Case Overview
In UFI Filters S.p.A. v. Agenzia delle Entrate (Cass. n. 10499/2024), the Italian Supreme Court (Sez. V) dismissed the Revenue Agency’s appeal in an IRAP transfer pricing case for FY2009. The Court agreed with the Lombardy Regional Tax Commission (CTR) that the Agency’s cost-plus benchmarking was not adequately supported because the selected external “comparables” were not shown to be sufficiently comparable—especially from a functional (FAR) perspective.
Key Facts
- Tax year / tax: IRAP, FY2009.
- Controlled transactions: UFI Filters S.p.A. purchased automotive filters from two Chinese related parties (Ufi Filters (Shanghai) Co. Ltd and Sofima Automotive Filter (Shanghai) Co. Ltd).
- Adjustment amount: the Agency disallowed EUR 11,422,242.74 of costs by re-characterising the purchase prices to “normal value” under Art. 110(7) TUIR.
- Method used by the tax authority: cost-plus, based on a set of six Chinese companies selected from an initial pool of 44; the Agency highlighted mark-ups of 39.77% and 17.20% for the Chinese suppliers.
- Procedural history: CTP Milan rejected the taxpayer’s appeal; CTR Lombardy annulled the assessment; the Supreme Court dismissed the Agency’s appeal.
The Dispute
For FY2009 (pre-2017 wording), Art. 110(7) TUIR required related-party cross-border dealings to be valued at “valore normale” (normal value) where an increase in Italian taxable income results. The Revenue Agency argued the Chinese related suppliers charged UFI Italy prices above normal value, inflating deductible costs and shifting profit outside Italy.
To quantify the alleged deviation, the Agency applied the cost-plus method and built an external benchmark using six Chinese companies. The Agency’s comparable selection was described (by the CTR and then summarised by the Supreme Court) as based mainly on:
- the same activity code (even if not the same products),
- absence of intangible fixed assets,
- geographical area (China), and
- broadly comparable turnover.
UFI challenged the adjustment primarily on comparability grounds: the six companies produced materially different items (examples noted by the courts included cabins, tanks, gearboxes rather than filters), were in materially different locations/conditions within China (remote areas versus the Shanghai port area), and—critically—were not shown to have comparable functions, risks, and assets. The taxpayer also disputed the Agency’s attempt to reframe the case as turning mainly on burden of proof given information asymmetry.
Holding (and Why the Agency Lost)
The Supreme Court dismissed the Revenue Agency’s appeal, confirming that the CTR had correctly found the benchmark unreliable because the Agency did not substantiate comparability in a way consistent with transfer pricing principles.
1) OECD Guidelines as an interpretative reference (even for pre-2017 years)
The Court reiterated that Italian courts interpret Art. 110(7) TUIR consistently with the arm’s length principle reflected in OECD Transfer Pricing Guidelines, treated as relevant interpretative “soft law” (the judgment discusses the evolution of the domestic rule and cites OECD guidance on method selection and comparability).
2) Cost-plus still requires demonstrated comparability of the mark-up (FAR and conditions)
Drawing on OECD guidance for cost-plus (including the idea that differences must not materially affect the mark-up, or must be reliably adjusted), the Court agreed that the CTR identified “unexceptionable” inconsistencies in the Agency’s comparables and—above all—an evidentiary gap on functional comparability.
The CTR finding endorsed by the Supreme Court was blunt:
the Agency provided no demonstration and/or evidence of comparable “functions, risks or investments”… and “the only data made available are those of the balance sheet.” (Cass. n. 10499/2024, summarising the CTR’s reasoning)
3) Burden of proof: taxpayer bears the “proximity” burden, but the authority still needs a reasoned, evidenced adjustment
The Supreme Court restated its settled approach: Art. 110(7) TUIR is not an anti-avoidance clause requiring proof of an elusive purpose; however, the taxpayer—given “proximity of proof” under Art. 2697 Civil Code—bears the burden of proving arm’s length conditions. At the same time, the Agency must still set out and support a coherent adjustment, including a defensible comparables analysis. Here, the CTR could lawfully conclude that the Agency’s benchmark was not adequately demonstrated.
4) Cassation review limits
The Court also indicated the appeal was, in substance, an attempt to obtain a reassessment of the factual record (comparability), which is not the function of the Supreme Court in this posture.
Takeaways for Practitioners
- Comparable-company selection must be evidenced, not asserted: an activity code plus financial statements is not a functional analysis.
- Cost-plus is not “comparability-light”: even where product identity is not required, the authority (or taxpayer) must show FAR comparability and address material geographic/market differences or make adjustments.
- Build an audit-ready narrative: contracts, supply chain facts, and a clear FAR story can be decisive when challenging (or defending) a benchmark.
Related Content
For practical guidance on building defensible benchmarking analyses, see our Benchmarking Study Guide. For terminology used in screening and selecting benchmarks, see Comparable Companies. Practitioner commentary links this IRAP decision with the “twin” Cassation decision n. 10577/2024 (IRES) arising from the same audit.
Frequently Asked Questions
What transfer pricing method was applied in the audit?
The Revenue Agency applied the cost-plus method, benchmarking the Chinese suppliers’ mark-ups (noted as 39.77% and 17.20%) against mark-ups observed in a set of selected Chinese companies.
Why did the Court consider the comparable selection unreliable?
Because the selected companies showed clear product and location mismatches and, most importantly, the Agency did not demonstrate functional comparability (FAR) with any meaningful operational description—relying essentially on balance-sheet data.
Does cost-plus require identical products?
Not necessarily. However, the Supreme Court confirmed that material differences (including product and market/geographic conditions) still matter, and the cost-plus mark-up must be supported by a proper comparability analysis (or reliable adjustments).
Who bears the burden of proof in Italian transfer pricing litigation?
The Court reaffirmed that the taxpayer bears the ultimate burden (principle of “proximity of proof” under Art. 2697 Civil Code) to show arm’s length conditions, while the tax authority must still articulate and support a reasoned adjustment grounded in defensible comparables.
What does this case change for benchmarking practice?
It reinforces that Italian courts will scrutinise how comparables are selected and whether FAR comparability is actually demonstrated—especially when the adjustment hinges on a small set of external comparables.