CJEU, 9th chamber, extended composition, December 13, 2018, No T-851/14

THE GENERAL COURT (Ninth Chamber, Extended Composition),

I. Background to the dispute

1 The applicant, Slovak Telekom, a.s., is the incumbent telecommunications operator in Slovakia. Deutsche Telekom AG, the incumbent telecommunications operator in Germany and the company at the helm of the Deutsche Telekom group, acquired a 51% stake in the applicant on 4 August 2000, a shareholding which it held throughout the relevant period in this case. The remaining shareholding in the applicant was held by the Ministry of Economy of the Slovak Republic (34%) and the National Property Fund of the Slovak Republic (15%).

2 On 15 October 2014, the European Commission adopted Decision C(2014) 7465 final relating to proceedings under Article 102 TFEU and Article 54 of the EEA Agreement (Case AT.39523 - Slovak Telekom), rectified by its Decision C(2014) 10119 final of 16 December 2014 and by its Decision C(2015) 2484 final of 17 April 2015, which was addressed to the applicant as well as to Deutsche Telekom ('the contested decision'). On 24 December 2014, Deutsche Telekom brought an action also seeking annulment of the contested decision (Case T 827/14).

A. Technological, factual and regulatory context of the contested decision

3 The applicant, which is the indirect successor of the public post and telecommunications undertaking that ceased to exist in 1992, is the largest telecommunications operator and broadband provider in Slovakia. The legal monopoly it enjoyed on the Slovakian telecommunications market came to an end in 2000. The applicant offers a full range of data and voice services, and owns and operates fixed copper and fibre optic networks as well as a mobile telecommunications network. The copper and mobile networks cover almost the entire territory of Slovakia.

4 The contested decision concerns anticompetitive practices on the Slovakian market for broadband internet services. In essence, it relates to the conditions set by the applicant for unbundled access of other operators to the copper local loop in Slovakia between 2005 and 2010.

5 The local loop is the physical twisted metallic pair circuit (also known as 'the line') connecting the network termination point at the subscriber's premises to the main distribution frame or equivalent facility in the fixed public telephone network.

6 Unbundled access to the local loop allows new entrants - usually called 'alternative operators', as opposed to the incumbent operators of the telecommunications networks - to use the pre-existing telecommunications infrastructure belonging to those incumbent operators in order to offer various services to end users, in competition with the incumbent operators. The different telecommunications services that can be provided to end users through the local loop include high bit-rate data transmission services for fixed internet access and for multimedia applications based on digital subscriber line (DSL) technology.

7 Local loop unbundling was organised at EU level by, inter alia, Regulation (EC) No 2887/2000 of the European Parliament and of the Council of 18 December 2000 on unbundled access to the local loop (OJ 2000 L 336, p. 4), and Directive 2002/21/EC of the European Parliament and of the Council of 7 March 2002 on a common regulatory framework for electronic communications networks and services (OJ 2002 L 108, p. 33). Regulation No 2887/2000 required operators holding 'significant market power' to give access to unbundled local loops (ULLs) and to publish a reference unbundling offer. Those provisions were implemented in Slovakia by the Zákon z 3. decembra 2003 c. 610/2003 Z.z. o elektronických komunikáciách, v znení neskorších predpisov (Law No 610/2003 of 3 December 2003 on electronic communications), as amended, which entered into force, with certain exceptions, on 1 January 2004.

8 In essence, that regulatory framework required the operator identified by the national regulatory authority as the operator with significant market power, which is generally the incumbent operator, to grant alternative operators unbundled access to its local loop and to related services under transparent, fair and non-discriminatory conditions, and to maintain an updated reference offer for such unbundled access. The national regulatory authority was required to ensure that charging for unbundled access to the local loop, set on the basis of cost-orientation, fostered fair and sustainable competition. To that end, the national regulatory authority was entitled inter alia to require changes to be made to the reference offer.

9 Following a market analysis, on 8 March 2005 the Slovakian national regulatory authority for telecommunications ('TUSR') adopted a first-instance decision - Decision No 205/14/2005 - designating the applicant as an operator with significant power on the wholesale market for unbundled access to the local loop within the meaning of Regulation No 2887/2000. Consequently, TUSR imposed a number of obligations on the applicant, including requiring it to submit a reference offer within 60 days. That decision, which the applicant challenged, was confirmed by the Chairman of TUSR on 14 June 2005. Pursuant to that confirmatory decision, the applicant was required to grant all reasonable and justified requests for unbundling of its local loop in order to enable alternative operators to use that loop with a view, on that basis, to offer their own services on the 'retail mass market' for broadband services at a fixed location in Slovakia. The decision of 14 June 2005 also ordered the applicant to publish all intended changes to the reference unbundling offer at least 45 days in advance and to submit them to TUSR.

10 The applicant published its reference unbundling offer on 12 August 2005 ('the reference offer'). That offer, which was amended on nine occasions between that date and the end of 2010, sets out the contractual and technical conditions for access to the applicant's local loop. On the wholesale market, the applicant offers access to unbundled local loops in or next to a main distribution frame on which the alternative operator seeking access has rolled out its own backbone network.

11 According to the contested decision, the applicant's local loop network, which could be used to supply broadband services after the lines concerned have been unbundled from that operator, covered 75.7% of all Slovakian households between 2005 and 2010. That coverage extended to all local loops in the applicant's metallic access network that could be used to transmit a broadband signal. However, during that same period, only very few of the applicant's local loops were unbundled, as from 18 December 2009, and used only by a single alternative operator to provide retail broadband services to undertakings.

B. Procedure before the Commission

12 The Commission opened an investigation on its own initiative, inter alia, the conditions for unbundled access to the applicant's local loop. Following requests for information sent to alternative operators on 13 June 2008 and an unannounced inspection at the applicant's premises which took place on 13 and 15 January 2009, the Commission decided, on 8 April 2009, to open a procedure against that operator, within the meaning of Article 2 of Commission Regulation (EC) No 773/2004 of 7 April 2004 relating to the conduct of proceedings by the Commission pursuant to Articles [101] and [102 TFEU] (OJ 2004 L 123, p. 18).

13 Further steps were taken in the investigation consisting of additional requests for information sent to alternative operators and TUSR, as well as an announced inspection at the applicant's premises on 13 and 14 July 2009.

14 In several discussion documents sent to the Commission between 11 August 2009 and 31 August 2010, the applicant argued that there were no grounds for finding that it had infringed Article 102 TFEU in the present case.

15 In the course of the investigation, the applicant objected to the provision of information dating from the period prior to 1 May 2004, the date of accession of the Slovak Republic to the European Union. It brought an action for annulment, first, of Commission Decision C(2009) 6840 of 3 September 2009 relating to a proceeding pursuant to Articles 18(3) and 24(1) of Council Regulation (EC) No 1/2003 of 16 December 2002 on the implementation of the rules on competition laid down in Articles [101 and 102 TFEU] (OJ 2003 L 1, p. 1) and, second, of Commission Decision C(2010) 902 of 8 February 2010 relating to a proceeding pursuant to Articles 18(3) and 24(1) of Regulation No 1/2003. By judgment of 22 March 2012, Slovak Telekom v Commission (T 458/09 and T 171/10, EU:T:2012:145), the Court dismissed the actions brought against those decisions.

16 Following requests for information sent to Deutsche Telekom, the Commission decided on 13 December 2010 to open a procedure against that company pursuant to Article 2 of Regulation No 773/2004.

17 On 7 May 2012, the Commission sent a statement of objections to the applicant. That statement of objections was sent to Deutsche Telekom the day after. In that statement of objections the Commission concluded, on a preliminary basis, that the applicant might have infringed Article 102 TFEU on account of a practice resulting in the margin squeeze as regards unbundled access to local loops in its network as well as national and regional wholesale broadband access to its competitors, and may have refused alternative operators access to some wholesale products. It also set out the preliminary view that Deutsche Telekom might be liable for the infringement in its capacity as parent company of the applicant during the infringement period.

18 After accessing the investigation file, the applicant and Deutsche Telekom each replied to the statement of objections on 5 September 2012. A hearing was held on 6 and 7 November 2012.

19 On 21 June 2013, the applicant sent the Commission a description of possible commitments intended to address its objections from the standpoint of competition law and asked the Commission to adopt a commitments decision within the meaning of Article 9 of Regulation No 1/2003 instead of a prohibition decision. The Commission nonetheless considered those commitments to be insufficient and therefore decided to continue the procedure.

20 On 6 December 2013 and 10 January 2014 respectively, the Commission sent the applicant and Deutsche Telekom a letter of facts intended to afford them the opportunity to comment on the additional evidence collected after the statement of objections had been issued. The Commission stated that that evidence, to which the applicant and Deutsche Telekom were given access, could be used in a possible final decision.

21 The applicant and Deutsche Telekom replied to the letter of facts on 21 February and 6 March 2014 respectively.

22 At meetings held with the applicant on 16 September 2014 and with Deutsche Telekom on 29 September 2014, the Commission provided information on the decision it planned to adopt under Article 7 of Regulation No 1/2003.

C. Contested decision

23 In the contested decision, the Commission finds that the undertaking comprising the applicant and Deutsche Telekom committed a single and continuous infringement of Article 102 TFEU and Article 54 of the EEA Agreement concerning broadband services in Slovakia between 12 August 2005 and 31 December 2010 ('the period under consideration').

1. Definition of the relevant markets and dominant position of the applicant on those markets

24 In the contested decision the Commission identifies two relevant product markets, namely:

- the retail mass market for broadband services at a fixed location;

- the wholesale market for access to unbundled local loops.

25 According to the contested decision, the relevant geographical market is the entire territory of the Republic of Slovakia.

26 The Commission states that, during the period under consideration, the applicant held a monopoly position on the wholesale market for unbundled access to local loops and there were no direct constraints in the form of actual or potential competition or countervailing buying power limiting its market power. The applicant therefore held a dominant position on that market during the period under consideration. The Commission also finds that the applicant held a dominant position during that period on the retail mass market for broadband services at a fixed location.

2. The applicant's conduct

(a) Refusal to provide unbundled access to the applicant's local loops

27 In the first part of its analysis entitled 'Refusal to supply', the Commission observes that, although several alternative operators had expressed great interest in being granted access to the applicant's local loops in order to compete with it on the retail market for broadband services, that operator set unfair terms and conditions in its reference offer rendering such access unacceptable. The applicant thereby delayed, complicated or prevented entry on the retail broadband market.

28 In that regard, the Commission states that, first, unbundled access to the local loop by an alternative operator is based on the premiss that that operator must first obtain sufficient and adequate information concerning the incumbent operator's network. That information must enable the alternative operator concerned to assess its business opportunities and to prepare appropriate business plans for its future retail services based on unbundled access to the local loop. In the present case, the reference offer did not satisfy that information requirement with respect to alternative operators.

29 Accordingly, despite the requirements laid down in the relevant regulatory framework (see paragraphs 7 and 8 above), the reference offer does not provide any basic information regarding the locations of physical access sites and the availability of local loops in specific parts of the network. Alternative operators had access to such information only upon request, in exchange for payment of a fee, within five days of the entry into force of a confidentiality agreement with the applicant and solely after the provision of a bank guarantee. The Commission considers, in essence, that those requirements unreasonably delayed and complicated disclosure of the relevant information to alternative operators and thus discouraged those operators from accessing the applicant's local loops.

30 Even in the case of access upon request, the Commission finds that the information provided by the applicant was insufficient. In particular, the applicant did not provide any information on the availability of its local loops, even though such information was crucial to enable alternative operators to prepare their business models in time and to identify the business potential of unbundling. The Commission takes the view that the applicant should have disclosed not only the list of main distribution frames and similar facilities, but also a description of their geographical coverage; information on the ranges of telephone numbers served by those exchanges; information on the actual use of cables (in %) for DSL technologies; information on the ratio of pulse code modulation (PCM) equipment deployment regarding the cables connected to the different main distribution frames; the names or functions of the distribution frames and information on how they are used in the applicant's technical and methodological regulations; and the maximum lengths of homogeneous local loops. The applicant was moreover well aware of the problems faced by alternative operators as a result of those terms of access to information and the information's limited scope. The Commission also points out that, although the applicant did not publish a template for the unbundling requests to be submitted by alternative operators until May 2009, its reference unbundling offer made provision at the very outset for the imposition of financial penalties if a request for access were deemed incomplete.

31 Secondly, according to the contested decision, the applicant unjustifiably reduced the scope of its obligation relating to unbundled access to its local loops.

32 Thus, in the first place, the applicant improperly excluded from that obligation 'passive' lines, namely lines which existed physically but were not in use. By doing so, the applicant reserved for itself a significant number of potential customers who were not yet purchasing its broadband services but to whom its network was available, even though the market was growing and the relevant regulatory framework did not restrict the unbundling obligation to active lines only. The restriction applied by the applicant was not, in the Commission's view, justified by any objective technical reasons.

33 In the second place, the applicant unjustifiably excluded from its unbundling obligation the services which it classed as 'conflicting services', namely services that it was likely to offer and which may have been in conflict with access to the local loop by an alternative operator. In addition to the fact that the actual concept of conflicting services is vague, the list of such services - drawn up unilaterally by the applicant - is open and, in consequence, creates uncertainty for alternative operators. That limitation deprived alternative operators of a large number of potential customers, customers which were reserved for the applicant and therefore withdrawn from the retail market.

34 In the third place, the Commission classifies as unjustified the rule imposed by the applicant in the reference offer, whereby only 25% of local loops contained in a multi-pair cable could be used for the provision of broadband services, in order to avoid cross-talk and interferences. That rule is not justified because it is of a general and abstract nature and does thus not take account of the characteristics of the cables and the specific combination of transmission techniques. The Commission points out, in that regard, that the practice followed in other Member States demonstrates that there are alternatives to such upstream abstract limitations on access, for instance the principle of 100% cable fill-in together with the a posteriori resolution of any specific problems stemming from spectrum interferences. Finally, the applicant applied to itself a maximum fill-in rule of 63%, which is less strict than the rule it applied to alternative operators.

35 Finally, thirdly, the applicant established a number of unfair terms and conditions in the reference offer concerning unbundled access to its local loops.

36 In that regard, in the first place, according to the contested decision, the applicant included unfair terms and conditions in the reference offer relating to collocation, defined in that offer as 'the provision of the physical space and the technical equipment necessary for the appropriate placement of the telecommunication equipment of the Authorised Provider with the purpose of provision of services to the end users of the Authorised Provider via access to the local loop'. The barrier thus erected for alternative operators was the result of, in particular, the following factors: (i) the conditions required a preliminary inquiry into the possibilities of collocation which was not objectively necessary; (ii) alternative operators were only able to challenge the determination of the form of collocation decided by the applicant by paying an additional fee; (iii) the consequence of the expiry of the reservation period after delivery to the alternative operator of the notice regarding the outcome of the preliminary or detailed inquiry, without any collocation agreement being concluded, was that the preliminary or detailed inquiry had to be repeated in full; (iv) the applicant was not bound by any deadlines in the event of additional detailed inquiries triggered by negotiations and was entitled to withdraw - without stating reasons and without any legal consequences - a proposed collocation agreement during the term for acceptance of the proposal by the alternative operators within the established deadlines; (v) the applicant was not bound by any precise time limit for implementing collocation; (vi) the applicant unilaterally imposed unfair and non-transparent fees for collocation.

37 In the second place, the Commission finds that, under the reference offer, alternative operators were required to submit forecasts of the requests for qualification of the local loop 12 months in advance for each collocation space, on a month-by-month basis, before being able to submit a request for qualification for access to the relevant local loop. The Commission considers that such a requirement obliges alternative operators to submit forecasts at a time when they are not in a position to estimate their needs in terms of unbundled access. It also criticises the fact that failure to comply with the forecasting terms triggered the payment of penalties and complains about the mandatory nature of the forecasting obligation and the lack of any deadline for the applicant to respond to a request for qualification in the event that such a request did not comply with the forecasted volume.

38 In the third place, the Commission considers that the mandatory qualification procedure, which was designed to enable alternative operators to determine whether a specific local loop was suitable for xDSL technology or any other broadband technology they might intend to use before placing a firm unbundling order, was such that those operators were deterred from requesting unbundled access to the applicant's local loops. Thus, while conceding that it is necessary to verify the suitability of local loops for unbundling or the basic preconditions for unbundling a specific line, the Commission states that the splitting of that qualification process from the very request for access to the local loop unnecessarily delayed unbundling and generated additional costs for alternative operators. Furthermore, a number of aspects examined in the context of the qualification process are superfluous. The Commission considers to be unjustified the validity period limited to 10 days applying to the positive qualification of a local loop, after which a request for access could no longer be made.

39 In the fourth place, according to the contested decision, the reference offer included disadvantageous terms as regards repairs, service and maintenance, due to (i) the lack of an appropriate definition of 'planned works' and 'unplanned works'; (ii) the unclear distinction between 'unplanned works' and straightforward 'defaults', liable to give rise to unjustified conduct on the part of the applicant; (iii) the very short deadlines for informing alternative operators of such works and for transmitting that information to their customers; and, finally, (iv) the shifting of responsibility to the alternative operator for service interruptions caused by repairs where that operator has been deemed to be uncooperative.

40 In the fifth place, the Commission regards as unfair several terms and conditions applying to the bank guarantee that must be provided by all alternative operators wishing to conclude a collocation agreement with the applicant and ultimately secure access to its local loops. Therefore, first of all, the applicant enjoys an overly wide discretion in deciding whether to accept or refuse a bank guarantee and is not subject to any deadline in that regard. Next, the amount of the guarantee, set at EUR 66 387.84, is disproportionate in the light of the risks and costs borne by the applicant. That is all the more so since the reference offer allowed the applicant to multiply that amount by up to 12 where it called on the guarantee. Furthermore, the applicant was able to call on the bank guarantee to cover not only the failure to pay for actual services it provided, but also to cover any claims for damages that it could submit. Moreover, the applicant was entitled to trigger the bank guarantee without having to show that it had first asked the debtor to pay and without the debtor having the option of raising a counterclaim. Lastly, the Commission notes that alternative operators do not benefit from any similar security even though they may incur losses as a result of the applicant's conduct in respect of unbundled access to local loops.

41 The Commission concludes that those aspects of the applicant's conduct, taken together, amounted to a refusal by that operator to provide unbundled access to its local loops.

(b) Margin squeeze of alternative operators in the provision of unbundled access to the applicant's local loops

42 In the second part of its analysis of the applicant's conduct, the Commission makes a finding of margin squeeze as a result of the behaviour of that operator in connection with unbundled access to its local loops, which constitutes an independent form of abuse of a dominant position. Accordingly, the spread between the prices charged by the applicant for the grant of such access to alternative operators and the prices charged to its own customers was either negative or insufficient for an operator as efficient as the applicant to cover the specific costs which it had to incur to supply its own products or service on the downstream market, namely the retail market.

43 Where the service portfolio under consideration includes only broadband products, the Commission notes that an equally efficient competitor would have been able, by means of unbundled access to the applicant's local loops, to replicate the entirety of the retail DSL offering of the applicant as it evolved over time. The 'period-by-period' margin analysis (namely the calculation of the available margins for each year of the period between 2005 and 2010) demonstrates that a competitor as efficient as the applicant faced negative margins and would not therefore have been able to replicate profitably the retail broadband portfolio offered by the applicant.

44 Where the portfolio examined includes voice telephony services in addition to broadband services by means of full access to the local loop, the Commission also concludes that a competitor as efficient as the applicant would not have been able, due to the prices charged by the applicant on the upstream market for unbundled access, to operate profitably on the relevant retail market during the period between 2005 and 2010. An equally efficient competitor would not therefore have been able to replicate profitably, over that same period, the applicant's portfolio. The addition to that portfolio of multi-play services, available as from 2007, would not alter that finding.

45 Since neither the applicant nor Deutsche Telekom put forward during the administrative procedure any objective justification for their exclusionary conduct, the Commission concludes that the applicant's conduct during the period under consideration constitutes an abusive margin squeeze.

3. Analysis of the anticompetitive effects of the applicant's conduct

46 The Commission considers that those two types of conduct on the part of the applicant, namely the refusal to provide unbundled access to the local loop and the margin squeeze of the alternative operators were likely to prevent alternative operators from relying on unbundled access in order to enter the Slovakian retail mass market for broadband services at a fixed location. According to the contested decision, its conduct made competition less effective on that market because there was no genuine profitable alternative for competing operators to wholesale access to xDSL broadband based on the unbundling of local loops. The impact of the applicant's conduct on competition was all the more significant because the retail market for broadband services showed strong potential for growth during the period under consideration.

47 The Commission also states, in essence, that according to the 'investment ladder' concept, that barrier to access to the unbundling of the local loop deprived alternative operators of a source of income which would have allowed them to make further investments in the network, particularly by developing their own access network to connect their customers directly.

48 The Commission concludes that the anticompetitive conduct of the applicant on the mass market for broadband services at a fixed location in Slovakia was likely to have negative effects on competition and, in the light of its geographical reach across the entire territory of the Slovak Republic, was able to affect trade between Member States.

4. Addressees of the contested decision and fines

49 According to the contested decision, not only was Deutsche Telekom in a position to exercise decisive influence over the applicant's commercial policy during the entire period under consideration, but it actually exercised such influence. Since the applicant and Deutsche Telekom form part of the same undertaking, both were held liable for the single and continuous infringement of Article 102 TFEU forming the subject matter of the contested decision.

50 As regards the penalty for that infringement, the Commission states that it set the amount of the fines by reference to the principles laid down in its Guidelines on the method of setting fines imposed pursuant to Article 23(2)(a) of Regulation No 1/2003 (OJ 2006 C 210, p. 2) ('the 2006 Guidelines').

51 First of all, the Commission calculates the basic amount of the fine by retaining 10% of the applicant's turnover on the market for unbundled access to the local loop and for fixed retail broadband in the last full financial year of its participation in the infringement, in the case at hand 2010, and by multiplying the resulting figure by 5.33 to take account of the duration of the infringement (5 years and 4 months). The basic amount of the fine is thus EUR 38 838 000. Under the first paragraph of Article 2(a) of the contested decision, this is the first fine imposed for the infringement in question and for which the applicant and Deutsche Telekom are held jointly and severally liable.

52 Next, the Commission applies a twofold adjustment to the basic amount. In the first place, it finds that when the infringement in question occurred, Deutsche Telekom had already been held liable for an infringement of Article 102 TFEU, on account of a margin squeeze in the telecommunications sector, in Decision 2003/707/EC of 21 May 2003 relating to a proceeding under Article 82 [EC] (Cases COMP/37.451, 37.578, 37.579 - Deutsche Telekom AG) (OJ 2003 L 263, p. 9), and that, when the decision was adopted, Deutsche Telekom held 51% of the applicant's shares and was in a position to exercise decisive influence over it. Consequently, the Commission finds that, for Deutsche Telekom, the basic amount of the fine should be increased by 50% on account of repeated infringement. In the second place, the Commission states that the worldwide turnover of Deutsche Telekom for 2013 was EUR 60 123 million and that, in order to give the fine sufficient deterrent effect, a coefficient multiplier of 1.2 should be applied to the basic amount. The product of that twofold adjustment, namely EUR 31 070 000, gives rise, under the first paragraph of Article 2(b) of the contested decision, to a separate fine imposed on Deutsche Telekom alone.

5. Operative part of the contested decision

53 Articles 1 and 2 of the contested decision read as follows:

'Article 1

(1) The undertaking consisting of Deutsche Telekom AG and Slovak Telekom a.s. has committed a single and continuous infringement of Article 102 of the Treaty and Article 54 of the EEA Agreement.

(2) The infringement lasted from 12 August 2005 until 31 December 2010 and consisted of the following practices:

(a) withholding from alternative operators network information necessary for the unbundling of local loops;

(b) reducing the scope of its obligations regarding unbundled local loops;

(c) setting unfair terms and conditions in its Reference Unbundling Offer regarding collocation, qualification, forecasting, repairs and bank guarantees;

(d) applying unfair tariffs which do not allow an equally efficient competitor relying on wholesale access to the unbundled local loops of Slovak Telekom a.s. to replicate the retail broadband services offered by Slovak Telekom a.s. without incurring a loss.

Article 2

For the infringement referred to in Article 1, the following fines are imposed:

(a) a fine of EUR 38 838 000 on Deutsche Telekom AG and Slovak Telekom a.s., jointly and severally;

(b) a fine of EUR 31 070 000 on Deutsche Telekom AG.

...'

II. Procedure and forms of order sought

[...]

71 The applicant claims that the Court should:

- annul Articles 1 and 2 of the contested decision in so far as that decision affects it;

- in the alternative, reduce the fine imposed on it under Article 2 of the contested decision;

- order the Commission to pay the costs;

- if the Court dismisses the action as inadmissible or unfounded, order each party to bear its own costs.

72 The Commission and the intervener contend that the Court should:

- dismiss the action in its entirety, and

- order the applicant to pay the costs.

III. Law

A. Admissibility

B. Substance

91 The applicant puts forward five pleas in law in support of both its principal claims seeking annulment of the contested decision and its alternative claims seeking a reduction of the fine imposed on it, alleging (i) manifest errors of law and assessment in the application of Article 102 TFEU as regards the applicant's abusive conduct; (ii) infringement of the applicant's rights of defence as regards the practice resulting in the margin squeeze; (iii) errors in the finding of that practice; (iv) error committed by the Commission in finding that the applicant and Deutsche Telekom were part of a single undertaking and were both liable for the infringement in question; and (v) errors in determining the amount of the fine.

1. First plea in law: manifest errors of law and assessment in the application of Article 102 TFEU as regards the applicant's abusive conduct

92 In support of its first plea, the applicant essentially objects to the legal test applied by the Commission in the contested decision in order to find that its practice constituted abuse of a dominant position within the meaning of Article 102 TFEU.

93 The first plea in law is essentially made up of five complaints, alleging (i) that the Commission failed to apply the condition relating to the indispensable nature of access to the applicant's copper DSL network to be able to operate on the retail market for broadband services in Slovakia, within the meaning of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569); (ii) misapplication of the judgment of 9 September 2009, Clearstream v Commission (T 30l/04, EU:T:2009:317); (iii) that the contested decision is inconsistent in terms of competition policy in so far as it concerns proof of the outright refusal of access and the refusal of access; (iv) errors of fact and of law as well as a defective statement of reasons in the justification for derogating from the conditions established in the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569); and (v) failure to demonstrate that access to the applicant's local loop is indispensable to competitors located on the downstream market.

94 The Commission and the intervener contest those complaints and contend that the present plea in law should be rejected.

(a) The first and fifth complaints

95 In the context of its first and fifth complaints, the applicant claims that the Commission, in essence, classified a range of its conduct during the period under consideration, referred to by the seventh part of the contested decision (recitals 355 to 821), as a 'refusal to supply' access to its local loop without having verified the indispensability of such access, for the purposes of paragraph 41 of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569).

96 By its first complaint, the applicant calls in question the Commission's findings in recitals 361 to 371 of the contested decision, according to which the circumstances of the present case are different from those in the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569). The applicant asserts that it is apparent from that judgment that a refusal to supply infringes Article 102 TFEU in particular where that refusal concerns a product the supply of which is indispensable for carrying on the relevant business (the 'indispensability condition'). In this case, the Commission wrongly failed to examine whether access to the applicant's network was indispensable to be able to operate on the retail market for broadband services in Slovakia. Accordingly, the applicant disputes the Commission's finding that the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), is authority for the proposition that, in the case of a constructive refusal of access, the Commission is not required to demonstrate that the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), particularly the indispensability condition, apply (recital 359 et seq. of the contested decision).

97 It follows from paragraphs 55 to 58 of the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), read together, that the practice resulting in the margin squeeze constitutes a standalone abuse under Article 102 TFEU which does not require prior proof of an obligation to deal meeting the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569). Since the Commission found that paragraph 55 of the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), covers not only the practice resulting in the margin squeeze but also a constructive refusal of access, as in this case, it was wrong to seek to expand significantly the narrow reasoning of that judgment.

98 In particular, according to the applicant, although it is apparent from the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), that the indispensability condition is not a requirement for all 'terms of trade' abuses under Article 102 TFEU, this does not, however, mean that the condition does not apply in the case of a refusal of access. Nowhere in the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), or in any other judgment, did the Court state that the indispensability condition established in the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), is limited to cases of outright refusal of access. On the contrary, that approach would undermine the effectiveness of Article 102 TFEU. Even if the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), involves an outright refusal to supply, the Court established in that judgment the general principles of a duty to assist competitors.

99 As regards the judgments cited by the Commission in the defence, the applicant considers that those judgments constitute a new approach compared to the contested decision. In any event, firstly, the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), assimilated the judgment of 6 March 1974, Istituto Chemioterapico Italiano and Commercial Solvents v Commission (6/73 and 7/73, EU:C:1974:18), which is authority for the proposition that indispensability is a legal prerequisite. Therefore, those two judgments are compatible.

100 Secondly, the case-law cited by the Commission, namely the judgments of 14 February 1978, United Brands and United Brands Continentaal v Commission (27/76, EU:C:1978:22), and of 16 September 2008, Sot. Lélos kai Sia and Others (C 468/06 to C 478/06, EU:C:2008:504), does not apply to the present case, since, first of all, the complaints raised in the context of those cases did not concern a refusal to deal but rather the fact that such a refusal was used as a way to bring about a different restriction of competition. Next, those cases did not involve the sale of an input to competitors on a downstream market, but the supply of finished goods for distribution or resale. Finally, in those cases, the dominant undertaking had decided to terminate the supply of goods that it previously supplied to the customers in question, while in this case - as in the case giving rise to the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), - the access seekers had never previously been supplied by the dominant undertaking.

101 Thirdly, the applicant considers, as regards the case-law cited by the Commission relating to the refusal to license intellectual property rights, namely the judgments of 5 October 1988, Volvo (238/87, EU:C:1988:477, paragraph 8), of 6 April 1995, RTE and ITP v Commission (C 241/91 P and C 242/91 P, EU:C:1995:98, paragraph 50), and of 29 April 2004, IMS Health (C 418/01, EU:C:2004:257, paragraph 35), that that case-law is compatible with the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), since that judgment refers to the judgment of 6 April 1995, RTE and ITP v Commission (C 241/91 P and C 242/91 P, EU:C:1995:98), which is itself cited in the later judgments. The fact that more stringent conditions, in particular that the input must be necessary to manufacture a 'new product', may be required in intellectual property cases does not mean that the Commission can do away with the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), in cases which are not related to that field.

102 Fourthly, the applicant claims that, regarding the application of the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), there is no reason to think that the Court wished to limit the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), to the strict circumstances of that case. There is a difference between stating, as was done in the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), that the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), do not apply to all 'terms of trade' cases and stating, as the Commission does, that the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), should not apply in any of those cases.

103 Fifthly, the decisions cited by the Commission do not support its proposition, since its analysis in Decision 2001/892/EC of 25 July 2001 relating to a proceeding under Article 82 [EC] (COMP/C 1/36.915 - Deutsche Post AG - Interception of cross-border mail) (OJ 2001 L 331, p. 40) is based on the fact that Deutsche Post's distribution network was indispensable to senders located in the United Kingdom. The Polaroid v SSI Europe case, cited as an example of an abusive constructive refusal of access, is not relevant to the case at hand.

104 By its fifth complaint, the applicant asserts that the contested decision does not demonstrate that access to its local loop is indispensable to downstream competitors. In that regard, it follows from the judgment of 29 April 2004, IMS Health (C 418/01, EU:C:2004:257, paragraph 28), that it is not sufficient to demonstrate that alternative approaches are less advantageous for other operators, but it is necessary to demonstrate the indispensability of the network concerned within the meaning of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569). The obligation to grant access to a facility arises if the refusal to supply objectively has a sufficiently serious effect on competition.

105 Furthermore, no relevance is to be attached to the questions examined by the Commission in section 7.3 of the contested decision and, in particular, in recitals 382 and 384 thereof, concerning whether the applicant's copper network was important and whether efficient wholesale access to xDSL based on the local loop was important for alternative operators in Slovakia. Accordingly, the Commission committed an error as regards the application of the criterion of indispensability. It is necessary for the Commission to consider whether access to the local loop is indispensable to enable the applicant's competitors to compete on the downstream retail market so that, without such access, competition would be impossible or excessively difficult. Since the vast majority of broadband access is based on technologies other than the applicant's copper network, such access is not indispensable in the sense of impossible or excessively difficult.

106 The Commission and the intervener dispute those arguments.

107 In that regard, according to established case-law, a dominant undertaking has a special responsibility not to allow its behaviour to impair genuine, undistorted competition on the internal market (see judgment of 6 September 2017, Intel v Commission, C 413/14 P, EU:C:2017:632, paragraph 135 and the case-law cited), and the fact that such a position has its origins in a former legal monopoly must, in that regard, be taken into account (judgment of 27 March 2012, Post Danmark, C 209/10, EU:C:2012:172, paragraph 23).

108 That is why Article 102 TFEU prohibits, in particular, a dominant undertaking from, among other things, adopting pricing practices that have an exclusionary effect on competitors considered to be as efficient as it is itself and strengthening its dominant position by using methods other than those that are part of competition on the merits. From that point of view, not all competition on price can be regarded as legitimate (see judgment of 6 September 2017, Intel v Commission, C 413/14 P, EU:C:2017:632, paragraph 136 and the case-law cited).

109 It has been held, in that regard, that abuse of a dominant position prohibited by Article 102 TFEU is an objective concept relating to the conduct of a dominant undertaking which, on a market where the degree of competition is already weakened precisely because of the presence of the undertaking concerned, through recourse to methods different from those governing normal competition in products or services on the basis of the transactions of commercial operators, has the effect of hindering the maintenance of the degree of competition still existing on the market or the growth of that competition (see judgments of 19 April 2012, Tomra Systems and Others v Commission, C 549/10 P, EU:C:2012:221, paragraph 17 and the case-law cited, and of 9 September 2009, Clearstream v Commission, T 301/04, EU:T:2009:317, paragraph 140 and the case-law cited).

110 Article 102 TFEU covers not only those practices that directly cause harm to consumers but also practices that cause consumers harm by interfering with the free play of competition (see, to that effect, judgments of 27 March 2012, Post Danmark, C 209/10, EU:C:2012:172, paragraph 20 and the case-law cited, and of 29 March 2012, Telefónica and Telefónica de España v Commission, T 336/07, EU:T:2012:172, paragraph 171).

111 The effect on competition referred to in paragraph 109 above does not necessarily relate to the actual effect of the abusive conduct complained of. For the purposes of establishing an infringement of Article 102 TFEU, it is necessary to show that the abusive conduct of the undertaking in a dominant position tends to restrict competition or, in other words, that the conduct is capable of having that effect (judgment of 19 April 2012, Tomra Systems and Others v Commission, C 549/10 P, EU:C:2012:221, paragraph 68; see, also, judgments of 9 September 2009, Clearstream v Commission, T 301/04, EU:T:2009:317, paragraph 144 and the case-law cited, and of 29 March 2012, Telefónica and Telefónica de España v Commission, T 336/07, EU:T:2012:172, paragraph 268 and the case-law cited).

112 Moreover, as regards the abusive nature of a practice resulting in a margin squeeze, it should be noted that the second paragraph of Article 102 TFEU expressly prohibits a dominant undertaking from directly or indirectly imposing unfair prices (judgments of 17 February 2011, TeliaSonera Sverige, C 52/09, EU:C:2011:83, paragraph 25, and of 29 March 2012, Telefónica and Telefónica de España v Commission, T 336/07, EU:T:2012:172, paragraph 173). Since the list of abusive practices contained in Article 102 TFEU is nevertheless not exhaustive, the list of abusive practices contained in that provision does not exhaust the methods of abusing a dominant position prohibited by EU law (judgments of 21 February 1973, Europemballage and Continental Can v Commission, 6/72, EU:C:1973:22, paragraph 26; of 17 February 2011, TeliaSonera Sverige, C 52/09, EU:C:2011:83, paragraph 26; and of 29 March 2012, Telefónica and Telefónica de España v Commission, T 336/07, EU:T:2012:172, paragraph 173).

113 In the present case, it should be pointed out that the argument presented by the applicant in the first plea in law refers only to the legal test applied by the Commission, in the seventh part of the contested decision (recitals 355 to 821), in order to classify a range of conduct of the applicant during the period under consideration as 'refusal to supply'. By contrast, the applicant does not dispute the existence itself of the conduct noted by the Commission in that part of the contested decision. As is apparent from recitals 2 and 1507 of the contested decision, that conduct, which contributed to the identification by the Commission of a single and continuous infringement of Article 102 TFEU (recital 1511 of the contested decision), consisted, firstly, in concealing from alternative operators information relating to the applicant's network, which is necessary for the unbundling of that operator's local loop, secondly, in a reduction by the applicant of its obligations relating to unbundling under the applicable regulatory framework and, thirdly, in the establishment by that operator of a number of unfair terms and conditions in its reference offer relating to unbundling.

114 Moreover, as the applicant confirmed during the hearing, the first plea in law does not call into question the analysis of the conduct which consisted in a margin squeeze carried out by the Commission in the eighth part of the contested decision (recitals 822 to 1045 of the contested decision). In its action, the applicant does not dispute that that type of conduct constitutes an independent form of abuse distinct from that of refusal to provide access and the existence thereof is therefore not subject to the criteria laid down in the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569) (see, to that effect, judgment of 10 July 2014, Telefónica and Telefónica de España v Commission, C 295/12 P, EU:C:2014:2062, paragraph 75 and the case-law cited). Therefore, in essence, by its first and fifth complaints, the applicant alleges that the Commission classified the conduct referred to in paragraph 113 above as 'refusal to supply' access to its local loop without having verified the 'indispensable' nature of such access, for the purposes of the third condition set out in paragraph 41 of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569).

115 In that judgment, the Court indeed considered that, in order for the refusal by a dominant undertaking to grant access to a service to constitute an abuse within the meaning of Article 102 TFEU, that refusal must be likely to eliminate all competition on the market on the part of the person requesting the service, such refusal must not be capable of being objectively justified, and the service must in itself be indispensable to carrying on that person's business (judgment of 26 November 1998, Bronner, C 7/97, EU:C:1998:569, paragraph 41; see, also, judgment of 9 September 2009, Clearstream v Commission, T 301/04, EU:T:2009:317, paragraph 147 and the case-law cited).

116 Moreover, it is clear from paragraphs 43 and 44 of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), that, in order to determine whether a product or service is indispensable for enabling an undertaking to carry on business in a particular market, it must be determined whether there are products or services which constitute alternative solutions, even if they are less advantageous, and whether there are technical, legal or economic obstacles capable of making it impossible or at least unreasonably difficult for any undertaking seeking to operate in the market to create, possibly in cooperation with other operators, alternative products or services. According to paragraph 46 of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), in order to accept the existence of economic obstacles, it must be established, at the very least, that the creation of those products or services is not economically viable for production on a scale comparable to that of the undertaking which controls the existing product or service (judgment of 29 April 2004, IMS Health, C 418/01, EU:C:2004:257, paragraph 28).

117 However, in the present case, since the legislation relating to the telecommunications sector defines the legal framework applicable to it and, in so doing, contributes to the determination of the competitive conditions under which a telecommunication undertaking carries on its business in the relevant markets, that legislation constitutes a relevant factor in the application of Article 102 TFEU to the conduct of that undertaking, in particular for assessing the abusive nature of such conduct (judgment of 14 October 2010, Deutsche Telekom v Commission, C 280/08 P, EU:C:2010:603, paragraph 224).

118 As the Commission correctly points out, the conditions referred to in paragraph 115 above were laid down and applied in the context of cases dealing with the question whether Article 102 TFEU could be such as to require the undertaking in a dominant position to supply to other undertakings access to a product or service, in the absence of any regulatory obligation to that end.

119 Such a context differs from that of the present case, in which TUSR, by a decision of 8 March 2005 confirmed by the director of that authority on 14 June 2005, required the applicant to grant all reasonable and justified requests for unbundling of its local loop, in order to enable alternative operators, on that basis, to offer their own services on the retail mass market for broadband services at a fixed location in Slovakia (paragraph 9 above). That requirement resulted from the public authorities' intention to encourage the applicant and its competitors to invest and innovate, whilst ensuring that competition in the market is maintained (recitals 218, 373, 388, 1053 and 1129 of the contested decision).

120 As is stated in recitals 37 to 46 of the contested decision, TUSR's decision, taken in accordance with Law No 610/2003, implemented in Slovakia the requirement of unbundled access to the local loop of operators with significant market power on the market for the provision of fixed public telephone networks, provided for in Article 3 of Regulation No 2887/2000. The EU legislature justified that requirement, in recital 6 of that regulation, by the fact that 'it would not be economically viable for new entrants to duplicate the incumbent's metallic local access infrastructure in its entirety within a reasonable time[, since a]lternative infrastructures ... do not generally offer the same functionality or ubiquity'.

121 Therefore, given that the relevant regulatory framework clearly acknowledged the need for access to the applicant's local loop, in order to allow the emergence and development of effective competition in the Slovak market for high-speed internet services, the demonstration, by the Commission, that such access was indeed indispensable for the purposes of the last condition set out in paragraph 41 of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), was not required.

122 It follows from the above that the Commission cannot be criticised for failing to establish the indispensable nature of access to the network at issue.

123 It should be added that such a complaint can also not be made against the Commission if it had to be considered that the implied refusal of access at issue was covered by the considerations in the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83). In that judgment, the Court held that it cannot be inferred from paragraphs 48 and 49 of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), that the conditions to be met in order to establish that a refusal to supply is abusive, which was the object of the first question referred for a preliminary ruling examined in that case, must necessarily also apply when assessing the abusive nature of conduct which consists in supplying services or selling goods on conditions which are disadvantageous or on which there might be no purchaser (judgment of 17 February 2011, TeliaSonera Sverige, C 52/09, EU:C:2011:83, paragraph 55). In that regard, the Court held that such conduct may, in itself, constitute an independent form of abuse distinct from that of refusal to supply (judgment of 17 February 2011, TeliaSonera Sverige, C 52/09, EU:C:2011:83, paragraph 56).

124 The Court, moreover, stated that a different interpretation of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), would amount to a requirement that any conduct of a dominant undertaking in relation to its terms of trade could be regarded as abusive, the conditions to be met to establish that there was a refusal to supply would in every case have to be satisfied, and that would unduly reduce the effectiveness of Article 102 TFEU (judgment of 17 February 2011, TeliaSonera Sverige, C 52/09, EU:C:2011:83, paragraph 58).

125 The applicant correctly notes, concerning that point, that the practice at issue in the main proceedings examined by the Court in the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), merely consisted, as is apparent from paragraph 8 of that judgment, in a possible margin squeeze by the historical Swedish fixed telephone network operator in order to discourage requests by alternative operators for access to its local loop. It cannot be deduced therefrom that the interpretation given by the Court in that judgment of the scope of the conditions set out in paragraph 41 of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), is limited to that sole form of abusive conduct and does not cover practices which are not strictly related to pricing such as those examined in the present case by the Commission in the seventh part of the contested decision (see paragraphs 27 to 41 above).

126 It should, first of all, be noted that, in paragraphs 55 to 58 of the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), the Court did not refer to the particular form of abuse constituted by the margin squeeze of competitor operators in a downstream market, but rather to the supply of 'services or selling goods on conditions which are disadvantageous or on which there might be no purchaser' and to 'terms of trade' fixed by the dominant undertaking. Such wording suggests that the exclusionary practices to which reference was therefore made concerned not solely a margin squeeze, but also other business practices capable of producing unlawful exclusionary effects for current or potential competitors, like those classified by the Commission as an implicit refusal to supply access to the applicant's local loop (see, to that effect, recital 366 of the contested decision).

127 That reading of the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), is supported by the reference made by the Court, in that part of its analysis, to paragraphs 48 and 49 of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569). Those paragraphs dealt with the second question for a preliminary ruling referred to the Court in that case and did not concern the refusal by the dominant undertaking at issue in the dispute in the main proceedings to give access to its home-delivery scheme to the publisher of a rival newspaper, examined in the context of the first question, but the possible classification as abuse of a dominant position of a practice which consisted for that undertaking in making such access subject to the condition that the publisher at issue entrust to it the carrying out of other services, such as sales in kiosks or printing.

128 In the light of the foregoing, it is necessary to conclude that the classification of the applicant's conduct examined in the seventh part of the contested decision as abusive practices for the purposes of Article 102 TFEU did not presuppose that the Commission establish that the access to the applicant's local loop was indispensable to the exercise of the activity of competitor operators in the retail market for fixed broadband services in Slovakia, for the purposes of the case-law cited in paragraph 116 above.

129 Therefore, the first and fifth complaints of the plea in law must be rejected as unfounded.

(b) The third complaint

130 By its third complaint, the applicant claims that the non-application of the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), in the case of a constructive refusal to supply access leads to inconsistency in terms of competition policy. In those circumstances, it is easier to prove a constructive refusal to supply access than an outright refusal, with the result that the more serious type of abuse would be treated less severely than the less serious abuse. In the present case, at least one of the applicant's competitors had access to its network, with the result that the refusal of access would not be outright (recital 408 of the contested decision). Outright refusal is more serious than constructive refusal of access, although the Commission's approach would result in the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), applying to the outright refusal and not to the constructive refusal of access.

131 However, the Commission offers no explanation as to why, in general terms, a constructive refusal of access must be treated more severely than an outright refusal of access or why, in particular, the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), no longer need to be satisfied in the former case.

132 The Commission and the intervener dispute those arguments.

133 In that regard, it suffices to note that that argument is based on an erroneous premiss, namely that the seriousness of an infringement of Article 102 TFEU consisting in a refusal by a dominant undertaking to supply a product or service to other undertakings depends solely on its form. The gravity of such an infringement is likely to depend on numerous factors independent of the explicit or implicit nature of that refusal, such as the geographic scope of the infringement, whether it is intentional, or further to its effects on the market. The 2006 Guidelines confirm that analysis when they state, in paragraph 20 thereof, that the assessment of the gravity of an infringement of Article 101 or Article 102 TFEU is made on a case-by-case basis for all types of infringement, taking account of all the relevant circumstances of the case.

134 Finally, it must be noted that, in paragraph 69 of the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), the Court observed that the indispensable nature of the wholesale product could be relevant in order to assess the effects of a margin squeeze. However, in the present case, it must be pointed out that the applicant invoked the Commission's obligation to show the indispensable nature of the unbundled access to the applicant's local loop only in support of its claim that the Commission failed to apply the legally appropriate criteria during its assessment of the practices examined in the seventh part of the contested decision (see, by analogy, judgment of 29 March 2012, Telefónica and Telefónica de España v Commission, T 336/07, EU:T:2012:172, paragraph 182), and not so as to call into question the Commission's assessment of the anticompetitive effects of those practices, carried out in the ninth part of that decision (recitals 1046 to 1109 of the contested decision).

135 Therefore the third complaint must be rejected as unfounded.

(c) The second complaint

136 By its second complaint, the applicant submits that the failure to apply the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), in the contested decision is at odds with the judgment of 9 September 2009, Clearstream v Commission (T 301/04, EU:T:2009:317), particularly paragraph 146 thereof, which, despite relating to a constructive refusal of sale, as mentioned in recital 360 of the contested decision, applies those conditions. The Commission is wrong because, in the Clearstream case, the de facto monopoly of the company at issue had statutory protection, even though the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), were satisfied. Unlike the case giving rise to the judgment of 9 September 2009, Clearstream v Commission (T 301/04, EU:T:2009:317), the Commission is, in this case, unable to demonstrate the indispensability of the applicant's DSL network. That is why it goes to such lengths to distinguish it from the Bronner and Clearstream cases.

137 The Commission and the intervener dispute those arguments.

138 In that regard, it must be noted, as the Commission correctly maintained, that there is no contradiction between the approach taken by the Commission in the case giving rise to the judgment of 9 September 2009, Clearstream v Commission (T 301/04, EU:T:2009:317), and in the present case, since, in that first case, there was no obligation for the dominant undertaking to supply the service at issue and the dominant undertaking had not developed its market position in the context of a legal monopoly.

139 As is apparent from the case-law referred to in paragraph 117 above, since the legislation relating to the telecommunications sector defines the legal framework applicable to it and, in so doing, contributes to the determination of the competitive conditions under which a telecommunications undertaking carries on its business in the relevant markets, that legislation is a relevant factor in the application of Article 102 TFEU to the conduct of that undertaking, in particular in assessing the abusive nature of such conduct.

140 Consequently, the second complaint must be rejected as unfounded.

(d) The fourth complaint

141 By its fourth complaint the applicant claims that the contested decision is vitiated by errors of fact and of law as well as an insufficient statement of reasons in recital 370. In that recital, the Commission defended the derogation from the conditions of the judgments of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), stating that those conditions did not apply to a refusal to supply access because, first, the applicant was subject to a regulatory obligation to grant access to the local loop under earlier rules and, secondly, because the applicant's network was developed under a former State monopoly.

142 In the first place, as regards the errors of fact and law relating to those two justifications, firstly, the applicant claims that the Commission committed errors by asserting that it was necessary to derogate from the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), due to the existence of the obligation to grant access to the local loop deriving from earlier rules.

143 In that regard, the applicant argues that such an obligation does not necessarily have implications as regards the conditions for the application of Article 102 TFEU, since they pursue different objectives. The Commission erred in law by failing to draw the distinction, flowing from paragraph 113 of the judgment of 10 April 2008, Deutsche Telekom v Commission (T 271/03, EU:T:2008:101), between the role of an ex ante regulatory obligation, which seeks to reduce the market power of dominant undertakings, and the role of ex post competition law, under which the main focus of the authorities is the specific conduct of undertakings and whether their use of any market power was abusive.

144 More particularly, an obligation to sell may be imposed by ex ante legislation where the Commission's power to impose such an obligation under Article 102 TFEU is exercisable only in exceptional circumstances. Although it is apparent from the case-law that the legislation relating to the telecommunications sector may be taken into account for the purpose of applying Article 102 TFEU to the conduct of a dominant undertaking (judgment of 14 October 2010, Deutsche Telekom v Commission, C 280/08 P, EU:C:2010:603, paragraphs 224 and 227), the Commission did not, in the contested decision, merely take into account the obligations imposed under such legislation; instead, it relied entirely on the assessment carried out by TUSR and did not conduct any examination of its own.

145 According to the applicant, the considerations in the judgment of 10 April 2008, Deutsche Telekom v Commission (T 271/03, EU:T:2008:101), from which it follows that secondary EU legislation 'may' be relevant under Article 102 TFEU, apply only in the context of that case, since the Court was required to examine whether the Commission had committed an error by noting the existence of a regulatory obligation under such legislation. It does not follow from that judgment or from the existence of a regulatory obligation that the Commission is able to sidestep the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569).

146 On the contrary, the applicant submits, and pointed out during the hearing, that Article 102 TFEU and the legislation at issue pursue different objectives, so that a national regulatory authority may decide to increase competition on the market while an obligation to deal can be imposed under Article 102 TFEU only to remedy an abusive refusal of access.

147 Moreover, the reference to Article 21(3) of Law No 610/2003, to which the Commission referred to argue that TUSR weighed up the interests involved, was not cited in that 'authority's earlier decisions. In any event, the general duty to conduct a balancing exercise under domestic law does not mean that the Commission is entitled to disregard the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569). It is for the Commission to demonstrate that, where there is an earlier regulatory obligation, the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), do not apply. The applicant acknowledges that, while it is true that there was no relevant regulatory obligation in the case which gave rise to that judgment, that does not lead to the conclusion which the Commission seeks to draw.

148 Secondly, as regards the justification according to which the applicant's network was developed in the context of a monopoly, the applicant claims that the case-law relied upon by the Commission, in the contested decision, does not allow that second justification to be rejected. First of all, paragraph 109 of the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), cited by the Commission, is not relevant. Next, it follows from paragraph 23 of the judgment of 27 March 2012, Post Danmark (C 209/10, EU:C:2012:172), to which the Commission refers, that the existence of a former State monopoly may be relevant when considering the conduct of an undertaking. Therefore, that judgment is not authority for the proposition that the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), do not apply.

149 The contention that the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), do not apply where the network at issue has its historical origins in a State monopoly is wrong because Article 102 TFEU does not envisage any special treatment for former State monopolies. On the contrary, the Commission has stated in the past that the historical nature of a monopoly was not relevant to the present-day assessment of abuse under Article 102 TFEU.

150 The Commission and the intervener dispute those arguments.

151 In that regard, it suffices, in order to reject those arguments, to note that the observations in paragraphs 117 to 121 above are not based on the premiss that the obligation imposed on the applicant to grant unbundled access to its local loop results from Article 102 TFEU, but merely notes, in accordance with the case-law cited in paragraph 117 above, that the existence of such a regulatory obligation is a relevant aspect of the economic and legal context in which it is necessary to assess whether the applicant's practices examined in the seventh part of the contested decision could be classified as abusive practices for the purposes of that provision.

152 Moreover, the reference made by the applicant to paragraph 113 of the judgment of 10 April 2008, Deutsche Telekom v Commission (T 271/03, EU:T:2008:101), to support the argument set out in paragraph 143 above, is not relevant. The Court admittedly held, in paragraph 113 of that judgment, that the national regulatory authorities operate under national law which may have objectives which differ from those of EU competition policy. That reasoning sought to support the Court's rejection of the applicant's argument invoked in that case, according to which, in essence, the ex-ante control of its charges by the German regulatory authority for telecommunications and post precluded that Article 102 TFEU could be applied to a possible margin squeeze resulting from its charges for unbundled access to its local loop. That paragraph was therefore unconnected with the question whether the existence of a regulatory obligation of access to the local loop of the dominant operator is relevant in order to assess the compliance of its access policy with Article 102 TFEU.

153 Moreover, it follows from settled case-law that the existence of a dominant position resulting from a legal monopoly must be taken into consideration in the context of the application of Article 102 TFEU (see, to that effect, judgment of 27 March 2012, Post Danmark, C 209/10, EU:C:2012:172, paragraph 23 and the case-law cited).

154 Therefore, in so far as it alleges errors of law and fact relating to the justifications put forward by the Commission in order to defend the derogation from the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), and relating to the applicant's obligation, deriving from the earlier rules, to grant access to the local loop and to the existence of a pre-existing State monopoly, the fourth complaint must be rejected as unfounded.

155 In the second place, the applicant alleges that the Commission failed to state adequate reasons relating to the justification that that institution put forward with a view to derogating from the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), and consisting in the need to make first-time access mandatory. The Commission did not conduct any examination as to the existence of an earlier regulatory obligation, did not analyse the content of that obligation and did not explain why first-time access should be mandatory or why the lack of such access would eliminate all effective competition. The Commission did not provide any evidence to support its assessment that national legislation had weighed up the incentives of the applicant for keeping its facilities for its own use against those of undertakings potentially wishing access to the local loop. The contested decision fails to demonstrate why the regulatory obligations concerned provide a sufficient basis for disregarding the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569). The Commission was required to give particularly clear, compelling and detailed reasons as to why first-time access should be mandatory and, in consequence, why the failure to grant access would eliminate all effective competition.

156 In the reply, firstly, the applicant adds that recitals 36 to 49 of the contested decision contain only a general description of the regulatory framework and a brief description of the earlier access obligation. That statement of reasons does not address the question whether that obligation justifies disregarding the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569).

157 Secondly, TUSR did not refer to the Slovakian regulatory provisions relating to the balancing exercise when it imposed the earlier obligations. In any event, the balancing exercise under the earlier rules is different from that under Article 102 TFEU. Moreover, an argument based on the balancing exercise allegedly carried out by TUSR cannot, according to the applicant, justify the complete lack of reasons as regards the other conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569).

158 Thirdly, when the Commission contends that the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), do not apply anyway in the present case, it is confusing the question of the merits of the decision with the obligation to state reasons.

159 Fourthly, the applicant claims that, as regards the reference to Section 9.3 of the contested decision relating to anticompetitive effects, that section does not state the reasons for the decision. The applicant claims, first of all, that the elimination of all effective competition is just one of the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), while the failure to state reasons exists as regards the other conditions of that judgment. Next, the examination of anticompetitive effects in the contested decision does not replace the need to provide specific reasons concerning the indispensability condition. Lastly, that section rebuts the Commission's argument in so far as it contains evidence demonstrating a lack of anticompetitive effects.

160 Moreover, as regards the justification based on the fact that the applicant's network was developed in the context of a monopoly, the applicant claims that the statement of reasons contained in recital 373 of the contested decision is not sufficient to explain why the Commission found the existence of a former State monopoly to be relevant to the consideration of abuse under Article 102 TFEU.

161 According to the applicant, since the Commission is subject to the principle of good administration, it is responsible for examining the specific circumstances of the former State monopoly that it seeks to rely on as the basis for disregarding the conditions of the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), circumstances which are highly relevant. The applicant also submits that it is not possible to describe as factual details of no particular relevance the information provided in recital 891 of the contested decision relating to the applicant's investments in broadband assets between 2003 and 2010. The applicant claims that, on the contrary, the Commission should have considered the nature and impact of those investments when compared with its historical position. The applicant concludes that if the statement of reasons contained in the contested decision was adequate, that would mean, in practice, that no limitation is imposed on the Commission where a State monopoly has existed in the past.

162 The Commission and the intervener dispute those arguments.

163 In that regard, the statement of reasons required by Article 296 TFEU must be appropriate to the act at issue and must disclose in a clear and unequivocal fashion the reasoning followed by the institution which adopted the measure, in such a way as to enable the persons concerned to ascertain the reasons for the measure and to defend their rights and the Court to exercise its power of review. In the case of a decision adopted pursuant to Article 102 TFEU, that principle requires that the contested decision mention facts forming the basis of the legal grounds of the measure and the considerations which led to the adoption of the decision (judgment of 9 September 2010, Tomra Systems and Others v Commission, T 155/06, EU:T:2010:370, paragraph 227).

164 Firstly, the applicant considers that the contested decision does not contain an examination concerning the existence of an earlier regulatory obligation, or an analysis of its contents, or evidence supporting the Commission's assessment that the national legislation had weighed up the incentives of the applicant for keeping its facilities for its own use against those of undertakings potentially wishing access to the local loop. It also does not put forward why the regulatory obligations allow the conditions of supply access arising from the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), to be disregarded. However, it must be noted that, first, the Commission set out, in the contested decision, the regulatory framework relating to unbundled access to the local loop in Slovakia in recitals 36 to 46 of the contested decision. Secondly, it set out the legal framework concerning abusive refusals to grant access in recitals 355 to 371 of the contested decision, explaining, more particularly, that it considered the present case to differ from the circumstances of the case giving rise to the judgment of 26 November 1998, Bronner (C 7/97, EU:C:1998:569), and that that judgment does not apply in the present case. Therefore, the argument put forward by the applicant must be rejected.

165 Secondly, as regards the applicant's claim that the Commission was required to give particularly clear, compelling and detailed reasons as to why the failure to grant access would eliminate all effective competition, it should be noted that, in the present case, the applicant invoked the indispensable nature of the unbundled access to its local loop only in support of its claim that the Commission failed to apply the legally appropriate criteria during its assessment of the practices examined in the seventh part of the contested decision (see, by analogy, judgment of 29 March 2012, Telefónica and Telefónica de España v Commission, T 336/07, EU:T:2012:172, paragraph 182), and not so as to call into question the Commission's assessment of the anticompetitive effects of those practices, carried out in the ninth part of that decision (recitals 1046 to 1109 of the contested decision). In any event, the reasoning contained in that part of the contested decision is clear and unequivocal concerning the negative effects on competition of the applicant's exclusionary conduct.

166 Thirdly, as regards the applicant's claim that the statement of reasons included in recital 373 of the contested decision does not suffice in order to explain the reasons why the Commission considers that the existence of a former State monopoly is relevant for the purposes of examining an abuse in the light of Article 102 TFEU, it should be noted that the Commission, first of all, stated in that recital, referring to the specific provisions of Articles 8 and 12 of Directive 2002/21 and Article 21(3) of Law No 610/2003, that the applicant's obligation to supply imposed by TUSR's decision took account of the applicant's incentives, as well as those of its competitors, to invest and innovate, while ensuring that competition in the market is maintained. The Commission added, in that recital 373, that it was possible that the imposition of an obligation to supply or grant access has no impact on incentives to invest or innovate, where the market position of the dominant undertaking had been developed under the protection of special or exclusive rights or had been financed by State resources, as happened in the present case. The Commission, next, referred to paragraph 23 of the judgment of 27 March 2012, Post Danmark (C 209/10, EU:C:2012:172), from which it follows that where the existence of a dominant position has its origins in a former legal monopoly, that fact has to be taken into account, explaining that that was the situation of the applicant in the present case. The Commission, finally, took care to explain, in recital 373 of the contested decision, that it followed from recital 3 of Regulation No 2887/2000 that one of the reasons why the local access network remained one of the 'least competitive segments of the liberalised telecommunications market' was that new entrants did not have widespread alternative network infrastructures, given that operators, such as the applicant, for a long time, had rolled out their local access networks over significant periods of time protected by exclusive rights and had been able, for decades, to fund investment costs from monopoly rents from the provision of voice telephony services and public funds.

167 Furthermore, the Commission noted, in recital 370 of the contested decision, that it followed from paragraph 109 of the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), that the structure of a market is also highly influenced by the former monopolistic structure.

168 In the light of all those reasons, it must be considered that the Commission gave sufficient reasons for its decision, where it concluded that the fact that the network at issue was developed under a monopoly constituted a relevant factor which it had to take into account in the context of the examination it conducted under Article 102 TFEU.

169 Therefore, it is necessary to reject the fourth complaint also in so far as it alleges a breach of the obligation to state reasons.

170 It follows from all of the foregoing that the first plea in law must be rejected as unfounded.

2. Second plea in law: infringement of the applicant's rights of defence as regards the assessment of the practice resulting in the margin squeeze

171 The second plea in law concerns the applicant's rights of defence and is divided into two parts. The first part alleges that the Commission committed procedural errors regarding the calculation of the applicant's long run average incremental costs ('LRAIC'), that is to say the costs which that operator would not have had to bear if it had not offered the corresponding services. The second part alleges that it was not possible for the applicant to adopt a position during the administrative procedure about the multi-period approach to the calculation of the costs borne by the applicant in order to assess the existence of a practice resulting in the margin squeeze.

(a) The first part, alleging procedural errors committed by the Commission concerning the calculation of the long run Average incremental costs (LRAIC)

172 The applicant complains that the Commission, first, changed the methods, the principles and the data for the purpose of conducting the analysis of the LRAIC and, secondly, failed, before the adoption of the contested decision, to communicate its objections to the information that the applicant had submitted to it in order to conduct that analysis. In the statement of objections, the Commission used solely the data from the applicant's internal cost reporting system, namely 'UCN' data (úcelové clenenie nákladov) (classification of specific costs) and the summaries of profitability provided by the applicant, since the applicant did not have specific data relating to the LRAIC. According to the applicant, that 'UCN' data was based on top-down, fully allocated historical costs. That data relies on straight-line depreciation that does not allow for the recovery of costs over time. In the statement of objections (paragraph 1038), the Commission itself acknowledged the limitations of that data for the purposes of the assessment of the practice resulting in the margin squeeze and stated that those data were unsatisfactory. Accordingly, following the statement of objections, the applicant submitted new data, on the basis of the consultancy report prepared in February 2013 and sent to the Commission in the annex to the response to the statement of objections. That new data inter alia adjusted the historical costs. Therefore, the Commission accepted the re-evaluation and asset costs and depreciation put forward by the applicant (recital 894 of the contested decision).

173 The applicant notes that, by accepting a significant part of that data, the Commission considers the consultancy report to be credible. Likewise, the Commission did not raise objections, with respect to the principles, methodology and data provided by the applicant, before the adoption of the contested decision. However, in the contested decision, it rejected part of those principles, that methodology and that data (recital 899 of the contested decision). The applicant considers that the Commission should have sent notification, before the adoption of the contested decision, of the detailed objections relating to the principles, methodology and data which it had set out in that decision. The lack of such notification constitutes an infringement of the rights of the defence. According to the applicant, the Commission was required to set out in full the methodology, principles and cost-related data on which it intended to rely as part of its burden of proving the infringement and was also required to communicate its view to the applicant. In addition, the applicant unsuccessfully raised those procedural problems with the Hearing Officer.

174 Furthermore, the Commission itself acknowledges that, at the time of the statement of objections, it was not in possession of any data on the LRAIC, while the contested decision relies on those costs, meaning that the Commission changed its approach between those two documents. According to the applicant, since the Commission changed its approach after sending the statement of objections, the onus was on it to send a fresh statement of objections or a new statement of facts to the applicant.

175 Moreover, according to the applicant, the tables relating to the margin squeeze calculation provided by the Commission during the state of play meeting of 16 September 2014 did not support the relevant passages of the contested decision or observe its rights of defence. In that context, the applicant describes those tables as superficial, being only four pages long and lacking any explanation in support of the data contained therein. Similarly, the applicant claims that the Commission did not send it the tables until the meeting on the state of play of the file held on 16 September 2014, one month before publication of the contested decision, showing that the Commission's position at that point in time was already fixed. The applicant noted, at the hearing, that, during that meeting, the Commission had indicated that it was in the process of drafting a negative decision regarding it. In any event, the disclosure of those tables demonstrates that the Commission felt compelled to send, after the statement of objections, a document setting out its margin squeeze calculation.

176 The Commission contests those arguments.

177 As is apparent from recital 862 of the contested decision, the Commission requested the applicant to supply the information necessary in order to calculate the costs relating to additional inputs which are necessary to transform its wholesale services into retail services. Before the statement of objections, the applicant sent the Commission calculations of the costs for 2003 to 2010 in 'UCN' spreadsheets and several spreadsheets containing additional calculations. In the context of the first part of its second plea in law, the applicant alleges, in essence, an infringement of its rights of defence in so far as the objections raised by the Commission concerning the methodology, the principles and the data which the applicant submitted were highlighted for the first time in recitals 860 to 921 of the contested decision.

178 In that regard, it should be recalled that observance of the rights of the defence in the conduct of administrative procedures relating to competition policy constitutes a general principle of EU law whose observance the European Courts ensure (see judgment of 18 June 2013, ICF v Commission, T 406/08, EU:T:2013:322, paragraph 115 and the case-law cited).

179 That principle requires that the undertaking concerned must have been afforded the opportunity, during the administrative procedure, to make known its views on the truth and relevance of the facts and circumstances alleged and on the documents used by the Commission to support its claim that there has been an infringement of the rules on competition. To that end, Article 27(1) of Regulation No 1/2003 provides that the parties are to be sent a statement of objections. That statement must set out clearly all the essential elements on which the Commission is relying at that stage of the procedure (judgment of 5 December 2013, SNIA v Commission, C 448/11 P, not published, EU:C:2013:801, paragraphs 41 and 42).

180 That requirement is satisfied where the final decision does not allege that the persons concerned have committed infringements other than those referred to in the statement of objections and only takes into consideration facts on which the persons concerned have had the opportunity of stating their views in the course of the procedure (see, to that effect, judgments of 24 May 2012, MasterCard and Others v Commission, T 111/08, EU:T:2012:260, paragraph 266, and of 18 June 2013, ICF v Commission, T 406/08, EU:T:2013:322, paragraph 117).

181 However, the essential facts on which the Commission is relying in the statement of objections may be set out summarily and the decision is not necessarily required to be a replica of the statement of objections, because that statement is a preparatory document containing assessments of fact and of law which are purely provisional in nature (see, to that effect, judgments of 17 November 1987, British American Tobacco and Reynolds Industries v Commission, 142/84 and 156/84, EU:C:1987:490, paragraph 70; of 5 December 2013, SNIA v Commission, C 448/11 P, not published, EU:C:2013:801, paragraph 42 and the case-law cited; and of 24 May 2012, MasterCard and Others v Commission, T 111/08, EU:T:2012:260, paragraph 267). Thus, it is permissible for the Commission to supplement the statement of objections in the light of the parties' replies, whose arguments show that they have actually been able to exercise their rights of defence. The Commission may also, in the light of the administrative procedure, revise or supplement its arguments of fact or law in support of its objections (judgment of 9 September 2011, Alliance One International v Commission, T 25/06, EU:T:2011:442, paragraph 181). Consequently, until a final decision has been adopted, the Commission may, in view, in particular, of the written or oral observations of the parties, abandon some or even all of the objections initially made against them and thus alter its position in their favour or decide to add new complaints, provided that it affords the undertakings concerned the opportunity of making known their views in that respect (see judgment of 30 September 2003, Atlantic Container Line and Others v Commission, T 191/98 and T 212/98 to T 214/98, EU:T:2003:245, paragraph 115 and the case-law cited).

182 It results from the provisional nature of the legal classification of the facts made in the statement of objections that the Commission's final decision cannot be annulled on the sole ground that the definitive conclusions drawn from those facts do not correspond precisely with that provisional classification (judgment of 5 December 2013, SNIA v Commission, C 448/11 P, not published, EU:C:2013:801, paragraph 43). The taking into account of an argument put forward by a party during the administrative procedure, without it having been given the opportunity to express an opinion in that respect before the adoption of the final decision, cannot as such constitute an infringement of its rights of defence, where taking account of that argument does not alter the nature of the complaints against it (see, to that effect, order of 10 July 2001, Irish Sugar v Commission, C 497/99 P, EU:C:2001:393, paragraph 24; judgments of 28 February 2002, Compagnie générale maritime and Others v Commission, T 86/95, EU:T:2002:50, paragraph 447, and of 9 September 2011, Alliance One International v Commission, T 25/06, EU:T:2011:442, paragraph 182).

183 The Commission is required to hear the addressees of a statement of objections and, where necessary, to take account of their observations made in response to the objections by amending its analysis specifically in order to respect their rights of defence. The Commission must therefore be permitted to clarify that classification in its final decision, taking into account the factors emerging from the administrative procedure, in order either to abandon such objections as have been shown to be unfounded or to amend and supplement its arguments, both in fact and in law, in support of the objections which it raises, provided however that it relies only on facts on which those concerned have had an opportunity to make known their views and provided that, in the course of the administrative procedure, it has made available the evidence necessary for their defence (see judgments of 3 September 2009, Prym and Prym Consumer v Commission, C 534/07 P, EU:C:2009:505, paragraph 40 and the case-law cited, and of 5 December 2013, SNIA v Commission, C 448/11 P, not published, EU:C:2013:801, paragraph 44 and the case-law cited).

184 Finally, it should be recalled that, according to established case-law, the rights of the defence are infringed where it is possible that, as a result of an error committed by the Commission, the outcome of the administrative procedure conducted by the latter might have been different. An applicant undertaking establishes that there has been such an infringement where it adequately demonstrates, not that the Commission's decision would have been different in content, but rather that it would have been better able to ensure its defence had there been no error, for example because it would have been able to use for its defence documents to which it was denied access during the administrative procedure (see judgments of 2 October 2003, Thyssen Stahl v Commission, C 194/99 P, EU:C:2003:527, paragraph 31 and the case-law cited, and of 24 May 2012, MasterCard and Others v Commission, T 111/08, EU:T:2012:260, paragraph 269 and the case-law cited; judgment of 9 September 2015, Philips v Commission, T 92/13, not published, EU:T:2015:605, paragraph 93).

185 In the present case, the Commission received, during the investigation, the data on costs in the 'UCN' spreadsheets, which constitute an accounting instrument of the applicant, representing, by commercial service and by service family, the total revenues, the total operating costs, the capital employed, the total cost of capital, the operating profit and the economic profit (recitals 863 and 864 of the contested decision). It follows from the contested decision that the costs included in the 'UCN' spreadsheets are based on fully allocated historical costs and differ from the LRAIC (recital 875 of the contested decision). The Commission also obtained presentations showing how the costs were grouped and tables and descriptions relating to the costs of each of the services (recitals 865 to 867 of the contested decision). The Commission requested the applicant to provide profitability data for broadband services, recalculated by using the LRAIC methodology (recitals 868 and 869 of the contested decision). Since the applicant confirmed that it did not calculate the profitability figures according to the LRAIC methodology for broadband services, the Commission used, at the statement of objections stage, the data at its disposal, namely the 'UCN' data and the explanations relating to costs, by adapting the individual costs (recitals 870 to 875 of the contested decision). According the Commission, at that stage, in the absence of data on the LRAIC, the figures contained in the 'UCN' spreadsheets were the best available source for the margin squeeze calculations (recital 875 of the contested decision). On that basis, in the statement of objections, the Commission established that an equally efficient competitor with access to the applicant's local loop would have faced significant negative margins if it had tried to replicate the applicant's retail broadband portfolio over the years 2005 to 2010 (paragraphs 1203 and 1222 of the statement of objections).

186 In the first place, as regards the complaint put forward by the applicant that it was not heard in relation to the principles, the methodology and the data relating to the calculation of the LRAIC, it should be noted that the applicant had the opportunity to respond to the arguments set out in the statement of objections and that it made full use of that possibility. Therefore, in its response to the statement of objections, on the basis of the consultancy report, the applicant submitted a methodology based on current cost accounting, by means of the estimation of the downstream costs for the period between 2005 and 2010 based on data from 2011 (recital 881 of the contested decision). In particular, the applicant claimed, in that response, that it was necessary, during the calculation of the LRAIC, first, to re-evaluate the assets and, secondly, to take into account the inefficiencies of its network for broadband provision. As regards, in particular, the taking into account of those inefficiencies, the applicant proposed to make optimisation adjustments, namely, firstly, the replacement of existing assets with their modern equivalents, which are more efficient and less costly (modern asset equivalent), secondly, to maintain as far as possible technical coherence and, thirdly, asset reduction on the basis of currently used capacity as opposed to the installed capacity (together, 'the optimisation adjustments').

187 In the contested decision, the Commission agreed to include in particular the applicant's asset re-evaluation in its margin squeeze analysis and to remove, as concerns the specific fixed costs, the joint and common costs. By contrast, it rejects the optimisation adjustments (recitals 894, 903, 904 and 910 of the contested decision). Therefore, in the contested decision, the Commission identified margins which were different from those calculated in the statement of objections.

188 However, it should be noted that the changes made in the contested decision in relation to the statement of objections, concerning the calculation of the margin squeeze, resulted from taking into consideration data and calculations provided by the applicant itself in response to the statement of objections. That taking into consideration appears therefore, in particular, in recitals 910, 945, 963 and 984 of the contested decision. It is moreover apparent from recitals 946 (footnote No 1405) and 1000 of the contested decision that the Commission took account, during the adoption of that decision, of the update of the margin squeeze calculations provided by the applicant in its response to the letter of facts (paragraph 21 above).

189 In the second place, as regards the complaint put forward by the applicant that the Commission changed the principles, methodology and data relating to the calculation of the LRAIC, without having heard the applicant on that issue, first of all, it is apparent from the examination of the statement of objections and the contested decision that the Commission did not put forward any new objections in the contested decision with respect to its assessment of the margin squeeze. In those two documents, the Commission considered that an equally efficient competitor using access to the local loop on the applicant's wholesale market would face significant negative margins if it offered the applicant's portfolio of broadband services via the local loop (paragraph 1203 of the statement of objections and recital 1023 of the contested decision). In both documents, the Commission considered that that would remain the case even if additional services in a downstream portfolio, namely voice services, internet television services (IPTV) and multi-play services, were taken into consideration (paragraph 1222 of the statement of objections and recital 1023 of the contested decision). Next, it should be noted that the infringement period applied by the Commission is shorter in the contested decision than that applied by the Commission in the statement of objections. In both documents, the date of the beginning of the infringement was 12 August 2005. By contrast, the date of the end of the infringement was 8 May 2012 in the statement of objections (paragraph 1546 of the statement of objections) and is set at 31 December 2010 in the contested decision (recital 1516 of the contested decision). Finally, as regards the methodology for calculating the margins, the Commission relied, in both documents, on the LRAIC. Therefore, both in paragraphs 996 to 1002 of the statement of objections and in recitals 860 and 861 of the contested decision, the Commission set out the guidelines for the calculation of costs on the basis of the LRAIC.

190 More particularly, as regards the methodology for calculating the margins, it should be noted that the Commission applied the same method at the stage of the statement of objections and of the contested decision. Firstly, tables 48 and 78 to 80 of the statement of objections and tables 21 to 24 of the contested decision show the wholesale charges for access to the local loop. It must be noted that the Commission nevertheless was careful to explain in recitals 935 to 938 of the contested decision the reasons why it considered that there was a difference between the figures provided by the applicant and the figures presented in the calculations which it carried out. Secondly, it should be noted that table 81 of the statement of objections corresponds to table 25 of the contested decision, both tables showing the network costs. Table 25 is based on data provided in the applicant's response to the statement of objections. Thirdly, it should be noted that table 82 of the statement of objections corresponds to table 26 of the contested decision, which shows the ISP (internet service provider costs) recurrent costs. The calculations of those costs are based on data provided by the applicant. Moreover, the Commission responds, in recitals 964 and 697 of the contested decision, to the applicant's arguments in that regard presented in its response to the state of objections. Fourthly, it must be pointed out that table 83 of the statement of objections and table 27 of the contested decision relate to the local loop set-up fees and are identical. Fifthly, both table 86 of the statement of objections and tables 29 and 30 of the contested decision concern the depreciation of subscribers' acquisition costs, table 29 taking into account a period of depreciation over three years and table 30 a period of depreciation over four years, in accordance with the applicant's suggestion in its response to the statement of objections. Sixthly, it must be noted that table 87 of the statement of objections is identical to table 31 of the contested decision relating to revenues from the applicant's bundled DSL Access and DSL internet services. Seventh, it should be pointed out that the results of the calculations of the margin squeeze are set out in table 88 of the statement of objections and in tables 32 and 33 of the contested decision, table 32 being based on the depreciation over three years and table 33 on the depreciation over four years.

191 It follows that the method and the principles that the Commission applied to examine the applicant's margins were, in essence, identical in the statement of objections and in the contested decision. Consequently, the complaint put forward by the applicant, according to which the Commission changed those methods and those principles before adopting the contested decision, without having heard the applicant on that issue, must be rejected.

192 As regards the data on which the calculations of the margins are based, as was explained in recitals 875 to 877 of the contested decision, admittedly, at the stage of the statement of objections, those calculations were based on the 'UCN' tables which reflected the fully allocated costs. However, as follows from recitals 885 to 894 of the contested decision, the Commission accepted the adjustments made by the applicant concerning current cost accounting. The Commission thus took into account the adjustments proposed in that regard by the applicant and modified the costs of network assets, so that they constitute a more precise estimate of an equally efficient competitor's costs. The taking those adjustments into account was specifically designed to satisfy the requirements noted in paragraph 183 above and the right of the parties to be heard during the administrative procedure therefore did not require that they be granted a new possibility to make known their point of view about the calculations of margins prior to the adoption of the contested decision.

193 In the light of the above, it is necessary to reject the first part relating to procedural errors concerning the calculation of the LRAIC.

(b) The second part, alleging the inability to take a position, during the administrative procedure, about the multi-period approach to the calculation of the costs borne by the applicant in order to assess the existence of a practice resulting in the margin squeeze

194 The applicant notes that, in the statement of objections, the Commission applied a method consisting in analysing the costs, by not taking into account the positive margin found in 2005, although, in the contested decision, the approach taken involved a multi-period analysis. By failing to give the applicant the opportunity to submit observations on that approach, the Commission infringed its rights of defence. The applicant considers that, in contrast to the Commission's assertions, it cannot be inferred from the response to the statement of objections that it itself proposed the multi-period approach. By contrast, the applicant proposed the discounted cash flow analysis, which was moreover applied by the Commission in its Decision C(2007) 3196 final, of 4 July 2007, relating to a proceeding under Article 82 [EC] (Case COMP/38.784 - Wanadoo España v Telefónica). The discounted cash flow analysis is justified by the duration of a customer's subscription or a contract.

195 The applicant claims that, in addition, in the context of the discounted cash-flow analysis, the Commission should not have commenced the evaluation in 2005 and terminated it in 2010 simply because that period corresponded to that examined in the context of the 'period-by-period' approach.

196 Specifically, the multi-period approach used in the contested decision leads to a finding of a positive margin in 2005 and to the extension of the period of the alleged abuse compared with the period set out in the statement of objections. In that regard, the Commission, according to the applicant, ignored that positive margin when it stated, in recital 998 of the contested decision, that entry for four months in 2005 could not be considered entry 'on a lasting basis'. The change of approach between the statement of objections and the contested decision transforms a positive margin in one year into a negative margin by selecting subsequent years of profitability and concluding that the net arithmetical difference was negative overall. Thus, the multi-period approach makes it impossible for a dominant undertaking to foresee the outcome of applying such an approach. Furthermore, the multi-period approach also leads to arbitrariness, since one or more periods could have both positive and negative margins at the same time depending on which years are used in that approach.

197 The Commission disputes the applicant's arguments and contends that the present part should be rejected.

198 In that regard, the applicant complains, in essence, that the Commission used the multi-period approach in order to extend the period of infringement set out in the statement of objections, since such an approach was not envisaged in the statement of objections, and that it infringed its rights of defence, by failing to give it the possibility of presenting its observations relating to that approach.

199 It should be noted that, in paragraph 1012 of the statement of objections, the Commission initially stated its intention to use the period-by-period approach in relation to the examination of the applicant's margins. The margin squeeze calculations included in paragraphs 1175 to 1222 of the statement of objections were made on a year-by-year basis during the period under consideration. In the contested decision, in order to assess the possible margin squeeze, the Commission adopted the 'period-by-period' approach consisting in determining the profits or losses realised over periods equivalent to one year (recital 851 of the contested decision). It should be noted that the summary of the results of the analysis is included in recitals 1007 to 1012 of the contested decision, from which it is apparent that the Commission bases its conclusions on the 'period-by-period' approach.

200 In paragraph 1281 of its response to the statement of objections, the applicant nevertheless opposed the sole use of the 'period-by-period' method, which had been used by the Commission in the statement of objections. The applicant claims in essence that, in the telecommunications sector, operators assessed their ability to obtain a reasonable return over a period exceeding one year. It thus proposed, inter alia, that the examination of a margin squeeze be supplemented by an analysis over several periods, in which the total margin would be evaluated over several years.

201 The Commission therefore decided to use, in addition, as is apparent from recital 859 of the contested decision, a multi-period approach so as to take account of that objection and in order to establish whether that approach altered its conclusion that the rates charged by the applicant for alternative operators to gain unbundled access to its local loop resulted in a margin squeeze between 2005 and 2010.

202 In the context of that additional examination, the result of which is included in recitals 1013 and 1014 of the contested decision, the Commission identified a total negative margin relating to each portfolio of services, first, for the period between 2005 and 2010 (table 39 in recital 1013 of the contested decision) and, secondly, for the period between 2005 and 2008 (table 40 in recital 1014 of the contested decision). The Commission inferred therefrom, in recital 1015 of the contested decision, that the multi-period analysis did not alter its finding of the existence of a margin squeeze resulting from a 'period-by-period' analysis.

203 It results from the above that, first, in order to establish the applicant's margins in the contested decision, the multi-period analysis followed the objection, made by the applicant, in its response to the statement of objections, relating to the 'period-by-period' method of calculation. Secondly, the multi-period analysis of the margins for unbundled access to the applicant's local loop, in the contested decision, aimed to complement the 'period-by-period' analysis. Moreover, the additional multi-period analysis led the Commission to consolidate its finding concerning the existence of a margin squeeze on the Slovak market for broadband internet services between 12 August 2005 and 31 December 2010.

204 Therefore, as the Commission, in essence, maintains, the multi-period analysis did not result in the applicant being held accountable for facts in respect of which the latter did not have the opportunity to express its views during the administrative procedure, by altering the nature of the objections made against it, but merely led to an additional analysis of the margins resulting from the applicant's pricing practices for unbundled access to its local loop, in the light of an objection raised by the applicant in its response to the statement of objections.

205 As regards the argument that the Commission used the multi-period analysis in order to establish the infringement period and to substitute a negative margin for 2005 for a previously positive margin, it should be noted that, following the 'period-by-period' analysis, the Commission had already reached the conclusion that a competitor as efficient as the applicant could not have, between 12 August 2005 and 31 December 2010, replicated profitably the applicant's retail portfolio comprising broadband services (recital 1012 of the contested decision). It is apparent in particular from recital 998 of the contested decision that, according to the Commission, the existence of a positive margin between August and December 2005 does not prevent that period from being included in the infringement period in the form of a margin squeeze, since operators consider their ability to obtain a return over a longer period. In other words, the Commission established the duration of the practice resulting in the margin squeeze on the basis of the 'period-by-period' approach and the multi-period approach was used solely by way of addition. In any event, it must be noted that that argument, in reality, seeks to contest the merits of that approach and cannot, therefore, be considered to be validly raised in support of an alleged infringement of the applicant's rights of defence. In reality, that complaint relates to a disagreement with the methodology used by the Commission to find a margin squeeze for the period from 12 August to 31 December 2005.

206 Concerning the applicant's argument that the method of calculating the margin squeeze applied by the Commission in the context of that additional examination does not correspond to the method suggested by the applicant in its response to the statement of objections and allegedly based on the Commission's decision-making practice, it is apparent from paragraphs 1498 to 1500 of that response that the applicant proposes to examine the 'cumulated profits' over a period between 2005 and 2008. The Commission however noted that the multi-period analysis suggested by the applicant was distinct from the backward-looking analysis of the discounted expected cash flow, which was based on different input data and a different methodology (recital 858 of the contested decision). It nevertheless took note of the applicant's suggestion concerning a multi-period analysis by undertaking, by way of addition, a multi-period examination, by analysing, in recital 1013 of the contested decision (table 39), the cumulated profits over a period between 2005 and 2010, and, in recital 1014 of that decision (table 40), the cumulated profits over a period between 2005 and 2008.

207 It is apparent from the case-law cited in paragraph 183 above that respect for the applicant's right to be heard merely required the Commission to take account, for the purposes of adopting the contested decision, of the criticism concerning the calculation of margins presented by the applicant in response to the statement of objections. By contrast, that right in no way implied that the Commission must necessarily reach the conclusion desired by the applicant when it made that criticism, namely the finding that there was no margin squeeze between 12 August 2005 and 31 December 2010.

208 Finally, as regards the document including the calculation of the margins sent by the Commission during the state of play meeting of 16 September 2014, the applicant claims, in essence, first, that that document was presented to it late, since the Commission had announced that the contested decision was being drafted and, secondly, since the Commission felt obliged to disclose its final calculations of margins before sending the contested decision to it.

209 However, for the reasons set out in in paragraphs 183 and 199 to 204 above, the Commission was not obliged to disclose its final calculations of margins before sending the contested decision to the applicant. Moreover, the fact that it organised a 'state-of-play meeting' does not invalidate that assessment. As the Commission maintained in its pleadings and during the hearing, such meetings are organised between the Commission and the parties under investigation in the interests of sound administration and transparency and to inform them of the state of play of the procedure. Nevertheless, those 'state-of-play meetings' are distinct from the formal meetings required in accordance with Regulations Nos 1/2003 and 773/2004 and are complementary to them. Therefore, the fact that it organised a state of play meeting on 16 September 2014 does not allow it to be concluded that the Commission was obliged to permit the applicant to present, on that occasion, its observations relating to the examination of margins, and that all the more so since the applicant was informed about all of the material elements of the calculation of margins made by the Commission and was given the opportunity to present its observations prior to the adoption of the contested decision.

210 It follows that the second part of the second plea in law and that plea in law must be rejected.

3. The third plea in law, alleging errors committed in the finding of the margin squeeze

211 By its third plea in law, the applicant essentially claims that the Commission's finding of the practice resulting in the margin squeeze was incorrect. This plea is divided into two parts. The first part alleges that no account was taken of the optimisation adjustments proposed by the applicant in the calculation of its LRAIC. The second part alleges error on the part of the Commission in its margin squeeze calculation due to the consolidation of revenue and costs for the entire infringement period, as well as infringement of the principle of legal certainty.

(a) First part alleging that no account was taken of the optimisation adjustments in the calculation of the applicant's LRAIC

212 In support of the first part of its third plea in law, the applicant contests the Commission's decision for not having accepted, in recitals 895 and 903 of the contested decision, the optimisation adjustments in order to calculate the margin squeeze. The inclusion of those optimisation adjustments would have reduced the upstream costs used in calculating the margin squeeze. Therefore, the reasons for that rejection set out by the Commission in recitals 894 and 900 to 902 of the contested decision are misconceived. Consequently, the Commission overestimated the applicant's actual downstream costs which, as a result, had material consequences for the findings relating to the margin squeeze, given that there was no margin squeeze in 2005 and 2007.

213 According to the applicant, its proposals were not additional adjustments, but in issue was its calculation of the LRAIC. The Commission's approach was inconsistent. On the one hand, it accepted the current cost accounting and, on the other hand, rejected the optimisation adjustments, which are compatible with the calculations of the LRAIC. As regards the adjustments of the network costs, the applicant considers that those adjustments, which are necessary in order to estimate the LRAIC, took into account a certain level of spare capacity included in the requirements for retail broadband services.

214 That approach is borne out by the case-law. Relying on the judgments of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), and of 10 April 2008, Deutsche Telekom v Commission (T 271/03, EU:T:2008:101), the applicant notes that, in some circumstances, it may be wise to take account of competitors' costs rather than those of the dominant undertaking. That is so in the present case, since the applicant did not have easy access to data to establish the LRAIC.

215 The Commission disputes the applicant's arguments.

216 The applicant complains, in essence, that the Commission committed an error during the calculation of the LRAIC, by refusing, in recitals 895 to 903 of the contested decision, to adjust the latter to the level of costs which would have been incurred by an efficient operator building an optimal network adapted to meet current and future demand on the basis of information available at the time of the assessment carried out by the Commission.

217 Therefore, as was stated in paragraph 186 above, in its response to the statement of objections, the applicant, relying on the consultancy report, proposed a methodology based on current cost accounting, by means of the estimation of the downstream costs for the period between 2005 and 2010 based on data from 2011 (recital 881 of the contested decision). In particular, the applicant submitted, in that response, that it was necessary, during the calculation of the LRAIC, first, to re-evaluate the assets and, secondly, to take account of the inefficiencies of its network for broadband provision. As regards, in particular, taking those inefficiencies into account, the applicant proposed the optimisation adjustments described in paragraph 186 above.

218 In its own calculations of the LRAIC, the applicant thus adjusted the capital cost of the assets and their depreciation values in the years 2005 to 2010, as well as the operating costs of those assets, by relying on the weighted average adjustment factor calculated by the author of the consultancy report for 2011 (recital 897 of the contested decision). The applicant claims that the suggested optimisation adjustments reflected the spare capacity identified in the elements of that network, namely assets removed from that network because they were not in productive use, but which had not yet been sold by that operator (recital 898 of the contested decision).

219 The Commission nevertheless refused to make those optimisation adjustments in the contested decision.

220 In the first place, as regards the replacement of existing assets with their more modern equivalents, the Commission stated, in recital 900 of the contested decision, that such a replacement could not be accepted since it amounts to adjusting the costs without properly adjusting the depreciations. The Commission referred to that paragraph in recitals 889 to 893 of the contested decision, in which it expressed doubts about the adjustment, as it was proposed by the applicant, of the costs of assets for the period between 2005 and 2010. Moreover, the Commission considered, in recital 901 of the contested decision, that such a replacement of existing assets was not compatible with the equally efficient competitor criterion. The case-law confirmed that the abusive nature of the pricing practices of a dominant operator is in principle determined in relation to its own position. In the present case, the adjustments of the LRAIC suggested by the applicant is based on a collection of hypothetical assets and not on the same assets as those held by that operator.

221 In the second place, as regards the taking into account of the excess capacity of the networks on the basis of 'actually' used capacity, the Commission concluded, in recital 902 of the contested decision, that, since investments are based on a forecast of demand, it was inevitable that, in the context of an ex post examination, a certain capacity remains sometimes unused.

222 None of the complaints put forward by the applicant against that part of the contested decision can be upheld.

223 Firstly, the applicant is wrong to claim that there is a contradiction between, on the one hand, the rejection of the optimisation adjustments of the LRAIC and, on the other hand, the acceptance, in recital 894 of the contested decision, of the asset re-evaluation which it proposed. The applicant also cannot claim that the Commission should have accepted the optimisation adjustments proposed by it on the ground that, as for the asset re-evaluation, the Commission did not have reliable historic costs concerning the optimisation adjustments.

224 The re-evaluation of the assets was based on the assets held by the applicant in 2011. With respect to that re-evaluation and as is apparent from recitals 885 to 894 of the contested decision, the Commission noted that it did not have at its disposal data better reflecting the applicant's incremental broadband asset costs for the period between 2005 and 2010. As a result, in the analysis of the margin squeeze in the contested decision, the Commission included the applicant's current assets proposed by the latter. However, the Commission pointed out that that re-evaluation was capable of leading to an underestimation of downstream asset costs.

225 By comparison, as is apparent from recital 895 of the contested decision, the optimisation adjustments proposed by the applicant consisted in adjusting the assets to the approximate level of an efficient operator that would build an optimal network adapted to satisfy future demand based on 'today's' information and demand predictions. Those adjustments were based on a forecast and on an optimal network model, and not on an estimate reflecting the incremental costs of the applicant's existing assets.

226 It follows that the optimisation adjustments, in general, and the replacement of existing assets by their more modern equivalents, in particular, had a different objective from the re-evaluation of assets proposed by the applicant. Furthermore, the taking into consideration, by the Commission, of the re-evaluation of current assets proposed by the applicant, due to the absence of other more reliable data on the LRAIC of that operator, did not suggest that the Commission necessarily accepted the optimisation adjustments of the LRAIC. The Commission was thus justified in treating differently, on the one hand, the replacement of existing assets by their more modern equivalents and, on the other hand, the re-evaluation of assets proposed by the applicant.

227 Secondly, the applicant cannot be followed when it disputes the conclusion, in recital 901 of the contested decision, that the optimisation adjustments would lead to a calculation of the LRAIC on the basis of the assets of a hypothetical competitor and not its own assets.

228 In that regard, it should be noted that, according to settled case-law, the assessment of the lawfulness of the pricing policy applied by a dominant undertaking, in the light of Article 102 TFEU, requires that reference be made, in principle, to pricing criteria based on the costs incurred by the dominant undertaking and on its strategy (see judgments of 17 February 2011, TeliaSonera Sverige, C 52/09, EU:C:2011:83, paragraph 41 and the case-law cited, and of 29 March 2012, Telefónica and Telefónica de España v Commission, T 336/07, EU:T:2012:172, paragraph 190; see also, to that effect, judgment of 10 April 2008, Deutsche Telekom v Commission, T 271/03, EU:T:2008:101, paragraph 188 and the case-law cited).

229 In particular, as regards a pricing practice resulting in a margin squeeze, the use of such analytical criteria can establish whether, in accordance with the equally efficient competitor test referred to in paragraph 108 above, that undertaking would have been sufficiently efficient to offer its retail services to end-users otherwise than at a loss if it had first been obliged to pay its own wholesale prices for the intermediary services (judgments of 14 October 2010, Deutsche Telekom v Commission, C 280/08 P, EU:C:2010:603, paragraph 201; of 17 February 2011, TeliaSonera Sverige, C 52/09, EU:C:2011:83, paragraph 42; and of 29 March 2012, Telefónica and Telefónica de España v Commission, T 336/07, EU:T:2012:172, paragraph 191).

230 The validity of such an approach is reinforced by the fact that it also conforms to the general principle of legal certainty, since taking into account the costs and prices of the dominant undertaking enables that undertaking, in the light of its special responsibility under Article 102 TFEU, to assess the lawfulness of its own conduct. While a dominant undertaking knows its own costs and prices, it does not as a general rule know those of its competitors (judgments of 14 October 2010, Deutsche Telekom v Commission, C 280/08 P, EU:C:2010:603, paragraph 202; of 17 February 2011, TeliaSonera Sverige, C 52/09, EU:C:2011:83, paragraph 44; and of 29 March 2012, Telefónica and Telefónica de España v Commission, T 336/07, EU:T:2012:172, paragraph 192).

231 The Court admittedly stated, in paragraphs 45 and 46 of the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), that it could not be ruled out that the costs and prices of competitors may be relevant to the examination of the practice resulting in the margin squeeze. It is apparent however from that judgment that it is only where it is not possible, in the light of the particular circumstances, to refer to the prices and costs of the dominant undertaking that the prices and costs of competitors on the same market should be examined, which the applicant has not claimed in the present case (see, by analogy, judgment of 29 March 2012, Telefónica and Telefónica de España v Commission, T 336/07, EU:T:2012:172, paragraph 193).

232 In the present case, first, the replacement of existing assets by their more modern equivalents sought to adjust the costs of assets by retaining the value of 'current' assets, without however properly adjusting the depreciations (recital 900 of the contested decision). That replacement led to a calculation of the margin squeeze on the basis of hypothetical assets, namely assets which do not correspond with those held by the applicant. The costs relating to the applicant's assets were thus underestimated (recitals 893 and 900 of the contested decision). Secondly, taking into consideration the excess capacity of the networks on the basis of the 'currently' used capacity would result in excluding the applicant's assets which are not in productive use (see paragraph 218 above).

233 Therefore, in the light of the principles noted in paragraphs 228 to 231 above, the Commission was able to conclude without committing an error that the optimisation adjustments of the LRAIC proposed by the applicant would have resulted, during the calculation of the margin squeeze, in the costs incurred by that operator itself between 12 August 2005 and 31 December 2010 being disregarded.

234 Finally, the applicant cannot be followed when it claims that, in the contested decision, the Commission infringed the principle that the examination of a margin squeeze must be based on an effective competitor, when it concluded in essence that it was inevitable that there sometimes remains unused capacity (recital 902 of the contested decision). It follows from the principles referred to in paragraphs 230 and 231 above that the examination of a pricing practice resulting in a margin squeeze consists, in essence, in assessing whether a competitor as efficient as the dominant operator is capable of offering the services concerned to final customers otherwise than at a loss. Such an examination is therefore not carried out by reference to a perfectly efficient operator in the light of market conditions at the time of such a practice. If the Commission had accepted the optimisation adjustments linked to excess capacity, the calculations of the applicant's LRAIC would have reflected the costs associated with an optimal network corresponding to demand and not affected by the inefficiencies of that operator's network, namely the costs of a competitor more efficient than the applicant. Therefore, in the present case, although it is not disputed that part of the applicant's relevant assets remained unused between 12 August 2005 and 31 December 2010, the Commission was able without committing an error to include that part of the assets, in other words the excess capacity, in the calculation of the LRAIC.

235 The Commission was thus correct to reject the optimisation adjustments and, therefore, to analyse the abusive nature of the applicant's pricing practices, in particular, on the basis of the applicant's costs.

236 Thirdly, contrary to what is claimed by the applicant, the rejection of the optimisation adjustments is not incompatible with the findings in the judgments of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), and of 10 April 2008, Deutsche Telekom v Commission (T 271/03, EU:T:2008:101), according to which it may be wise to take account of competitors' costs rather than the costs of the dominant undertaking.

237 First, regarding the case giving rise to the judgment of 10 April 2008, Deutsche Telekom (T 271/03, EU:T:2008:101, paragraph 210), the termination fees at issue constituted wholesale charges billed by the dominant undertaking to its competitor and making up part of the total cost borne by that competitor. Those fees had therefore to be included in the calculation of the costs of an equally efficient competitor. However, those fees were different from a forecast as well as from an optimal network model, which did not reflect the incremental costs of the applicant's existing assets (see paragraph 225 above).

238 Secondly, as regards the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83), as referred to in paragraphs 230 and 231 above, it follows from that judgment that it is only when it is not possible, in the light of the circumstances, to refer to the prices and costs of the dominant undertaking that it is appropriate to examine those of the competitors on the same market. However, that is not the situation in the present case, since the costs of the applicant's assets could be established on the basis of a later revaluation, and they constituted an indicator in order to evaluate the costs of an equally efficient competitor.

239 It follows from the above that the first part of the third plea in law must be rejected as unfounded.

(b) The second part, alleging an error in the margin squeeze calculation due to the consolidation of revenue and costs for the entire period under consideration, as well as infringement of the principle of legal certainty

240 The applicant contests the Commission's resorting to the multi-period approach (multi-annual approach set out in recital 1013 of the contested decision). The Commission adopted that approach, which does not exist in the statement of objections, in order to convert the positive margins into negative margins. It obtained positive margins, during the administrative procedure, by adopting the 'period-by-period' approach. However, by adopting the multi-period (multi-annual) approach, the Commission extended the period of the infringement. In particular, the 'period-by-period' (year-by-year) approach led to a finding of negative margins for each year during the period between 2005 and 2010. The negative margin of 2005, found in the statement of objections, was, however, transformed into a positive margin in the contested decision. Consequently, by applying the 'period-by-period' (year-by-year) approach, in actual fact, there was no practice resulting in the margin squeeze in 2005. Relying on a numerical example, the applicant concludes that, in accordance with the multi-period (multi-annual) approach used by the Commission, a margin squeeze could be found for the entire period, whereas that is not the case when each year is consolidated.

241 Therefore, the multi-period (multi-annual) approach is arbitrary and incompatible with the principle of legal certainty, since the infringement period depends entirely on the period over which the margins are aggregated and compared.

242 The applicant considers that, if those errors were corrected, there would be no basis for a finding of the practice resulting in the margin squeeze and the Commission would not have discharged its burden of proving the infringement. Annex A.21 to the application demonstrates the existence of a material error in the Commission's costs and revenue analysis.

243 The Commission's argument that a finding of a practice resulting in the margin squeeze may be made, notwithstanding the existence of a positive margin, is at odds with the case-law, as the legal test, in order to establish that a pricing practice resulting in a margin squeeze is abusive for the purposes of Article 102 TFEU, is whether the undertaking itself or an undertaking as efficient as it would have been in a position to offer its services to subscribers otherwise than at a loss. A positive margin does not necessarily lead to an abuse. The necessary precondition for a finding of a practice resulting in an abusive margin squeeze is the existence of a negative margin at an equally efficient competitor, which was not demonstrated in the present case for 2005.

244 Moreover, the Commission's contention that the multi-period (multi-annual) approach was proposed by the applicant is incorrect, since, in reality, it suggested the updating method referred to in paragraph 194 above.

245 In the first place, the Commission notes that it is apparent from recitals 997 and 998 of the contested decision that the 'period-by-period' (year-by-year) approach allowed it to be demonstrated that an equally efficient competitor using, on the wholesale market, access to the applicant's local loop faced negative margins and could not replicate profitably the applicant's retail broadband portfolio. That conclusion is not undermined by the fact that the margin was positive for the last four months of 2005. It is only after reaching that conclusion that the Commission strengthened its analysis with the 'multi-period' approach. As regards the arguments relating to the validity of that multi-period (multi-annual) approach, the Commission refers back to the arguments it submitted in the context of the second part of the second plea in law.

246 In the second place, the Commission maintains that it follows from the judgment of 17 February 2011, TeliaSonera Sverige (C 52/09, EU:C:2011:83, paragraphs 74 and 75), that an abusive squeeze may exist even where margins remain positive, where the dominant undertaking's practices are likely to make it at least more difficult for the operators concerned to trade on the market by reason, for example, of artificially reduced profitability, where such practices are not economically justified. Consequently, the fact that the margin was positive for the last four months of 2005 does not automatically mean that the applicant's conduct did not amount to abuse during that period. On the contrary, the Commission argues that such conduct constitutes abuse if the applicant's pricing policy was likely to have an exclusionary effect for competitors that are at least as efficient as the applicant by making access to the market concerned more difficult or even impossible for them. In addition, when determining the lawfulness of the pricing policy applied by a dominant undertaking, reference should be made to that undertaking's strategy which, in the case at hand, indicates that the applicant knew that it was setting higher prices than its average revenue for wholesale access at local loop level and that it could carry out a margin squeeze.

247 In the third place, as regards the criticisms of the multi-period (multi-annual) approach, the Commission repeats its arguments that the infringement period was already determined using the 'period-by-period' approach. On the basis of that approach, the Commission concluded that that infringement period commenced on 12 August 2005. The period adopted for the multi-period approach was determined by the infringement period which had already been established in the context of the 'period-by-period' approach. Moreover, the Commission submits that while it was aware of the shortcomings of the 'multi-period' (multi-annual) approach, that approach was proposed by the applicant in paragraphs 1388 and 1389 of its response to the statement of objections.

248 Finally, the applicant's contention that the multi-period (multi-annual) approach could be based on the duration of a customer's subscription or a contract is not supported by the case-law since, in the Telefónica case, as in the present case, the multi-period (multi-annual) analysis covered approximately five years, corresponding both to the length of the infringement and the lifetime of the relevant assets.

249 The intervener points out that the Commission's approach, in calculating the margin squeeze, was prudent and to the applicant's advantage since, with a view to avoiding purely hypothetical assumptions of those costs, collocation fees, which constituted, for alternative operators, an unknown amount and, for the applicant, a significant part of the costs connected with the local loop, were not included in the LRAIC.

250 By the second part of its third plea in law, the applicant accuses the Commission, in essence, of having applied the multi-period (multi-annual) approach solely in order to extend the infringement period to the last four months of 2005 during which, according the period-by-period (year-by-year) approach, there was a positive margin. The Commission thus wrongly concluded that there was a margin squeeze during 2005 and infringed the principle of legal certainty.

251 In that regard, it should be pointed out that the Commission concluded, relying on the 'period-by-period' (year-by-year) approach, that the applicant had engaged in margin squeeze practices from 12 August 2005. It is apparent from recital 997 of the contested decision that, on the basis of an analysis relating to every year during the period under consideration, an equally efficient competitor using wholesale access to the applicant's local loop was faced with negative margins and could not replicate profitably the retail broadband portfolio of the applicant. In recital 998 of the contested decision, the Commission pointed out that the fact that there is a positive margin for four months in 2005 does not disprove that conclusion, given that an entry over four months could not be considered as entry on a lasting basis. According to the Commission, operators consider their ability to earn a reasonable return over a longer period, which extends over several years (recital 998 of the contested decision). On that basis, the Commission concluded, in recital 1012 of that decision, that, in the period between 12 August 2005 and 31 December 2010, a competitor as efficient as the applicant could not have replicated profitably that operator's retail portfolio.

252 However, as was noted in paragraph 228 above, in order to assess the lawfulness of the pricing policy applied by a dominant undertaking, it is necessary, in principle, to refer to pricing criteria based on the costs incurred by the dominant undertaking itself and on its strategy.

253 In particular, as regards a pricing practice resulting in a margin squeeze, the use of such analytical criteria can establish whether that undertaking would have been sufficiently efficient to offer its retail services to end-users otherwise than at a loss if it had first been obliged to pay its own wholesale prices for the intermediary services (see paragraph 229 above and the case-law cited).

254 First, the validity of such an approach is all the more justified by the fact that it also conforms to the general principle of legal certainty, since taking into account the costs and prices of the dominant undertaking enables that undertaking, in the light of its special responsibility under Article 102 TFEU, to assess the lawfulness of its own conduct. While a dominant undertaking knows what its own costs and charges are, it does not, as a general rule, know what its competitors' costs and charges are. Secondly, an exclusionary abuse also affects potential competitors of the dominant undertaking, which might be deterred from entering the market by the prospect of a lack of profitability (see paragraph 230 above and the case-law cited).

255 It follows therefrom that, in order to establish the constituent elements of the practice of a margin squeeze, the Commission, in recital 828 of the contested decision, correctly had recourse to the equally efficient competitor criterion, by demonstrating that the dominant undertaking's downstream operations could not trade profitably on the basis of the wholesale price applied in respect of its downstream competitors and on the retail price applied by the downstream arm of that undertaking.

256 As is apparent from tables 32 to 35 of the contested decision, the analysis carried out by the Commission resulted, in all the scenarios envisaged and as the latter itself acknowledged in recital 998 of that decision, in a positive margin for the period between 12 August and 31 December 2005.

257 In such circumstances, the Court has already held that, to the extent that a dominant undertaking sets its prices at a level covering the great bulk of the costs attributable to the supply of the goods or services in question, it is, as a general rule, possible for a competitor as efficient as that undertaking to compete with those prices without suffering losses that are unsustainable in the long term (judgment of 27 March 2012, Post Danmark, C 209/10, EU:C:2012:172, paragraph 38).

258 It results therefrom that, during the period between 12 August and 31 December 2005, a competitor as efficient as the applicant had, in principle, the possibility to compete with the latter on the retail market for broadband services in so far as unbundled access to the local loop was granted to it, and without suffering losses that were unsustainable in the long term.

259 The Court has indeed held that, if a margin is positive, it is not ruled out that the Commission can, in the context of the examination of the exclusionary effect of a pricing practice, demonstrate that the application of that practice was, by reason, for example, of reduced profitability, likely to have the consequence that it would be at least more difficult for the operators concerned to trade on the market concerned (see, to that effect, judgment of 17 February 2011, TeliaSonera Sverige, C 52/09, EU:C:2011:83, paragraph 74). That case-law can be read in conjunction with Article 2 of Regulation No 1/2003, according to which, in any proceedings for the application of Article 102 TFEU, the burden of proving an infringement of that article rests on the party or the authority alleging the infringement, namely, in the present case, the Commission.

260 However, in the present case, it must be noted that the Commission did not demonstrate in the contested decision that the applicant's pricing practice, during the period between 12 August and 31 December 2005, resulted in such exclusionary effects. However, such a demonstration was required particularly given the presence of positive margins.

261 The mere claim, in recital 998 of the contested decision, that the operators consider their ability to earn a reasonable return over a longer period, which lasted several years, cannot constitute such proof. Such a fact, assuming it is established, is based on a prospective examination of profitability, which is necessarily hypothetical. Furthermore, in the present case, those positive margins appeared at the very beginning of the period under consideration, at a time when no negative margin could yet have been found. In those circumstances, it must be concluded that the reason set out in recital 998 of the contested decision does not satisfy the requirement arising from the principle of legal certainty noted in paragraph 230 above, according to which a dominant undertaking must be in a position to assess the conformity of its conduct with Article 102 TFEU.

262 For that same reason, the finding of the negative margins, by means of the application of the multi-period approach, cannot undermine that assessment, since, in the present case, that approach resulted in such a finding only by means of a weighting of the positive margins for 2005 with the negative margins found respectively for the years 2006 to 2010 (recital 1013 of the contested decision) and 2006 to 2008 (recital 1014 of the contested decision).

263 Moreover, in recital 1026 of the contested decision, on the basis of the documents established by applicant's regulatory department in April 2005 and relating to a strategy of submission of the reference offer concerning unbundled access to the local loop and ULL prices, the Commission considered that the latter knew, from 1 August 2005, that the prices for wholesale access at local loop level were squeezing the margins of alternative operators.

264 Nevertheless, it should be noted that, in the light of positive margins between 12 August and 31 December 2005, the Commission was subject to a specific obligation with regard to the proof of exclusionary effects of the practice of a margin squeeze alleged against the applicant during that period (see the case-law referred to in paragraph 259 above).

265 Therefore, the Commission's allegation and the documents invoked in support thereof are not sufficient to demonstrate the exclusionary effect of the practice of a margin squeeze alleged against the applicant and, for example, a reduction of profitability, likely to make it at least more difficult for the operators concerned to exercise their activities on the market at issue.

266 Moreover, sections 9 and 10 of the contested decision, which deal with the anticompetitive effects of the applicant's conduct, do not contain any examination of the effects of the practice of a margin squeeze alleged during the period between 12 August and 31 December 2005.

267 Therefore, in the light of settled case-law according to which any doubt in the mind of the Court must operate to the advantage of the undertaking to which the decision finding an infringement was addressed (judgments of 8 July 2004, JFE Engineering and Others v Commission, T 67/00, T 68/00, T 71/00 and T 78/00, EU:T:2004:221, paragraph 177, and of 12 July 2011, Hitachi and Others v Commission, T 112/07, EU:T:2011:342, paragraph 58), it must be concluded that the Commission has not provided proof that the practice leading to a margin squeeze by the applicant had begun before 1 January 2006. Since the contested decision is, consequently, vitiated by an error of assessment on that point, it is not necessary to examine whether that approach also infringed Article 23 of Regulation No 1/2003, as the applicant claims.

268 In the light of the above, the second part of the third plea in law invoked by the applicant must be partially upheld and Article 1(2)(d) of the contested decision must be annulled in so far as it declares that, over the course of the period between 12 August and 31 December 2005, the applicant imposed unfair tariffs which do not allow an equally efficient competitor relying on wholesale access to its unbundled local loops to replicate the retail broadband services offered by the applicant without incurring a loss.

5. The fifth plea in law, raised in the alternative, alleging errors in the determination of the amount of the fine

427 By its fifth plea in law, raised in the alternative, the applicant claims that the Commission committed errors in the calculation of the amount of the fine imposed on it. That plea is divided into two parts. The first part alleges a manifest error of assessment where the Commission took account of the applicant's turnover for the 2010 financial year for the purpose of calculating the amount of the fine. The second part alleges a manifest error of assessment relating to the date on which the infringement period commenced.

(a) First part alleging a manifest error of assessment as a result of the applicant's turnover for the 2010 financial year being taken into account for the purpose of calculating the amount of the fine

428 The applicant considers that the Commission committed a manifest error of assessment in finding, in accordance with point 13 of the 2006 Guidelines, that the basic amount of the fine had to be calculated on the basis of the turnover for the last full year of the infringement, namely, in particular, the applicant's turnover on the unbundled local loop market and fixed broadband market in 2010.

429 In so doing, the Commission diverged from its own decision-making practice, namely Decision C(2011) 4378 final, of 22 June 2011, relating to a proceeding under Article 102 TFEU (Case COMP/39.525 - Polish telecommunications) ('the Polish Telecommunications decision'). In recital 896 of that decision, the Commission found that it was appropriate to take the average annual sales due, first, to the significant growth in sales during the period in question in the relevant market, particularly wholesale sales and, secondly, to the fact that the market was still developing and hence growing beyond normal market growth rates at the time of the infringement. That consideration should be applied to the present case since the Commission admitted in the contested decision that the applicant's turnover had grown by 133% between 2005 and 2010. Therefore, relying on that decision, the applicant argues that the basic amount of the fine should have been calculated on the basis of the average of the five years of infringement found by the Commission.

430 By relying on the last financial year, the Commission applied stricter standards to the applicant than in the Polish Telecommunications decision. The applicant also states that although the Commission enjoys a margin of discretion in setting the amount of fines, it cannot act in an arbitrary and inconsistent manner.

431 The Commission, supported by the intervener, contests those arguments.

432 It is necessary, first of all, to note that Article 23(3) of Regulation No 1/2003 provides that, in fixing the amount of the fine, regard must be had both to the gravity and to the duration of the infringement.

433 Moreover, it should be noted that, according to point 13 of the 2006 Guidelines, 'in determining the basic amount of the fine to be imposed, the Commission will take the value of the undertaking's sales of goods or services to which the infringement directly or indirectly relates in the relevant geographic area within the EEA' and that for that purpose, it 'will normally take the sales made by the undertaking during the last full business year of its participation in the infringement ...'.

434 It is apparent, moreover, from the case-law that the proportion of the turnover accounted for by the goods or services in respect of which the infringement was committed gives a proper indication of the scale of the infringement on the relevant market, since the turnover in those goods or services constitutes an objective criterion giving a proper measure of the harm which that practice does to normal competition (see, to that effect, judgment of 28 June 2016, Portugal Telecom v Commission, T 208/13, EU:T:2016:368, paragraph 236 and the case-law cited).

435 Point 13 of the 2006 Guidelines thus aims, as regards an infringement of Article 102 TFEU, to adopt as the starting point for the calculation of the amount of the fine imposed on the undertaking at issue an amount which reflects the economic significance of the infringement (see, to that effect, judgments of 11 July 2013, Team Relocations and Others v Commission, C 444/11 P, not published, EU:C:2013:464, paragraph 76; of 12 November 2014, Guardian Industries and Guardian Europe v Commission, C 580/12 P, EU:C:2014:2363, paragraph 57; and of 23 April 2015, LG Display and LG Display Taiwan v Commission, C 227/14 P, EU:C:2015:258, paragraph 53).

436 However, it should also be noted that the self-limitation of the Commission's discretion arising from the adoption of the 2006 Guidelines is not incompatible with that institution maintaining a substantial margin of discretion. Those guidelines display flexibility in a number of ways, enabling the Commission to exercise its discretion in accordance with the provisions of Regulation No 1/2003, as interpreted by the EU Courts (see, to that effect, judgment of 11 July 2013, Team Relocations and Others v Commission, C 444/11 P, not published, EU:C:2013:464, paragraph 96 and the case-law cited), and with other rules and principles of Union law. In particular, point 13 of the 2006 Guidelines itself states that the Commission must 'normally' use the sales made by the undertaking during the last full business year of its participation in the infringement for the calculation of the amount of the basic fine (see, to that effect, judgment of 9 September 2015, Samsung SDI and Others v Commission, T 84/13, not published, EU:T:2015:611).

437 In the present case, it is apparent from recitals 1490 to 1495 of the contested decision that, in order to determine the basic amount of the fine imposed jointly and severally on the applicant and on Deutsche Telekom, the Commission took account of the sales made by the applicant during the last full business year of its participation in the infringement, namely the turnover achieved by that operator on the market for access to unbundled local loops for fixed retail broadband in 2010. The Commission thus applied point 13 of the 2006 Guidelines.

438 The applicant cannot be followed when it claims that the Commission committed a manifest error of assessment by not deviating from that rule in the present case, despite the sharp increase in its turnover during the period under consideration.

439 First, although the applicant claims that, between 2005 and 2010, its turnover increased by 133%, from EUR 31 184 949 to EUR 72 868 176, it nevertheless does not put forward any evidence capable of establishing that that latter turnover, achieved over the last full calendar year of the infringement, did not constitute, at the time when the Commission adopted the contested decision, an indication of its true size, of its economic power on the market and of the scope of the infringement at issue.

440 Secondly, the applicant cannot be followed when it complains that the Commission ignored its Polish Telecommunications decision and, as a result, disregarded its previous practice and imposed a criterion different from that provided for in point 13 of the 2006 Guidelines.

441 It is apparent from settled case-law that the Commission's practice in previous decisions does not itself serve as a legal framework for the fines imposed in competition matters and that decisions in other cases can give only an indication for the purpose of determining whether there might be discrimination since the facts of those cases, such as markets, products, the undertakings and periods concerned, are not likely to be the same (see judgments of 24 September 2009, Erste Group Bank and Others v Commission, C 125/07 P, C 133/07 P and C 137/07 P, EU:C:2009:576, paragraph 233 and the case-law cited; of 16 June 2011, Heineken Nederland and Heineken v Commission, T 240/07, EU:T:2011:284, paragraph 347; and of 27 February 2014, InnoLux v Commission, T 91/11, EU:T:2014:92, paragraph 144).

442 Therefore, previous decisions by the Commission imposing fines can be relevant from the point of view of observance of the principle of equal treatment only where it is demonstrated that the facts of the cases in those other decisions, such as markets, products, the countries, the undertakings and periods concerned, are comparable to those of the present case (see judgments of 13 September 2010, Trioplast Industrier v Commission, T 40/06, EU:T:2010:388, paragraph 145 and the case-law cited; of 29 June 2012, E.ON Ruhrgas and E.ON v Commission, T 360/09, EU:T:2012:332, paragraph 262 and the case-law cited; and of 9 September 2015, Philips v Commission, T 92/13, not published, EU:T:2015:605, paragraph 204 and the case-law cited).

443 In the present case, the applicant has not put forward any evidence capable of establishing that the facts of the case which gave rise to the Polish Telecommunications decision were comparable to those of the present case. The Commission stated in its pleadings that, in that case, it had taken into account the average of the turnover of 2005 to 2009 on the ground that the relevant turnover of the period concerned had grown exponentially, namely an increase of 2 800% for the period from 2006 to 2007, an increase of 370% for the period from 2007 to 2008 and an increase of 160% for the period from 2008 to 2009. It follows from that turnover, the correctness of which is not called into question by the applicant, that, first, the rate of growth of the turnover was much higher in the case giving rise to the Polish Telecommunications decision than that of the applicant's turnover in the present case and, secondly, that that turnover evolved in a less stable way than the turnover observed in the present case.

444 It follows from the above that, by taking into account in the present case the turnover achieved by the applicant during the year ending 31 December 2010, namely the last full business year of participation in the infringement, and by thus complying with its own rule set out in point 13 of the 2006 Guidelines, the Commission did not exceed the limits of its discretion concerning the determination of the amount of fines.

445 The first part of the fifth plea in law must therefore be rejected as unfounded.

(b) Second part, alleging a manifest error of assessment relating to the date on which the infringement period commenced

446 By its second part, the applicant claims that the contested decision is vitiated by a manifest error of assessment in so far as it finds the existence of an infringement commencing on 12 August 2005, the date on which the reference offer was published. As a framework contract, that offer was meant to evolve, particularly in the course of negotiations with third parties or following advice provided by national regulatory authorities.

447 More particularly, the applicant notes, first, that that offer was the first the applicant had ever drawn up, which made clarifications and amendments through negotiations all the more necessary.

448 Secondly, the Commission's position according to which that infringement starts with the publication of the reference unbundling offer is not consistent with its decision-making practice. By way of example, in its Decision C(2004) 1958 final, of 2 June 2004 (Case COMP/38.096 - Clearstream) ('the Cleastream decision'), the Commission concluded that Clearstream had abused its dominant position by constructively refusing to supply Euroclear with primary clearing and settlement services for registered shares. The Commission, nevertheless, acknowledged that it was necessary to give the parties some time to enable them to negotiate the terms and conditions of the contracts (recital 341 of the Clearstream decision). Likewise, in its Polish Telecommunications decision, the Commission did not take the date of publication of the reference offer as the starting date of the infringement, but the day of the beginning of its first negotiations with the other operators.

449 The applicant states that a refusal to supply infringement can only exist after access negotiations have failed due to the unreasonableness of the conditions set by the network holder. Furthermore, according to the applicant, the onus was on the Commission to prove at what point in time the negotiations failed due to the applicant's unreasonable demands. Moreover, account needs to be taken of the fact that access negotiations are by definition long and cumbersome given the complexity of the matter.

450 In the alternative, the applicant considers that the alleged refusal to grant access starts either at the end of a reasonable period that granting access would normally require involving preparations from both sides (recital 341 of the Clearstream decision) or on the start date of the initial access negotiations with alternative operators (recital 909 of the Polish Telecommunications decision).

451 The Commission, supported by the intervener, disputes those arguments.

452 In that regard, it is not disputed that, by his decision of 14 June 2005, the Chairman of TUSR required the applicant to provide unbundled access to its local loop under fair and reasonable conditions, and that it is with a view to fulfilling that obligation that the applicant published, on 12 August 2005, a reference unbundling offer (see paragraphs 9 and 10 above).

453 Moreover, the applicant does not dispute the description of the contents of its reference offer made in section 7.6 of the contested decision ('ST's unfair terms and conditions'), according to which the Commission concluded, in recital 820 of that decision, that the terms and conditions of that offer had been set so as to render unbundled access to the local loop unacceptable for the alternative operators.

454 It is apparent from that part of the contested decision that the abusive practices which are there classified as 'refusal to supply' by the Commission resulted, essentially, from the reference offer itself.

455 Therefore, as regards, firstly, the lack of disclosure to alternative operators of information relating to the applicant's network, which is necessary for the unbundling of the local loop, it follows first of all from recital 439 of the contested decision that the Commission considered that the reference offer did not contain the basic information regarding the locations of physical access sites and the availability of local loops in specific parts of the access network. Furthermore, in recitals 443 to 528 of the contested decision, the Commission admittedly examined the information relating to the network provided by the applicant at the request of an alternative operator with a view to unbundling. However, it follows also from that part of the contested decision that the arrangements for access to such information which were considered by the Commission to be unfair and, therefore, dissuasive for alternative operators resulted from the reference offer itself. The Commission in particular criticised the fact, in the first place, that the reference offer had not determined the exact scope of the information relating to the network that the applicant would place at the disposal of alternative operators, by specifying the categories of information concerned (recital 507 of the contested decision), in the second place, that that offer provided access to information from the non-public information systems only after the conclusion of the framework agreement on access to the local loop (recital 510 of the contested decision) and, in the third place, that that offer made such access to information relating to the applicant's network subject to the payment by the alternative operator of large fees (recitals 519 and 527 of the contested decision).

456 As regards, secondly, the reduction of the scope of its regulatory obligation concerning unbundled access to the local loop, first of all, it follows from recitals 535 and 536 of the contested decision that the restriction of that obligation to active lines only (see paragraph 32 above), alleged of the applicant by the Commission, resulted from paragraph 5.2 of the introductory part of its reference offer. Next, it follows in particular from recitals 570, 572, 577, 578 and 584 of the contested decision that it is in the light of stipulations contained in Annex 3 to that reference offer that the Commission inferred that the applicant had unjustifiably excluded conflicting services from its obligation relating to unbundled access to the local loop (see paragraph 33 above). Finally, it follows from recital 606 of the contested decision that the 25% cable use limitation rule imposed by the applicant for unbundled access to the local loop and regarded by the Commission as unjustified (see paragraph 34 above), resulted from Annex 8 to the reference offer.

457 As regards, thirdly, the setting by the applicant of unfair conditions relating to unbundling regarding collocation, qualification, forecasting, repairs, service and maintenance and bank guarantee, they all resulted, as is shown in section 7.6.4 of the contested decision, from the reference offer published by that operator on 12 August 2005. Also, the terms which are considered by the Commission to be unfair were included respectively in Annexes 4, 5, 14 and 15 to that offer as regards collocation (recitals 653, 655 and 683 of the contested decision), in Annexes 12 and 14 as regards the alternative operators' forecasting obligation (recitals 719 and 726 to 728 of the contested decision), in Annex 5 as regards the qualification procedure of the local loops (recitals 740, 743, 767, 768 and 774 of the contested decision), in Annex 11 as regards the terms and conditions relating to repairs, service and maintenance (recitals 780, 781, 787, 790 and 796 of the contested decision), and in Annexes 5 and 17 as regards the bank guarantee required from the alternative operator which is a candidate for unbundled access (recitals 800, 802 to 807, 815 and 816 of the contested decision).

458 It follows that, even assuming that some of those terms of access were liable to be relaxed in the context of bilateral negotiations between the applicant and operators seeking access, which the applicant merely affirms without corroborating evidence, the Commission correctly concluded that the reference offer published on 12 August 2005 was capable of dissuading as from that date applications for access made by alternative operators, due to the unfair terms and conditions which it contained.

459 In those circumstances, the Commission did not commit an error in finding that, as a result of the terms of access in its reference offer published on 12 August 2005, the applicant had compromised the entry of alternative operators onto the mass (or general public) retail market for broadband services at a fixed location in Slovakia, in spite of the obligation imposed on it in that regard under TUSR's decision, and that that conduct was therefore capable of having such negative effects on competition from that date (see, inter alia, recitals 1048, 1050, 1109, 1184 and 1520 of the contested decision).

460 That conclusion is not undermined by the applicant's claim that the Commission infringed its own decision-making practice, namely the approach adopted in the Clearstream decision and in the Polish Telecommunications decision. It suffices to note that those decisions were made in a context which is different from that of the present case and that they are therefore not capable of establishing that in the contested decision the Commission diverged from its previous decision-making practice.

461 Therefore, in the first place, as regards the Clearstream decision, it suffices to note that that decision, unlike the contested decision in the present case, was taken in a context characterised by the absence of any regulatory obligation for the undertaking which owns the infrastructure at issue to grant other undertakings access to that infrastructure, and by the absence of an obligation imposed on that undertaking to publish a reference offer setting out the terms and conditions of such access.

462 Moreover, the period of four months, which the General Court considered to correspond to the reasonable time limit for the provision of the primary clearing and settlement services by Clearstream, had been established by comparing the examples in which Clearstream granted access to its Cascade RS system. Therefore, it should be noted that, in that case, there were several examples in which Clearstream had granted access, which allowed the Commission, and later the Court, to reach the conclusion that the period of four months was reasonable to grant such access (judgment of 9 September 2009, Clearstream v Commission, T 301/04, EU:T:2009:317, paragraph 151). However, in the present case, since the applicant gave access to its local loops only to one other operator on 18 December 2009, there was no example which could serve as a reference, so that the Commission could not fix such a 'reasonable period'. It follows that the circumstances of the present case are not comparable with those of the case giving rise to the judgment of 9 September 2009, Clearstream v Commission (T 301/04, EU:T:2009:317).

463 In the second place, as regards the Polish Telecommunications decision, the Commission found that the incumbent operator at issue had abused its dominant position on the Polish wholesale market for broadband access and unbundled access to the local loop, by refusing to give access to its network and to supply wholesale products from those markets in order to protect its position on the retail market. Moreover, the context of the Polish Telecommunications case was characterised by a regulatory obligation to grant access similar to that imposed on the applicant in the present case, and by the obligation imposed on the Polish Telecommunications operator at issue to publish a reference offer for unbundled access to its local loop. Nevertheless, it is apparent from a detailed analysis of the Polish Telecommunications decision that the approach taken in that decision is not inconsistent with that taken in the contested decision. In the Polish Telecommunications decision, the Commission noted that the dominant operator's anticompetitive strategy essentially materialised only in the course of negotiations with alternative operators which were candidates for the grant of unbundled access to the local loop and wholesale access to the dominant operator's broadband services. Therefore, the unreasonable access conditions resulted from proposals for access contracts made by the dominant operator at issue in the context of negotiations with alternative operators. Furthermore, the delay in the process of negotiating access agreements could not, by hypothesis, have been identified upon the publication of the dominant operator's first reference offer. In addition, the limitation of access to its network carried out by the dominant operator took place at a later stage than the conclusion of the wholesale access agreements with the alternative operators. Moreover, the limitation of effective access to subscriber lines took place after the alternative operator concerned obtained access to a collocation space or the authorisation to install a correspondence cable. Finally, the problems of access to reliable and accurate general information which is necessary for alternative operators to take decisions relating to access were manifest at every step of the procedure for access to the dominant operator's network. The conduct of the dominant operator in the Polish Telecommunications case was therefore different from the practices which were classified as 'refusal to supply' by the Commission in the contested decision, practices which, as is apparent from the analysis in paragraphs 455 to 459 above, resulted essentially from the reference offer for unbundled access to the local loop of the applicant itself. In contrast to the Polish Telecommunications decision in which the starting point of the infringement of Article 102 TFEU was fixed at the date on which the first access negotiations between the dominant operator at issue and an alternative operator had begun, several months after the publication of the first reference offer (recital 909 and the footnote to page 1259 of the contested decision), those differences justify the Commission identifying, in the present case, 12 August 2005, that is to say the date of publication of the reference offer, as the date of commencement of the implied refusal of access to the local loop.

464 For the same reason, it is necessary to reject the applicant's argument that the refusal to supply infringement can be established only after the access negotiations have failed due to the unreasonableness of the conditions set by the network holder. In addition, it is not certain that the negotiations could have led to the unfair terms and conditions in the reference offer being removed.

465 Concerning the applicant's allegation that the Commission must bear the burden of proof in relation to the point in time the negotiations failed due to the applicant's unreasonable demands, first, for the same reasons as those indicated in paragraphs 461 to 464 above, that date cannot be relevant in order to determine the start of the infringement. Secondly, as the intervener contended, the exact date of the failure of the negotiations cannot be objectively determined, so that the Commission is not required to provide such proof.

466 As regards the arguments presented in the alternative, in so far as the applicant considers that the alleged refusal to grant access should start at the end of a reasonable period that granting access would normally require involving preparations from both sides (recital 341 of the Clearstream decision), it should be noted that such a reasonable period does not exist in the present case for the reasons set out in paragraphs 460 to 462 above. Therefore, that argument must be rejected. In so far as the applicant intends the infringement to start on the date of the initial access negotiations with alternative operators (recital 909 of the Polish Telecommunications decision), as was considered, in essence, in paragraphs 463 and 464 above, the negotiations were not relevant for the purposes of determining the start of the infringement in this case. Consequently, that argument must also be rejected.

467 The second part, alleging an error committed by the Commission when it found that the implied refusal of access to the local loop began on 12 August 2005, must therefore be rejected as unfounded.

468 It should be added that the applicant does not dispute the classification of a single and continuous infringement adopted by the Commission with respect to all of the practices mentioned in Article 1(2) of the contested decision, namely (a) withholding from alternative operators network information necessary for the unbundling of local loops; (b) reducing the scope of its obligations regarding unbundled local loops; (c) setting unfair terms and conditions in its reference unbundling offer regarding collocation, qualification, forecasting, repairs and bank guarantees; (d) applying unfair tariffs which do not allow an equally efficient competitor relying on wholesale access to the unbundled local loops of the applicant to replicate the retail broadband services offered by the applicant without incurring a loss.

469 In those circumstances, and in so far as the second part of the present plea in law, alleging an error committed by the Commission when it found that the implied refusal of access to the local loop began on 12 August 2005, was rejected (see paragraph 467 above), the Commission was justified in finding that the single and continuous infringement which is the object of the contested decision had started on 12 August 2005.

470 Therefore, it is necessary to reject the fifth plea in law in its entirety.

471 It follows from all the foregoing that Article 1(2)(d) of the contested decision must be annulled in so far as it declares that, during the period between 12 August and 31 December 2005, the applicant applied unfair tariffs which do not allow an equally efficient competitor relying on wholesale access to its unbundled local loops to replicate the retail broadband services offered by it without incurring a loss (see paragraph 268 above). As a result, Article 2 of that decision must also be annulled in so far as it concerns the applicant. The remainder of the claim for annulment of the contested decision must be dismissed.

IV. The conclusions, put forward in the alternative, seeking variation of the amount of the fine

472 The applicant also requests the Court, in the alternative, to reduce the amount of the fines which were imposed on it by the contested decision.

473 It should be noted, in that regard, that, according to settled case-law, the review of legality provided for in Article 263 TFEU entails the EU judicature conducting a review, in respect of both the law and the facts, of the contested decision in the light of the arguments relied on by an applicant, which means that it has the power to assess the evidence, annul the decision and to alter the amount of the fines (see, to that effect, judgments of 3 September 2009, Prym and Prym Consumer v Commission, C 534/07 P, EU:C:2009:505, paragraph 86 and the case-law cited; of 26 January 2017, Duravit and Others v Commission, C 609/13 P, EU:C:2017:46, paragraph 30 and the case-law cited; and of 27 March 2014, Saint-Gobain Glass France and Others v Commission, T 56/09 and T 73/09, EU:T:2014:160, paragraph 461 and the case-law cited).

474 The review of legality is supplemented by the unlimited jurisdiction which Article 31 of Regulation No 1/2003, in conjunction with Article 261 TFEU, confers on the EU judicature. That jurisdiction empowers the Courts, beyond carrying out a mere review of legality with regard to the penalty, to substitute their own appraisal for the Commission's and, consequently, to cancel, reduce or increase the fine or penalty payment imposed (judgments of 8 December 2011, Chalkor v Commission, C 386/10 P, EU:C:2011:815, paragraph 63, and of 8 December 2011, KME Germany and Others v Commission, C 389/10 P, EU:C:2011:816, paragraph 130; see, also, judgment of 26 January 2017, Duravit and Others v Commission, C 609/13 P, EU:C:2017:46, paragraph 31 and the case-law cited).

475 It should be noted that the exercise of unlimited jurisdiction does not amount to a review of the Court's own motion, and that proceedings before the Courts of the European Union are inter partes. Therefore, with the exception of pleas involving matters of public policy which the Courts are required to raise of their own motion, it is for the applicant, in principle, to raise pleas in law against the decision under appeal and to adduce evidence in support of those pleas (see, to that effect, judgment of 10 July 2014, Telefónica and Telefónica de España v Commission, C 295/12 P, EU:C:2014:2062, paragraph 213 and the case-law cited).

476 It is in the light of those principles that it is necessary to determine whether the amount of the fines imposed by the Commission in the contested decision must be modified.

477 As is apparent from paragraphs 267, 268 and 471 above, the Commission has not provided evidence that that practice leading to a margin squeeze committed by the applicant could have started before 1 January 2006 and, consequently, Article 1(2)(d) of the contested decision must be annulled to the extent that it concerns the applicant and that it includes a margin squeeze which was committed between 12 August and 31 December 2005 in the single and continuous infringement.

478 As regards the impact of that error on the basic amount of the fine for which the applicant is jointly and severally liable, the Court considers, in the exercise of its unlimited jurisdiction, that it is necessary to reduce the proportion of the applicant's relevant sales applied by the Commission and to establish that proportion at 9.8% instead of 10%. Since the applicant achieved over the course of the last full year of the infringement a relevant turnover of EUR 72 868 176, the amount which must be used to calculate the basic amount of the fine for which the applicant is jointly and severally liable is EUR 7 141 081.20. The basic amount of that fine corresponds to the multiplication of that amount by a coefficient of 5.33, reflecting the duration of the infringement, and must therefore be set at EUR 38 061 963. The applicant's application for the amount of the fine to be reduced is dismissed as to the remainder.

479 As regards the Commission's application, brought in the alternative during the hearing, for the amount of the fine imposed jointly and severally on the applicant and on Deutsche Telekom to be increased, the Court considers, without it even being necessary to rule on the admissibility of such an application, that, in the light of the circumstances of the case, it is not necessary to modify the amount fixed in paragraph 478 above.

V. Costs

480 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. Furthermore, under Article 134(3) of the Rules of Procedure, the parties are to bear their own costs where each party succeeds on some and fails on other heads. However, if it appears justified in the circumstances of the case, the Court may order that one party, in addition to bearing its own costs, pay a proportion of the costs of the other party. In the present case, the Commission and the intervener have failed on some heads. Nevertheless, the applicant has not applied for the intervener to be ordered to pay the costs, but only for the Commission to be so ordered.

481 In those circumstances, the applicant must be ordered to bear four fifths of its own costs, as well as four fifths of the costs of the Commission and the intervener, in accordance with the form of order sought by them. The Commission is to bear one fifth of its own costs and of those of the applicant. The intervener is to bear one fifth of its own costs.

On those grounds,

THE GENERAL COURT (Ninth Chamber, Extended Composition)

hereby:

1. Annuls Article 1(2)(d) of Commission Decision C(2014) 7465 final of 15 October 2014 relating to proceedings under Article 102 TFEU and Article 54 of the EEA Agreement (Case AT.39523 - Slovak Telekom) in so far as it declares that, over the course of the period between 12 August and 31 December 2005, Slovak Telekom, a.s., imposed unfair tariffs which do not allow an equally efficient operator relying on wholesale access to its unbundled local loops to replicate the retail broadband services offered by it without incurring a loss;

2. Annuls Article 2 of Decision C(2014) 7465 final in so far as it fixes the amount of the fine imposed jointly and severally on Slovak Telekom at EUR 38 838 000;

3. Fixes the amount of the fine imposed jointly and severally on Slovak Telekom at EUR 38 061 963;

4. Dismisses the action as to the remainder;

5. Orders Slovak Telekom to bear four fifths of its own costs, four fifths of the costs of the European Commission and four-fifths of the costs of Slovanet, a.s.;

6. Orders the Commission to bear one fifth of its own costs and one fifth of the costs incurred by Slovak Telekom;

7. Orders Slovanet to bear one fifth of its own costs.

Van der Woude Gervasoni Madise

da Silva Passos Kowalik-Banczyk

Delivered in open court in Luxembourg on 13 December 2018.

E. Coulon S. Gervasoni

Registrar President

* Language of the case: English.

1 Only the paragraphs of the present judgment which the Court considers it appropriate to publish are reproduced here.