The General Court last week handed down its long awaited decision in the Apple case, concerning tax rulings given by the Irish Revenue Commissioners to two Apple companies in 1991 and 2007 (Cases T-778/16 and T-892/16). The outcome is a surprise: not just because the court found against the Commission, but also because of the comprehensive manner in which it did.
The two Apple companies ASI and AOE, by virtue of the combination of Irish law and US at the time, were treated as neither resident in Ireland nor the US. As such, the only taxes due in Ireland would be those that could be attributed to the branches of these companies (Taxes Consolidation Act 1997 s 25). The rulings provided formulae for determining what these amounts would be in any given year.
In a nutshell, the Commission’s argument was that Ireland granted state aid to Apple because the rulings were not in line with the applicable tax rules in terms of how much taxable profit could be said to have arisen in Ireland. Far more profit ought to have been attributed and taxed. The court rejected the Commission’s formulation of this argument on the basis that it has misconceived and misapplied the Irish tax rules and the OECD transfer pricing guidance.
There were two strands to the Commission’s central argument. The first was that the profits from Apple’s intellectual property, for which ASI and AOE held licences, should be taxable in Ireland. Critically, however, the General Court found that the Irish tax rules attributed profit from intellectual property to a branch of a non-resident company where the branch actually controls the intellectual property (S Murphy (Insp of Taxes) v Dataproducts (Dub) Ltd [1988] IR10). The Commission had assumed it did not need to demonstrate control on the part of the branches because the head offices of the companies had no physical presence or employees. By failing to demonstrate whether the branches controlled the intellectual property, the Commission had failed to demonstrate that the Irish branches were not sufficiently taxed (see paras 184-186, in particular).
By way of similar reasoning, the court found that the Commission did not prove that the Irish tax treatment was out of step with the OECD guidance because it erroneously assumed the profit from the intellectual property was attributable to the branches without an assessment of the functions carried out by those branches (see para 242). Indeed, the judgment makes clear just how limited Apple’s operations are in Ireland: for instance, only one person was responsible for quality control in ASI in Ireland (para 268); no branch staff were responsible for R&D facilities management (para 272); and no staff assigned to marking (para 274). Thus, these operations play only a limited role in generating value for the Apple group.
The second strand of the Commission’s argument was that, even if the profits from the intellectual property were not taxable in Ireland, the branches of ASI and AOE were insufficiently taxed given the value that these companies generated for the Apple Group. In a similar manner to the Starbucks judgment (Case T-636/16), the court found that although there were some errors on the part of the Irish Revenue Commissioners in granting the rulings, the Commission had failed to prove a causal link between the errors and the provision of an advantage (see para 480). An alternative line of reasoning (that the Irish Revenue Commissioners inconsistently applied the law on attributing profit) fell short on that same basis (paras 489–504).
Though a lengthy judgment and hence one that will still take time to digest, there are some preliminary comments which spring to mind.
First, much has been made of the €13bn headline figure that was at stake in the case. But the problem with that figure is that it assumed that all of Apple’s profits were taxable in Ireland. The Commission did not challenge the Irish rules on residency themselves and so had the Herculean task of somehow arguing that by misapplying the rules of profit attribution, the Irish Revenue Commissioners had undervalued the huge role that these small branches in Ireland played in Apple’s activities. A different route would have been possible if the Commission had instead argued that the Irish tax rules on residency gave rise to State aid because they were geared specifically to ensure that companies like Apple could be treated as non-resident in Ireland and the US. The articulation of that argument would rely upon digging down into the weeds of the exceptions to the exceptions in the old Irish rules on corporate residency. The time has passed for that argument, but it at least would have been more consistent with the figure of €13bn because it would have based on the notion that the profits of ASI and AOE were taxable in Ireland and not just the profits of their branches.
Second, buried in the judgment, the General Court made expressly clear that there was no EU arm’s length principle by which the transactions and arrangements in the case could be judged. Instead, ‘the Commission cannot, however, contend that there is a freestanding obligation to apply the arm’s length principle arising from [TFEU article 107] obliging member states to apply that principle horizontally and in all areas of their national tax law’ (para 221). A different finding would have had profound constitutional implications.
Third, the General Court was, in its judgment, pretty kind to the Irish Revenue Commissioners. Commenting on the process that led to the rulings being granted, the court noted that the methodology was ‘regrettable’ (para 348), ‘the explanations ... were brief’ (para 432), the Commissioners were ‘[un]able to [explain] the indicators and figures used’ (para 433), the process was ‘incomplete and occasionally inconsistent’ (para 479) and the rulings were ‘insufficiently documented’ (para 500). State aid law should focus not on whether the tax rules have been correctly applied, but instead whether the tax authorities have followed good processes and acted reasonably when engaging with taxpayers. Essentially, the question should not be whether the tax authority made a mistake, but rather whether it acted lawfully in public law terms. What the General Court has implicitly acknowledged, and what is revealed from a thorough reading of the Commission’s decisions in relation to Apple, is that there are serious questions to be answered by the Irish Revenue Commissioners as to the lawfulness of the 1991 and 2007 rulings (for instance, the length of time for which they ran, whether they were based upon sufficient information, and whether they were driven by employment consideration). But the Commission simply did not deal in any detail with the effects of these purported improprieties.
Prior to reading the judgment I would have thought there was an incredibly high likelihood that the case would be appealed. After reading it, I am less sure (though it is still a pretty good call that it will).
The General Court last week handed down its long awaited decision in the Apple case, concerning tax rulings given by the Irish Revenue Commissioners to two Apple companies in 1991 and 2007 (Cases T-778/16 and T-892/16). The outcome is a surprise: not just because the court found against the Commission, but also because of the comprehensive manner in which it did.
The two Apple companies ASI and AOE, by virtue of the combination of Irish law and US at the time, were treated as neither resident in Ireland nor the US. As such, the only taxes due in Ireland would be those that could be attributed to the branches of these companies (Taxes Consolidation Act 1997 s 25). The rulings provided formulae for determining what these amounts would be in any given year.
In a nutshell, the Commission’s argument was that Ireland granted state aid to Apple because the rulings were not in line with the applicable tax rules in terms of how much taxable profit could be said to have arisen in Ireland. Far more profit ought to have been attributed and taxed. The court rejected the Commission’s formulation of this argument on the basis that it has misconceived and misapplied the Irish tax rules and the OECD transfer pricing guidance.
There were two strands to the Commission’s central argument. The first was that the profits from Apple’s intellectual property, for which ASI and AOE held licences, should be taxable in Ireland. Critically, however, the General Court found that the Irish tax rules attributed profit from intellectual property to a branch of a non-resident company where the branch actually controls the intellectual property (S Murphy (Insp of Taxes) v Dataproducts (Dub) Ltd [1988] IR10). The Commission had assumed it did not need to demonstrate control on the part of the branches because the head offices of the companies had no physical presence or employees. By failing to demonstrate whether the branches controlled the intellectual property, the Commission had failed to demonstrate that the Irish branches were not sufficiently taxed (see paras 184-186, in particular).
By way of similar reasoning, the court found that the Commission did not prove that the Irish tax treatment was out of step with the OECD guidance because it erroneously assumed the profit from the intellectual property was attributable to the branches without an assessment of the functions carried out by those branches (see para 242). Indeed, the judgment makes clear just how limited Apple’s operations are in Ireland: for instance, only one person was responsible for quality control in ASI in Ireland (para 268); no branch staff were responsible for R&D facilities management (para 272); and no staff assigned to marking (para 274). Thus, these operations play only a limited role in generating value for the Apple group.
The second strand of the Commission’s argument was that, even if the profits from the intellectual property were not taxable in Ireland, the branches of ASI and AOE were insufficiently taxed given the value that these companies generated for the Apple Group. In a similar manner to the Starbucks judgment (Case T-636/16), the court found that although there were some errors on the part of the Irish Revenue Commissioners in granting the rulings, the Commission had failed to prove a causal link between the errors and the provision of an advantage (see para 480). An alternative line of reasoning (that the Irish Revenue Commissioners inconsistently applied the law on attributing profit) fell short on that same basis (paras 489–504).
Though a lengthy judgment and hence one that will still take time to digest, there are some preliminary comments which spring to mind.
First, much has been made of the €13bn headline figure that was at stake in the case. But the problem with that figure is that it assumed that all of Apple’s profits were taxable in Ireland. The Commission did not challenge the Irish rules on residency themselves and so had the Herculean task of somehow arguing that by misapplying the rules of profit attribution, the Irish Revenue Commissioners had undervalued the huge role that these small branches in Ireland played in Apple’s activities. A different route would have been possible if the Commission had instead argued that the Irish tax rules on residency gave rise to State aid because they were geared specifically to ensure that companies like Apple could be treated as non-resident in Ireland and the US. The articulation of that argument would rely upon digging down into the weeds of the exceptions to the exceptions in the old Irish rules on corporate residency. The time has passed for that argument, but it at least would have been more consistent with the figure of €13bn because it would have based on the notion that the profits of ASI and AOE were taxable in Ireland and not just the profits of their branches.
Second, buried in the judgment, the General Court made expressly clear that there was no EU arm’s length principle by which the transactions and arrangements in the case could be judged. Instead, ‘the Commission cannot, however, contend that there is a freestanding obligation to apply the arm’s length principle arising from [TFEU article 107] obliging member states to apply that principle horizontally and in all areas of their national tax law’ (para 221). A different finding would have had profound constitutional implications.
Third, the General Court was, in its judgment, pretty kind to the Irish Revenue Commissioners. Commenting on the process that led to the rulings being granted, the court noted that the methodology was ‘regrettable’ (para 348), ‘the explanations ... were brief’ (para 432), the Commissioners were ‘[un]able to [explain] the indicators and figures used’ (para 433), the process was ‘incomplete and occasionally inconsistent’ (para 479) and the rulings were ‘insufficiently documented’ (para 500). State aid law should focus not on whether the tax rules have been correctly applied, but instead whether the tax authorities have followed good processes and acted reasonably when engaging with taxpayers. Essentially, the question should not be whether the tax authority made a mistake, but rather whether it acted lawfully in public law terms. What the General Court has implicitly acknowledged, and what is revealed from a thorough reading of the Commission’s decisions in relation to Apple, is that there are serious questions to be answered by the Irish Revenue Commissioners as to the lawfulness of the 1991 and 2007 rulings (for instance, the length of time for which they ran, whether they were based upon sufficient information, and whether they were driven by employment consideration). But the Commission simply did not deal in any detail with the effects of these purported improprieties.
Prior to reading the judgment I would have thought there was an incredibly high likelihood that the case would be appealed. After reading it, I am less sure (though it is still a pretty good call that it will).
