Transalliance Iberica SA had priced its controlled transactions for the years 2008-2013 by comparing the gross margin achieved on an overall basis with the gross margins of comparable companies.
Following an audit, the tax authorities issued a notice of assessment rejecting the method used by the company due to differences in the treatment of cost items and thus issues of comparability at a gross margin level. Instead, the tax authorities applied the TNMM. The profit was outside the interquartile range and an adjustment to the median was made.
Transalliance lodged an appeal.
Judgement of the Court
The Court largely ruled in favor of the tax authorities, but according to the Court, an adjustment to the median could only be made where the tax authorities established the existence of comparability defects. Since such defects had not been established, the adjustment was reduced to the lower quartile.
Excerpt
“Of the points that are dealt with, the appellant focuses the discussion on the application of the median. In particular the Guidelines – 3.62 – state that “where the range comprises highly reliable and relatively equal results, it can be argued that any one of them satisfies the arm’s length principle. Where some defects in comparability persist, as discussed in paragraph 3.57, it may be appropriate to use measures of central tendency that allow this point to be determined (e.g. median, measure or weighted mean, depending on the speciï¬c characteristics of the data) in order to minimise the risk of error caused by defects in comparability that persist but are not known or cannot be quantiï¬ed”.
Applying this rule, p. 109 of the Agreement states that “the normal practice in such cases is to use the median as the most signiï¬cant value of the interquantile range, as it avoids the problems that extreme values cause in the calculation of the arithmetic mean”.
The appellant – p. 120 of the agreement – argued that the administration could not apply the median “mechanically”, as such automatism is not required by the Guidelines. Therefore, it argued that it is sufficient to apply the “lower quartile of the interquartile range” instead of the median. To which the Agreement replied that ‘the preference for the median must be justified on statistical grounds: it is a robust statistic, which is not influenced by extreme values in the sample of purchasables’. Both the TEAC and the AbogacÃa del Estado insist on the argument.
The Chamber’s position in this regard is described in our SAN (2nd) of 6 March 2019 (Rec. 353/2015 ) – the appeal was rejected by order of 14/11/2019 – and 4 February 2021 (Rec. 658/2017), which hold that “it is legitimate to resort to what the Guideline calls “measures of central tendency”, but whoever resorts to them has the burden of reasoning and setting out the reasons that lead to their application”. This solution has been accepted in the Resolution of the TEAC of 23 November 2021 (4881/2019), which states, with a correct interpretation of the position of this Chamber, “that in order to resort to the median, there must be defects of comparability. In the event that such defects are not highlighted by the inspection, the adjustment would be made to the lower quartile”.
Well, what the Inspectorate has done is to “automatically” apply the median -also the TEAC and the AbogacÃa del Estado-, without explaining and reasoning the concurrence of “defects of comparability”, a burden that corresponds to it and that the Chamber should not replace. This means that the lower inter-quantile range must be applied and not the median, as the appellant claims.
On this point, the appeal is also upheld.”
Click here for English translation
Click here for other translation