Internal TNMM for freight-forwarding transactions prevails where AE and non-AE segments are materially comparable and adjustments target AE dealings only.
Internal TNMM may benchmark international freight-forwarding transactions where the enterprise's non-AE segment is separately identifiable and materially comparable in functions, assets, risks, contractual terms, economic conditions and shipment composition. Entity-level losses alone do not make segmental results unreliable; where enterprise-specific variables are neutralised, internal comparables are more reliable than external TNMM. Arm's-length income under sections 92 and 92C must be determined only for transactions with associated enterprises. Accordingly, external-TNMM adjustments cannot be computed on entity-wide results that include uncontrolled non-AE transactions.
Issues: (i) Whether internal Transactional Net Margin Method (TNMM) based on the assessee's non-AE freight-forwarding segment was a reliable method for benchmarking the AE transactions, rather than external TNMM; (ii) Whether a transfer-pricing adjustment under external TNMM could be computed on entity-wide results including non-AE transactions.
Issue (i): Whether internal Transactional Net Margin Method (TNMM) based on the assessee's non-AE freight-forwarding segment was a reliable method for benchmarking the AE transactions, rather than external TNMM.
Analysis: Rule 10B permits comparison with the enterprise's own comparable uncontrolled transaction as well as an unrelated enterprise's transaction, subject to functional, asset, risk, contractual and economic comparability. Entity-level losses did not by themselves establish unreliability of segmental results. The AE and non-AE freight-forwarding segments both incurred losses, but the AE segment's margin was better; the non-AE segment also constituted a substantial and separately identifiable part of the business. No material difference in functions, assets, risks, air-sea shipment composition, or profitability was demonstrated. An internal comparable was consequently more reliable because enterprise-specific variables were neutralised.
Conclusion: Internal TNMM based on the non-AE segment was accepted and external TNMM could not replace it. The finding is in favour of the assessee.
Issue (ii): Whether a transfer-pricing adjustment under external TNMM could be computed on entity-wide results including non-AE transactions.
Analysis: Sections 92 and 92C confine determination of arm's length income to international transactions with associated enterprises. Uncontrolled transactions with independent parties cannot be subjected to an adjustment merely because entity-wide margins are used for benchmarking.
Conclusion: Any adjustment under external TNMM had to be confined to AE international transactions and could not be computed on entity-wide results. The finding is in favour of the assessee.
Final Conclusion: The international freight-forwarding transactions satisfied the arm's-length principle under internal TNMM, rendering the transfer-pricing adjustment of Rs. 3,37,76,000 unsustainable.
Ratio Decidendi: Transfer-pricing benchmarking must use reliable comparable uncontrolled transactions, with internal TNMM preferred where the same enterprise's controlled and uncontrolled segments are materially comparable, and any adjustment confined to AE international transactions.