Rule 10AA and 10AB Safe Harbours in India: Conditions and Strategy

India’s safe harbour regime (section 92CB) entered in 2015 with the service harbours at its centre: Rule 10AA (the specified intra-group services) and Rule 10AB (the KPO and IT-enabled services tiering). The regime has since been extended and its circumstances table consolidated in Rule 10TD — and the 10AA/10AB labels remain the reference point for the service harbours’ design: the prescribed cost-plus for the routine intra-group services, and the employee-cost-tiered OP/OC for the knowledge- intensive services. This guide is the jurisdiction mechanics of those two provisions; the current full circumstances table (all eligible transactions, post the 2025 amendment) and the election mechanics are in the safe harbour guide.

The design: what the service harbours do

The service harbours answer the question the Indian service economy keeps asking: what is the arm’s length return for a routine service, without the comparables fight? The design, in the two provisions:

Provision	The fact pattern	The prescribed position
Rule 10AA — the specified intra-group services	The routine intra-group services — the administration, the support, the back-office — where the service is genuinely routine and low-value-adding	A prescribed cost-plus markup (3%) on the eligible services: cost plus the fixed margin, no benchmark, no matrix
Rule 10AB — the KPO / IT-enabled services tiering	The knowledge-intensive services — KPO, ITES — where the return tracks the knowledge content of the function	A prescribed OP/OC margin, tiered by the employee-cost ratio (the share of employee cost in the operating cost): the more knowledge-intensive the cost structure, the higher the tier — the 18–24% band in the current circumstances

The two provisions are the same idea at two resolutions: 10AA fixes the return for the routine service (a cost-plus, because the routine service earns a routine mark-up on cost); 10AB prices the knowledge-intensive service on the profitability indicator (OP/OC, because the knowledge function’s return is a margin on the cost base it earns on), with the tier tracking the knowledge content. Both are the standardized return concept — the OECD’s Amount B logic, in the Indian prescriptive form, predating and informing the OECD’s 2022 codification.

The conditions: what makes a service eligible

The harbour runs on the eligible service, not the invoice label. The conditions, in the order the examination applies them:

1. The service is within the eligible category. The specified intra- group services (10AA) or the KPO/ITES categories (10AB, as reflected in the current 10TD table — software development, ITeS, KPO, contract R&D). The category is the service’s function, determined by the function analysis — not by what the agreement calls it. A “service” that is really a licensing of technology, or a distribution, is outside the harbour whatever the label.
2. The service is routine (10AA) or matches the tier’s profile (10AB). The 10AA services are the low-value-adding, no-intangibles, no-significant- risk services — the LVAS profile. The 10AB tiering conditions run on the cost structure: the employee-cost ratio that places the entity in the tier, computed on the year’s actuals.
3. The transaction is within the value limits. The harbour circumstances carry the transaction-value limits (the current table: the services categories with their ₹-limits — the threshold raised to ₹300 crore for the software/ITeS/KPO/contract-R&D categories by the 2025 amendment). Above the limit, the category falls out of the harbour for that transaction volume, and the benchmarking applies.
4. The election is made. The option under the regime’s election rule (Form 3CEFA, exercised per assessment year) — the harbour protects the elected position. The service priced at the harbour margin without the election is a benchmarked position that happens to equal the harbour, not a safe-harboured one.

The strategy: when the harbour beats the benchmark

The decision is the safe harbour guide’s decision, run on the service facts:

The position	The harbour case	The benchmark case
The entity’s actual return is above the harbour margin	The election leaves return on the table — the KPO at 30% OP/OC electing at the 24% top tier, the service centre earning above the 3% cost-plus. The harbour is a floor-and-settlement instrument, not an optimisation tool — the benchmark stands, and the higher return is defensible on its own matrix	The benchmark carries the position; the harbour is the documented fallback (“the position is within the harbour in any event”)
The entity’s return is inside the harbour band	The election buys the certainty: the declared price is accepted, the matrix is not examined, the TPO’s pool does not exist for this transaction. For the routine service with a noisy pool, the certainty is worth the band	The benchmark is run anyway (the decision is made knowing where the range sits), and the election is made where the range’s defence is weaker than the harbour’s certainty
The entity’s return is below the harbour margin	The election is the fix: the price moves to the harbour, the under-charge is cured by the election, and the position is the prescribed one. (The direction to watch: the harbour is a floor for the eligible transaction — a position below it is a position the examination corrects anyway, and the election makes the correction the prescribed one)	The benchmark’s range, if below the harbour, is the position — but the harbour floor usually governs the eligible transaction

The strategic through-line: the harbour is the certainty instrument for the eligible, routine fact pattern, and the benchmark is the position instrument for everything else — and the decision between them is made on both numbers, in the same file, before the election is filed. Running the benchmark anyway is the discipline: the election made in the dark (without knowing where the range sits) is the election the later examination re-examines.

The interplay with the rest of the regime

• The benchmarked position and the harbour, side by side. The Local File for an eligible service carries both: the benchmark (the pool, the range, the position) and the harbour margin — the file that shows both, with the election decision, is the file the safe harbour guide describes.
• The value-limit crossings. The year the transaction volume crosses the category’s limit: the harbour falls away for the excess, and the benchmark is the position — the year-on-year monitoring (the TP policy trigger) is what catches the crossing before it is a problem.
• The TDS follows the character. The service fee’s withholding character (the royalty vs the fee-for-service question) is separate from the pricing position — see TDS on TP payments. The harbour prices the transaction; the TDS characterizes the payment.

See also

• The safe harbour guide — current circumstances
• Shared services — the LVAS fact pattern
• KPO & BPO in India
• Amount B — the OECD standardized return
• The compliance calendar