1

In early 2016, the Polish Government planned to introduce a new tax on the retail goods sector. Although various consultations had to be held on certain aspects of its implementation, the principle was of a tax whose basis of assessment would be turnover and which would be progressive in nature.

2

When the European Commission learned of that plan, it sent requests for information to the Polish authorities and, referring to the position it had taken in July 2015 on an amendment to the food chain inspection fee in Hungary, which was also based on the principle of progressive taxation of turnover, it stated:

‘The rates of progressive turnover taxes paid by undertakings are in fact linked to the size of the undertaking and not to its profitability or solvency. They cause discrimination between undertakings and may seriously disrupt the market. In so far as they introduce a difference in treatment between undertakings, they have been found to be selective. Since all the conditions set out in Article 107(1) TFEU are met, [they give rise to State aid under that article].’

3

On 6 July 2016, the Sejm Rzeczypospolitej Polskiej (lower chamber of the Parliament of the Republic of Poland) adopted the Law on the tax on the retail sector, whose essential characteristics were ultimately as follows. The industry concerned is that of the retail sale of goods to consumers who are natural persons. All retailers, regardless of their legal status, must pay the tax. The basis of assessment is monthly turnover above 17 million Polish zlotys (PLN), approximately EUR 4 million. The tax rates are 0.8% for the portion of monthly turnover between PLN 17 million and PLN 170 million and 1.4% for the portion of monthly turnover above that. The law in question entered into force on 1 September 2016.

4

After some discussion between the Polish authorities and the Commission, the latter initiated the procedure laid down in Article 108(2) TFEU in respect of the measure at issue by Decision of 19 September 2016 on the State aid SA.44351 (2016/C) (ex 2016/NN) (‘the decision to initiate the procedure, the first contested decision’). In that decision, the Commission not only requested the interested parties to submit their observations, but also ordered the Polish authorities, pursuant to Article 13(1) of Council Regulation (EU) 2015/1589 of 13 July 2015 laying down detailed rules for the application of Article 108 [TFEU] (OJ 2015 L 248, p. 9), to suspend immediately ‘the application of progressive rates to its tax, until the Commission has taken a decision on the compatibility of [the Law on the tax on the retail sector] with the internal market’.

5

Throughout the procedure, the Polish authorities — which suspended the application of the measure at issue — challenged its classification as State aid within the meaning of Article 107(1) TFEU.

6

At the same time as continuing discussions with the Commission, the Polish Government applied to the General Court to have the decision to initiate the procedure, the first contested decision, annulled (Case T‑836/16).

7

The Commission closed the procedure by adopting Decision (EU) 2018/160 of 30 June 2017 on the State aid SA.44351 (2016/C) (ex 2016/NN) implemented by Poland for the tax on the retail sector (OJ 2018 L 29, p. 38) (‘the final decision, the second contested decision’). The Commission stated in that decision that the measure at issue constituted State aid which was incompatible with the internal market and that it had been unlawfully put into effect. The Polish authorities had to cancel permanently all payments suspended pursuant to the decision to initiate the procedure, the first contested decision. Since the measure at issue had not in actual fact been implemented, the Commission considered that there was no need to recover aid from beneficiaries.

8

The Polish Government also requested the General Court to annul the final decision, the second contested decision (Case T‑624/17).

9

In essence, in the decision to initiate the procedure, the first contested decision, and in the final decision, the second contested decision, (collectively, ‘the contested decisions’) — but with its line of argument being augmented in certain regards in the final decision — the second contested decision, the Commission essentially justified the classification of the measure at issue as State aid as follows, in the light of the definition contained in Article 107(1) TFEU.

10

First, as regards the imputability of the measure at issue to the State and its financing from State resources, the Commission took the view that some of the undertakings concerned, namely those with a low turnover, were granted favourable tax treatment by the Law on the tax on the retail sector in comparison with other undertakings required to pay that tax, and that the waiver by the State of the financial resources which it would have collected if all undertakings were subject to the same average effective tax rate entailed a transfer of resources from the State to the favoured undertakings.

11

As regards the existence of an advantage, the Commission noted that, just like positive benefits, measures which mitigated the charges normally borne by the undertakings provided an advantage. In the present case, average tax rates at zero or at a lower level for undertakings with a low turnover in comparison with higher average tax rates for undertakings with a higher turnover gave the former an advantage. In the final decision, the second contested decision, the Commission added that retail structures based on the franchise principle were advantaged in comparison with integrated retail structures, since turnover was divided into a number of parts corresponding to the number of franchisees for the former but was determined on the basis of the entire undertaking’s monthly turnover for the latter.

12

As to whether the advantage identified favoured certain undertakings (selectivity criterion), the Commission stated that in respect of a tax advantage, the assessment had to be carried out in several stages. First of all, the reference tax system had to be identified, then it had to be determined whether the measure at issue constituted a derogation from that system in the sense that it differentiated between undertakings which, in light of the intrinsic objectives of the system, were in a comparable factual and legal situation, and lastly, if the answer was in the affirmative, it had to be established whether that derogation was justified by the nature or general scheme of the reference tax system. A negative answer at the second stage or, as the case may be, a positive answer at the third ruled out a selective advantage in favour of certain undertakings, whereas a positive answer at the second stage and a negative answer at the third stage, on the other hand, led to the conclusion that there was a selective advantage.

13

In the present case, the Commission first of all considered that the reference system was the turnover tax on the retail sector, including in respect of undertakings with a turnover of less than PLN 17 million, but that the progressive structure of the tax did not form part of that reference system (rates of 0% for the band of turnover that was not taxable and of 0.8% and 1.4% for the associated bands of turnover).

14

To that extent, the Commission considered, next, that the progressive structure of the tax, in so far as it entailed not only marginal tax rates but also average tax rates which differed between undertakings, constituted a derogation from the reference system which was considered to be applied with a single tax rate. In the final decision, the second contested decision, the Commission provided a specific example of the taxation of three retail undertakings, the first with a monthly turnover of PLN 10 million, the second of PLN 100 million and the third of PLN 750 million. The average tax rate for the first was calculated to be zero, that of the second 0.664% and that of the third 1.246%.

15

The Commission considered, lastly, that the derogation from the reference system entailed by the progressive structure of the tax was not justified by the nature or general scheme of the system. In the decision to initiate the procedure, the first contested decision, the Commission stated that sectoral policy objectives, such as regional policy, environmental or industrial policy, could not be taken into account in that respect. As the Polish authorities had emphasised the redistributive purpose of the progressive tax structure, justified by the fact that undertakings with higher turnovers enjoy economies of scale, better conditions of supply and tax strategies that are not available to smaller undertakings, the Commission stated that such a redistributive purpose was not consistent with a turnover tax which was only levied on undertakings relative to their volume of activity and not relative to their charges, profitability, ability to pay or facilities which, according to the Polish authorities, only large undertakings can use. For the Commission, a progressive tax levied on turnover could be justified in order to offset or deter the occurrence of certain negative effects likely to be generated by the activity concerned (negative externalities), which were more significant the larger the turnover, but such a situation had in no way been established in the present case.

16

Furthermore, the Commission stated that the measure at issue distorted or threatened to distort competition and affected trade between Member States. In that regard, it found, in particular, that the Polish retail market was open to competition, that the undertakings from other Member States participated in that market and that undertakings benefiting from the lowest tax rates therefore received operating aid. The Commission viewed the assertion by the Polish authorities that the progressive structure of the tax allowed small-scale retailers to be preserved against large format retail as evidence that those authorities were seeking to influence the structure of competition in the market.

17

The Republic of Poland submitted the application for annulment of the decision to initiate the procedure, the first contested decision, on 30 November 2016 (Case T‑836/16).

18

The Commission lodged its defence on 21 February 2017.

19

On 17 March 2017, Hungary sought leave to intervene in support of the Republic of Poland. That request was approved by decision of the President of the Ninth Chamber of the Court of 27 April 2017.

20

The Republic of Poland, Hungary and the Commission lodged a reply, a statement in intervention and a rejoinder on 11 May, 19 June and 2 August 2017 respectively.

21

The Republic of Poland submitted the application for annulment of the final decision, the second contested decision, on 13 September 2017 (Case T‑624/17).

22

The Republic of Poland and the Commission each submitted observations on the statement in intervention of Hungary in Case T‑836/16 on 20 October 2017.

23

By letter of 21 November 2017, the Republic of Poland submitted a reasoned request for a hearing to be held in Case T‑836/16.

24

The Commission lodged its defence in Case T‑624/17 on 29 November 2017.

25

On 30 November 2017, the Commission requested the joinder of Cases T‑836/16 and T‑624/17 for the oral part of the procedure.

26

On 15 December 2017, Hungary sought leave to intervene in support of the Republic of Poland in Case T‑624/17. That request was approved by decision of the President of the Ninth Chamber of the Court of 12 January 2018.

27

On 20 February 2018, Hungary submitted its statement in intervention in Case T‑624/17. The Republic of Poland and the Commission submitted their observations thereon on 9 and 19 April 2018 respectively.

28

By letter of 15 May 2018, the Republic of Poland submitted a reasoned request for a hearing to be held in Case T‑624/17.

29

On hearing the report of the Judge-Rapporteur, the Court decided to open the oral procedure in Cases T‑836/16 and T‑624/17. The Court also decided to put a question to the parties to be answered during this phase.

30

Acting upon a proposal of the Ninth Chamber, the Court decided, pursuant to Article 28 of the Rules of Procedure of the General Court, to refer the cases to a Chamber sitting in extended composition.

31

By decision of the Court of 4 July 2018, Cases T‑836/16 and T‑624/17 were joined for the purposes of the oral part of the procedure pursuant to Article 68(2) of the Rules of Procedure.

32

The parties presented oral argument and their answers to the questions put by the Court at the hearing on 26 September 2018. On that occasion, after hearing the parties, the President of the Ninth Chamber, Extended Composition, of the General Court decided that Cases T‑836/16 and T‑624/17 would likewise be joined for the purposes of the decision closing the proceedings.

33

In Case T‑836/16, the Republic of Poland claims that the Court should:

34

In Case T‑624/17, the Republic of Poland claims that the Court should:

35

In Cases T‑836/16 and T‑624/17, the Commission contends that the Court should:

36

In Cases T‑836/16 and T‑624/17, Hungary contends that the applications should be granted.

37

In Case T‑836/16, Hungary contends that the Commission should be ordered to bear the costs incurred by the Hungarian Government.

38

In Case T‑836/16, the Polish Government raises four pleas in law against the decision to initiate the procedure, the first contested decision: first of all, alleging an error in the legal characterisation of the measure at issue as State aid within the meaning of Article 107(1) TFEU; next, alleging an infringement of Article 13(1) of Regulation No 2015/1589 and of the principle of proportionality because of the injunction to suspend immediately ‘the application of progressive rates to its tax, until the Commission has taken a decision on the compatibility of [the Law on the tax on the retail sector] with the internal market’; and lastly, alleging an erroneous and inadequate statement of reasons.

39

In Case T‑624/17, the Polish Government raises two pleas in law against the final decision, the second contested decision: first, alleging an error in the legal characterisation of the measure at issue as State aid within the meaning of Article 107(1) TFEU and, secondly, alleging an erroneous and inadequate statement of reasons.

40

In the present case, the Court considers it appropriate to examine, first, the pleas in law alleging that the contested decisions erred in the legal characterisation of the measure at issue as State aid within the meaning of Article 107(1) TFEU.

41

The Polish Government submits that the Commission erred in finding that the tax on the retail sector constituted a selective measure favouring certain undertakings owing to the progressive rates applied to the basis of assessment, turnover. The Polish Government asserts that, on the contrary, it is a general, non-selective measure, or a measure which might possibly be regarded prima facie as selective, but is ultimately not selective since it is justified by the nature and general scheme of the tax system in question.

42

According to the Polish Government’s submissions in a first set of arguments, the tax on the retail sector cannot be considered inherently selective in nature as its structure, which causes it to be selective in the Commission’s view, does not derogate from the reference system of which that tax is a part, since selectivity is a component of that system. More specifically, the Polish Government argues as follows.

43

The progressive nature of the rates of the tax on the retail sector, which the Commission considers as the sign of a selective advantage favouring certain undertakings, is instead an integral part of the reference system, which is constituted by the aforementioned tax with its characteristics in respect of its tax base, the taxable persons, the taxable event and the structure of tax rates. The progressive nature of the rates cannot, therefore, be regarded as causing a derogation from the reference system. The Commission wrongly limited the reference system to the tax in question without its rate structure, leading to the odd situation where the tax reference system that it identified did not include a ‘normal’ tax rate in comparison with which it could be ascertained whether there was a selective advantage, as is apparent from recitals 26 and 51 of the decision to initiate the procedure, the first contested decision, and from recitals 47 and 49 of the final decision, the second contested decision. The Commission simply took the view that there should be only one tax rate, which could be set by the Polish authorities at the maximum marginal rate of 1.4% or at the highest average effective tax rate to which the taxable persons were subject, as the case might be.

44

However, the rates of taxation, including where there is a progressive scale, necessarily form part of any tax, as was, moreover, stated by the Commission in paragraph 134 of its Notice on the notion of State aid as referred to in Article 107(1) [TFEU] (OJ 2016 C 262, p. 1, ‘the Notice on the notion of State aid’). In its desire to impose a single taxation rate for a tax, the Commission is, moreover, encroaching on the Member States’ fiscal powers.

45

The Polish Government states that in this case the progressive scale is comprehensible and clear, and that tax rates are set at relatively low levels and in a linear manner, the top rate of 1.4% being only 1.75 times greater than the first rate of 0.8%. There is no threshold effect since, irrespective of the turnover of the undertakings concerned, all are given a tax exemption on monthly turnover of up to PLN 17 million, a rate of 0.8% being applied for the portion of monthly turnover between PLN 17 million and PLN 170 million and a rate of 1.4% applied for the portion of monthly turnover above PLN 170 million. The system is neither discriminatory nor discretionary; there is no aspect that constitutes a derogation. The Polish Government also submits that the structure of the tax on the retail sector cannot be equated with the full exemption enjoyed by offshore companies in Gibraltar, examined in the judgment of 15 November 2011, Commission and Spain v Government of Gibraltar and United Kingdom (C‑106/09 P and C‑107/09 P, EU:C:2011:732), which was contrary to the objective of the taxes concerned, which was to introduce a general taxation system for all companies; the structure in the present case is, however, similar to the capping mechanisms on those taxes at 15% of profit for all companies, which were considered in the same judgment as not entailing a selective advantage.

46

The Polish Government adds that the tax on the retail sector as it was designed meets the dual objective of raising tax revenue for the State while splitting the tax burden fairly between taxpayers according to their ability to pay, this having a redistributive aim, an aspect which itself aims to ensure receipt of tax revenue. Contrary to the Commission’s contention in recital 29 of the decision to initiate the procedure, the first contested decision, and recital 49 of the final decision, the second contested decision, the purpose of that tax is not confined to raising tax revenue or even to ‘taxing the turnover of all undertakings in the retail sector’. This confirms that the associated tax rates and tax thresholds form part of the reference system. In addition, while the way in which retail chains chose to be organised could indeed affect the level of the tax that they would have had to pay, each was free to adopt the most favourable organisational structure in that regard, including through franchising. In particular, the Carrefour Group, like other large, foreign-owned retailers, makes extensive use of franchising, while some large taxpayers with integrated structures are Polish companies.

47

The Commission begins its response to those submissions by making some preliminary comments. It notes that it considered that all retail undertakings were in a comparable factual and legal situation in the light of the objective of the tax in question and that the progressive structure of its rates led to discrimination between those undertakings according to their size, which was not justified by the purpose or nature of the tax, since undertakings with a low turnover were subject to an average effective rate of zero or a lower average effective rate than undertakings with a higher turnover. Thus, almost all small and medium-sized independent retailers were in practice exempted or taxed at an average effective rate of less than 0.8% on their total turnover, while large format retailers, such as integrated chains of hypermarkets, were subject to an average effective rate closer to the maximum rate of 1.4%, which would significantly reduce their profits. Polish-owned retail undertakings generally benefit from the system while foreign-owned undertakings are taxed at a higher average rate. The Commission notes in this respect that according to various publicly available reports, out of nearly 200000 shops or retail undertakings, only about one hundred would have been liable to the tax in September 2016, the proceeds of which would have been PLN 114 million, of which approximately PLN 80 million would have been owed by the 10 largest undertakings. Only 12 undertakings would have reached the 1.4% tax band. According to the Commission, various political declarations in Poland also clearly stated that the tax aimed to re-balance the terms on which small undertakings and international retail chains compete. Moreover, a retail chain that is organised on a franchise model pays little or no tax, while an integrated distribution chain generating the same turnover pays far more. The Commission provides the example of the Carrefour Group, which is partly organised on an integrated model and would be taxed for that part at an average rate of 1.2%, while the Polish retail chain Lewiatan, which operates on a franchise model and is itself divided into 16 companies, with a higher total turnover than Carrefour, would be taxed at an average rate of almost zero. In that regard, although foreign-owned chains, such as the Carrefour Group, also use franchising, the franchisees are precisely the local Polish undertakings who are advantaged by the tax measure in question. However, at the hearing, in respect of establishing the selectivity of the advantages generated by the progressive structure of the rates of the tax on the retail sector, the Commission emphasised that the contested decisions were not adopted on the basis that discrimination according to the national origin of the taxpayers had been identified.

48

The Commission adds, referring to the judgment of 15 November 2011, Commission and Spain v Government of Gibraltar and United Kingdom (C‑106/09 P and C‑107/09 P, EU:C:2011:732), that it is not sufficient, when determining whether a tax measure is selective to the advantage of some undertakings, to examine whether there is a derogation from the reference system’s rules as defined by the Member State concerned itself, but that it must also be ascertained whether the limits or structure of that reference system were defined consistently or, on the contrary, in a clearly arbitrary or biased manner so as to favour those undertakings, as is the case here, in the Commission’s view. The Commission states that, in that judgment, the Court held that the selective advantage enjoyed by certain companies resulted from the very design of the tax concerned. The judgment of 21 December 2016, Commission v World Duty Free Group and Others, C‑20/15 P and C‑21/15 P, EU:C:2016:981), confirms that approach.

49

In so far as the Polish Government justifies the tax measure in question by the need to take into account the ability to pay of undertakings, the Commission also contends that a tax base corresponding to turnover is irrelevant in that regard, since high turnover may be associated with loss making, and vice versa. The fact that an undertaking is large does not mean that it is capable of paying a sizeable amount in tax. The desire to combat tax optimisation and avoidance, which the Polish Government also mentions, is also misplaced, since the risk of reduction of the tax base arises only for profit taxes.

50

The Commission explains that its analysis does not undermine the fiscal autonomy of Member States. The Republic of Poland will remain sovereign, subject to compliance with the rules of the FEU Treaty relating to State aid.

51

More specifically in relation to the discussion on determining the reference system, the Commission states that, in order to establish the selectivity of an advantageous tax measure, it is necessary to identify that system, composed of a consistent set of rules that generally apply on the basis of criteria applicable to all undertakings falling within its scope as defined by its objective, and then to show that the measure at issue derogates from that system in so far as it distinguishes between undertakings that are in a comparable factual and legal situation in the light of that objective. In the present case, since the subject of the tax in question is retail turnover and the taxable persons are retailers, in the light of the objective of that tax, all retailers, irrespective of their size, are in a comparable legal and factual situation. The reference system is hence the taxation of turnover generated by retail sales.

52

However, as in the case leading to the judgment of 15 November 2011, Commission and Spain v Government of Gibraltar and United Kingdom (C‑106/09 P and C‑107/09 P, EU:C:2011:732), the reference system as presented by the Polish Government is deliberately designed to be selective, which cannot be justified by the objective of the tax, which is to raise revenue for the State. The Commission did not overlook the fact that the same rates and the same bands apply to all retail undertakings, but, despite that, local retailers will benefit from an average effective rate of zero or much lower than that paid by retailers with a high turnover. In that respect, the Commission advances the figures included in the final decision, the second contested decision, to which reference is made in paragraph 14 above. In the absence of valid justification by the Polish authorities, the reason for setting the tax bands can only be to favour small retailers and to make the largest undertakings in the sector pay.

53

The Polish Government’s argument that the progressive nature of the tax on the retail sector is justified by the dual objective of raising tax revenue for the State while sharing the tax burden fairly among taxpayers according to their ability to pay is not, in the Commission’s view, relevant to the stage of identifying the reference tax system, but is a justification to be provided, where appropriate, after a derogation from that system is identified. In any event, the intrinsic objective of the tax to be taken into account is not to generate tax revenue, which is the objective of any tax, but to tax retail turnover, in the same way as the objective of a profit tax is to tax profits. As was stated in paragraph 49 above, nor can the objective be to take into account the ability to pay of undertakings involved in retail.

54

The reference system was hence correctly determined in the contested decisions as the taxation of turnover generated from retail sales without a progressive scale, but without a specific single rate having been set, contrary to what the Hungarian Government claims.

55

The arguments summarised above must be examined.

56

Article 107(1) TFEU provides that, save as otherwise provided for in the Treaties, any aid granted by a Member 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 is, in so far as it affects trade between Member States, incompatible with the internal market.

57

It is apparent from settled case-law that the aid referred to in Article 107(1) TFEU is not limited to subsidies, given that it includes not only positive benefits, such as subsidies themselves, but also State measures which, in various forms, mitigate the charges which are normally included in the budget of an undertaking and which thus, without being subsidies in the strict sense of the word, are similar in character and have the same effects (see, to that effect, judgments of 23 February 1961, De Gezamenlijke Steenkolenmijnen in Limburg v High Authority, 30/59, EU:C:1961:2, p. 39; of 2 July 1974, Italy v Commission, 173/73, EU:C:1974:71, paragraph 33; of 15 March 1994, Banco Exterior de España, C‑387/92, EU:C:1994:100, paragraph 13; and of 15 November 2011, Commission and Spain v Government of Gibraltar and United Kingdom, C‑106/09 P and C‑107/09 P, EU:C:2011:732, paragraph 71).

58

Consequently, in tax matters, a measure by which the public authorities grant certain undertakings favourable tax treatment which, although not involving the transfer of State resources, places the recipients in a more favourable financial position than other taxpayers amounts to State aid within the meaning of Article 107(1) TFEU (see, to that effect, judgments of 15 March 1994, Banco Exterior de España, C‑387/92, EU:C:1994:100, paragraph 14; of 15 November 2011, Commission and Spain v Government of Gibraltar and United Kingdom, C‑106/09 P and C‑107/09 P, EU:C:2011:732, paragraph 72; and of 21 December 2016, Commission v World Duty Free Group and Others, C‑20/15 P and C‑21/15 P, EU:C:2016:981, paragraph 56).

59

In order to demonstrate the existence of favourable tax treatment reserved for certain undertakings, or in other words characterising the measure at issue as selective, requires assessment of whether, under a particular legal regime, that measure is such as to favour certain undertakings in comparison with others which, in the light of the objective pursued by that regime, are in a comparable factual and legal situation (see, to that effect and by analogy, judgments of 2 July 1974, Italy v Commission, 173/73, EU:C:1974:71, paragraph 33; see also judgment of 15 November 2011, Commission and Spain v Government of Gibraltar and United Kingdom, C‑106/09 P and C‑107/09 P, EU:C:2011:732, paragraph 75 and the case-law cited).

60

More specifically, according to the method of analysis upheld in the case-law, for a tax measure to be classified as ‘selective’, it is necessary to begin by identifying and examining the common or ‘normal’ tax system (see, to that effect, judgments of 21 December 2016, Commission v World Duty Free Group and Others, C‑20/15 P and C‑21/15 P, EU:C:2016:981, paragraph 57, and of 28 June 2018, Andres (faillite Heitkamp BauHolding) v Commission, C‑203/16 P, EU:C:2018:505, paragraph 88 and the case-law cited).

61

It is in relation to that tax system that it must, secondly, be assessed and, where appropriate, determined whether any advantage granted by the tax measure at issue may be selective by showing that the measure derogates from that ‘normal’ system in that it differentiates between economic operators who, in light of the objective assigned to the common or ‘normal’ tax system applicable, are in a comparable factual and legal situation (see, to that effect, judgments of 8 September 2011, Paint Graphos and Others, C‑78/08 to C‑80/08, EU:C:2011:550, paragraph 49, and of 21 December 2016, Commission v World Duty Free Group and Others, C‑20/15 P and C‑21/15 P, EU:C:2016:981, paragraph 57). On the other hand, if it is apparent that the tax advantage (in other words, the differentiation) is justified by the nature or general structure of the system of which it forms part, it cannot constitute a selective advantage (see, to that effect, judgments of 8 November 2001, Adria-Wien Pipeline and Wietersdorfer & Peggauer Zementwerke, C‑143/99, EU:C:2001:598, paragraph 42; of 15 December 2005, Unicredito Italiano, C‑148/04, EU:C:2005:774, paragraphs 51 and 52; of 6 September 2006, Portugal v Commission, C‑88/03, EU:C:2006:511, paragraph 52; of 22 December 2008, British Aggregates v Commission, C‑487/06 P, EU:C:2008:757, paragraph 83; and of 21 December 2016, Commission v World Duty Free Group and Others, C‑20/15 P and C‑21/15 P, EU:C:2016:981, paragraphs 58 and 60).

62

It is apparent from the case-law that when reference is made to the nature of the ‘normal’ system, it is the objective attributed to that system which is being referred to, whereas when the general structure of the ‘normal’ system is mentioned, reference is being made to its rules of taxation (see, to that effect, judgments of 6 September 2006, Portugal v Commission, C‑88/03, EU:C:2006:511, paragraph 81, and of 7 March 2012, British Aggregates v Commission, T‑210/02 RENV, EU:T:2012:110, paragraph 84). It must be noted that the concept of objective or nature of the ‘normal’ tax system mentioned above refers to the basic or guiding principles of that tax system and refers neither to the policies which may, as the case may be, be financed by resources which it provides (like family policy measures in the present case), nor to the aims which might be sought by establishing derogations from that tax system.

63

In the present case, the Court must consider, first, the issue of the determination of the ‘normal’ tax system against which the existence of a selective advantage must as a rule be examined.

64

The Court points out, in so far as the Commission refers in particular in the contested decisions to the judgment of 15 November 2011, Commission and Spain v Government of Gibraltar and United Kingdom (C‑106/09 P and C‑107/09 P, EU:C:2011:732), that the three taxes the subject of the cases giving rise to that judgment constituted together the general taxation scheme for all companies established in Gibraltar, whereas, in the present case, the measure described by the Commission as State aid is part of the framework of a specific sectoral tax concerning the retail sale of goods to individuals. The ‘normal’ tax system cannot, therefore, in any event, exceed that sector (see, to that effect and by analogy, judgment of 21 December 2016, Commission v Hansestadt Lübeck, C‑524/14 P, EU:C:2016:971, paragraphs 54 to 63).

65

The Polish Government correctly maintains that rates of taxation cannot be excluded from the content of a tax system, as the Commission did (see recitals 22 and 29 of the decision to initiate the procedure, the first contested decision, and recitals 46 and 49 of the final decision, the second contested decision). Whether tax is levied at a single rate or at a progressive rate, the tax rate forms part of the fundamental characteristics of a tax levy’s legal regime, just as the basis of assessment, the taxable event and the group of taxable persons do. As the Polish Government argues, the Commission itself states, in point 134 of the Notice on the notion of State aid that, ‘in the case of taxes, the reference system is based on such elements as the tax base, the taxable persons, the taxable event and the tax rates’. In the absence of the tax rate enabling the structure of the ‘normal’ system to be determined, it is indeed impossible to examine whether there is a favourable derogation to the advantage of certain undertakings (see, to that effect, judgments of 6 September 2006, Portugal v Commission, C‑88/03, EU:C:2006:511, paragraph 56, and of 7 March 2012, British Aggregates v Commission, T‑210/02 RENV, EU:T:2012:110, paragraph 52). That is why if, in the context of the same tax, certain undertakings are subject to different tax rates, including different exemptions, from other undertakings, it is necessary to determine the ‘normal’ situation relevant, which forms part of the ‘normal’ system, without whose identification the method referred to in paragraphs 60 and 61 above cannot be applied.

66

It is clear moreover from the contested decisions and the Commission’s arguments in defence that the latter sought to identify a ‘normal’ system involving a tax structure, which it could refer to. It is apparent, in particular, from recitals 26 and 32 of the decision to initiate the procedure, the first contested decision, and from recitals 47, 49 and 54 of the final decision, the second contested decision, that, for the Commission, that system has to be one in which retailers’ turnover is taxed at a single rate from the first PLN (linear). The Commission shows moreover that it regretted that the Polish authorities failed to indicate a value for that single rate to it (recital 26 of the decision to initiate the procedure, the first contested decision, and recital 47 of the final decision, the second contested decision) and it even suggested adopting the maximum rate of 1.4% or the highest average effective tax rate observed among undertakings subject to the tax (recital 51 of the decision to initiate the procedure, the first contested decision). It must, however, be stated that the ‘normal’ single-rate system referred to by the Commission in some passages of the contested decisions is a hypothetical system which could not be sustained. The analysis of whether a tax advantage is selective, which occurs at the second stage of the method referred to in paragraphs 60 and 61 above, must be carried out in the light of the actual features of the ‘normal’ tax system of which it forms part, identified during the first stage of that method, not in the light of assumptions not accepted by the competent authority.

67

Consequently, the Commission identified, in the contested decisions, a ‘normal’ system which was either incomplete, without any tax rate, or hypothetical, with a single tax rate, which constitutes an error of law.

68

Having regard to the sectoral nature of the tax at issue and the absence of differentiated scales of rates for certain undertakings, the only ‘normal’ system which could be chosen in the present case was, as the Polish Government maintains, the tax on the retail sector itself, with its structure including its scale of progressive rates and its bands, including, however, contrary to that government’s submissions, the reduction of the tax base specified for the band of turnover from PLN 0 to 17 million P; this is because that reduction de facto forms part of the structure of taxation and, although it is exempt from the tax, the corresponding activity falls within its sectoral scope of application.

69

However, even though the Commission erred in the identification of the relevant ‘normal’ tax system, it must be ascertained whether the conclusion it reached is justified by other grounds in the contested decisions which would enable the existence of a selective advantage in favour of certain undertakings to be identified.

70

The Commission did not simply consider that the progressive structure of the tax at issue derogated from a ‘normal’ system, in this case identified incompletely or hypothetically, but it also, in essence, based the existence of a selective advantage in favour of undertakings with a low level of turnover on the judgment of 15 November 2011, Commission and Spain v Government of Gibraltar and United Kingdom (C‑106/09 P and C‑107/09 P, EU:C:2011:732), which concerned a tax system in itself discriminatory in the light of the objective it was supposed to pursue, that is to say in the light of its nature. In the present case, the Commission considered that the structure of the tax on the retail sector, with its progressive rates and successive bands, was contrary to the objective pursued by that tax and produced in that regard discriminatory effects between undertakings in that sector. The Court must, therefore, examine whether that assessment is well founded.

71

Thus, in recital 23 of the decision to initiate the procedure, the first contested decision, and in recital 46 of the final decision, the second contested decision, the Commission stated that, ‘it [was] also necessary to evaluate whether the boundaries of that system have been designed by the Member State in a consistent manner or, conversely, in a clearly arbitrary or biased way, so as to favour certain undertakings over others’. Recital 47 of the final decision, the second contested decision, states that ‘[the progressive rate structure subjects] undertakings with lower turnover to a lower average effective tax rate than undertakings with a higher turnover … although both types of undertaking are engaged in the same activity’. In recitals 28 and 29 of the decision to initiate the procedure, the first contested decision, the Commission observed that ‘the stated objective of the tax [wa]s to collect revenue for the general budget’, that ‘in light of that objective, the Commission consider[ed] all retail operators to be in a comparable legal and factual situation, regardless of … their level of turnover’, that ‘it appear[ed] that Poland ha[d] deliberately designed the tax in such a manner so as to arbitrarily favour certain undertakings’ and that ‘the reference system [wa]s therefore selective by design in a way that is not justified in light of the objective of the tax’. Recital 49 of the final decision, the second contested decision, includes similar assessments, which are, however, accompanied by, as in recital 44 of the same decision, the assertion that the objective of the tax is ‘to tax the turnover of all retail operators’.

72

However, in the first place, the objective identified in recitals 28 and 29 of the decision to initiate the procedure, the first contested decision, namely to raise revenue for the general budget, is, as the Commission itself says in its defences, common to all unallocated taxes, which account for the bulk of the taxation systems, and is insufficient, in itself, to determine the nature of the various taxes, for example according to the type of taxable person concerned, whether the taxes are general or sectoral, or according to a specific objective they may pursue, for example as regards taxes seeking to reduce certain damage to the environment (ecotaxes). Moreover, the progressive structure of a tax rate cannot as such be contrary to the objective of collecting budgetary revenue.

73

In the second place, the objective identified in recitals 44 and 49 of the final decision, the second contested decision, namely to tax the turnover of all undertakings in the relevant sector, could not be sustained either. There is nothing in the file to indicate that the Polish legislature had that intention. On the contrary, both the explanatory memorandum of the Law on the tax on the retail sector (see, in that regard, the section entitled ‘Tax liability and rates’) and the observations of the Polish authorities during the administrative procedure leading to the final decision, the second contested decision (see, in that regard, recital 27 of that decision), show that the objective was to put in place a sectoral tax adhering to a principle of fiscal redistribution.

74

More specifically, it is apparent from the evidence in the file that the Law on the tax on the retail sector introduced a tax on retailers’ turnover, irrespective of their legal status, in respect of their sales of goods to individuals, coupled with a redistributive purpose. Although the tax in question was presented as allowing family policy measures to be financed, it had to raise revenue for the general budget. No other specific aim, for example seeking to offset or deter the occurrence of negative effects likely to be caused by the activity at issue, was put forward.

75

Furthermore, contrary to the Commission’s submissions, the scheme of the tax on the retail sector, characterised by a progressive tax structure, was a priori consistent with that objective, even though the tax at issue was a turnover tax. It may reasonably be presumed that an undertaking which achieves a high turnover may, because of various economies of scale, have proportionately lower costs than an undertaking with a smaller turnover — because fixed unit costs (buildings, property taxes, plant, staff costs for example) and variable unit costs (raw material supplies for example) decrease with levels of activity — and that it may, therefore, have proportionately greater disposable revenue which makes it capable of paying proportionately more in terms of turnover tax.

76

What the Polish Government submits, in essence, must therefore be confirmed: namely, that the objective of that tax was to introduce a sectoral tax with a redistributive purpose on retailers’ turnover.

77

In the present case, the Commission therefore made a further error in selecting an objective of the tax on retail trade that was different to the one put forward by the Polish authorities.

78

That second error is indeed linked to the first that the Commission made, since the objective of taxing the turnover of ‘all undertakings’ in the sector concerned that it selected in fact evidenced, in its view, the lack of a reduction in the tax base and the existence of a uniform rate of taxation, which corresponds to the hypothetical tax system which the Commission sought to identify, as shown by the identical final sentences of recital 32 of the decision to initiate the procedure, the first contested decision, and of recital 54 of the final decision, the second contested decision, reproduced below:

‘The reference system [consists in] the imposition of a single (flat) rate tax on … all undertakings involved in the retail trade in Poland.’

79

At the present stage of the analysis, the issue is whether the Commission was still able, notwithstanding the two errors identified above as regards the definition of the reference system and its objective, to discern correctly elements showing the existence of selective advantages in the tax on the retail sector taking into account the reference system and that tax’s objective mentioned in paragraphs 68 and 76 above, as resulting from the Polish legislation. More specifically, the issue is whether the Commission has shown that the tax structure chosen by the Polish authorities was contrary to that system’s objective.

80

It must be borne in mind that the EU Courts have on numerous occasions ruled on whether there are selective advantages within tax systems, or more generally compulsory contribution systems, which were characterised by rules varying those contributions according to the situation of the person liable. In that regard, the fact that a tax is characterised by a progressive tax structure, deductions, ceilings or by other variation mechanisms, and that different effective levels of taxation result therefrom depending on the size of the taxpayer’s taxable amount or the parameters of the variation mechanisms invoked, does not necessarily imply the existence of a selective advantage in favour of certain undertakings, as is apparent from the case-law referred to in paragraphs 58 to 62 above.

81

That statement may be illustrated in particular by various specific examples related to the question formulated in paragraph 79 above, which show the circumstances in which the existence of a derogation from the application of the ‘normal’ system may be identified because a measure varying the tax at issue fails to have regard to the nature of that system, that is to say its objective.

82

Consequently, such a derogation was identified in the judgments of 8 November 2001, Adria-Wien Pipeline and Wietersdorfer & Peggauer Zementwerke (C‑143/99, EU:C:2001:598, paragraphs 49 to 55); of 22 December 2008, British Aggregates v Commission (C‑487/06 P, EU:C:2008:757, paragraphs 86 and 87); of 26 April 2018, ANGED (C‑233/16, EU:C:2018:280); of 15 November 2011, Commission and Spain v Government of Gibraltar and United Kingdom (C‑106/09 P and C‑107/09 P, EU:C:2011:732, paragraphs 85 to 108); and of 21 December 2016, Commission v World Duty Free Group and Others (C‑20/15 P and C‑21/15 P, EU:C:2016:981, paragraphs 58 to 94, read in conjunction with paragraph 123 thereof), with the first judgment concerning a limitation, the following three judgments exemptions and the last of the judgments reductions in the taxable amount. The Court has thereby held, in the light of the objectives of the taxes concerned — which sought (i) to combat negative externalities, in particular environmental ones, as regards the first three judgments, (ii) the establishment of a general taxation system for all undertakings as regards the following judgment, and (iii) the amortisation, with regard to corporation tax, of the goodwill resulting from the acquisition of company assets in certain circumstances as regards the last of the judgments — that the advantages which were reserved to some of the undertakings, but not others, in a similar situation in relation to those objectives, were, therefore, selective.

83

It is apparent from those judgments that, regardless of whether the objective of the tax includes a purpose linked to the impact of the activity of the undertakings liable to tax, or the advantage concerns a specific economic sector in relation to the other undertakings subject to tax or a specific form of operating companies, or even whether the advantage is potentially open to any undertaking subject to tax, if that advantage leads to differences in treatment which are contrary to the objective of the tax, it is selective. However, the objective of a tax may itself include a variation seeking to apportion the tax effort or limit its impact. Specific situations which distinguish certain taxable persons from others may also be taken into account without the tax’s objective being disregarded.

84

In that respect, in the judgment of 8 November 2001, Adria-Wien Pipeline and Wietersdorfer & Peggauer Zementwerke (C‑143/99, EU:C:2001:598, paragraphs 33 to 36), referred to in paragraph 82 above, the Court stated that the partial rebate of taxes on the energy consumed by undertakings, applicable when those taxes exceeded a certain threshold of the net value of what those enterprises produced, did not constitute State aid if it benefited all undertakings subject to those taxes regardless of their activity, while it could lead to different levels of taxation between undertakings consuming the same amount of energy.

85

Similarly, in the judgment of 15 November 2011, Commission and Spain v Government of Gibraltar and United Kingdom (C‑106/09 P and C‑107/09 P, EU:C:2011:732, paragraphs 77 to 83), referred to in paragraph 82 above, the Court held that the advantages which could arise from a generalised capping of two taxes on undertakings, not based on profit, to 15% of profit, leading to undertakings with the same tax base potentially paying different tax, were established on the basis of objective criteria irrespective of the choices of the undertakings concerned and were not, therefore, selective.

86

In the judgment of 8 September 2011, Paint Graphos and Others (C‑78/08 to C‑80/08, EU:C:2011:550, paragraphs 48 to 62), the Court ruled that, in the context of the tax on company profits, which constituted the ‘normal’ system in that case, the total exemption enjoyed by cooperative societies did not constitute a selective advantage because they were not in a comparable factual and legal situation to that of commercial companies, provided that it was verified that they did indeed act under the conditions inherent to being a cooperative, implying, in particular, a profit margin considerably lower than that of capital companies.

87

In the judgment of 29 March 2012, 3M Italia (C‑417/10, EU:C:2012:184, paragraphs 37 to 44), the Court held, taking into account also the specific situation of certain undertakings, that a mechanism for concluding at a standard rate old tax proceedings, available to undertakings meeting objective criteria not placing them in a factual and legal situation comparable to that of other undertakings, did not entail a selective advantage, even if it could lead to the beneficiaries of that mechanism paying less tax, all other things being equal elsewhere, than other undertakings.

88

Similarly, in the judgment of 26 April 2018, ANGED (C‑233/16, EU:C:2018:280), referred to in paragraph 82 above, the Court stated that, in the context of a tax on retail establishments, whose basis of assessment was essentially constituted by the sales area and which sought to offset negative externalities for the environment and town and country planning, the 60% reduction or total exemption enjoyed by establishments carrying on certain activities and those whose sales area was below a given threshold did not constitute State aid if it was verified that those various establishments were indeed in a different situation from those of the other establishments subject to tax, having regard to the impacts which the tax at issue sought to correct and offset, that is to say in the light of the objectives of that tax.

89

Those examples confirm that there are taxes whose nature does not preclude them from being accompanied by variation mechanisms, which may extend as far as exemptions, without those mechanisms leading, however, to selective advantages being granted. In short, there is no selectivity if those differences in taxation and the advantages which may flow therefrom, even if justified only by the purpose governing the apportionment of tax between taxpayers, stem from the straightforward application, without derogation, of the ‘normal’ system, if comparable situations are treated comparably and if those variation mechanisms do not misconstrue the objective of the tax concerned. Similarly, special provisions laid down for certain undertakings by reason of situations specific to them, causing them to benefit from a variation in, or even an exemption from, tax, must not be analysed as constituting a selective advantage if those provisions do not contravene the objective of the tax in question. In that regard, the fact that only taxpayers meeting the conditions for the application of a measure can benefit from the measure cannot, in itself, make it into a selective measure (see judgment of 21 December 2016, Commission v World Duty Free Group and Others, C‑20/15 P and C‑21/15 P, EU:C:2016:981, paragraph 59 and the case-law cited). Such mechanisms fulfil the condition of complying with the nature and general structure of the system of which they form part, referred to in paragraph 61 above.

90

On the other hand, if undertakings in a comparable situation in the light of the objective of the tax or the purpose justifying a variation thereof are not treated equally in that regard, that discrimination gives rise to a selective advantage which may constitute State aid if the other conditions laid down in Article 107(1) TFEU are met.

91

Accordingly, in particular, progressive tax structures, including significant reductions to the basis of assessment, which are not exceptional in the Member States’ tax systems, do not in themselves imply the existence of State aid. In its Notice on the notion of State aid, the Commission states in that regard, in point 139, that the progressive nature of income tax may be justified by its redistributive purpose. However, there is no basis for limiting that type of assessment, as the Commission did in recitals 58 and 59 of the final decision, the second contested decision, to taxes on income and to exclude that assessment for taxes applying to the undertakings’ activity, not their net revenue or profit. It is not apparent from the case-law referred to in paragraphs 58 to 62 above that, in order for a measure varying a tax not to be characterised as a selective advantage, a Member State could have recourse only to variation criteria limited to certain aims, such as the redistribution of wealth or the offsetting or deterrence of negative externalities that may be caused by the activity concerned. What is necessary to that end is that the intended variation must not be arbitrary, contrary to what occurred in the case giving rise to the judgment of 22 December 2008, British Aggregates v Commission (C‑487/06 P, EU:C:2008:757), mentioned in paragraph 82 above, that it must be applied in a non-discriminatory manner and must remain consistent with the objective of the tax concerned. For example, the variation mechanisms referred to in paragraphs 84, 85 and 87 above, which were not held to be selective by the Court, did not reflect a taxation purpose proportional to negative externalities, nor indeed a redistributive purpose, but other aims. In addition, as pointed out in paragraph 75 above, it cannot be excluded that a redistributive purpose may also justify the progressivity of a turnover tax, as the Polish Government rightly maintains in the present case. A redistributive purpose may indeed even justify a total exemption for some undertakings as shown by the case mentioned in paragraph 86 above.

92

Consequently, as regards a turnover tax, a variation criterion taking the form of progressive taxation above a certain threshold — even if that threshold is a high one — which may reflect the wish to tax an undertaking’s activity only when that activity reaches a certain level, does not in itself imply the existence of a selective advantage.

93

It follows, therefore, from paragraphs 79 to 92 above that the Commission was not entitled to infer the existence of selective advantages accompanying the tax on the retail sector solely from the progressive structure of that new tax.

94

However, if it were proven by the Commission in the contested decisions that the progressive taxation structure actually chosen was adopted in a manner which largely deprives the objective of the tax in question of its substance, it could be considered that the advantage which may be derived by undertakings benefiting from zero or low taxation compared with other undertakings is selective.

95

It must, therefore, be further ascertained whether the Commission provided such proof in the contested decisions.

96

However, it must be concluded that, in the contested decisions, the Commission merely considered that it was the principle of progressive taxation itself which gave rise to a selective advantage (recitals 32 and 37 of the decision to initiate the procedure, the first contested decision, recitals 47, 49 and 54 of the final decision, the second contested decision), which, in the light of what is stated in paragraph 92 above, constitutes an error of law.

97

It is only in recital 51 of the final decision, the second contested decision, that the Commission advanced evidence that might be capable of supporting the proposition that the progressive structure chosen in this case for the tax on the retail sector was incompatible with its objective as stated in paragraph 76 above. In that recital, the Commission stated that it had essentially inferred from various publicly available reports that in September 2016, only 109 out of 200000 undertakings operating in the retail sector would have passed the threshold of monthly turnover of PLN 17 million, which is around EUR 4 million, above which turnover was taxed.

98

However, that fact alone, which, as the main parties confirmed at the hearing, was not discussed with the Polish authorities during the administrative procedure, was not combined with reasoning other than that directed at the very principle of progressive taxation and is therefore in any event insufficient to constitute reasoning capable of establishing that the progressive structure chosen in the present case for the tax on the retail sector was incompatible with its objective.

99

In addition, the Commission did indeed state in the contested decisions that the progressive taxation structure of the tax on the retail sector led to undertakings in a comparable factual and legal situation being treated differently, in other words that it led to discriminatory treatment. However, despite giving specific examples, it relied principally in that regard only on the fact that the undertakings’ average effective rate and the marginal rate of tax had to vary according to their turnover (recitals 24, 25, 27, 28, 32 and 37 of the decision to initiate the procedure, the first contested decision, and recitals 47, 49, 53 and 54 of the final decision, the second contested decision). That variation in the average effective rate and marginal rate according to the size of the taxable amount is an integral part of any taxation system with a progressive structure and such a system is not, as set out in paragraph 92 above, as such and by virtue of that fact alone, such as to give rise to selective advantages. Moreover, when a tax’s progressive taxation structure reflects the objective pursued by that tax, it cannot be considered that two undertakings with a different taxable amount are in a comparable factual situation in the light of that objective.

100

In the contested decisions, the Commission also referred to the circumstances that, de facto, the tax on the retail sector would place a greater burden on foreign-owned undertakings than Polish-owned undertakings and would place a greater burden on retail networks organised according to an integrated model than retail networks which make extensive use of franchising.

101

In respect of the first of those circumstances, disputed by the Polish Government, it suffices to note, as indicated in paragraph 47 above, that the Commission itself noted at the hearing with regard to proving the selective nature of the advantages entailed by the rate structure of that tax, that the contested decisions had not been taken on the basis that discrimination on the ground of the national origin of the taxable persons had been identified. Furthermore, it must be pointed out that, even if they are proven, the circumstances described in paragraph 100 above are merely the corollary of the application of a progressive tax structure corresponding to the objective and general scheme of the tax in question, and if the various undertakings liable to fall under the scope of the tax are free to choose how they are organised, those circumstances cannot also lead to the conclusion that comparable factual and legal situations are treated differently, or the reverse. Furthermore, as argued by the Polish Government in its applications without being challenged by the Commission, franchising is practised in Poland by both foreign-owned retail chains and Polish retail chains. Moreover, the situation of a franchised shop is different from that of an integrated shop. The first is, as a rule, legally and financially independent from its franchiser, which is not the case for an integrated shop in respect of the undertaking which controls it, whether it is a subsidiary or a branch of a retail network.

102

Consequently, the Commission failed to establish in the contested decisions the existence of a selective advantage which differentiated between economic operators who, in light of the objective attributed by the Polish legislature to the tax on the retail sector, were in a comparable factual and legal situation. The errors the Commission made with regard to defining the ‘normal’ tax system, with regard to its objective and with regard to the inherent existence, in its view, of selective advantages in a structure of progressive taxation on turnover, did not allow it to ascertain whether the progressive structure actually selected led, in the light of the objective of the tax in question, to different treatment of undertakings in a comparable factual and legal situation, for instance to ascertain properly whether or not the sectoral tax in question did in fact affect a greatly inadequate share of the activity that it was supposed to apprehend, thus entailing a selective advantage for the undertakings which were not involved in that share but were still active to a significant extent in that area.

103

The final decision, the second contested decision, must therefore be annulled on the basis of the plea in law concerning the existence of an error in the legal characterisation of the measure at issue as State aid within the meaning of Article 107(1) TFEU, and there is no need to examine the Polish Government’s other pleas and submissions.

104

With regard to the same plea in law, in respect of the decision to initiate the procedure, the first contested decision, it must be pointed out that the Court has consistently held that when the Commission examines aid measures under Article 107 TFEU in order to determine whether they are compatible with the internal market, it is required to initiate the procedure under Article 108(2) TFEU where, after the preliminary examination, it has been unable to overcome all the difficulties involved in determining whether those measures are compatible with the internal market. The same principles apply where the Commission also entertains doubts as to the actual classification as aid, within the meaning of Article 107(1) TFEU, of the measure examined. The Commission cannot, therefore, in principle be criticised for having initiated that procedure on the basis of, in particular, doubts as to whether the measures which are the subject of that procedure can be qualified as aid in the sense referred to above (see, to that effect, judgment of 10 May 2005, Italy v Commission, C‑400/99, EU:C:2005:275, paragraph 47).

105

However, in view of the consequences of initiating the procedure provided for in Article 108(2) TFEU with regard to measures treated as new aid subject to the Commission’s preliminary authorisation pursuant to Article 108(3) TFEU (‘new aid’), where the Member State concerned contends during the preliminary examination that those measures do not constitute aid within the meaning of Article 107(1) TFEU, the Commission must, before initiating that procedure, undertake a sufficient examination of the question on the basis of the information notified to it at that stage, even if the outcome of that examination is not definitive (see, to that effect, judgment of 10 May 2005, Italy v Commission, C‑400/99, EU:C:2005:275, paragraph 48). That initiation of the procedure usually entails, notably, the suspension of the measures examined, particularly when, as in this case, the Commission orders the Member State to suspend those measures on the basis of Article 13(1) of Regulation 2015/1589.

106

In that regard, if the provisional classification by the Commission as new aid results from factual or economic uncertainties as to the nature, the content or the effects of the measure at issue or its context, and even if it is ultimately apparent that that classification was erroneous in the light of new evidence which was subsequently produced, the decision to initiate the procedure is still justified having regard to the legitimate doubts entertained by the Commission when that decision was adopted (see, to that effect, judgment of 10 May 2005, Italy v Commission, C‑400/99, EU:C:2005:275, paragraphs 48 and 49). In that regard, it has been held that review by the General Court of a decision to initiate the formal investigation procedure had necessarily to be limited, and that when the applicants challenged the Commission’s assessment of whether a contested measure constitutes State aid, review by the EU judicature was confined to ascertaining whether the Commission did not commit a manifest error of assessment in considering that it could not overcome all the difficulties on that point during its initial examination of the measure concerned (judgments of 21 July 2011, Alcoa Trasformazioni v Commission, C‑194/09 P, EU:C:2011:497, paragraph 61, and of 9 September 2014, Hansestadt Lübeck v Commission, T‑461/12, EU:T:2014:758, paragraph 42).

107

Thus, if, in view of the evidence that was already available to the Commission at the time when the procedure was initiated, it appeared that the classification of the measure at issue as new aid clearly had to be dismissed at that stage, the decision to initiate the procedure in respect of that measure must be annulled (see, to that effect, judgment of 10 May 2005, Italy v Commission, C‑400/99, EU:C:2005:275, paragraph 48).

108

The same applies in the present situation, where the Commission essentially based its provisional classification as new aid on an analysis of information in its possession which appears manifestly incorrect. The decision to initiate the procedure, the first contested decision, was not justified, in respect of the issue of whether new aid existed, by legitimate doubts in the light of the evidence in the file, but by a view supported by legal reasoning which did not allow that decision to be legally justified, as is apparent from paragraphs 63 to 102 above. That the Commission held its view as a matter of principle, according to which view, in particular, progressive taxation applied to a turnover tax inherently gives rise to selective advantages is, indeed, confirmed by the fact that its reasoning appears not to differ substantially between that decision and the final decision, the second contested decision.

109

The decision to initiate the procedure, the first contested decision, must therefore also be annulled, including the suspension injunction concerning the ‘application of the progressive rate for the tax’ contained therein, because such an injunction assumes that the State measure that it targets was correctly classified as unlawful new aid in a provisional analysis conducted under the conditions set out in paragraphs 104 to 108 above, as is apparent from Article 13(1) of Regulation No 2015/1589, which provides that ‘the Commission may, after giving the Member State concerned the opportunity to submit its comments, adopt a decision requiring the Member State to suspend any unlawful aid until the Commission has taken a decision on the compatibility of the aid with the internal market ’. That provision covers only unlawful new aid, as referred to in Article 1(f) of the same regulation, that is to say measures which must, in particular, be capable of corresponding, according to the provisional analysis mentioned above, to the definition of State aid provided for in Article 107(1) TFEU (see, to that effect, judgment of 25 April 2018, Hungary v Commission, T‑554/15 and T‑555/15, under appeal, EU:T:2018:220, paragraphs 30, 153 and 154). Thus, in the present case, the fate of the suspension injunction is not severable from the fate of the decision to initiate the procedure and must be annulled, without it being necessary to examine whether, considered in isolation, the plea put forward by the Polish Government alleging infringement of Article 13(1) of Regulation No 2015/1589 is or is not well-founded in view of the submissions supporting it.

110

In the light of the foregoing, there is likewise no need to examine the other pleas and submissions advanced by the Polish Government against the decision to initiate the procedure, the first contested decision.

111

It follows from all the considerations set out above that the two actions for annulment brought by the Republic of Poland must be upheld.

112

Under Article 134(1) of the Rules of Procedure, the unsuccessful party is to be ordered to pay the costs if they have been applied for in the successful party’s pleadings. Since the Commission has been unsuccessful, it must be ordered to pay the costs incurred by the Republic of Poland, in accordance with the form of order sought by the latter.

113

Under Article 138(1) of the Rules of Procedure, the Member States which have intervened in the proceedings are to bear their own costs. Hungary must, therefore, bear its own costs.

On those grounds,

THE GENERAL COURT (Ninth Chamber, Extended Composition)

hereby: