In the Keladis cases (Keladis I C-72/24 and Keladis II C-73/24), the Court of Justice of the European Union (‘CJEU’) has finally ruled on the use by customs authorities of “acceptable minimum prices” derived from European statistical databases.
For several years, OLAF has made available to national administrations tools based on aggregated EU customs averages in order to identify declared values considered “too low”. The issue was a sensitive one: can such data be used not only to detect a risk, but also to recalculate the customs value and recover additional duties?
Following the Opinion of the Advocate General Norkus of 8 May 2025 – which we analysed in a previous brief dated 9 June 2025 and which had already called for a strict and subsidiary framework governing reliance on statistical databases – the Court, on the questions referred for a preliminary ruling, adopts a nuanced position. It largely confirms that approach by firmly recalling that “acceptable minimum prices” may only be used on a strictly exceptional basis and cannot constitute an autonomous basis for determining customs value.
The Court first confirms that such statistical data can under no circumstances constitute an autonomous ground for rejecting the transaction value. The hierarchy of valuation methods remains intact: the transaction value is the rule, and alternative methods are the exception, applicable only where no acceptable transaction value exists.
However, where all successive methods prove inapplicable, in particular in the context of a post-clearance audit without the possibility of physical verification and in the presence of imprecise documentation, the Court accepts that “acceptable minimum prices” calculated on the basis of aggregated European data may be used as a last resort under the so-called fallback method.
The significance of the judgment therefore lies less in recognising the possibility of exceptional reliance on acceptable minimum prices than in the procedural safeguards provided to operators ensuring that such reliance remains genuinely exceptional.
In that regard, the Court specifies that the statistical data relied upon must relate to imports carried out at the same time, or approximately the same time, as those under review. By analogy with the rules applicable to the deductive method, the Court refers to a 90-day period, which may be adjusted with “reasonable flexibility”, in order to prevent averages too remote in time from serving as a reference.
Moreover, the Court addresses the issue of simplified declarations through tariff grouping. It holds that the use, at the time of importation, of the procedure allowing several goods to be grouped under the highest tariff heading does not prevent the administration, in the event of a post-clearance audit, from reassessing the customs value according to that same logic of simplification. In other words, the declarant’s choice logically continues to produce its effects in the context of a reassessment.
Finally, the judgment emphasises the obligation to state reasons and the importer’s right to understand the method applied. Even where the statistical databases used form part of anti-fraud tools and are not fully disclosable, the administration must set out the essential elements that led to the determination of the value. This requirement is central to safeguarding the rights of defence and preventing reliance on opaque statistics from depriving operators of any meaningful opportunity to challenge the assessment.
The Keladis judgment therefore does not close the door to the use of statistical databases, but it does strictly circumscribe the conditions under which they may be relied upon. It is nevertheless to be feared that EU customs authorities may view this ruling as encouragement to make more intensive use of such databases.
For businesses, the message is clear: only robust documentation, supported by an individualised economic analysis of the declared prices, can effectively shield them from reassessments based on European statistical comparisons.