Non-Discrimination in European and Tax Treaty Law

I. S. 310Introduction

State aid is defined as an advantage in any form whatsoever conferred on a selective basis to undertakings by national public authorities. The favorable treatment that is entailed in the process of offering advantages may be direct subsidies or fiscal reliefs, only to certain beneficiaries; this reflects the concept of selectivity. Those discriminatory interventions of States by which they favor “certain undertakings or the production of certain products” pose the same threat for preserving a level playing field in a market as the discriminatory obstacles that occur in legal contexts such as the fundamental freedoms and the non-discrimination prohibition in the context of the OECD Model Tax Convention. The relevance of this contribution in a collection on the general topic of non-discrimination is explained not only because of the theoretical and conceptual proximity among the various set of rules, but also because the examination of selectivity, still a controversial and blurred stage in the State aid review, seems to encompass in its methodology a comparability analysis examination, similar to the one used in the aforementioned legal contexts.

The contribution will examine the evolution of the assessment of the condition of selectivity in the area of taxation (fiscal State aid) in the light of recent developments. Further, it will try to reveal and critically assess the issues that emerged in that period and which are related to the methodological framework by which selectivity is examined.

II. State Aid Rules in the Context of the Internal Market

The establishment of the internal market has been the cornerstone of the European Union’s policy since its early stages. The ultimate aim of that policy is the integration of the national markets of the Member States into a single European market, without internal frontiers where persons, goods, services and capital can move freely. For the achievement of this key task the Treaties of the European Union contain a specific system of rules. These are the rules laid down in Part Three of the current TFEU, under the title “Union Policies and Internal Actions”, and they constitute a core element of the European Union’s legal order. That system operates at two levels: prohibiting actions of the Member States that are “incompatible with the internal market” (negative integration), on the one hand, and S. 311adoption of measures that promote the establishment and “the functioning the internal market” (positive integration), on the other. The fundamental freedoms, the competition rules and the Monetary Union rules are prominent parts of that system.

Competition rules, laid down in the chapter entitled “Common Rules on Competition, Taxation and Approximation of Laws”, are an integral part of the system that serves the establishment of the internal market. On the one hand, they promote the achievement of a high level of competitiveness of the European Union’s economy, and, on the other, and most importantly, they ensure that the objective of an integrated market will not be hindered by obstacles related to the area of competition, caused either by private firms or by Member States. This way, by securing undistorted competition within the internal market they contribute to its establishment.

State aid rules are part of the competition rules that are addressed to the Member States. Their aim is to tackle state interventions that may cause situations “incompatible with the internal market”, in a similar way that the fundamental freedoms tackle state obstacles with regard to the functioning of the internal market, but for the former, from the specific perspective of competition (area) policy. What they finally intend to achieve is a level playing field for all operators in the European single market. As stated in the Report on Competition Policy 2013 “State Aid policy protects the internal market from distortions and helps to steer public resources towards competitiveness-enhancing objectives”. At the end of the day, both State aid rules and fundamental freedoms serve the same overall purpose.

Within this framework, State aid rules, laid down in Article 107 TFEU, declare incompatible with the internal market any state financial interventions in the market that have the effect of providing an advantage in any form whatsoever to undertakings on a selective basis. According to Article 107(1) TFEU, “[s]ave as otherwise provided in this Treaty, any Aid granted by aMember State or through State resources in any form whatsoever which distorts or threatens to distort competition by favouring certain undertakings or the production of certain goods shall, insofar as it affects trade between Member States, be incompatible with the common market”. Nonetheless, a State aid may be compatible with the internal market on the grounds laid down in Article 107(2) and 107(3) TFEU.

S. 312The ECJ, along with the practice of the European Commission, is involved in a continuous process of developing the notion of State aid by setting out its general principles, its special features and by dictating its implementation in specific areas (social security, health care) through a series of actions. Part of this development concerns the area of taxation, since tax expenditure constitutes a typical instrument of intervention that states use extensively. In recent years, that relationship has proven to be of great significance and its importance is growing. However, what triggered a more systematic approach to the relationship between State aid and taxation were the wide-ranging discussions on addressing harmful tax competition among the Member States in the late nineties that led to the adoption of the Code of Conduct for Business Taxation, a non-binding instrument that is the background for addressing harmful tax competition. This is the context in which the European Commission based its work in State aid and taxation. The first prominent outcome of that work was the Commission Notice on the application of the State aid rules to measures relating to direct business taxation, which clarifies the application of State aid rules on tax incentives. Later, it continued with the 2004 Report on the Implementation of the Commission Notice on the Application of the State Aid rules to Measures relating to Direct Business Taxation and, ultimately, with the recently published Draft Commission Notice on the notion of State aid pursuant to Article 107(1) TFEU.

Against that background, State aid rules scrutinize tax measures that may distort competition and are thus incompatible with the internal market. A typical example of the working of the State aid prohibition in the area of taxation is the case where a State intervenes in a market and offers a subsidy (in the form of tax relief) only to certain beneficiaries. The effect of this is that a given group of companies enjoys a better position in the market, in terms of competition. Provided that these companies are competing, the result is a distortion of competition and creation of an obstacle to the internal market. It is thus apparent why the harmful eleS. 313ment for the competition is selectivity (the relief is given “only to certain undertakings”).

III. Selectivity Assessment in Fiscal State Aid

A. Introduction

The wording of Article 107(1) TFEU prohibits aids that, among other criteria, favor “certain undertakings or the production of certain goods”. As is usually the case, the wording of a broad and general provision is far from sufficient to reveal its actual content. What can initially be inferred from it, as a first point, is that measures that apply openly and indiscriminately to all the undertakings are outside the scope of Article 107 TFEU. On the other hand, measures that are designed and by their statutory nature (de jure) restrict their application only to certain beneficiaries (a single company or even a whole sector) are found to be selective. However, seemingly general measures, designed in broad and objective terms, can in practice (de facto) favor only certain beneficiaries. In this respect, the selectivity examination should set the boundary between what can be regarded as a genuine general measure and what is selective. Trying to set that boundary between genuine and seemingly general (but in practice selective) measures becomes a challenging task in the complex area of taxation. Apart from those first remarks, it is a starting assumption that the notion of selectivity clearly implies that it is all about the differential treatment of certain undertakings or products that takes place when States introduce tax measures that favor (provide advantages to) certain undertakings.

The burden of determining the content and the scope of that phrase, i.e. the selectivity criterion, lies with the European Commission and the ECJ. While the European Commission has the exclusive competence to run the State aid review purS. 314suant to Article 108 TFEU, its acts are subject to the judicial review of the ECJ, while the latter also responds to preliminary ruling requests from national courts. Therefore, at the end of the day, even if the Commission through its practice contributes to the development of the methodology by which selectivity is assessed, the ECJ will, eventually, determine it. Indeed, the notion has been developed widely through both institutions, however with serious instances of inconsistencies and uncertainties.

It can be argued that two elements comprise the complexity of the selectivity examination in fiscal aid. First, the notion of de facto selectivity, which requires a delicate and thorough examination of the effect of a tax measure in a market. The beneficiaries can be not only some precisely indicated undertakings, but also a whole sector and the large number of beneficiaries does not rule out that they may be a part of a certain beneficiary group that enjoys selective advantages. Second, in the area of taxation, rather than direct subsidies, the Member States intervene with measures that mitigate the “normal” charges on undertakings by forgoing potential revenues from their budgets. The notion of “normal” charge that is calculated against a “normal” level of taxation has produced severe ambiguity in the State aid review. These two observations give an indication why assessing selectivity is not a straightforward exercise. In fact, it is the very nature of tax measures that makes selectivity one of the most challenging parts of the State aid review.

In this context, the European Commission and the ECJ gradually developed a methodology according to which the selectivity of a tax measure will be assessed: the selectivity test. In the development of formulating the selectivity condition, one can distinguish among three different periods. The first period starts from the early judgments in De Gezamenlijke and in Italian Republic vCommission of the European Communities and continues with the first stage of the European Commission’s practice, which, eventually, ends in the Commission Notice on the application of the State Aid rules to measures relating to direct business taxation in 1998, by which the European Commission in the Notice formulated the case law findings in the light of the political dictates of the Code of Contact and proposed a specific selectivity test. The second period is highlighted by the judgment in the Adria-Wien case where new elements were inserted into the notion of selectivity. The third period covers the years 2008 to 2015, where some important S. 315cases (judgments in Gibraltar and Paint Graphos) moved the notion of selectivity further.

Nevertheless, among many legal commentators, as well as the Advocates General themselves, there is a general consensus that the notion of selectivity in the area of fiscal State aid is far from being clear and coherent. The overview of the court case law and the Commission’s practice that follows will support that view.

B. Early Period and the Notice of 1998

This first stage can be regarded as the most important for the evolution, not only of the selectivity test, but for the broader review of fiscal State aid. In one of the very first cases in which the ECJ dealt with tax measures, the Italian Republic vCommission of the European Communities case, which concerned a reduction of social charges only for the textile sector, the ECJ held that the measure at stake “is ameasure intended partially to exempt undertakings of aparticular industrial sector from the financial charges arising from the normal application of the general social security system, without there being any justification for this exemption on the basis of the nature or general scheme of this system”. This statement introduced two important notions encompassed in the selectivity test, namely the “exception” from a “normal application of the general (…) system” and the “justification on the basis of the nature or general scheme”. Years later, in an Opinion for the Sloman Neptun case, Advocate General Darmon merged those observations into a formula that is called the “derogation test”.

The Notice of 1998 accumulated the findings of the Court’s and the European Commission’s practice until that time and formulated the application of Article 107(1) TFEU to tax measures. According to the Notice, there are four criteria to label a tax measure incompatible with the internal market: (a) the measure must confer on recipients an advantage which relieves them of charges “normally” borne by their budgets, (b) the advantage must be granted by the State or through State resources, (c) the measure must affect competition and trade between Member States, (d) the measure must be specific or selective in that it favors “certain undertakings or the production of certain goods”.

S. 316The Notice also went a step further. Except for summarizing some basic features of the selectivity condition derived from the case law of the ECJ it laid down, based on the derogation test proposed by Advocate General Darmon, a three-step formula on how Article 107(1) TFEU applies to tax measures. Accordingly, the formula consists of three steps: (a) identification of the common system applicable (or the reference system), (b) the existence of an exception to the application of that common system, (c) the possible justification “by the nature or general scheme” of the system.

This analytical framework, known as the derogation test (or rule/exception test or three-step approach) is until now, although with some modifications, the basis of the fiscal State aid review of the condition of selectivity. According to that method, the selective nature of a measure can be seen in the fact that while all undertakings are subject to a common applicable system (which could be the corporate income Tax, the VAT system) there are certain beneficiaries that enjoy a favorable exception, provided that this exception is not justified “by the nature or general scheme” of the system.

The first step according to that formula is the identification of the common applicable system (also called “reference system” or “normal system of taxation”). This will serve as a benchmark that will make the identification of an exception possible. This is a vital part of the test, since the existence of an exception or a derogation can only be conceivable if there is a point of departure. It is also a step that poses challenges for the general limits of the fiscal State aid review in tax matters, as will be discussed later. At the same time, as will be also discussed later, it is a point that has produced severe confusion in the formulation, since it is related both to the examination of the criterion of advantage (mitigation of the “normal” charges that arise from the common applicable system) and to the selectivity test (exception to the common applicable or “normal” system). The common system is then a set of rules that allow the “normal” boundaries of that system to be drawn and an exception to be identified.

The exception (or derogation) element is plainly the offering of a favorable exception to that common applicable system and thus, resulting in offering a competitive advantage to certain undertakings. This is conceptually the step where the examination of the differential treatment takes place. In other words, the exception means that the undertakings that are affected by the measure are treated more favorably than the undertakings subject to the common system. If an exception or a derogation can be established, then the measure is considered prima facie selective. Ultimately, the selective or general nature of the measure will be established after the third step of the test.

S. 317The third step consists of the examination of a possible justification of the differential treatment (or of the exception from the common rule) because of the “nature or general scheme of the system”. According to the Notice, these are measures whose economic rationale makes them necessary for the functioning and the effectiveness of the tax system or interventions that serve the objectives of the common system and its proper functioning and effectiveness. Objectives outside the framework of the general scheme, such as social or ecological goals, are not regarded as grounds that could justify an exception for the purposes of Article 107(1) TFEU.

The test described above and applied extensively by the Courts and the European Commission entailed, from the beginning, several controversial issues. The review in the literature of that three-step approach mainly concerned the difficulties arising from how to identify the “reference system”. In this respect, Wolfgang Schön reasonably questions how one can determine a “normal” level of taxation in instances where a State sets two different rates for polluting and non-polluting cars. It seems that no answer can be given to which is the “normal” and which the derogation scheme or whether the State has introduced an incentive in buying environmentally-friendly cars or a disincentive in buying ecologically harmful cars. On the other hand, the Court regularly recalls that “the very existence of an advantage may be established only when compared with ‘normal’ taxation”. While this seems to refer to the examination of the condition of the advantage, it also coincides with the reference system which is the benchmark to establish the selectivity of a measure. The Court not only has failed to give clear answers on this, it further seems to blur the concepts of advantage and selectivity, as will be discussed below. Moreover, there is also some ambiguity as to how to determine the right reference system when a measure introduces a sectoral, or a specific new system of taxation, that forms part of the broader tax system of the State. The ECJ has also failed to give concrete answers.

C. The Adria-Wien Pipeline Case

Ιn the Adria-Wien Pipeline case the ECJ applied a different kind of wording and, perhaps, reasoning. The ECJ was asked whether a rebate on a newly introduced S. 318energy tax granted only to undertakings whose activity is in the production of tangible goods, but not to those in the production of services, favored certain undertakings. The appeal in the Austrian courts, based on the breach of the equal treatment principle, was filed by the Adria-Wien Pipeline company whose activity was regarded a service-providing activity and according to the scheme was thus not entitled to the rebate. According to the European Commission, that rebate was granted only to certain undertakings, and was an exception to the normal regime, and it was not justified by the nature of the system. In his Opinion Advocate General Mischo claimed that it was doubtful whether the rebate was a derogation from a normal rule. In fact, he took the view that, provided that the reference system was the general system of taxation (not the newly introduced system of energy taxation), then the rebate just restored the tax burden of the manufacturing sector undertakings back to the normal level of the general system, so no derogation could be conceivable, admitting in the same time that if Member States were to be allowed to make use of that pattern it would open the door to similar abuses.

The ECJ’s reasoning followed a new line of description of the selectivity condition and contributed an important element to the evaluation of the selectivity test. Now, on the question whether measures that apply only to undertakings of the manufacturing industry can be regarded as State aid, it responded that the “[t]he only question to be determined is whether, under aparticular statutory scheme, aState measure is such as to favour ‘certain undertakings or the production of certain goods’ within the meaning of Article 92(1) of the Treaty in comparison with other undertakings which are in alegal and factual situation that is comparable in the light of the objective pursued by the measure in question”. It concluded that undertakings of the services sector, just like those of manufacturing sector, can be both major consumers of energy and that in the light of the ecological objective of the measure, both manufacturers and service providers damaged the environment equally; thus, they were in comparable legal and factual circumstances. In the third step, the ecological considerations raised by the Austrian authorities were not found to justify that treatment, so insofar as that measure was not found to be justified by the general scheme it was declared selective.

The reasoning in Adria-Wien Pipeline incorporated into the selectivity test a comparability analysis, according to which if a measure favors certain beneficiaries over others that are in a comparable legal and factual situation in the light of S. 319the objective of the measure, it can be regarded as prima facie selective. The comparability exercise introduced by the ECJ inserted two areas of ambiguity. The first is that now the Court must make the comparison in the light of a comparability factor, i.e. an objective of a system. In that first case it was the objective of the measure. However, in other cases the Court looks to the objective of the system producing uncertainty in determining the boundaries of the comparison. In the second area, it created some uncertainty on the architecture of the three-step approach (or the derogation test). Is that examination “the only question to be determined” or is the comparability exercise the tool to establish the exceptional or derogating character of the measure?

In its judgment in Adria-Wien Pipeline, the Court focused directly on the question of selectivity, without even applying the criteria of Article 107(1) TFEU. However, in paragraph 39 of the judgment the Court accepted that there was a normally applicable rate and maintained that the assessment must be made with regard to a “particular scheme”. Under those circumstances, if one were to try to apply the rule/exception formula, thus take as reference the same “particular scheme”, indentify the exception of the undertakings of the manufacturing sector and consider a possible justification, the result would be the same. That led some scholars to the conclusion that the pattern proposed by the ECJ on Adria-Wien Pipeline did not alter the basic structure of the selectivity test; it just incorporated a comparability exercise into the selectivity examination.

D. Recent Developments

During the period after the Adria-Wien Pipeline judgment the application of the selectivity criterion seems to become more blurred and uncertain. Many of the important cases that developed the notion of selectivity arose after the ECJ overturned the General Court’s judgments that entailed a different understanding of the selectivity criterion. That period also illustrates a seemingly inconsistent application of the selectivity examination. Reasonably, commentators inquired whether there is a single examination that just evolved, or two separate ones, or just a confused application by the courts. This is seen in the reasoning of the following cases.

In the judgment in Unicredito Italiano SpA in 2005 on an apparent selective measure the Court held that “Article 87(1) EC prohibits Aid which ‘favours certain undertakings or the production of certain goods’, that is to say, selective Aid. Aid may be selective in the light of that provision even where it concerns a whole economic sector (see, in particular, Case C-75/97 Belgium v Commission [1999] ECR S. 320I-3671, paragraph 33). In the present case, the tax reduction applies to the banking sector. It does not benefit undertakings in any other economic sectors. In addition, within the banking sector it benefits only undertakings which carry out the operations referred to. 48 Without it being necessary to determine in addition whether, as stated by the Commission in point 33 of the grounds of the contested decision, the tax reduction is more advantageous to large undertakings, it is therefore clear that that measure is selective in relation to other economic sectors and within the banking sector itself”. And according to the reasoning that follows “[i]t is in fact a departure from the ordinary tax scheme. The undertakings concerned enjoy tax relief to which they would not be entitled under the normal application of that scheme and to which undertakings in other sectors which carry out similar operations, or undertakings in the banking sector which do not carry out operations such as those referred to, are not entitled”.

In the judgment in Kingdom of Belgium and Forum 187 ASBL, the Court, following the Adria-Wien Pipeline pattern, held that “[a]ccording to settled case-law, Article 87(1) EC requires that it be determined whether, under a particular statutory scheme, a State measure is such as to favour ‘certain undertakings or the production of certain goods’ in comparison with others which, in the light of the objective pursued by the system in question, are in a comparable legal and factual situation”. In the next paragraph it continued by stating that “it is clear from the above analysis that the exemptions from property and withholding tax, from capital duty and the granting of a notional withholding tax constitute derogations from the ordinary Belgian tax regime”. From the above reasoning it seems that, regardless of the different introductory wording, the Court applied the same rule/exception (or derogation) approach. The Court looked for an exception or derogation from the reference system that enterprises carrying out similar operations did not receive, since they are subject to the common applicable reference system.

On a similar pattern, in the judgment in European Commission vKingdom of the Netherlands the ECJ proceeded with a comparability analysis, but concluded with a derogate type of reasoning:

S. 321Reasoning such as that above led some commentators to argue that the comparability analysis was incorporated into the rule/exception test in such a way that the comparability analysis constitutes the method to identify the exception in the framework of the second step of the three-step analysis.

However, in certain cases, the ECJ found itself in applying a pure comparability analysis, similar to that of the fundamental freedoms. In this regard the Gibraltar case is of great interest. According to the facts, Gibraltar introduced a new corporate tax system applicable to all undertakings in its territory that included three taxes: a payroll tax, a business property occupation tax, and a registration fee. The payroll tax and the business occupation tax were capped at 15 % of profits, a design that had the effect of companies being liable for these taxes only if they made profits and in an amount not exceeding 15 % of profits. Initially, the Commission found that tax selective for two main reasons. The 15 % cap on the two taxes favored companies with profits that were low in relation to the number of employees and occupation of business property and second and most importantly, the imposition of the first two taxes favored offshore companies that were left essentially outside the scope of the new corporate tax system, since they did not have any physical presence in Gibraltar. Interestingly, the General Court when assessed the selective nature of the measure made reference to the paragraph 16 of the Notice of 1998 and the three-step methodology finding that the European Commission failed to appreciate what was the commonly applicable system, and thus a derogation from it, in the new Gibraltar corporate income tax system. Consequently, in the view of the General Court, no derogation from the system could be established and therefore the selectivity criterion could not be met.

The ECJ overturned that reasoning. It held that “[a]s regards appraisal of the condition of selectivity, it is clear from settled case-law that Article 87(1) EC requires assessment of whether, under a particular legal regime, a national measure is such as to favour ‘certain undertakings or the production of certain goods’ in comparison with others which, in the light of the objective pursued by that regime, are in a comparable factual and legal situation”. It further ruled that since based on a settled principle that the measures at stake are defined in relation to their effects, when the reference system of taxation is designed in clearly arbitrary way in order to circumvent the State aid review by rendering it impossible to determine what constitutes the reference system, it proceeded with a pure comparability analysis: “In view of the features of that regime, outlined in the preceding paragraph, it is S. 322apparent that the regime at issue, by combining those bases, even though they are founded on criteria that are in themselves of a general nature, in practice discriminates between companies which are in a comparable situation with regard to the objective of the proposed tax reform, namely to introduce a general system of taxation for all companies established in Gibraltar”. In that instance, the ECJ departed from the three-step formula, by which a derogation must be assessed against a defined reference system, and proceeded to make an independent comparability analysis outside the scope of a three-step test.

Similarly, in the judgment in Sardegna of 2009, selectivity was declared after a pure comparability examination:

IV. Different approaches to the nature of the Selectivity test

That diversity in the reasoning and the wording in the ECJ case law has been interpreted variously by legal commentators, and differently understood by various Advocates General and even by the European Commission. In particular, it is being debated whether there is a new trend being developed by the ECJ towards a pure comparability analysis test, or whether the analysis of the case law nevertheless reflects the same derogation-based formula, or, even, whether the analysis of selectivity is adjusted, case by case, to reach policy choices according the specific circumstances of the case.

S. 323Advocate General Leger understood the concept as follows:

Advocate Leger implies that the comparability analysis is the methodological tool that will determine the existence of the derogation or exception.

On the same line, C. Micheau supports the view that despite the various reasoning and terminologies used by the Courts there is only one test in substance, the derogation test. According to that test, a measure is considered selective when it constitutes a derogation from the standard application of the tax system. The way the derogation is established is by comparing the recipients of the measure with other undertakings in a similar situation that are excluded from the application of the measure. In other words, she too notes that a measure can be regarded as derogating if it differentiates between undertakings in comparison with others which are legally and factually comparable in the light of the objective pursued by the scheme.

P. Nicolaides and I.E. Rusu describe the derogation approach formula in its first step (the determination of the reference framework) as being the “establishing of the benchmark or the general system of reference for undertakings which are in comparable legal and factual situation in the light of the objective pursued by the measure in question” in a sense that in the second step, that of derogation or exception, the beneficiaries will be compared with that group of undertakings that compose the reference framework. A derogation can exist if the beneficiaries enjoy, by derogation, a favorable differentiation with regard to a group of comparable undertakings. Nicolaides confirms that that the comparability exercise has been inserted in the selectivity test and he highlights the effect of that development: Now a measure may be found selective not only if grants exceptions to a certain category of undertakings but also if a tax measure is narrower in the scope than its stated objective and it does not include any potential recipients that are in comparable legal and factual situation. The same approach is taken by M. Perk S. 324and S. Lefevre. They also link the comparability exercise to the identification of the reference system. They comment that “that is where the objective of the measure takes apart, in the sense that it helps delimit the circle of undertakings whose treatment should be taken into consideration when analyzing the effects of the measure at stake”.

On the other hand, R. Szudoczky, even though admitting that a trend towards synthesizing the two tests has emerged, considers that there are two sufficiently distinct approaches. The Adria-Wien Pipeline reasoning that introduced the comparison approach is conceptually different from the derogation test. However, she finds that in the Paint Graphos judgment the ECJ required the examination of both the derogation from a common tax system and the comparability exercise in the same step with the latter serving as a confirmation of the existence of the derogation. She further argues that, as long as the derogation from a common tax system by itself is also a part of the second step, then the three-step approach reach its limits in situations where a derogation from the common tax system is unable to be established, such as under the facts of the Gibraltar case. Under those circumstances, when the scope of a tax regime is designed on an arbitrary manner so as to exclude certain undertakings that should be covered in the light of the objective of the regime, she suggests that there should be no need to identify a common system and a derogation from it but rather to proceed only with the comparability examination. On the contrary, the derogation approach can be used when the measure “apparently” forms part of the broader system.

Interestingly, W. Schön, while admitting a “growing trend” regarding the selectivity examination based on discrimination analysis, considers that it is not a sufficient approach to achieve a result. Discriminatory taxation, in his words, cannot by its own fulfill the requirements of Article 107(1) TFEU. The existence of the distinct condition of advantage must always be identified.

S. 325On the other hand, M. Lang considers that the latest judgments suggest the ECJ clearly makes an examination of comparability. He cites the Gibraltar case to stress that an identification of the general system or a reference system and derogations to it are irrelevant. He heavily criticizes the derogation or rule/exception formula as totally arbitrary by arguing that at certain times it is impossible to determine what is the rule and what the exception and also disregards the strict approach of the three-step formula. According to M. Lang, a selective advantage will take place only when undertakings that are in a comparable factual and legal situation are treated differently under tax provisions. However, he further proposes that, in order to examine whether two beneficiaries are actually in comparable factual and legal situations, it is not the objective of the measure or the system that must be taken into account, but the level of competition between the beneficiaries. He notes that companies that are not sufficiently competitive can be treated differently and that can be assessed properly with an inclusion of a proportionality review into the broader examination. Opponents of this approach note, however, that even if different treatments do not distort domestic competition within a Member State it will still potentially distort the intra-State competition, which is the primary goal of Article 107(1) TFEU. They further add that the issue of the distortion of competition is dealt under the other condition of Article 107(1) TFEU.

That diverging understanding and interpretation of the selectivity criterion illustrates the common ground of all the commentators that the ECJ has failed to provide a coherent and lucid concept of material selectivity.

V. Critical analysis

A critical analysis of the issue should take as a starting point the examination of what can be regarded as a derogation analysis, what as a comparability analysis and how do they fit into the context of the three-step approach. The derogation analysis (or what is known as the rule/exception test) is the analysis in which one has to examine whether there is an exception to the application of the normal (reference) tax system. Yet, that exception can only be determined after a comparison of the treatment of the alleged reference or normal tax system with the tax S. 326treatment under the exceptional/derogatory regime. A comparability analysis, on the other hand, is an analysis that can be derived from the context of the fundamental freedoms review, according to which is examined whether two undertakings in a similar or comparable factual and legal situation, in the light of a specific object or purpose, are treated equally or not.

There is no doubt that the latter type of exercise is executed in several late cases. However, according to the author, that exercise seems to be carried out by the ECJ in the context of identifying the derogation or the exception from a reference system. In other words, if we consider it in the context of the three-step formula, where the second step is the step of the identification of the exception, then the comparability exercise can be considered the tool for establishing the exception. The subtle difference from the formula without any comparability exercise is that the second step would be established by just comparing the treatment with the alleged reference or normal tax system and the tax treatment with the exceptional/derogatory regime.

In the context of that analysis, it is important to clarify where the reference system plays a role. The starting point of the Fiscal Aid review is that States provide tax incentives by mitigating the normal tax burden of undertakings. The identification of that normal burden requires the identification of the normal tax system, i.e. the reference tax system. However, in the author’s opinion, the reference system, and thus a derogation from it, is conceptually related mostly to the identification of the condition of the advantage. The ECJ has repeatedly held that “[t]he Court admittedly held in paragraph 56 of Portugal vCommission that the determination of the reference framework has aparticular importance in the case of tax measures, since the very existence of an advantage may be established only when compared with ‘normal’ taxation”. However, it is not irrelevant for the comparability analysis at all. The identification of the reference framework, the common applicable tax system, is necessary to provide the objective “assigned to the tax system of the Member State” in light of which the comparison will be made.

Interestly, the above considerations are reflected in one of the most recent ECJ judgments the Ministerio de Defensa and Navantia SA vConcello de Ferrol given S. 327in 2014. What is of great interest is that the ECJ applied a derogation analysis, once under the examination of the condition of advantage and then again in the context of the second step, along with a comparability analysis, in the three-step analysis of the condition of selectivity. The preliminary question addressed to the ECJ asked if an exemption from the property tax of a plot of land belonging to the State and made available to an undertaking whose capital is wholly State-owned and which produces, from that plot of land, goods and services that may be traded between Member States on markets open to competition constitutes State aid prohibited by that provision.

The ECJ developed its reasoning as follows. It started by stating the conditions of Article 107(1) TFEU, not including the condition of selectivity among them, but only that of the advantage. According to its reasoning, the examination of the advantage is the first step necessary, i.e. assessing whether the measure at issue must be regarded as conferring an advantage on its recipient. Thus, it starts its reasoning by stating that “[i]t follows that ameasure by which the public authorities grant certain undertakings favourable tax treatment, which, although not involving the transfer of State resources, places the recipients in afinancial position more favourable than that of other taxpayers, amounts to State Aid within the meaning of Article 107(1) TFEU”. After reviewing the relevant tax provisions, the ECJ found that “[i]t must accordingly be held that the property tax constitutes atax normally payable by Navantia and that the exemption enjoyed by that undertaking has the effect of mitigating directly, without any other measure being necessary, the charges that would ordinarily be borne by an undertaking in the same situation. Consequently, it appears that atax exemption of that nature confers an economic advantage on Navantia”.

After having established the existence of an (economic) advantage it proceeds by stating that “[i]t should be noted, moreover, that Article 107 TFEU prohibits Aid ‘favouring certain undertakings or the production of certain goods’, that is to say, it prohibits selective Aid”. Subsequently, the ECJ proceeds with the wording of the Adria-Wien judgment by stating that “[i]n that regard, it is clear from the Court’s settled case-law that Article 107(1) TFEU requires it to be determined whether, under aparticular statutory regime, aState measure is of such anature as to favour ‘certain undertakings or the production of certain goods’ as compared with others which, in the light of the objective pursued by the regime in question, are in acomparable legal and factual situation”.

S. 328However, while in the Adria-Wien Pipeline judgment the above thought was laid down as the “only question to be determined” the Court now proceeds by describing the three-step formula:

The analysis that comes next is of great interest because it only re-assesses what was already assessed with respect to the condition of the (economic) advantage. In paragraph 36, the Court determines that the reference framework is the property tax of Spain.

Then, it repeats the result of the analysis of paragraphs 25-27 where it elaborated on the existence of the economic advantage that derived from the exemption from the normal regime of the property tax.

After identifying what was already examined in the discussion of the advantage criterion, it proceeds with a comparability analysis in the light of the objective of the property tax regime:

Ultimately, the Court established that Navantia was indeed favored in relation with undertakings that, in the light of the objective of the property tax regime, were in comparable legal and factual situation.

S. 329The above reasoning of the Court seems to illustrate two facts. First, the derogation analysis described above is applied for the identification of the (economic) advantage and, second, the three-step approach, as presented, merges the examination of the (economic) advantage and that of selectivity (advantages provided only to certain undertakings). Ultimately, in the second step of the measure the ECJ applies both the derogation and the comparability examination to find selectivity in the light of the objective of the system of reference.

One could argue that in a more coherent approach the Court should proceed, first, with the identification of the existence of the (economic) advantage by identifying the reference tax system and then circumvent paragraphs 36-37, that are only a repeat of the economic advantage examination that ends to the same result. It should proceed only with the comparability analysis in the light of the objective of the tax system. This could clarify that, as soon as the advantage is a separate criterion of Article 107(1) TFEU, then the selectivity could be established only with a comparability analysis. Alternatively, the criterion of advantage would not need to be identified separately but in the context of the three-step framework, as in the Paint Graphos judgment.

It is interesting to consider the Gibraltar case in light of the above. In the Gibraltar case the identification of an economic advantage, the one that occurs from the mitigation of the “normal” tax burden, was difficult to be identified because the tax regime was deliberately designed in such a way that no “normal taxation” could be established. Under the Gibraltar tax regime, offshore companies that have neither premises nor personnel were left effectively untaxed. Since the “normal” level was unable to be determined and the criterion of advantage could not be met, the General Court rejected the Commission Decision because it failed to identify the common or reference system and thus any derogation from it, hence failing to comply with the “analytical framework” dictated by the Notice of 1998.

The ECJ, however, overturned that approach and proceeded by only examining whether offshore and typical companies “are in acomparable situation with S. 330regard to the objective of the proposed tax reform, namely to introduce ageneral system of taxation for all companies established in Gibraltar”. Even if in the Gibraltar system the evaluation of the normal level of taxation was difficult to be calculate, its objective was declared, thus permitting the ECJ to proceed only with the comparability analysis. Advocate General Jaaskinen objected to that approach in his Opinion, before the judgment by the ECJ, by commenting that not only does that approach disregards the separate condition of advantage, which cannot be established without a reference system, it also produces the side-effect of rendering it impossible to calculate the potential amount of recovery.

VI. Draft Notice and Further Challenges

In January 2014, the European Commission made public a Draft Communication Notice for the notion of State Aid. The purpose of the new Notice, as part of the Commission’s State Aid modernization program is to clarify and explain the content of Article 107 TFEU in all areas that are affected by the State aid provision, including the area of taxation. The criterion of selectivity for measures mitigating the normal charges is particularly discussed in a separate subchapter that encompasses all the latest developments in the relevant case law.

Accordingly, the approach followed by the ECJ and applied in the later case law is, although with some divergences in the conditions of Article 107(1) TFEU, the one followed in that draft Notice. Starting from the requirements of Article 107(1) TFEU, according to the draft a measure: (a) must confer an advantage on the recipient, (b) must be financed from state resources, (c) must favour certain undertakings or the production of certain goods and (d) potentially affect competition and trade within the Union. An advantage is to be considered any mitigation of charges normally included in the budget of the undertakings whose existence is based on an analysis of the financial situation (economic advantage), not the competitive situation with respect to other undertakings (competitive advantage). On selectivity, the draft reproduces the three-step formula as follows: First, the system of reference must be identified. Second, it should be determined whether a given measure constitutes a derogation from that system insofar as it differentiates between economic operators who, in light of the objectives intrinsic to the system, are in a comparable factual and legal situation. Assessing whether a S. 331derogation exists is the key element of this part of the test and allows a conclusion to be drawn as to whether the measure is prima facie selective. If the measure in question does not constitute a derogation from the reference system, it is not selective. However, if it does (and therefore is prima facie selective), it needs to be established, in the third step of the test, whether the derogatory measure is justified by the nature or the general scheme of the (reference) system.

That approach seems, again, to unify under the second step the derogation wording with the comparability analysis as already discussed above, in such a way that (as seems to happen in the Navantia judgment where that approach is executed exactly as presented) the criterion of advantage and that of selectivity seem to be confused, with the former being examined twice. However, the Notice explains specifically that a “derogation” from the system of reference is evaluated by determining whether the measure is liable to favor certain beneficiaries as compared with others that are in a similar factual and legal situation, in light of the intrinsic objective of the system of reference.

The reference to the objective of the system of reference, however, leads to another reasonable question: what will be the objective of the reference system based on which the comparison will be made? The European Commission states in the draft notice that the reference system is composed of a consistent set of rules that generally apply to all undertakings falling within its scope “as defined by its objective”. Until now, the comparison made by the ECJ sometimes refers to the “objective pursued by the measure in question”, while other times to the objective pursued by the general applicable system of the State concerned. Of course. the outcome of the selectivity test will not be the same in the two situations. If the comparison is executed in the light of the objective of a specific measure, which introduces an exemption or a subsystem of taxation, then it will be convenient for the legislator to adjust the objective of a measure in a manner that renders certain undertakings non-comparable with respect to others. For example, in the light of an environmental objective, a high-polluting industrial company is not comparable with a number of lesser or non-polluting companies. However, those two companies may be totally comparable in the light of objective of the general applicable tax system, usually the corporate income tax system of a State, that seeks to tax the companies according to their ability to pay. The European Commission notes in the Draft Notice that the comparison should be made in the light of the “intrinsic objective of the system of reference” and further states that, in general, all undertakings having an income are considered to be in a similar legal and factual situation from the perspective of direct company taxation. Because of the unclear S. 332practice of the ECJ this issue triggered noticeable attention in the relevant literature before the Notice Communication. Some commentators propose that only the objective of the system of the Member State should be taken into consideration for the purpose of the comparison, since under the measure objective’s comparison the selectivity test could easily be avoided; while others argue that a distinction should be drawn between permissible and non-permissible objectives that Member States can pursue by their measures.

Interestingly, in an early 2015 decision the ECJ ruled on the question about what objective should be used in light of which the comparison should be made. The ECJ replied that “[f]urther, it must be stated that the identification of the objective pursued is, in principle, amatter within the prerogative of the competent national public authorities alone and they must have adegree of discretion both as regards whether it is necessary, in order to achieve the regulatory objective pursued, to forgo possible revenue and also as regards how the appropriate criteria for the granting of the right, which must be determined in advance in atransparent and non-discriminatory manner, are to be identified”. The case was not in the context of Fiscal Aid. Even it is difficult to see how this view can be aligned with the field of direct taxation it could be, “in principle”, a workable context. It could balance the national prerogative of states in the field of taxation and the power to design their tax system. In addition, the European Commission should contribute clarity on what cannot be regarded as a legitimate objective and what cannot be relied upon to draw the comparison.

VII. Concluding remarks

There is a consensus among the legal scholars and various commentators that the selectivity examination is far from clear. Neither the body of case law nor the European Commission’s acts have succeeded in presenting and implementing a coherent examination. Instead, the test is blurred by subtle terminology-related confusions and conceptual incoherencies. The fact that notable judgments in recent years emerged from General Court’s judgments that were overturned by the ECJ is highly indicative of that general ambiguity. Given the significant developments in the area of Fiscal State aid and its emerging role in tackling harmful tax measures concerted efforts should be made to enhance clarity and legal certainty.