The compliance of transfer pricing policies with the arm’s-length principle is frequently assessed through an analysis of the profitability results achieved by entities comparable to the tested party.
If there are no internal comparable transactions to use as a possible benchmark, the testing phase of the transfer pricing policy generally occurs through comparison to independent external comparable entities using the ‘additive’ and ‘deductive’ approaches.
The additive and deductive approaches
Through the additive approach, comparables are selected (after verifying compliance with the comparability criteria set by the OECD) from those identified as such by the company (in Italy, generally by the Administrative Department and the Commercial Department) or by the entity conducting the benchmarking analysis (through, for example, consultation and collection of information from certain trade associations).
Through the deductive approach, the selection of comparables is carried out through the use of commercial databases. Paragraph 3.42 of the OECD Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations states that the deductive approach begins with the examination of a broad list of companies that operate in the same industry, perform broadly similar functions, and do not exhibit clearly different economic characteristics. The list is refined using selection criteria and publicly available information (e.g., databases, websites, information about the taxpayer's competitors).
According to the OECD, no preference is given to one approach over the other, provided that the selection result is transparent and verifiable. It is common for these methodologies to be used in conjunction.
Refining the list of comparable companies
Once a list of companies potentially comparable to the tested party is identified, based on the available information, the sample is refined, with the possible elimination of:
Companies that carry out completely different activities from those of the tested party;
Companies that produce or distribute goods that are entirely non-comparable to those of the tested party;
Companies for which there is insufficient information to assert that they are comparables; and/or
Companies that are part of multinational groups or otherwise not independent.
The crucial point in a benchmarking analysis, therefore, lies in identifying the correct reference sample, leading to the identification of entities that:
Operate at the same stage (consider, for example, the functional difference between wholesale and retail, or companies at different stages of the production process);
Operate in the same sector as the tested party, and under the same operating conditions (consider companies that conduct their business with multiple clients, as opposed to those with a single, exclusive client, or companies that conduct retail through one or a few stores, as opposed to those with a large number of stores); and
Produce/distribute comparable products (in terms of characteristics and volumes).
All this is in order to avoid the phenomenon of cherry picking.
The selection of loss-making comparables in Italy
Italian transfer pricing legislation does not provide specific guidelines for the preparation of benchmarking analyses but generally follows OECD criteria.
One of the debated criteria concerns the acceptance of loss-making companies in the set of comparable entities.
The Italian tax authorities have generally rejected loss-making comparables because such results can reflect an extraordinary situation for the companies that cannot be used to demonstrate the compliance of transfer pricing policies with the arm’s-length principles.
On July 16 2024, however, the Supreme Court (judgment No. 19512/2024) stated that loss-making comparables cannot be rejected solely due to negative results achieved.
Implications of the Italian Supreme Court decision
In the case at hand, the company determined the markup applied to call centre services through the preparation of a group report, as well as through a country file that included a benchmarking analysis.
The analysis led to a 5% markup value, positioned in the upper quartile of the arm's-length range. The Italian tax authorities, however, applied a higher value of 7.42% as a result of an analysis that rejected, from each tax period concerned, companies without accounting data or with negative operating results in at least two out of three periods.
The Supreme Court ruling highlights that the criterion used by the tax authority does not appear to be consistent with the OECD guidelines, as it is not possible to pre-emptively exclude certain potentially comparable companies solely because they have reported reduced or negative results. In fact, the OECD guidelines do not provide for the outright elimination of loss-making companies or those with reduced or absent accounting values, if such results are achieved with the aim of obtaining better ones in the future.
Paragraph 10.67 of the OECD guidelines indicates that the exclusion of companies in “special circumstances”, such as those in start-up phases or those that have gone bankrupt, is permitted.
Therefore, loss-making comparables that satisfy a comparability analysis should not be rejected on the sole basis that they suffer losses. It is necessary to understand and analyse the economic reasons behind such losses to evaluate if they are a genuine managerial outcome that must be considered in the analysis of transfer pricing policies.
