At a glance. In Gap France, the Paris Administrative Court of Appeal refused to substitute a comparable set built for the resale price method for the set required to apply the transactional net margin method. The court also found that comparability was not established because the proposed companies did not reflect Gap France’s start-up conditions. A benchmark remains tied to its method, profit indicator and economic circumstances.
The decision of 6 May 2026 (CAA Paris, no. 24PA04535) addresses a recurring audit practice. A tax authority may reject the taxpayer’s selected method and propose another. It must then support the substituted method with comparables that are suitable for that method and the tested transaction. Data produced by the taxpayer do not become universally usable merely because they appear in the transfer pricing documentation.
The dispute and the competing methods
Gap France distributed clothing under the Gap and Banana Republic brands. During the audited years, the company reported operating losses while paying intra-group service charges and royalties and incurring expenditure connected with the launch of Banana Republic in France.
The taxpayer’s documentation applied the resale price method. That method starts from the price at which goods purchased from a related party are resold to independent customers and deducts an appropriate gross margin. Its economic test therefore operates at gross-margin level.
The tax authority considered that this approach did not capture expenses borne by Gap France in connection with the development of group brands. It substituted the transactional net margin method, using operating margin as the profit indicator and a wider set of comparable companies. The adjustment sought to bring Gap France to a positive net margin.
The attempted substitution on appeal
After the first-instance court rejected the transfer pricing adjustment, the Minister asked the Court of Appeal to retain seven companies originally selected by Gap France. The argument was that the taxpayer had itself proposed those companies during the audit and could not later challenge their relevance.
The Court rejected that reasoning. The seven-company set had been assembled to apply the resale price method and determine an arm’s-length range at gross-margin level. It could not simply replace the set needed for a net-margin analysis. The change of method altered the economic question, the profit indicator and the information needed to establish comparability.
A comparable set belongs to its analytical framework
A comparable search is not an independent asset that can be detached from the analysis for which it was designed. The search strategy reflects the transaction, tested party, method, profit-level indicator, functional profile and available accounting data. A company suitable for a gross-margin comparison may not be suitable for an operating-margin comparison.
This does not prevent the tax authority from changing the method. It means that a change requires a new demonstration. The authority must explain why the substituted method is appropriate, identify reliable data for that method and examine whether material differences can be adjusted with sufficient accuracy.
Start-up conditions were also decisive
The Court identified a second defect. None of the proposed companies operated in a start-up phase, while the Banana Republic brand had been introduced to the French market in December 2011. A new market launch may involve exceptional marketing expenditure, initial operating inefficiencies and a gradual build-up of sales.
Under a net-margin method, differences in competitive position and commercial experience can materially affect profitability even where the companies perform broadly similar distribution functions. The Court therefore required the comparison to reflect whether the tested activity was mature or still being launched.
The judgment did not establish a fixed start-up period or grant an automatic tolerance for losses. It required the tax authority to account for the actual economic conditions before treating low profitability as evidence of an advantage.
The burden of proof under Article 57
The relationships between Gap France and the relevant foreign group companies were not disputed. The remaining question was whether the tax authority had established an advantage granted by the French company.
Where no advantage is established by its nature, the administration may rely on comparison with independent enterprises. That comparison must be reliable. The Court held that the proposed set did not meet this condition because it had been constructed for another method and did not include companies in comparable start-up circumstances. The Minister’s appeal on the transfer pricing reassessments was therefore rejected.
Documentation is evidence but remains contextual
The decision is important for the use of taxpayer documentation during an audit. Documentation may reveal inconsistencies, provide admissions about functions and risks and supply data relevant to an alternative analysis. It does not make every component interchangeable.
This point remains relevant under the strengthened French rules applicable since the 2024 Finance Act. A difference between the method applied and the method described in the documentation can have significant consequences. Even so, the tax authority must respect the economic framework in which a comparable set was constructed and establish the advantage required by Article 57.
A separate point on penalties
The outcome was not a complete success for Gap France on every issue. The Court rejected the Minister’s appeal concerning the transfer pricing adjustments, but it also rejected Gap France’s cross-appeal against the 40% penalty applied to separate service-charge adjustments. The judges relied on the amounts involved, the accumulation of losses and the company’s continued reduction of its French tax base despite its deteriorating financial position.
The distinction matters. The methodological failure of one transfer pricing adjustment does not automatically neutralise penalties attached to different transactions whose factual and evidential record is less favourable.
Practical implications for benchmark design
- Define the method before launching the search. The search criteria and financial data must answer the question posed by the selected method.
- Retain the original analytical record. Preserve why each company was accepted or rejected and which profit indicator was tested.
- Identify economic phases. Record launches, restructurings, exceptional investments and other conditions that may affect profitability.
- Test the accounting data. Gross-margin and operating-margin analyses require different levels of reliable segmentation and consistent classification.
- Challenge substitutions explicitly. If the authority changes the method, require a new comparability analysis rather than a mechanical reuse of the taxpayer’s set.
- Explain losses with evidence. Budgets, launch plans, actual-versus-forecast data and the duration of exceptional costs should support the narrative.
- Review the whole file for consistency. The functional analysis, method, comparable search and actual pricing must describe the same transaction.
Expert perspective
The key contribution of Gap France is methodological discipline. A comparable set is not validated merely because the taxpayer produced it. Its reliability depends on the question it was designed to answer. When the administration moves from gross margin to net margin, it must rebuild the analysis around that new indicator and the tested company’s real economic circumstances.
References
- CAA Paris, 2nd Chamber, 6 May 2026, no. 24PA04535, SAS Gap France.
- French Tax Code, Article 57.
- OECD Transfer Pricing Guidelines 2022, Chapters II and III.
Current at 5 October 2026. General information only; it is not legal or tax advice. The implications of a judgment depend on its facts, procedural status and subsequent case law.