defence.
It is also not enough on its own. The Indian entity is assessed on its own record, by officers applying Indian requirements to Indian documentation. A group policy that specifies a margin for shared-service entities does not, by itself, demonstrate that the margin is arm’s length in India, benchmarked against comparables that an Indian assessing officer would accept. The policy needs to be localised: same principle, evidenced locally.
The related trap is the passage of time. Assessments do not arrive promptly. A transfer pricing question may be raised on a year the business has long since closed, reported to its board, and possibly staffed with entirely different people. The file assembled at the time is what answers it — because the people who could have explained the reasoning from memory have generally moved on.
The practical position
For a subsidiary in its first years, the sensible standard is not an elaborate compliance programme. It is a short, honest, annually refreshed file that a stranger could read and follow: here is what we transact with the group, here is how we priced it, here is the evidence that price is reasonable, and here is the signed agreement that governs it.
That file takes a few days a year to maintain. Reconstructing it under assessment takes considerably longer, costs more, and produces weaker evidence — which is the whole argument for doing it while the facts are still current.