Irish Court of Appeal

THE COURT OF APPEAL

Record Number: 2022/189

High Court Record Number: 2021/9 MCA

Neutral Citation Number: [2022] IECA 300

Haughton J.

Collins J.

Allen J.

 

BETWEEN

THREE IRELAND (HUTCHISON) LIMITED

AND

THREE IRELAND SERVICES (HUTCHISON) LIMITED

Appellants/Respondents

AND

COMMISSION FOR COMMUNICATIONS REGULATION

                                                Respondent/Appellant

AND

VODAFONE IRELAND LIMITED

EIR LIMITED

TESCO MOBILE IRELAND LIMITED

Notice Parties

 

JUDGMENT of Mr. Justice Maurice Collins delivered on 21 December 2022

PRELIMINARY

 

1.                  This judgment concerns the appeal of Commission for Communications Regulations (“ComReg”) against the Judgment and Order of the High Court (McDonald J) of 21 July 2022 staying the commencement by ComReg of the “Main Stage” of the auction process provided for in ComReg Decision D11/20 of 18 December 2020 [1] (“the Decision”) and ComReg’s Information Memorandum and Draft Regulations of 16 April 2021 [2] pending the determination of the appeal brought by Three Ireland (Hutchison) Limited and Three Ireland Services (Hutchison) Limited (collectively “Three”) from that Decision or until further order.

 

2.                  That stay was granted by the High Court pursuant to Regulation 7(2) of the European Communities (Electronic Communications Networks and Services) (Framework) Regulations 2011 (hereafter, “the Framework Regulations”). [3] The application for a stay - which was vigorously contested by ComReg - was heard over two days in the High Court (on 7 and 8 July 2022) with the Judge giving a detailed judgment on 20 July 2022. I will make further reference to his judgment below.

 

3.                  ComReg’s appeal was given an early hearing in this Court and was listed for a one day hearing on 19 October 2022. It did not conclude that day however and the hearing resumed, and concluded, on 25 October 2022.

 

4.                  On 8 November 2022, the Court indicated to the parties that it had decided that the stay granted by the High Court should be varied so as to permit ComReg to commence the “Main Stage” of the auction process and to complete the auction process up to (but not including) notifying the Winning Bidders of their entitlement to apply for licences as provided for in paragraph 3.259 of the Information Memorandum, with ComReg being restrained from taking that or any subsequent steps in the auction process pending the determination of the proceedings in the High Court and/or further Order. The Court indicated that it would give its reasons for its decision later. This judgment sets out my reasons for that decision.

 

 

COMREG DECISION D11/20

 

5.                  ComReg is a statutory body established by Part 2 of the Communications Regulation Act 2002 (to which, as amended, I shall refer as “the 2002 Act”). ComReg has extensive and important functions in the area of telecommunications regulation: section 10 of the 2002 Act. Section 12 of the 2002 Act identifies the objectives of ComReg in carrying out those functions, which include the promotion of competition and the promotion of the interests of users of electronic communications within the EU. ComReg is the national regulatory authority (NRA) for the purposes of the many EU legislative measures that, to a significant extent, harmonise telecommunications regulation within the European Union and that are now consolidated in Directive (EU) 2018/1972 of the European Parliament and of the Council of 11 December 2018 establishing the European Electronic Communications Code (Recast) (“the Recast Directive”)

 

6.                  One of ComReg’s functions is to manage the radio frequency spectrum and make decisions regarding the allocation of rights of use of radio spectrum for electronic communications services. [4] Ensuring the efficient management and use of the radio frequency spectrum is one of ComReg’s statutory objectives under section. It is common case that decisions relating to spectrum allocation have significant implications for service providers, users and for the wider economy.

 

7.                  Following a very lengthy consultation process, Decision D11/20 was adopted by ComReg in December 2020. The Decision is concerned with the allocation of long-term rights of use across four spectrum bands, all of them suitable for mobile and wireless broadband (WBB). [5] According to ComReg (and this is not in dispute) this spectrum - and in particular the 700 MHz Duplex [6] - is critically important to the rollout of 5G mobile services in the State. That, in turn, is said by ComReg to be critically important to the future economic welfare of the State. Again, that is not in controversy. [7]

 

8.                  The Decision is lengthy and complex. Including the Response to Consultation, the formal Decision Document and the many annexes, it runs to more than 900 pages. It references many previously published ComReg documents addressing various aspects of the award process. What follows is a brief, and inevitably simplistic, summary of its principal aspects insofar as they appear to be relevant to this appeal.

 

9.                  As already mentioned, the Decision is concerned with the allocation of long-term rights of use across four spectrum bands, the 700 MhZ Duplex, 2.1 GHz, 2.3 GHz and 2.6 GHz Bands. These bands are harmonised at EU level for the provision of WBB services.

 

10.              ComReg proposes to award 470MHz of spectrum rights in aggregate, which represents a significant increase in the spectrum assigned to WBB services. It includes 6 “blocks” of 700 MHz Duplex spectrum (each block consisting of 2 x 5 MHz of spectrum).

 

11.              Sub-1 GHz spectrum (spectrum in the 700 MHz Duplex, 800 MHz and/or 900 MHz bands) is critical to the provision of mobile telephony services in Ireland because low frequency spectrum enables operators to achieve wide-area geographic coverage, including in rural areas and areas of low density population, on a financially sustainable basis. Otherwise, the cost of achieving such coverage would be prohibitive. The 700 MHz band is described in the Decision as a “coverage band”.

 

12.              Higher frequency spectrum is typically used to provide/improve capacity and performance and the 2.1 GHz, 2.3 GHz and 2.6 GHz Bands are described in the Decision as “performance bands”.

 

13.              There are three mobile network operators (“MNOs”) in the State, Three, Vodafone and Eir. Each already holds sub-1 GHz spectrum, with Three holding the largest amount (50 MHz (5 x 2 x 5 MHz blocks) as against 40 MHz (4 blocks) held by Vodafone and Eir). Three also holds more spectrum overall (i.e. including supra 1 GHz spectrum) than Vodafone or Eir (280 MHz as against 225 MHz (Vodafone) and 185 MHz (Eir)).

 

14.              It is evident from the Decision that ComReg had concerns that the award process might result in “extreme asymmetries" that could damage competition. In particular, ComReg was concerned that Three and Vodafone could engage in “strategic bidding” that might result in Eir being denied further 700 MHz spectrum which would undermine its capacity to compete effectively in the mobile market. As a result, ComReg decided to impose a spectrum cap of 70 MHz (7 blocks) in the sub-1 GHz band. That spectrum cap included the MNOs’ existing holdings. Given that Three already holds 50 MHz (5 blocks) of sub-1 GHz spectrum, the effect of that cap was to limit the sub-1 GHz spectrum that Three could win to 20 MHz or 2 additional blocks, whereas Vodafone and Eir would be permitted to win up to 30 MHz or 3 blocks. In effect, the auction rules preclude Three from bidding for more than 2 blocks, a point of some importance in its appeal. I will refer to this spectrum cap as “the Sub-1 GHz spectrum cap”. As will appear, that cap is challenged by Three.

 

15.              ComReg also decided to impose an aggregate spectrum cap of 375MHz across the 700 MHz, 800 MHz, 900 MHz, 1800 MHz, 2.1 GHz, 2.3 GHz, 2.6 GHz and 3.6 GHz Bands. That cap is uncontroversial and it will not be necessary to say anything further about it.

 

16.              Certain other aspects of the award process must also be noted. The Decision envisaged a number of stages in the process and further detail was given about these in ComReg’s Information Memorandum and Draft Regulations which was published on 16 April 2021 (“the Information Memorandum”). After the application stage (in which applicants were required to make initial bids), there was to be a qualification stage, in which the qualification of applicants to participate in the award process as bidders would be determined. The qualification stage would also establish whether an auction was necessary, on the basis of the demand expressed in the initial bids. In the event that demand exceeded supply, there would be a “Main Stage”, involving a Combinatorial Clock Auction (“CCA”).

 

17.              Auction theory is a complex field, with a significant literature underpinning it. It is concerned with how bidders act in an auction and how the auction process may be designed so as to maximise the prospects of achieving the objectives of the seller (which may not always be simply to achieve the highest price). For a number of reasons, spectrum auctions present particular challenges and a number of different auction formats have been developed specifically for such auctions. The Simultaneous Multiple Round Auction (“SMRA”) format was invented in the 1990s by two Stanford economists, Robert Wilson and Paul Milgrom and has since been used around the world (including by ComReg). Milgrom was also involved in the subsequent development of the Combinatorial Clock Auction format. The CCA auction format had also been used by Governments and regulators around the world (including by ComReg) since the mid-late 2000s.

 

18.               A combinatorial auction is one in which the participants bid on combinations of items or packages - here, packages of lots of spectrum. That is the fundamental difference between the CCA auction format and SMRA auction format, in which bidders bid for individual lots of spectrum. The CCA format involves two stages. I take the following from the Affidavit of Niamh Hodnett grounding Three’s appeal:

 

“60….. The first stage (the ‘Clock Rounds’) progresses in a similar manner to the SMRA, with bidding proceeding over multiple rounds with rising prices and ends once demand in all categories has been reduced to a level that is equal [to] or less than supply. Bidders then have an opportunity to submit additional bids in a second stage (known as the ‘Supplementary Round’) which is a single round sealed bid. The bid amounts that bidders can express for packages in the Supplementary Round are constrained by activity rules, which require a certain consistency with their stated preferences between packages in the Clock Rounds. In bidding in the Supplementary Round, bidders have an insight into their rivals’ incentives from the bids made in the Clock Rounds.

 

61. In the CCA, the spectrum is then allocated to the highest value combination of package bids, taking at most one package bid from each bidder. Unlike in an SMRA, winning bidders do not necessarily pay the amount of their winning bid but rather pay an amount that may be less than or equal to this level, as determined by the ‘Second Price Rule’  ..” [8]

 

19.              The “Second Price Rule” is a standard feature of CCA auctions. In simple terms, it involves the successful bidder paying the price bid by the second placed bidder. That price represents the “opportunity cost” of denying the second placed winner. Again, there is a substantial literature exploring the rationale for the use of the second-price rule in auctions which, happily, it is not necessary to consider here.

 

20.              One further aspect of the auction process should be mentioned. It is touched upon by Ms Hodnett in the passage set out above, where she refers to bidders gaining an insight into their rivals’ incentives from the bids made in the “Clock Rounds”. The CCA auction format requires bidders to reveal their valuation for packages of spectrum as the auction progresses. That process of “price discovery” is an intended and important feature of the auction design. As the Information Memorandum explains, the “open, multiple Round, structure of the Primary Bid Rounds is intended to allow Bidders to learn about Aggregate Demand for Lots and to provide an opportunity for Bidders to revise their assessment of the value of Lots in light of this information.” [9] There appears to be no dispute but that such price discovery assists in ensuring that the spectrum being auctioned is appropriately valued. Indeed, as is noted below, one of the complaints made by Three in its substantive appeal is that the spectrum cap will operate to subvert the process of price discovery.

 

21.              But, Three says, the price discovery process in the Main Stage will involve the disclosure of commercially sensitive information about bidder valuations and strategies and, it says, once the Main Stage is conducted the auction process cannot be re-run or any similar bid process conducted, without undermining the integrity of the process as bidders will have an awareness of each other’s bidding strategies. That is strenuously contested by ComReg.

 

22.              Finally, I should explain that the Decision provides for the imposition on successful bidders of very onerous network coverage and roll-out obligations (Decision, section 8.4) and quality of service obligations (Decision, section 8.5). Compliance with these obligations will require very substantial investment by the successful bidders, of a scale - so ComReg says - many multiples of the cost of acquiring the spectrum in the first place. Again, that does not appear to be a matter of controversy.

 

THREE’S APPEAL

 

23.              Regulation 4(1) of the Framework Regulations provides that a user or undertaking affected by a decision made by ComReg may appeal to the High Court against the decision. Regulation 6(1) then provides that, having heard and determined such appeal, the court “may make such orders as it considers appropriate” and Regulation 6(2) provides that the orders that may be made include an order affirming or setting aside the whole or any part of the Regulator’s decision and an order “remitting the case to the Regulator to be reconsidered, either with or without the hearing of further evidence, in accordance with the directions of the High Court.”

 

24.              The Framework Regulations gave effect in the State to 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 (the “Framework Directive”). Regulation 4(1) of the Framework Regulations gave effect to the provisions of Article 4(1) of the Framework Directive, which required Member States to ensure that users or undertakings affected by a decision taken by the national regulatory authority had a right of appeal against that decision. Article 4(1) was subsequently amended by Directive 2009/140/EC (the “Better Regulation Directive”), as further explained below. [10] Thus the right of appeal exercised here by Three has its origins in Union law. [11]

 

25.              Regulation 4(1) of the Framework Regulations is silent as to the scope of the appeal provided by it and the standard of review. Those issues were closely analysed by the High Court (Cooke J) in Vodafone Ireland Ltd v Commission for Communications Regulations [2013] IEHC 382 and he expressed his conclusions thus:

 

“32  Drawing all of these observations and considerations together, it seems to the Court that its approach to determination of an appeal under Regulation 4 should be as follows. If it is established that the decision under appeal is vitiated by a material error of law including a significant failure to comply with a mandatory requirement of the Regulations or by a misinterpretation or misapplication of the Regulations, the Court can and should intervene to make an order appropriate to the effect of that deficiency. It can and should also intervene where it is established that the decision is vitiated by a serious and significant error or series of errors of the kind described by Keane C.J. in Orange Limited. Having regard to the apparent purpose of the appeal in requiring the merits to be taken into account, the Court is also obliged to consider whether the decision is "wrong" in the sense contended for by an appellant. To be wrong in that sense, however, the Court must be satisfied that there has been a serious, significant and material mistake such that the operation or implementation of the decision as it stands would be manifestly unreasonable, disproportionate or incompatible with the outcome sought to be achieved by the exercise of the regulatory remedies which ComReg is entitled to impose. …”

 

26.              As already noted, the Framework Directive has been repealed and replaced by the Recast Directive. Article 31 of the Recast Directive substantially re-enacts Article 4 of the Framework Directive. Certain parts of the Recast Directive, including Article 31, are to be implemented by the Communications (Regulation) Bill 2022 when enacted. The Bill is currently before the Dáil. Clause 17 of the Bill will implement Article 31. Clause 17 is drafted in more detailed and prescriptive terms than Regulation 4(1) of the Framework Regulations and Clause 17(15) appears, in substance, to enshrine the “serious and significant error” standard applied in Vodafone Ireland Ltd v Commission for Communications Regulations (and which has its origins in the decision of the Supreme Court in Orange Communications Limited v Director of Telecommunications Regulation (No 2) [2000] 4 IR 159).

 

27.              In any event, on 14 January 2021 Three appealed against the Decision. Three’s Originating Notice of Motion sought a number of reliefs. The first was an order pursuant to Regulation 6 of the Framework Regulations setting aside those parts of ComReg’s formal Decision Document that provide for ComReg to (i) proceed with the proposed release of the award spectrum (para 3.1 of the Decision Document); (ii) grant licences to operators on receipt of an application (as provided for in para 3.6 of the Decision Document); (iii) select the parties who will be eligible to be granted licences by means of a competitive selective procedure by way of auction (as provided for in para 3.11 of the Decision Document) to make rights of use available (as provided for in paragraphs 3.13 and 3.14 of the Decision Document) subject to the elements particularised in para 3.15 of the Decision Document and (iv) incorporate into the competitive selection procedure the elements particularised in para 3.15 of the Decision Document.

 

28.              Paragraph 3.15 of the Decision Document (pages 499 - 502 of the Decision) encapsulates all of the principal elements of the proposed auction process, including (but not limited to) the CCA auction and the Sub-1 GHz spectrum cap and the first relief sought in the Three’s Notice of Motion thus effectively comprises a full-frontal challenge to the Decision.

 

29.              The notice of motion also seeks an order pursuant to Regulation 6 remitting the Decision to ComReg for reconsideration in accordance with the directions of the High Court. Three also seeks damages.

 

30.              Three’s substantive appeal is not of course before the Court but we have been provided with a considerable volume of material relating to it. In very broad terms, it involves the following specific grounds of challenge to the Decision:

 

·         Three says that ComReg made a serious error in deciding to impose the Sub-1 GHz spectrum cap. According to Three, the analysis in the Decision is insufficient to justify the cap and ComReg’s conclusion that the cap was necessary to avoid a distortion of competition has no adequate or plausible evidential basis.

 

·         The Sub-1 GHz spectrum cap constitutes an undue and disproportionate restriction on Three’s ability to bid for and obtain spectrum.

 

·         ComReg failed to give any or any adequate consideration to the combined effect of the Sub-1 GHz spectrum cap and CCA auction format and in particular the impact of that combination on Three’s competitive position.

 

·         The combination of the Sub-1 GHz spectrum cap and CCA auction format undermines the prospects of the auction process resulting in an efficient allocation of spectrum in that it gives a substantial and inefficient gain to Vodafone and Eir and increases the risk of inefficient strategic bidding by them, to the detriment (so it is said) of Three and contrary to the interests of ComReg and the public in the efficient allocation of spectrum.. That, Three says, is because Vodafone and Eir each stands to win a block of 700 MHz spectrum at the reserve price, which is lower than the real value of that spectrum and lower than the price that Three will have to pay for 700 MHz spectrum. It appears to be common case that, in the absence of a new entrant MNO - and it is accepted that there is little or no prospect of a new entrant MNO– Vodafone and Eir will each win a block of 700 MHz spectrum at the reserve price. ComReg says that, if Three’s complaint about this aspect of the auction rules is upheld, it can be compensated by being given an equivalent “discount” on the price that it pays. It will be necessary to refer to this issue further in discussing McDonald J’s analysis around the issue of “strategic bidding”.

 

It will in due course be necessary to look at certain of these grounds in more detail for the purpose of determining the stay appeal.

 

31.              Three’s appeal was opposed by ComReg. It was also opposed by Vodafone. After exchanges of affidavits and expert reports, the appeal was heard by O’Moore J in the High Court over six hearing days in June 2021. The hearing proceeded on affidavit, without oral evidence or cross-examination. Following the conclusion of the hearing, the parties, at the direction of O’Moore J, identified the particular portions of the evidence (affidavits, exhibits and expert reports) on which they sought to rely. O’Moore J then directed the parties to address the question of whether, on the assumption that the evidence was disputed, and that it was necessary for the court to decide on the disputed evidence, he could properly decide the appeal in the absence of cross-examination. [12] In response, Three made it clear that it was not asking the High Court to resolve any disputed evidence and was content to rely on “the text of the decision and the undisputed parts of ComReg’s own expert evidence”. [13] It follows that, for the purposes of its substantive appeal, Three does not seek to rely on its own affidavit evidence, including its expert evidence and reports, to the extent that such evidence is disputed.

 

32.              In November 2021, the High Court reserved judgment on the appeal. When Three’s stay application came before McDonald J in the High Court, it was unclear when O’Moore J would be able to give judgment. In advance of the hearing of this appeal, the Court suggested that the parties might ask O’Moore J whether he might be in a position to give an indication of when he anticipated giving judgment and the learned judge has indicated that he intends to give judgment on 31 January 2023.

 

THREE’S APPLICATION FOR A STAY

 

33.              Bringing an appeal to the High Court from a decision of ComReg does not of itself affect the operation of that decision or prevent its implementation: Regulation 7(1) of the Framework Regulations. However, Regulation 7(2) provides that where such an appeal is lodged:

 

“the High Court or a Judge of the High Court may make such order staying or otherwise affecting the operation or implementation of the decision of the Regulator, or a part of that decision, as the High Court or Judge of the High Court considers appropriate for the purpose of securing the effectiveness of the hearing and determination of the appeal.” (my emphasis)

 

34.              Again, these provisions of the Framework Regulations reflect provisions of the Framework Directive. When first adopted in 2002, the final sentence of Article 4(1) of the Directive had provided that “[p]ending the outcome of any such appeal, the decision of the national regulatory authority shall stand, unless the appeal body decides otherwise.” A new Article 4(1) was substituted by the Better Regulation Directive. It included a reworded provision regarding interim measures in the following terms: “[p]ending the outcome of the appeal, the decision of the national regulatory authority shall stand, unless interim measures are granted in accordance with national law.” [14]

 

35.              When it brought its appeal, Three did not seek any order staying any part of the Decision. In April 2021, ComReg published the Information Memorandum, which gave detailed guidance as to the how the auction process was to proceed, including a detailed indicative timeline up to and including qualification stage. [15] The application stage was duly completed, followed by the qualification stage. On 31 May 2022, ComReg indicated that a Main Stage would be required (in other words, demand exceeded the spectrum available so that the need for an auction was triggered) and on 3 June 2022 it gave notice to bidders that the Main Stage would commence on 11 July 2022. [16]  

 

36.              On 8 June 2022 Three moved in the High Court for a stay on the commencement of the Main Stage. Its application was grounded on an affidavit sworn by Tom Hickey, Three’s Head of Regulatory Affairs, as well as an affidavit sworn by Richard Marsden, an economist and Managing Director of NERA Economic Consulting (“NERA”), exhibiting an expert report prepared by him.

 

37.              Mr Hickey’s affidavit set out the basis for the stay sought. He averred that there were arguable grounds that the Decision was invalid and/or that there were serious doubts as to its validity. For the purpose of the stay application, that was accepted by ComReg and it is not necessary to say anything further on that issue here. Mr Hickey went on to assert that there was a likelihood of serious and irreparable harm being done to Three in the absence of a stay. His evidence on that issue was substantially based on Mr Marsden’s report. Mr Hickey referred to the potential for harm via mispricing or misallocation of spectrum because of the design of the auction process, arising from the de facto reservation of the first lots of 700MHz spectrum for Vodafone and Eir at the reserve price (or, as it was put in submission, these lots were “in the bag” from Vodafone’s and Eir’s point of view ). He went on to refer to the possibility that Eir and Vodafone might decide to engage in strategic bidding in which event spectrum could be mis-priced and allocated inefficiently. In its appeal (so Mr Hickey explained), Three had identified three ways in which it could be harmed as a result of the auction rules: (1) Three pays more than Vodafone or Eir for the same amount of spectrum as a result of Vodafone and Eir each acquiring one block of 700 MHz spectrum at the reserve price that would not be available to Three; (2) Three pays more than it would have otherwise done because rival bidders (presumably Vodafone and Eir) bid more aggressively for incremental spectrum that they do not win, thus setting the price for Three (because of the “Second Price Rule”) and (3) Three wins less spectrum than it would otherwise have done because rival bidders are encouraged by the auction rules to bid more aggressively for incremental spectrum, thereby winning more spectrum at the expense of Three (para 108).

 

38.              Mr Hickey went on to explain Three’s view that any ex post remedies that might be available to Three would be inadequate. In the first place the auction process could not be re-run without undermining the integrity of the process and potentially prejudicing some or all bidders (para 112). Secondly, it was unlikely to be possible to re-run the process. Once bidders were awarded spectrum, they would undertake network investment and incur costs and any attempt to cancel the auction would be “contentious” (para 114). Mr Hickey went to suggest that, if the auction proceeded on the basis of the auction rules and those rules were ultimately set aside by the High Court, “the inefficient spectrum allocation which the Appeal has sought to avoid would already have occurred and there would be no means to remedy that.” (para 120). Mr Hickey also suggested that there would be significant difficulty in the calculation of damages (para 119). The balance of interests, he suggested, favoured the granting of a stay. ComReg had put in place a temporary licensing regime for parts of the 700 MHz and 2.1 GHz bands and could extend that regime pending the determination of Three’s appeal. That, Mr Hickey suggested, undermined any suggestion that there was any pressing urgency in proceeding with the auction process (paras 124 - 134) A further factor weighing in favour of the stay, Mr Hickey suggested, was the unavailability of damages and/or their inadequacy as a remedy (paras 135-139).

 

39.              In his report, Mr Marsden covered essentially the same territory, though in somewhat greater detail. At paragraph 21 of his report he expressed the opinion that, if the auction proceeded under the current rules, there was “potential for harm via mispricing or misallocation of spectrum” by reason of “missing bids” (the effect of the auction rules not permitting Three to bid - or, as Mr Marsden put it, “express value” - for a third lot of 700 MHz spectrum), “strategic bidding” (the possibility that Vodafone and Eir “may be tempted to engage in price driving” based on the knowledge that they were positioned to win a first lot of 700 MHz spectrum at the reserve price whereas Three would have to pay full opportunity costs on all lots it won) and “subversion of price discovery” (the risk that if there were missing bids and/or if bids were distorted by aggressive bidding strategies, the benefits of price discovery would be eroded and bidders “are likely to make mistakes when updating their valuations and strategies, and this in term may lead to misallocation and mispricing”). Mr Marsden essentially made the same points as Mr Hickey regarding the inadequacy of damages and the difficulty of calculation of damages, the impossibility of re-running the auction process and the lack of urgency having regard to the temporary licensing regime that ComReg had put in place.

 

40.              Three’s application was opposed by ComReg and it filed a number of affidavits in response, an affidavit from George Merrigan, Director of the Market Framework Division in ComReg as well as affidavits from Dan Maldoom (a partner in DotEcon Limited, economic consultants retained by ComReg) and Peter Clinch (Chairman of EnvEcon Decision Support Limited, also retained by ComReg) exhibiting expert reports. I will refer to certain aspects of these affidavits and reports below.

 

41.              Notably, the stay application was supported by Vodafone and, it seems, by Eir also, with Vodafone filing an affidavit from its Head of Regulation, Andrew Corcoran, which had a significant impact on the High Court’s assessment of the application.

 

42.              In due course, after the delivery of further affidavits from Mr Hickey and Mr Marsden, the stay application came on for hearing before McDonald J in the High Court on 7 and 8 July 2022.

 

THE HIGH COURT JUDGMENT

 

43.              The Judge gave his judgment on 20 July 2022. I will set out the principal aspects of his judgment here and will return to consider some of these in more detail later, when discussing the arguments and issues in the appeal.

 

·         The Judge did not consider that the application should be dismissed on grounds of delay (page 6).

 

·         As to the criteria governing the application for a stay, the Judge referred to his judgment in Eircom Limited v Commission for Communications Regulation [2022] IEHC 165 and expressed his understanding that such an application was to be assessed by reference to the following criteria, which derive in large measure from the decision of the Court of Justice in Joined Cases C-143/88 and C-92/89 Zuckerfabrik Süderdithmarschen AG v Hauptzollamt Itzehoe [1991] ECR I-415 (“Zuckerfabrik”) and subsequent EU case law:

 

Ø   The applicant had to raise “serious doubts” as to the validity of the decision being challenged. That criterion was, the Judge observed, probably no more onerous than the arguable ground standard applied in the context of interlocutory injunctions as a matter of domestic law but it was unnecessary to consider that issue further given that ComReg had agreed that, for the purposes of the stay application, this first criterion was satisfied.

 

Ø   The applicant had to demonstrate that a stay was required as a matter of urgency, in the sense that a stay was necessary to avoid serious and irreparable damage to the applicant. EU case law indicated that it was necessary for an applicant to demonstrate, as a matter of probability, that it would be exposed to such damage before the court should proceed to any balancing of interests (in contradistinction to the approach propounded by the Supreme Court in Merck Sharp & Dohme Corporation v Clonmel Healthcare Limited [2019] IESC 65, [2020] 2 IR 1).

 

Ø   If the applicant gets over the serious and irreversible loss hurdle, it was then necessary to consider where the balance of interests lies, involving a weighing of the interests of the applicant (including the interest in securing the effectiveness of the appeal) against the interests of all those who would be adversely affected by the stay and the public interest in the orderly implementation of the decision under appeal) (at pages 8-9).

 

·         The availability of a Francovich claim for damages, [17] notwithstanding its limitations, was sufficient to establish that any claimed harm of a pecuniary nature, or harm which was otherwise capable of being remedied by an award of compensation, was in principle, reparable (at page 13). In so holding, the Judge rejected Three’s argument, based on the decision of this Court in Word Perfect Translation Services Ltd v Minister for Public Expenditure and Reform [2018] IECA 35, [2019] 2 IR 503, that the availability of Francovich damages would not constitute an adequate remedy in this context (at pages 12-13).

 

·         However, the availability of a Francovich claim did not always trump a claim that, absent a stay, an appellant would suffer serious and irreversible harm. Pecuniary loss might, in certain circumstances, be sufficient to meet the serious and irreversible harm standard. Furthermore, where the harm alleged was not capable of being remedied by an award of pecuniary compensation, the relevance of a Francovich claim simply fell away (ibid).

 

·         The onus was on Three to prove serious and irreparable loss as a matter of probability (at page 26).

 

·         The Judge did not accept that the risk that, as a consequence of the de facto set aside of spectrum for Vodafone and Eir at the reserve price, Three would have to pay more than Vodafone or Eir for the same spectrum, amounted to serious and irreparable harm. That head of loss appeared to be exclusively pecuniary in nature and there was nothing to suggest that it gave rise to any existential problem for Three (at page 27). Furthermore, any concern that Three might suffer loss was sufficiently addressed by ComReg’s undertaking to hold part of the proceeds of the auction pending judgment on the appeal (at pages 27-28). [18]

 

·         However, the Judge was persuaded that the de facto set aside of 700 MHz spectrum at the reserve price for Vodafone and Eir gave rise to a risk that they would engage in strategic bidding with “a view to either pushing up the price which Three has bid for spectrum, or with a view to reducing the amount of spectrum that Three can secure.” Either outcome would make it more costly or more difficult for Three to service its customers, which could “only be a commercially desirable outcome for its rivals.” In the circumstances, the Judge was satisfied that “such bidding is more likely than not to occur” (at page 32). As to the harm to Three, while it was possible in many instances to put a value on a loss of opportunity, the Judge considered that such an exercise presented particular difficulty here in circumstances where the loss of opportunity contended for arose in the context of a market where the demand for spectrum outstrips supply and there is significant ongoing competition between the three MNOs operating in the market. It was, in his view, very difficult to see how it would be possible reliably to estimate the extent of any losses that Three would suffer as a consequence of strategic bidding (at pages 33-34). The Judge was fortified in that conclusion by evidence from Mr Maldoom (on behalf of ComReg) as to the difficulty of establishing strategic bidding (at pages 34-35). Given the acknowledged importance of the availability of spectrum, such harm to Three was undoubtedly of a serious nature (at page 36).

 

·         Separately, the Judge concluded that Three would suffer serious and irreparable harm in the event that the auction proceeded, Three subsequently succeeded in its appeal and the auction had then to be re-run. Relying largely on the affidavit evidence of Mr Corcoran, the Judge was of the view that any re-run of the process would be “irreversibly compromised” by reason of the disclosure of the positions of bidders within the first auction process which would enable bidders “to construct a realistic view of the other bidders’ valuations” (at pages 36-41).

 

·         As to the balance of interests, the Judge identified a number of interests weighing in favour of the granting of a stay. These were Three’s interests in avoiding the harm to which it would be exposed in the event that a stay was refused (at page 42), the interests of the other MNOs who would be exposed to irreversible harm through the compromise of the auction process in the event that a stay was refused and the High Court’s decision on the appeal required a re-running of the auction (at page 42) and the public interest in maintaining the effectiveness of the appeal process (at page 42). Weighing the other way (i.e. against the grant of a stay) was the public interest in ensuring that effect was given to a regulatory decision of the kind at issue, an interest emphasised in the Irish and EU cases. That factor carried even more weight in light of the “undoubted importance” of the Decision in terms of improving 4G and 5G services, for the benefit of consumers (at pages 42-43). The interests of consumers weighed strongly against a stay although, the Judge noted, those interests did not all point in the same direction in that consumers also had an interest in ensuring that the auction process was not compromised in the manner described by Mr Corcoran (at page 43). A further factor weighing against a stay was the fact that there were likely to be other bidders interested in the 2.3 GHz and 2.6 GHz bands whose access to that spectrum would be delayed by a stay (at page 43). ComReg had also highlighted the damage to the economy generally of a delay in the award, though in the Judge’s view that was based on what he considered to be an unrealistic estimate of the likely period of delay of two to three years. The only relevant delay was between the granting of a stay and the High Court’s ultimate judgment on the appeal (at page 44).

 

·         The Judge clearly considered that these factors were closely balanced. If the stay sought was likely to be substantial in duration, he considered that it might be very difficult to conclude that the public interest in securing the effectiveness of Three’s appeal outweighed the public interest in the implementation of the Decision. There were, however, a number of factors which significantly lessened the weight to be given to the public interest in the immediate implementation of the Decision. First any stay was likely to be of a relatively short duration and was likely to be in place for no more than a few months. There was no hard evidence that a delay of that short duration was likely to cause any serious harm. Second, 5G roll-out would not be frozen in the event of a stay, as ComReg was likely to be able to issue temporary licences which could ameliorate the impact of the delay caused by a stay. Third, the risk to consumers and MNOs if the auction proceeded and was subsequently annulled had to be kept in mind. That was, in his view, a “significant factor” that, in combination with the relatively short duration of the stay and the likely availability of temporary licences during the period of such stay, tilted the balance in favour of a stay (subject to Three giving an undertaking in damages) (at pages 45 - 48).

 

·         In reaching the conclusions he did, the Judge appears to have discounted the evidence of Mr Marsden (Three’s expert) because he had, in the Judge’s view, crossed the line between expert and advocate (at page 38). The Judge expressed himself “very surprised” that “material of that kind” (i.e. expert reports expressing views in the terms used by Mr Marsden) had been put before the Court (at page 38). [19]

 

44.              Reference should also be made in this context to the Judge’s earlier judgment in Eircom Limited v Commission for Communications Regulation [2022] IEHC 165, given that the Judge effectively adopted its analysis of the criteria for the granting of a stay pursuant to Regulation 7(2) of the Framework Regulations in his Judgment here.

 

45.              In Eircom Limited v Commission for Communications Regulation, the Judge noted that Regulation 7 reflected the terms of Article 4 of the Framework Directive. As already noted, Article 4(1) of the Framework Directive required Member States to ensure that users or undertakings affected by a decision taken by the national regulatory authority would have a right of appeal against that decision and (as amended by the Better Regulation Directive) also provided that, pending the outcome of the appeal, the decision of the NRA should stand “unless interim measures are granted in accordance with national law.”

 

46.              Notwithstanding the reference to “national law”, the Judge concluded that in considering Eircom’s application for stay, he should apply the approach to interim measures taken by the Court of Justice in cases such as Zuckerfabrik. He reached that conclusion largely by reference to recitals (14) and (15) in the Better Regulation Directive and the equivalent recitals in the Recast Directive, namely recitals (77) and (78).

 

47.              Recital (14) of the Better Regulation Directive stated (inter alia) that “[i]nterim measures suspending the effect of the decision of a national regulatory authority should be granted in urgent cases in order to prevent serious and irreparable damage to the party applying for those measures and if the balance of interests so requires.” Recital (15) then provided that “[t]here has been a wide divergence in the manner in which appeal bodies have applied interim measures to suspend the decisions of the national regulatory authorities. In order to achieve greater consistency of approach common standards should be applied in line with Community case-law.” Recitals (77) and (78) of the Recast Directive are in similar terms, though recital (78) differs somewhat from its predecessor in that it refers to the need to apply common standards “in line with the case-law of the Court of Justice” rather than “in line with Community case-law”. Although noting that Ireland had not taken any measures to transpose the Recast Directive, the Judge considered that he was required to have regard to its provisions. [20] In any event, in his view there was no difference in substance between the relevant recitals in the Better Regulation Directive and the equivalent recitals in the Recast Directive (para 22).

 

48.              Applying the Zuckerfabrik criteria, the Judge considered that the evidence established that, in the absence of a stay, and on the assumption that Eircom was ultimately successful in its appeal, it would suffer loss in the range of €4.5 million to €7 million per year by reason of the fact that it would be compelled to provide broadband services to its wholesale customers (principally though not solely the Notice Parties) at a significantly reduced price. It would not be able to recover damages from ComReg or the Notice Parties and, on that basis, the Judge concluded that the refusal of a stay would result in irrecoverable loss to Eircom which, in his view, was of a scale which had to be said to be serious. There was no discussion of the possibility of a Francovich claim by Eircom against ComReg because, as the Judge explained in the course of the hearing in this case, that issue was not raised by any of the parties. Nevertheless, the stay sought was refused, inter alia because of an undertaking offered by the Notice Parties to account to Eircom for the difference between the price previously paid by them and the reduced price.

 

49.              One other aspect of the judgment in Eircom Limited v Commission for Communications Regulation should be noted. The Judge rejected Eircom’s submission that, having regard to the Supreme Court’s decision in RAS Medical Ltd v Royal College of Surgeons [2019] IESC 4, [2019] 1 IR 63, its evidence could not be called into question in the absence of cross-examination. Citing IBB Internet Services v Motorola [2013] IESC 53, the Judge observed that cross-examination would not be appropriate in an interlocutory application such as an application for a stay under Regulation 7(2) and expressed the view that the court had to do the best it could on the basis of the affidavits presented. It followed that ComReg and the Notice Parties were entitled to comment adversely on Eircom’s evidence without having to seek to cross-examine (paras 39-41).

 

 

APPEAL AND CROSS-APPEAL

 

50.              ComReg says that the Judge was correct to apply the Zuckerfabrik test but says that he erred in his application of that test, both as regards the issue of serious and irreparable harm and in his assessment of the balance of interests. ComReg’s criticisms of the Judge’s findings are discussed in detail below.

 

51.              Conversely, Three seeks to uphold the Judge’s findings about serious and irreparable harm and the balance of interests. As an additional ground on which it says that the Judge’s decision should be affirmed, Three says that all of the harm that it would suffer in the event that the auction proceeded is serious and irreversible in circumstances where the only means available to it to recover in respect of the harm done to it would be through a Francovich-type damages action. Three says that the mere possibility of such an action being available is not sufficient to prevent pecuniary harm being regarded as serious and irreversible. Three also says that it is not necessary for the potential financial losses to Three to be of such a scale as to pose an existential threat to it in order for the harm to be considered to be serious and irreversible.

 

52.              Three also cross-appeals. The essential point made on that cross-appeal is that the Judge erred in his application of the Zuckerfabrik test by taking into account EU case-law relating to the granting of interim measures against EU institutions by EU courts. In doing so - so Three says - the Judge failed to have regard to the principle that remedies were a matter for national legal systems and national courts and erred in his interpretation and application of para 29 of the judgment in Zuckerfabrik. In concrete terms, Three’s fundamental complaint is that the Judge erred in concluding that the mere possibility of bringing a Francovich-type claim for damages against ComReg in the event that Three was successful in its appeal meant that any financial or pecuniary harm that it might suffer would be reparable (assuming that the harm was such as to be capable of being remedied by an award of pecuniary compensation). Three also says that the Judge erred in taking the view that the undertaking offered by ComReg was sufficient to address any concern that Three would have to pay more for spectrum in the auction.

 

53.              Vodafone was represented by solicitor and counsel at the hearing of the appeal but other than indicating that it continued to rely on Mr Corcoran’s affidavit, it did not make submissions, written or oral. Eir did not appear or make any submissions on the appeal.

 

DEVELOPMENTS SUBSEQUENT TO THE HEARING

OF THE STAY APPLICATION IN THE HIGH COURT

 

54.              Shortly before the appeal was to be heard, the Court was furnished with a further affidavit from ComReg’s Mr Merrigan (his fifth affidavit in the proceedings, sworn on 14 October 2022). Mr Merrigan exhibited a redacted version of a communication that ComReg had received from an unidentified bidder expressing the view that the auction should be postponed until the final determination of Three’s appeal (including any further appeals from the High Court). “Only then”, the bidder explained, “can there be confidence that there is a final set of auction rules and that the Main Stage, with its inevitable disclosure of confidential strategic information, will not need to be re-run compromising the optimal outcome for all bidders”. [21] The bidder considered that postponing the auction would provide the stability required for a successful outcome and a “level playing field for all bidders” as well as facilitating “the climate required to make such a significant investment in spectrum.” The bidder suggested that the auction should be formally be postponed until at least April 2023.

 

55.              Mr Merrigan also exhibited a letter which ComReg had sent to all bidders on 7 October 2022. The letter noted that applications submitted by bidders remained binding offers for 6 months from the date of submission, with the earliest application received by ComReg remaining valid until 24 November 2022. The letter went on to put bidders on notice that, if following judgment in the substantive appeal or in this Court (in this appeal) it was in a position to advance to the Main Stage by 24 November 2022, it intended to do so promptly by giving bidders the 10 working days’ notice required by paragraph 3.22 of the Information Memorandum. If ComReg was not in a position to advance to the Main Stage by 24 November 2022 (the letter went on), it expected to return the bidders’ deposits to them shortly after that date.

 

56.              Mr Merrigan explained that the current applications which had been submitted by bidders (which include deposits and bids) remained valid for six months from when binding offers were submitted by each bidder. That period would expire in respect of certain bidders on 24 November 2022 and, in respect of all remaining bidders, shortly after that. Paragraph 3.22 of the Information Memorandum provided for 10 working days’ notice to bidders of the commencement of the Main Stage of the auction. In order to continue with the current award process, Mr Merrigan explained, notice of the commencement of the Main Stage would have to be given by no later than 8 November 2022, notifying bidders that the Main Stage would commence on 23 November 2022.

 

57.              Unless the Main Stage was commenced before 24 November 2022, then (so Mr Merriman continued), even if Three’s appeal was ultimately dismissed by the High Court (or this Court subsequently lifted the stay granted by the High Court), there would be significant further delay - potentially “between nine to eleven months” - in getting to the point where a new Main Stage could commence. The Information Memorandum would require updating, which would involve further public consultation. A new application stage would have to be run, followed by a new qualification stage, to bring the auction process back to the point where the Main Stage could be commenced. 

 

58.              On this basis, Mr Merrigan stated that, in the event that this Court were to indicate that it anticipated delivering its judgment by 22 November 2022, ComReg would then give notice of the commencement of the Main Stage on 23 November 2023, conditional on the Court deciding to lift the stay. ComReg accepted that this course of action would involve putting bidders to expense in the event that the Court decided to uphold the stay.

 

59.              By letter from its solicitors Matheson of 18 October 2022 (the eve of the appeal hearing) Three took issue with Mr Merrigan’s assertion that, in the event that the Main Stage did not commence before 24 November 2022, that would give rise to significant delay in commencing that Stage. [22] Three also took issue with the suggestion that the only option open to ComReg was to commence the auction before 24 November 2022. That was (the letter went on) because the Information Memorandum and ComReg’s auction rules did not prevent it from asking applicants to extend the validity of their applications (and at a later point the letter indicated that Three would be willing to extend the validity period of its application “if requested to do so on reasonable terms”). Three also questioned whether the auction could proceed without the consent of the applicants or renewal of their applications/bids, even if the auction commenced before 24 November 2022, referring in that context to paragraphs 3.66 and 3.165 of the Information Memorandum. Three also expressed concern at ComReg’s apparent willingness to commence the Main Stage in circumstances where it was possible that the judgment of O’Moore J “will land during the auction process.”

 

60.              Unsurprisingly, Mr Merrigan’s affidavit (and the exhibits to it), as well as Matheson’s letter of 18 October 2022, were the subject of debate at the commencement of the appeal on 19 October 2022. Counsel for ComReg confirmed that it was her client’s position that, for the reasons explained by Mr Merrigan, unless the Main Stage was commenced before 24 November 2022, there would be significant delay in restarting the auction process, even if the High Court dismissed Three’s appeal on 31 January 2023. That position was disputed by Counsel for Three. While allowing that the Information Memorandum did not expressly permit ComReg to extend the period of validity of applications, he nonetheless suggested that para 5.31 of the Information Memorandum provided a basis on which it could ask applicants to agree to do so and re-iterated his client’s willingness to agree such an extension.

 

61.              Paragraph 5.31 (which is headed “Amendments”) provides that “ComReg reserves, at its discretion, the right, at any time until the conclusion or termination of the Award Process, to amend or modify this Information Memorandum or Award Process in any respect, including the shortening or extension of any and all timelines, by way of clarification, addition, deletion or otherwise.” Asked by the Court whether that provision could be read as giving ComReg the power to simply tell applicants that their applications had been extended, Counsel for Three indicated that on the face of the Information Memorandum it appeared to be possible for ComReg to do so. However, he quite properly indicated that there might be an issue about that, given that the applications were effectively offers from the applicants. In any event, in his submission, ComReg could ask applicants to extend and it was very likely that they would agree to do so.

 

62.              Vodafone also indicated, through Counsel, its willingness to extend the validity period of its application “to overcome the cliff edge .. of the 23^rd November.”

 

63.              In response, Counsel for ComReg emphasised the uncertainty and risk that would be involved in seeking the agreement of bidders to extend the period of validity of their applications. There were bidders other than Three and Vodafone who might be unwilling to agree and, from ComReg’s perspective, that was not a viable or workable solution.

 

64.              There was also discussion about the likely duration of the Main Stage and whether, in the event that ComReg was permitted to commence the Main Stage prior to 24 November 2022, the auction process might then be suspended, or run only to a certain point, until the High Court gave judgment on the appeal. It was clear from the discussion that both ComReg and Three had serious difficulties with any approach that would involve the suspension or interruption of the Main Stage which, it was common case, was designed to run as a continuous process. Nonetheless, the discussion had the benefit of clarifying that, in the event that ComReg was permitted to commence the Main Stage before 24 November 2022, that stage, and the remaining stages of the award process up and including the award of the licences, could possibly have concluded before the High Court gave its judgment on 31 January 2023. As I shall explain, that prospect caused concern to the Court and led it to impose a stay in the terms set out in its order of 8 November 2022.

 

65.              The evidence set out in Mr Merrigan’s fifth affidavit was not before the High Court. In my view, it was and is highly material to assessing the balance of interests. The Judge attached significant weight to the fact that any stay granted by him would be of a “relatively short duration” and “likely to be in place for no more than a few months”. At that point, of course, the Judge did not know when O’Moore J might be in a position to give judgment on Three’s appeal. By the time that ComReg’s appeal came before this Court for hearing, O’Moore J had indicated that he anticipated giving judgment at the end of January 2023 and thus it was apparent that the stay would in fact be in place for a period in excess of 6 months, absent intervention by this Court. More significantly, on the basis of the evidence in Mr Merrigan’s fifth affidavit, it was apparent that there was - at a minimum - a very significant risk that, if the stay was left in place, its actual effect would be to delay the auction process for a substantially longer period. Even on the hypothesis that O’Moore J dismissed Three’s appeal and forthwith discharged the stay granted by the Judge (and that no further stay was sought or granted pending any appeal by Three), in such a scenario (according to Mr Merrigan), it could be late 2023 or early 2024 before the Main Stage could commence. Such a scenario is radically different to that presenting when the Judge made his decision to grant a stay.

 

66.              In this context, I do not disregard the fact that Three made it clear that it did not accept what was stated by Mr Merrigan. Three did not, however, submit any contrary evidence.  While Three made the point that bidders might agree to extend the period of validity of their respective applications, it is difficult to disagree with the point by ComReg in response, namely that the possibility of such agreement was too uncertain to provide a workable solution. The only bidders before the Court were Three and Vodafone. There may be many other bidders for different lots of spectrum (bearing in mind that the auction is not limited to the 700 MHz band), any of whom might have been unwilling to agree to an extension. Equally, while there may be an argument that, under the relevant provisions of the Information Memorandum ComReg was entitled to unilaterally extend the period of validity of applications, as Counsel for Three fairly accepted, there is considerable scope for argument the other way also. Again, it is difficult to take issue with ComReg’s point that any such course of action would be fraught with uncertainty.

 

67.              In any event, I do not consider that it is necessary or appropriate to reach a conclusion as a matter of probability here. It is, in my view, sufficient to observe that, if the stay granted by the Judge was to be left in place, the evidence establishes a very significant risk that, even if the stay was discharged on or immediately after 31 January 2023, the commencement of the Main Stage of the auction would nevertheless be delayed for a further substantial period, potentially until late 2023 or early 2024. That further period of delay would be attributable to the stay.

 

THE PRINCIPLES GOVERNING APPLICATIONS FOR A STAY

 UNDER REGULATION 7(2) OF THE FRAMEWORK REGULATIONS

 

68.              Three’s cross-appeal notwithstanding, there is (or at least appeared to be) a significant measure of common ground between the parties as to the principles that apply to applications for a stay under Regulation 7(2).

 

69.              Three’s written submissions on appeal put its position as follows:

 

“25. Three did not dispute before the High Court (nor does it dispute now) that the relevant legal test to be applied in respect of a stay pursuant to [Regulation 7(2)] of the Framework Regulations is that identified by McDonald J in Eircom v ComReg which is based on the decision of the CJEU in [Zuckerfabrik].”

 

70.              Consistently with that stated position, Three explicitly accepted that an applicant for a stay must satisfy the following three requirements: (1) the applicant must establish that there are “serious doubts as to the validity of the decision under challenge”; (2) the applicant must show that “a stay is required as a matter of urgency in the sense that it is necessary that a stay be put in place before a decision on the merits in order to avoid serious and irreversible damage to the party seeking the stay” and (3) the court must be satisfied that the “balance of interests” lies in favour of granting a stay. [23] These are, of course, the requirements identified by the Court of Justice in Zuckerfabrik and applied by the High Court in Eircom v ComReg.

 

71.              Three also accepts that the onus is on the applicant for a stay to establish, as a matter of probability, that it would suffer serious and irreparable loss unless a stay is granted. That was expressly accepted by Counsel for Three in the High Court [24] and Three did not seek to resile from that position in the appeal.

 

72.              The other area of common ground is that the three requirements for a stay derived from Zuckerfabrik are cumulative and distinct. It follows that, if an applicant fails to establish a probability of serious and irreparable harm, the application for a stay fails, without any necessity to consider the balance of interests. [25]

 

73.              The area of disagreement is limited or at least that appeared to be so before the hearing of the appeal.

 

74.              ComReg maintains (and the High Court Judge agreed) that, in assessing whether a probability of serious and irreparable harm has been established, the availability to Three of a Francovich claim for damages against ComReg means that any pecuniary loss that it may suffer absent a stay is, in principle, to be regarded as reparable. That, ComReg says, follows from the Zuckerfabrik jurisprudence.

 

75.              There seems to be little doubt that decisions such as Case C-282/21 P(R) Symrise AG v European Chemicals Agency indicate that the availability of an action against an EU institution for non-contractual liability on the basis of Articles 268 and 340 TFEU means that alleged damage of a pecuniary nature can be remedied and cannot, at least in the absence of exceptional circumstances, be regarded as irreparable: para 423. Reference may also be made in this context to Case T-95/09 R, United Phosphorus Ltd v Commission, at para 73 (Articles 235 and 288 EC are now Articles 268 and 340 TFEU).

 

76.              While the Court of Justice in Zuckerfabrik did not refer to Articles 235 and 288 EC, its statement that purely financial damage cannot be regarded in principle as irreparable must, in my view, be understood in the context of those provisions. Zuckerfabrik was, after all, concerned with the alleged invalidity of a Union legislative measure (Council Regulation (EEC) No 1914/87) which introduced a “special elimination levy” in the sugar sector. If Zuckerfabrik suffered financial damage as a result of the Regulation, and the Regulation was then ruled invalid, the only means of obtaining compensation for such damage would appear to have been an action against the Council under Articles 235 and 288 EC.

 

77.              There is a high threshold for establishing non-contractual liability on the part of an EU institution under Article 340 TFEU (ex-Article 288) : see the discussion in Craig & de Búrca, EU law, Text, Cases & Materials (7^th ed, 2020), chapter 17.

 

78.              A similar threshold applies to establishing state liability under Francovich: Cases C-46 and 48/93 Brasserie du Pêcheur SA v Germany, para 53. The entitlement to Francovich damages is “highly conditional and limited”: per Hogan J (Hedigan and Gilligan JJ agreeing) in Word Perfect Translation Services Ltd v Minister for Public Expenditure and Reform [2018] IECA 35, [2019] 2 IR 503, at para 55.

 

79.              In Word Perfect, the applicant was challenging a public procurement award by the Minister and sought to have the automatic suspension continued pending the determination of that challenge. Reversing the High Court, this Court held that the limitations of a Francovich claim were such that it could not be said that damages were an adequate remedy for the applicant (at para 60) and went on to conclude that the balance of justice was in favour of maintaining the suspension of the award as otherwise the applicant would be left without a real or effective remedy (at para 63).

 

80.              Relying on Word Perfect, but also on the decisions of the Supreme Court in Okunade v Minister for Justice, Equality and Law Reform [2012] IESC 49, [2012] 3 IR 152 and Dowling v Minister for Finance [2013] IESC 37, [2013] 4 IR 576, Three says that the availability of a Francovich claim does not mean that (in domestic law terms) damages can be regarded as an adequate remedy for it or (in the language of Zuckerfabrik) that the pecuniary loss it says it will suffer in the absence of a stay is to be regarded as reparable and therefore not capable of satisfying the requirement to establish serious and irreparable harm.

 

81.              In Okunade, the applicants sought to restrain their deportation pending the determination of their challenge to the deportation decision, raising “the question of what the appropriate criteria are or, to put it another way, what the test is for the grant of a stay or injunction which has the effect of preventing an otherwise valid measure or order from having effect pending trial” (per Clarke J at para 62). Whatever the form of order sought, there was “an inevitable risk that, with the benefit of hindsight, and after a full trial has been conducted, an injustice may be seen to be done” (para 65). On the one hand, a party may be subject to a challenged measure only to find that the measure is held to be invalid after a full trial. On the other hand, if a stay or injunction is granted suspending the effect of a measure that is ultimately found to be valid, an injustice may also arise (para 66). In each case, “the problem stems from the fact that the court is being asked, on the basis of limited information and limited argument, to put in place a temporary regime pending trial in the full knowledge that the court does not know what the result of the trial will be” (para 67). In such circumstances, “the underlying principle must be that the court should put in place a regime which minimises the overall risk of injustice” (ibid).

 

82.              Clarke J (with whose judgment Denham CJ and Hardiman, Fennelly and O’ Donnell JJ agreed) went to address whether and how the Campus Oil test applied in the context of a challenge to a public law measure or decision. He was satisfied that the fair or arguable issue to be tried leg of the test also applied in public law proceedings (para 86). As to the second leg - the adequacy of damages as a remedy - Clarke J considered that it was less likely to be of significant relevance in public law proceedings. Firstly, the potential harm to an applicant was often non-financial in nature. Secondly - and this is the aspect of Okunade relied on by Three - even where the harm relied on was commercial or financial in nature, “damages may well not be recoverable and thus, by definition, not be capable of providing adequate compensation” (at para 89). Citing Kennedy v Law Society of Ireland (No 4) [2005] IESC 23, [2005] 3 IR 228 and Glencar Exploration plc v Mayo County Council (No 2) [2002] 1 IR 84, Clarke J noted there could be little doubt that the circumstances in which damages could be recovered in public law challenges were “limited” (ibid). Finally, it was unlikely that applicants would be able meet an undertaking in damages or, indeed, that any rational basis for calculating damages to be paid by an applicant who obtained a stay or injunction pending trial but who ultimately failed could be shown to exist.

 

83.              Clarke J went on to consider how the balance of justice was to be weighed in the context of a public law challenge to a public law measure or decision. This aspect of his analysis is potentially relevant here also. It is notable for its recognition of the significant public interest in permitting prima facie valid measures “to be carried out in a regular and orderly way” (at para 92). Clarke J went on:

 

“[92] .. Regulators are entitled to regulate. Lower courts are entitled to decide. Ministers are entitled to exercise powers lawfully conferred by the Oireachtas. The list can go on. All due weight needs to be accorded to allowing the systems and processes by which lawful power is to be exercised to operate in an orderly fashion. It seems to me that significant weight needs to be attached to that factor in all cases. Indeed, in that context it is, perhaps, appropriate to recall what was said by O'Higgins C.J. in Campus Oil v. Minister for Industry (No. 2) [1983] I.R. 88. At p. 107 of the report he said the following:-

‘The order which is challenged was made under the provisions of an Act of the Oireachtas. It is, therefore, on its face, valid and is to be regarded as a part of the law of the land, unless and until its invalidity is established. It is, and has been, implemented amongst traders in fuel, but the appellant plaintiffs have stood aside and have openly defied its implementation.’

It is clear, therefore, that the apparent prima facie validity of an order made by a competent authority was a factor to which significant weight was attributed. While the comments of O'Higgins C.J. were directed to a ministerial order made under an Act of the Oireachtas it seems to me that there is a more general principle involved. An order or measure which is at least prima facie valid (even if arguable grounds are put forward for suggesting invalidity) should command respect such that appropriate weight needs to be given to its immediate and regular implementation in assessing the balance of convenience.

[93]  It is also, in my view, appropriate to take into account the importance to be attached to the operation of the particular scheme concerned or the facts of the individual case in question which may place added weight on the need for the relevant measure to be enforced unless and until it is found to be unlawful.

[94]  That is not to say, however, that there may not also be weighty factors on the other side. It is necessary for the court to assess the extent to which, in a practical way, there is a real risk of injustice to an applicant for judicial review in being forced to comply with a challenged measure in circumstances where it may ultimately be found that the relevant measure is unlawful. The weight to be attached to such considerations will inevitably vary both from type of case to type of case and by reference to the individual facts of the case in question.”

84.              In Dowling, the plaintiffs sought to injunct the sale of Irish Life by the Minister for Finance, pending the determination of challenges they had brought to the Minister’s acquisition of the shares in Irish Life pursuant to the Credit Institutions (Stabilisation) Act 2010, inter alia on the basis that the Minister had acted in breach of EU law. The High Court (Laffoy J) refused to grant an injunction. On appeal, the plaintiffs contended that the test in Okunade failed to comply with the principles of equivalence and effectiveness. In that context, Clarke J (Denham CJ and Murray J agreeing) considered the jurisprudence of the Court of Justice concerning interim measures, including Zuckerfabrik. He emphasised that the situation with which the Court of Justice was concerned in Zuckerfabrik – where what was at issue was the validity of a European Union measure - was different to the situation with which the court was concerned in Dowling, where it was alleged that a national measure taken on foot of national legislation was in breach of EU law. It was, Clarke J stated, “clear that different considerations apply in such cases” (at para 84). It was precisely because the validity of an EU measure was at issue in Zuckerfabrik that the Court of Justice had considered it appropriate that there should be “a single, uniform test by reference to which intervention by means of interim measures is to be adopted”, that test being one identified in Zuckerfabrik (at para 85).

 

85.              Equally (per Clarke J at para 86) it was clear from Case C-432/05 Unibet (London) Ltd v Justitiekanslern [2007] ECR I - 2271 that, where there is a challenge to a national measure:

 

“national procedural rules are to apply subject to the principles of equivalence and effectiveness. It follows that it is possible that different results might arise in different member states, in the event of a challenge to similar national measures, because of the applicability of different procedural regimes, provided always that each such regime must provide an effective remedy.”

 

86.              Notwithstanding those differences (Clarke J went on), the Zuckerfabrik test was not irrelevant. It was clearly intended by the Court of Justice to provide an effective remedy to persons asserting that their EU rights had been breached by the adoption of an invalid EU measure. Thus while the court should apply the test in Okunade, it should also have regard to the question of whether the application of that test might deprive the plaintiffs of an effective remedy and, in that context, the court should have regard to Zuckerfabrik (at paras 88 and 89). Clarke J noted that the first element of the Zuckerfabrik test arguably required a greater level of doubt as to the validity of the measure than applies under Okunade; thus, if anything, Irish national rules provided greater protection under this heading for applicants (at para 90). As regards the requirement in Zuckerfabrik that a risk of “serious and irreparable harm” be established, Clarke J referred to his judgment in Okunade, observing that if, in accordance with the test in Okunade, damages were truly an adequate remedy, it was difficult to see how any irreparable damage, in the Zuckerfabrik sense, could arise (at para 92). It was a well-established principle of EU law that “purely financial damage cannot be regarded, in principle, as irreparable”. In such cases, where the claimant succeeded in establishing an infringement of their rights, they could be “fully compensated in damages”, thus vindicating their rights. It was hardly surprising that there was such a similarity between the Zuckerfabrik test and the test that applied in many national systems such as the Irish one as “the problem is the same in all cases”, namely the “inevitable risk of injustice in either granting or refusing interim measures.” Where the alleged harm was “purely financial”, then at least in most cases there would be no risk of serious injustice “for a party who wins can be properly compensated in damages” (also at para 92).

 

87.              In observing that there was such a similarity between the Zuckerfabrik test and that set out in Campus Oil/Okunade, Clarke J does not appear to have adverted to the possibility that, where the Zuckerfabrik test was applicable, financial loss might be regarded as reparable even though the prospect of actually recovering monetary compensation was limited by reason of the restrictive rules governing the non-contractual liability of the Union and its institutions (rules which, as mentioned, apply also to the non-contractual liability of Member States under Francovich). In any event, in Dowling itself, Clarke J was satisfied that, if the plaintiffs succeeded in their actions, they could recover “full compensation” from the Minister in the event that the sale of Irish Life took place and the transaction could not be set aside because they would receive full value for their shares (at paras 134-136).

 

88.              As regards the balance of convenience, Clarke J noted that EU law similarly required the balancing of competing interests (para 94). For the reasons set out in his judgment - with particular emphasis on the fact that the plaintiffs would be fully compensated in the event that they were successful - Clarke J concluded that the balance weighed against the granting of the injunction sought. It appears from his discussion of this issue that he would have come to the same conclusion had he been applying the EU law balancing of interest test. There was therefore no question of the plaintiffs being deprived of an effective remedy (para 143-144).

 

89.              Brief reference should also be made in this context to the Supreme Court’s decision in Merck Sharp & Dohme Corporation v Clonmel Healthcare Limited [2019] IESC 65, [2020] 2 IR 1. While a decision of major importance regarding interlocutory injunctions generally, for present purposes its principal significance lies in its clarification that issues around the adequacy of damages are to be considered as part of the broader assessment of the balance of convenience/balance of justice, rather than as a separate free-standing element: per O’ Donnell J (as he then was) (Clarke CJ, McKechnie, Dunne and O’ Malley JJ agreeing) at para 36. As already noted, that is not the position under Zuckerfabrik.

 

90.              Three also invokes Unibet in this context. However, it seems to me that that simply highlights a fundamental inconsistency in Three’s position. An applicant for a stay pursuant to Regulation 7(2) might of course rely on Unibet in order to argue that, in the absence of applicable EU rules governing such stay applications, it is for the domestic legal system of each Member State to determine the conditions under which a stay is to be granted, subject to the principles of equivalence and effectiveness and that, in this jurisdiction, that means that such applications are governed by the principles set out in Okunade, Dowling and Merck Sharp & Dohme. But Three did not seek to make such an argument in the High Court. It did not contend for the application of Okunade. Rather, it accepted that the application was governed by Zuckerfabrik (while also contending that Zuckerfabrik did not require the court to conclude that apprehended pecuniary loss was in principle reparable if a Francovich claim would be available, however unlikely it might be that such a claim would actually lead to the recovery of damages).

 

91.              I have already referred to Three’s written submissions on appeal, which explicitly accept that the legal test to be applied was that identified by McDonald J in Eircom, in turn based on the decision of the Court of Justice in Zuckerfabrik.

 

92.              The Court was left in a position of real uncertainty as to Three’s actual position as to the applicable test and whether and to what extent the Court was being invited to depart from Zuckerfabrik. Accordingly it raised that issue with Counsel for Three in the course of his submissions. In response, Counsel explained that, but for the recitals in the Recast Directive, Three would never have accepted Zuckerfabrik as the applicable test. He emphasised that no EU measure was being impugned here; rather the measure being challenged was a decision of a national regulator. On that basis, Counsel submitted that the decisions of the General Court in cases such as Case T-181/02 Neue Erba Lautex v Commission and T-299/10 Babcock Noell GmbH and of the CJEU in cases such as Case-281/21 P(R) Symrise AG v European Chemicals Agency and Case C-471/00 P(R) Commission v Cambridge Healthcare Supplies have no relevance or application here and the Judge erred in having regard to them. That, it was said, had led the Judge to the erroneous conclusion that the possibility of a Francovich claim for damages against ComReg in the event that Three was successful in its appeal meant that any financial or pecuniary harm suffered by Three had to be regarded as being reparable, however unlikely it might appear that Three would actually be in a position to establish an entitlement to Francovich damages.

 

93.              Counsel properly accepted that Three had not made the case in the High Court that national law was applicable. [26] Three’s case was that there was no basis for applying “the full Zuckerfabrik standard with all of its strictness” given that the validity of an EU measure was not at issue. The recitals in the Recast Directive were, it was said, “too thin a support” for such an approach. On that basis, Counsel suggested, there was scope for diluting the Zuckerfabrik test by reference to national law principles. If that submission did not find favour, Three invited the Court to disregard the Zuckerfabrik test entirely and apply national law instead. As a further fall-back/alternative submission, Counsel said that even if Zuckerfabrik applies, there was a greater flexibility in the application of the “serious and irreparable harm” criterion than ComReg allowed, citing the decision of the General Court Case T-95/09R United Phosphorus v Commission. [27]

 

94.              In my view, it is far too late for Three now to contend, however conditionally, that the stay application is not governed by Zuckerfabrik and is instead governed by purely domestic law principles set out in decisions such as Okunade and Dowling. To permit Three to change its position diametrically on such a fundamental issue in the middle of the appeal hearing would be wholly inconsistent with the fair and efficient administration of justice and with the proper discharge of this Court’s functions as an appellate court: see Lough Swilly Shellfish Growers Co-Operative Society v Bradley [2013] IESC 16, [2013] 1 IR 227.

 

95.              If it had been open to Three to advance that argument, and in the event that it appeared that the outcome of the appeal might turn on the issue of the applicable test, the Court might have had to give consideration to making a reference to the CJEU pursuant to Article 267 TFEU. So far as the researches of the parties could ascertain, this aspect of Article 4(1) of the Framework Directive (now Article 31(1) of the Recast Directive) has never been considered by the Court of Justice. There can be little doubt that, as the Judge concluded in Eircom and as Counsel for Three accepted in argument here, the language of the relevant recitals appears clearly to point to the application of the Zuckerfabrik test. But recitals are not operative legislative provisions. [28] Recitals may, of course, be relied on to resolve ambiguity in related legislative provisions. [29] However, a notable feature of the relevant legislative provision here is that it expressly refers to the granting of interim measures “in accordance with national law” (my emphasis). That language was introduced by the Better Regulation Directive and has been retained in the Recast Directive. On their face, the recitals and the operative legislative provision appear to be in tension and there may be scope for argument that the recitals must yield to the terms of Article 4(1) in those circumstances. There are, of course, other factors to be considered also in this context, including the legal basis of the functions being carried out by ComReg and in particular the legal and regulatory rules for the spectrum allocation function being exercised here (the ultimate source of which is, according to Three, the Framework Directive). [30] The fact that Three’s right of appeal, and its entitlement to seek a stay on ComReg’s Decision, ultimately derive from EU legislation are also very relevant considerations in this context. [31] All these considerations might appear to indicate that the position here is closer to Zuckerfabrik than to Okunade or Dowling (or Unibet).

 

96.              With becoming frankness, Mr Kennelly argues that it should be easier to obtain interim relief when one is challenging a domestic measure and not impugning an EU measure. But that argument begs the question as to the appropriate characterisation of the impugned measure here. It appears wholly implausible to suggest that the Decision, made by the Irish NRA within a regulatory framework that is governed by EU law, can be characterised as a purely domestic law measure, such as was at issue in Dowling. Any such suggestion appears wholly at odds with Three’s acknowledgment that the ultimate source of ComReg’s decision-making competence here was the Framework Directive and that any claim for damages that Three might seek to make against ComReg in connection with the functions at issue are governed by Francovich principles. If the availability of a restrictive claim for non-contractual damages against the EU under the Treaties is, in general, an answer to an application for interim measures against the EU, why as a matter of principle shouldn’t the availability of a similar claim not be an answer to an application for interim measures against a national body exercising EU law functions?

 

97.              In any event, for the reasons already indicated, it appears to me that for the purposes of determining this appeal the Court should apply the Zuckerfabrik test. That is not, I emphasise, because the Court has decided that that is the applicable test as a matter of law; rather it is because that was the basis on which the High Court was asked to determine the stay application and it would not be appropriate for this Court to apply a different test in the circumstances already set out. In any event, as will become clear, the outcome of the appeal does not in my view turn on the applicable test. Regardless of whether the test is that set out in Zuckerfabrik or in Okunade, Three is not in my opinion entitled to any stay beyond the stay already granted by this Court, namely a stay on ComReg notifying the Winning Bidders of their entitlement to apply for licences as provided for in paragraph 3.259 of the Information Memorandum or taking any subsequent steps in the auction process pending the determination of the proceedings in the High Court.

 

98.              As for Three’s argument that this Court should apply some modified or hybrid version of Zuckerfabrik, involving Irish law elements relating to the adequacy of damages, in my view that argument is not plausible or persuasive. As I observed in the course of argument, it appears to me that either Zuckerfabrik applies or it does not. Three has not identified any basis on which the Court can, or should, seek to splice genetic material from Word Perfect or Okunade into Zuckerfabrik. It cannot be suggested that it is necessary to do so in order to secure an effective remedy for Three: as Clarke J observed in Dowling, the Court of Justice in Zuckerfabrik was clearly satisfied that the application of the criteria for granting interim measures identified in its decision ensured an effective remedy for claimants. Nor does the suggested “reading in” of Word Perfect or Okunade involve some minor modification of the Zuckerfabrik test. On the contrary, Three’s argument, if accepted, would effectively turn the second stage of the Zuckerfabrik test on its head, in that once a risk of pecuniary loss was established, the court would inevitably have to conclude that damages were not an adequate remedy because they were unlikely to be available under Francovich.

 

99.              Accordingly, in my view, this Court must determine ComReg’s appeal by reference to Zuckerfabrik. Three’s arguments as to the construction and application of the requirement in Zuckerfabrik for “serious and irreparable harm”, and in particular its contention that purely financial loss may, at least exceptionally, satisfy that requirement, remains to be considered.

 

APPLYING THE ZUCKERFABRIK TEST HERE

 

THE PROPER APPROACH OF THIS COURT TO THE FINDINGS MADE BY THE JUDGE

 

100.          Both parties addressed this issue in their written submissions. Citing my judgment (with which Whelan and McGovern JJ agreed) in Betty Martin Financial Services Ltd v EBS DAC [2019] IECA 327, ComReg submitted that, while the appeal was not a re-hearing, the ultimate decision on an appeal from an interlocutory order of the High Court was one for this Court. ComReg emphasised that the stay application had been heard on affidavit evidence only and also referred to the Judge’s criticisms of Three’s expert evidence from Mr Marsden as advocacy, not independent expert evidence.

 

101.          Three also cited Betty Martin Financial Services Ltd v EBS DAC, emphasising statements in my judgment to the effect that an appeal from an interlocutory order such as an injunction (or stay) was not a full de novo hearing and that, as a matter of principle, “great weight” was to be given to the views of the High Court judge. So much was not in dispute. However, Betty Martin was also cited as authority for the application by this Court of the principles in Hay v O’ Grady [1992] 1 IR 210 to findings of fact made after an interlocutory hearing entirely on affidavit.

 

102.          Hay v O’ Grady does not apply to findings of fact made after a hearing on affidavit, without oral evidence: Ryanair Ltd v Billigfluege.de GmbH [2015] IESC 11, per Charleton J (Hardiman, McKechnie, Clarke and MacMenamin JJ agreeing) at paras 3-4. Nothing in Betty Martin suggests that it does. Ryanair Ltd v Billigfluege.de GmbH does, however, make it clear that, where findings of fact are made on the basis of affidavit evidence, the appellant bears the burden of demonstrating that there is some error in such findings: at para 11. The approach to be taken by an appellate court was recently addressed by this Court (Murray J, Haughton and Barniville JJ agreeing) in AK v US [2022] IECA 65. Murray J explained that, where findings of fact based on the assessment of affidavit evidence were challenged on appeal:

 

“[53] …  the appellate court is free to correct errors of fact as well as of law, and mistaken inference as well as erroneous application of principle. It is thus not necessary for the appellant to establish that a judge has erred in law or in principle, the appellate court is not concerned to establish that the decision of the trial judge was not one that was reasonably open to him or her, nor will the appellate court be necessarily constrained to affirm a finding which is supported by credible evidence (although obviously where a judge has so erred or there is no credible evidence to support the finding the appellate court will interfere). Instead, the appellate court affords limited deference to the decision of the trial court by beginning its analysis from the firm assumption that the trial judge was correct in the findings or inferences he or she has drawn, and interfering with those conclusions only where it is satisfied that the judge has clearly erred in the findings made or inferences drawn in a material respect. This, I should observe, reflects the standard of review referred to by Finlay Geoghegan J. in D.E v. E.B [2015] IECA 104 at paras. 39 and 40, while taking account of the decision in Ryanair Ltd. v. Billigfluege.de GmbH to which she also referred. …”

 

 

104.          As regards the balance of interests, great weight must be given to the Judge’s assessment: Betty Martin, (at para 35) and Ryanair DAC v Skyscanner Ltd [2022] IECA 64, (per Murray J (Donnelly and Haughton JJ agreeing) at para 110.

 

 

            SERIOUS AND IRREPARABLE HARM

 

            General

 

105.          Zuckerfabrik emphasises that interim measures suspending the operation of a contested measure may only be granted “if it is necessary for them to be adopted and to take effect before the decision on the substance of a case, in order to avoid serious and irreparable damage to the party seeking them” (at para 28). The damage invoked by the applicant “must be liable to materialise before [the court] has been able to rule on the validity of the contested … measure” (Ibid, at para 29). The purpose of the procedure for interim relief is “to guarantee the full effectiveness of the future final decision, in on order to prevent a lacuna in the legal protection…” (Case T-849/16 R, PGNiG Supply & Trading GmbH v Commission, at para 29; see also Case C-282/21 P(R) Symrise AG v European Chemicals Agency, para 40.). The party seeking such measures “must demonstrate that it cannot await the outcome of the main proceedings without suffering serious and irreparable damage” (Ibid).

 

106.          If it is possible to safeguard the applicant’s position retroactively in the event of success in the main action, the requirement for irreparable damage will not be satisfied (Case T-13/99 R, Pfizer Animal Health SA/NV v European Council, at para 91, citing (inter alia) Case C29/66 R Gutmann v Commission). The same point was made by the Court of Justice in Zuckerfabrik - the court must consider “whether immediate enforcement of the measure which is the subject of the application for interim relief would be likely to result in irreversible damage to the applicant which could not be made good if the [measure] were to be declared invalid” (at para 29). Conversely, irreparable harm may be suffered if, even in the event of success in the main action, the judgment of the court would not be fully effective to vindicate the rights of the applicant.

 

107.          The fundamental purpose of any interim measures regime is, of course, to ensure the effectiveness of any judgment that the applicant may succeed in obtaining in the “main proceedings” (Factortame, at para 21). That is reflected in the express terms of Regulation 7(2) of the Framework Regulations, providing as it does for the granting of a stay or other order “for the purpose of securing the effectiveness of the hearing and determination of the appeal.” If, absent the grant of interim measures, there is likely to be an irreversible alteration of the status quo such that an effective remedy will not be available by the time the main action is determined, judicial intervention will be indicated.

 

108.          The decision of this Court in Word Perfect usefully illustrates that principle. The applicant challenged the decision to award the contract to a third party. Unless the award of the contract was stayed, the applicant would effectively achieve nothing of value even if it succeeded in its challenge: per Hogan J at para 64. While the court had power to declare a reviewable public contract “ineffective”, such a remedy was available only in very limited circumstances, which had no application on the facts. Annulment of the award decision was thus not an available remedy (at para 63). In such circumstances, Francovich damages would be the only available remedy but, given the limitations on the recovery of such damages, that was not an adequate remedy either and so the continued suspension of the award was necessary to ensure that Word Perfect had an effective remedy. If the remedy of annulment had been available, the position would clearly have been materially different.

 

109.          There may also be circumstances in which, absent interim measures, annulment of the challenged decision may not be sufficient, of itself, to protect and vindicate the rights of the applicant in a concrete way. An example is provided by Case T-95/09 R United Phosphorus Ltd v Commission, para 32, which is relied on by Three here and which is discussed in detail below. At this point, it suffices to note that the President of the Court of First Instance (now the General Court) was persuaded that the requisite urgency had been established by the applicant in circumstances where its (very detailed) evidence demonstrated that, even if it obtained annulment of the contested decision (which required the discontinuance of the use of napropamide as an active ingredient in plant protection products) and also obtained granted national authorisations for marketing napropamide, it would not in fact be able to resume production due to the unavailability of supply.

 

110.          Here we are concerned with the question of interim measures pending the High Court’s adjudication on Three’s appeal. Three has challenged ComReg’s Decision on a number of grounds as already discussed. In the event that O’Moore J finds that Three has failed to establish that the Decision was vitiated by a serious and significant error or series of such errors (as Three contends), then its appeal fails. On that hypothesis, any interim measures granted pending the High Court’s decision ought not to have been granted. If, on the other hand, Three succeeds in establishing that the Decision was vitiated by a serious and significant error or series of such errors (as it contends), then its appeal succeeds and it will be entitled to appropriate relief from the High Court. The only substantive relief sought by Three in its appeal is an order setting aside specified parts of ComReg’s Decision. While its Notice of Motion also sought damages, the issue of damages does not appear to have been pursued before O’Moore J. If Three persuades O’Moore J that ComReg erred in imposing the Sub-1 GHz spectrum cap and/or in opting for the CCA format, it would appear to follow that the relevant parts of the Decision should be set aside. That was certainly position adopted by Three at trial. [32] Of course, it will be a matter for O’Moore J to determine what relief, if any, Three may be entitled to. However, the key point is that, should he consider it necessary or appropriate to do so in order to vindicate Three’s rights, O’Moore J can make an order setting aside the offending parts of the Decision and remitting the Decision back to ComReg for reconsideration and repair.

 

111.          If, on the other hand, the High Court declines to make such an order, that will indicate that it was not persuaded that the Decision was vitiated by a serious error or series of errors and was not persuaded that an auction process conducted in accordance with the Decision would be inconsistent with ComReg’s statutory duties and/or the applicable requirements of EU law.

 

112.          In the event that Three is successful in its appeal, an order in the terms indicated above would appear to constitute an effective remedy for it. The position might be otherwise if, when such order came to be made, the auction process had concluded and the spectrum rights at issue had all been finally allocated. Such a scenario would involve an alteration of the status quo and it could be argued that the setting aside of the Decision would/should not, of itself, affect steps taken on foot of it, particularly in circumstances where decisions relating to the allocation of spectrum to third parties could potentially give rise to significant reliance interests (though see, a contra, PGNiG Supply & Trading v Commission, at para 38). But any such risk has been excluded by the stay granted by this Court, as a result of which the auction process will not proceed to the point of the allocation of spectrum prior to the High Court’s decision on Three’s appeal.

 

113.          In light of the above analysis, it appears to me that the first head of serious and irreparable harm identified by the Judge (the losses that Three was likely to suffer as a result of strategic bidding by Vodafone and/or eir) falls away as a potential basis for the stay granted by him. Annulment (setting aside) of the relevant parts of the Decision will avoid any such harm to Three. If, on the other hand, Three fails to persuade O’Moore J that the auction design adopted by ComReg is so flawed as to warrant the annulment (setting aside) of the relevant parts of the Decision, then it follows that no such stay ought to have been granted. I will come back to this point below.

 

 

 

The Applicable Threshold - A Probability of Serious and Irreparable Harm

 

114.          As I have already noted, it was common case that the Zuckerfabrik test requires the applicant for interim measures to establish serious and irreparable harm as a matter of probability. In Zuckerfabrik itself, the Court of Justice identified the fundamental issue as whether immediate enforcement of the challenged measure “would be likely to result in irreversible damage to the applicant which could not be made good if the [measure] were to be declared invalid.” In many subsequent decisions in applications for interim reliefs, the Union courts have emphasised that, while it is not necessary for the occurrence of the requisite harm to be “demonstrated with absolute certainty”, it being sufficient “to show that damage is foreseeable with a sufficient degree of probability”, applicants “are required to prove the facts forming the basis of their claim that serious and irreparable damage is likely” (Case C-335/99 P(R) HFB Holdings v Commission, at para 67. [33]

 

115.          A review of these decisions makes it clear that the threshold for establishing serious and irreparable harm is, in practice, a high one. In many - if not most - of the decided cases, applications for interim measures fall at this hurdle. A searching inquiry is to be undertaken. A probability of concrete harm must be demonstrated. Hypothetical or speculative assertions of harm will not justify the granting of interim measures: see, for example, Joined Cases T-195/01 R and T-207/01 R Government of Gibraltar v Commission, para 101. Furthermore, only harm to the applicant can be taken into account for the purposes of establishing urgency. Harm to third parties or wider considerations of harm to the public interest, can be taken into account only when balancing the interests at stake, assuming that the assessment reaches that stage: Case T-13/99 R, Pfizer Animal Health SA/NV v European Council, at para 136.

 

Harm of a Pecuniary Nature

 

116.          As already stated, the EU jurisprudence indicates that, in principle, alleged harm of a pecuniary nature can be remedied and cannot, at least in the absence of exceptional circumstances, be regarded as irreparable. That is so notwithstanding the limited circumstances in which a claim for non-contractual liability against an EU institution can be successfully maintained, limitations which are mirrored as regards claims for Francovich damages.

 

117.          As already noted, it appears to be common case that, insofar as Three might have a claim for damages against ComReg arising from the spectrum auction and its outcome, it would be a Francovich claim. [34]

 

118.          Three accepts that the harm that it apprehends is exclusively pecuniary harm or loss and it does not suggest that, in the absence of the stay sought, it would be liable to suffer any form of non-pecuniary loss. [35] It was right to do so. Any loss that would arise if Three was compelled to “overpay” for spectrum would obviously be pecuniary in character. Equally, any loss of market position/market share resulting from a failure to secure spectrum in the auction would be purely pecuniary in character: “[a] market share can thus be represented in financial terms, as the holder of that market share can benefit from it only in so far as it generates profits for him” (Case T-95/09 R United Phosphorus Ltd v Commission, para 64).

 

119.          Case T-95/09 R United Phosphorus Ltd v Commission is relied on by Three as demonstrating that, even with the framework of Zuckerfabrik, there is no absolute rule that purely pecuniary loss cannot constitute serious and irreparable harm.

 

120.          The factual position in United Phosphorus Ltd v Commission was complicated. United Phosphorus was active in the development, manufacture and distribution of plant protection products. It used napropamide as an active substance in the products made by it. Directive 91/414 EEC established a framework for the evaluation and approval of plant protection products and their active substances. Following a prolonged process of evaluation, the Commission adopted a decision in November 2008 excluding napropamide from the list of active substances approved for marketing and directing Member States to withdraw existing authorisations for products containing napropamide by 7 May 2009 and not to grant/renew any further such authorisations with effect from the date of publication of the decision. The decision allowed a grace period of up to 7 May 2010 for the marketing and use of existing stocks of napropamide-containing products. In December 2008, United Phosphorus resubmitted an application for assessment of napropamide on an accelerated basis. It also sought the annulment of the Commission’s decision and sought an order suspending the decision (without prejudice to the resubmission) and order directing the Commission to instruct Member States to re-instate any authorisations withdrawn or refused as a result of the decision.

 

121.          Being satisfied that the applicant had established a prima facie case that the decision was invalid, the President addressed the issue of urgency. He noted that damage had to be foreseeable with a sufficient degree of probability and that the party invoking damage is required to prove the facts forming the basis of its claim that serious and irreparable damage is likely (para 32). He also noted that it was “well-established” that damage of a purely financial nature cannot “save in exceptional circumstances” be regarded as irreparable or even as being reparable only with difficulty (para 33). On the assumption that such damage exists, interim measures would be justified “only if it appears that, without such a measure, the applicant would be in a position that could imperil its existence before final judgment is given in the main action” (para 34). Loss of market share would only be relevant in this context if it involved a serious and irremediable effect on market share, regard being had to the characteristics of the corporate group to which the applicant belonged (para 35).

 

122.          It is evident from the judgment that the applicant put a significant amount of confidential information about its market and financial position, and the impact of the decision on it, before the President. The information included information as to its turnover, the proportion of that turnover attributable to sales of napropamide and napropamide products and as to the turnover of the group of companies to which it belonged. The applicant also gave the court information about its supply source for napropamide.

 

123.          The evidence before the President allowed him to assess what the impact on the turnover of the group would be. On that basis, and on the basis of the general economic conditions which exacerbated the impact of that loss of turnover, the President was “in these specific circumstances” obliged to acknowledge that the applicant had established the gravity of the harm it would suffer absent the interim measures sought (paras 67-71).

 

124.          As to whether such financial harm would be of an irreparable nature, “that harm, in so far as it is not rectified solely by implementation of the judgment in the main action, may be made good by the means of redress provided for in Articles 235 EC and 288 EC … given that the mere possibility of bringing an action for damages is sufficient to show that such financial harm is ‘in principle reparable’” (Para 73). Any such “pecuniary reparation” would not, however, be obtained for several years. Furthermore, the President emphasised, “the judge dealing with the application for interim measures must not apply mechanically and rigidly the condition relating to the irreparable nature of the financial harm pleaded, but must take account of the factual and legal circumstances specific to each case … and determine, in the light of those specific circumstances, the manner in which those conditions of urgency are to be examined” (Ibid).

 

125.          As to whether there were “specific circumstances” which could justify a finding of urgency notwithstanding the “in principle reparable” nature of the harm asserted by the applicant, the President referred to the fact that the applicant had resubmitted an application for assessment of napropamide pursuant to the accelerated procedure. That procedure could last between 14 and 19 months at most but that period could be significantly shortened with the co-operation of the parties and the authorities (para 76). It followed that the resubmission procedure could be concluded only a few months after the 7 May 2008 deadline for the withdrawal of authorisations for napropamide-containing products. It was “not improbable” that the procedure would result in the marketing of napropamide being authorised and, in those circumstances, it “would be unreasonable to allow the prohibition” on its marketing to take effect (para 77). That factor, in conjunction with the fact that the evidence adduced by the applicant established that, absent a stay, the applicant would by reason of probable difficulties of supply and/or availability of manufacturing capacity face significant problems in resuming production, even if napropamide were to be included as a permitted active substance (paras 78 - 81), led the President to conclude that “that the present case is characterised by specific circumstances establishing the existence of urgency” (para 82). The order granted by the President suspended the operation of the Commission’s decision up to 7 May 2010 (when the grace period for the disposal of existing stocks would expire).

 

126.          Case T-95/09 R United Phosphorus Ltd v Commission does not call into question the general principle that damage of a purely financial nature cannot be regarded as irreparable. It does, however, caution against an excessively rigid or formulaic application of that principle. The specific circumstances of each case must be considered and may, exceptionally, justify departing from that general principle.

 

127.          That, no doubt, is potentially helpful as far as applicants for interim measures are concerned. But the circumstances in Case T-95/09 R United Phosphorus Ltd v Commission were very particular. As already noted, there was very detailed evidence as to the financial and market position of the applicant and of the group of which it was part, and of the impact of the withdrawal of napropamide. There was also detailed evidence to the effect that, in the absence of a stay, the applicant would face significant hurdles in resuming production, even if napropamide was subsequently authorised for use. There was also the additional factor that napropamide could be re-authorised relatively shortly after the cut-off date for the withdrawal of authorisations under the impugned decision. The facts and circumstances here are quite different and the evidence put before the court by Three falls significantly short of establishing any probable impact of the Decision on it that is in any way equivalent to the impact on the applicant demonstrated in Case T-95/09 R United Phosphorus Ltd v Commission.

 

            The First Head of Harm Identified by the Judge - Strategic Bidding by Vodafone          and/or eir

 

 

128.          The fundamental complaint made by Three in this regard, as I understand it, is that by reason of errors made by ComReg in its auction design, and in particular its imposition of the Sub-1 GHz spectrum cap, Vodafone and eir have the capacity and incentive to engage in strategic bidding for spectrum (bidding at a level in excess of the intrinsic value of the spectrum to the bidder in order to deny spectrum to their competitors and/or to increase the price that their competitors will be required to pay for spectrum) which will cause harm to Three, either because it will have to pay more for spectrum than it otherwise would have to (in the absence of the Sub-1 GHz spectrum cap) or because it will not succeed in obtaining the spectrum that it needs in order to maintain its competitive position (at page 29 of his Judgment, the Judge records that Three’s counsel had submitted that this element of Three’s case was not just about financial loss but about “the loss of competitive power”).

 

129.          As already discussed, if Three’s complaints about the design of the auction are made out in its appeal, the High Court can - and presumably will - grant it appropriate relief by way of setting aside the relevant aspects of the Decision. By reason of the stay granted by this Court after the hearing of the appeal, there is no risk that spectrum will be allocated prior to the High Court’s decision on the appeal. Annulment (setting aside) of the relevant parts of the Decision will avoid any potential harm to Three arising from the possibility that Vodafone and/or eir may engage in strategic bidding. The stay granted by the Judge cannot therefore be sustained by reference to this head of alleged harm.

 

130.          It is not, therefore, strictly necessary to address the Judge’s analysis on this issue. But it has been fully debated in argument and it may be useful to offer some brief observations on the Judge’s approach, given that stay applications of this kind are inevitably going to come before the courts again.

 

131.          The starting point is to emphasise that (as was common case) the onus was on Three to establish, firstly, that as a matter of probability such strategic bidding would occur and, secondly, that, as a matter of probability, it would cause serious and irreparable harm to Three.

 

132.          As to the likelihood of strategic bidding occurring, Three’s own evidence was notably circumspect. Significantly, it appears that neither Mr Hickey nor Mr Marsden was prepared to swear that strategic bidding was probable. Each discussed the prospect that such bidding might occur in highly conditional terms. [36] Indeed in his second report, [37] while noting that the Sub-1 GHz Spectrum Cap made “strategic price driving behaviour more attractive for Eir and Vodafone”, Mr Marsden appeared to be at pains to make it clear that he was not claiming “any special insight into whether Eir or Vodafone will engage in strategic bidding. They may do so or may not.”

 

133.          The fact that none of Three’s deponents was prepared to assert on oath that strategic bidding was probable might be thought to be a fatal obstacle to any finding to that effect. The Judge nonetheless made such a finding. He did so on the basis of ComReg’s analysis of the risk of strategic bidding in the Decision and on statements that had been made by ComReg’s expert, Mr Maldoom (Judgment, pages 29-31).

 

134.          As to the first of those elements, ComReg in its Decision had concluded that, under a cap level of 80 MHz there was a “real potential” for Three and Vodafone to bid strategically for 700 MHz spectrum for the purpose of denying 700 MHz spectrum to eir (para 6.180). “Real potential” is, it appears, the applicable threshold for ex ante regulatory intervention. Accordingly, it was not necessary for ComReg to be satisfied that such strategic bidding was probable and it did not purport to make any such finding conclusion. ComReg’s analysis identified a number of factors said to support its conclusion. While some of those factors relate to the market generally (para 6.180 (a) - (d)), the majority (para (e) - (j)) relate specifically to the particular position of eir in the mobile market. ComReg then undertakes a detailed assessment of the impact on eir if it was prevented out from winning any 700 MHz spectrum, concluding that it would significantly diminish eir’s capacity to compete effectively in the market and to operate as an effective competitive restraint on Vodafone and Three (6.183 - 6.211). That analysis led ComReg to impose the Sub 1 GHz spectrum cap fixed at 70 MHz rather than 80 MHz.

 

135.          In my opinion, this analysis - which is, of course, vigorously contested by Three - simply cannot do duty in the manner suggested by Three. It is clearly specific to eir and its vulnerability to strategic bidding in light of its particular market position. The analysis cannot fairly or meaningfully be transposed to Three, which is in a very different market position to eir (that is, of course, a key element of ComReg’s analysis and is dealt with in considerable detail in the Decision). There is nothing in the Decision to suggest that Vodafone and/or eir would have the same incentive to engage in strategic bidding in order to deny spectrum to Three (in fact, there appears to be nothing in the Decision to suggest that eir might engage in strategic bidding in any circumstance).  Furthermore, ComReg’s analysis (which was directed to establishing only whether there was a “real potential” for anti-competitive effects) was carried out on the assumed premise of a 80 MHz spectrum cap, whereas ComReg ultimately decided to impose a 70 MHz cap. That is another reason why the analysis in ComReg’s Decision cannot be transposed to support the contention (made, as noted, in argument rather than in Three’s evidence) that, in the event that the Main Stage was permitted to proceed (under the auction rules including the 70 MHz cap), it was probable that there would be strategic bidding by Vodafone or by eir causing harm to Three. 

 

136.          I therefore respectfully disagree with the Judge’s view that the concerns expressed in the Decision as to the perceived incentive of Three to engage in strategic bidding “apply with the same force to eir and Vodafone” (Judgment, page 31).

 

137.          There is a further, and significant, problem with this aspect of the Judgment in my view. In its appeal against the Decision, Three had frontally challenged ComReg’s analysis around the risk of strategic bidding by it. Three’s position was that the analysis was deeply flawed and that ComReg’s finding that there was a “real potential” for Three (and Vodafone) to bid strategically for 700 MHz spectrum was unsustainable and ought to be set aside. That was a central element of its appeal from the Decision. That being so, it is very difficult to understand how it could be permissible for Three to seek to rely on that same analysis for any purpose, still less as being probative of the probability of strategic bidding by Vodafone and/or eir. Such opportunistic approbation and reprobation ought not to be permitted.

 

138.          As for the statements made by Mr Maldoom in paragraphs 130 - 131 of his report of 20 June 2022 they do not, in my view, provide a sufficient evidential foundation for a finding that strategic bidding harmful to Three would probably occur in the event that ComReg was permitted to proceed with the auction. That part of Mr Maldoom’s report addressed the position of Three (not Vodafone or eir) in a hypothetical counterfactual scenario in which the pricing methodology was modified so to as guarantee Three a discount on the price that it would have to pay for any lots of 700 MHz spectrum awarded to it. Even Mr Marsden - whose evidence was largely discounted by the Judge because he had crossed the line between expert and advocate - did not go so far as to suggest that Mr Maldoom’s analysis established a probability of strategic bidding harmful to Three (see Mr Marsden’s second report, para 58-59). In my view the Judge erred in attaching the weight that he did to what had been stated by Mr Maldoom.

 

139.          I also have difficulty with the finding that serious and irreparable harm would be suffered by Three. On Three’s case, it was not even certain what harm it might suffer. On the one hand, it said that it might have to overpay for spectrum. The evidence suggested that any such overpayment would be quantifiable (being the difference between the price paid by Three for the lots of 700 MHz spectrum awarded to it and the reserve price of those lots). As a matter of principle, any loss resulting from such “overpayment” would quintessentially be a pecuniary loss. There was no evidence before the Court that could have allowed it to conclude that such loss would be “serious” as far as Three was concerned, given the evidence as to its financial position and the resources available to it. Moreover, ComReg had given an undertaking which, on its face, addressed this potential head of loss.

 

140.           On the other hand, Three said that it could be denied essential spectrum. Understandably, that scenario appears to have exercised the Judge more than the prospect of Three overpaying for spectrum. However, there was no evidence that Three was subject to any effective constraints on what it could pay for 700 MHz spectrum. On the contrary, the evidence established that Three had significant resources available to it, both internally and from its parent. Of course, Three might be outbid at auction but that is possible in any auction scenario. Three had to establish that, as a result of the Sub-1 GHz spectrum cap, it was likely to be out-bid for 700 MHz spectrum which, but for that cap, it would have secured. The evidence fell far short of establishing that scenario, which at all times remained wholly speculative and hypothetical. Nor was there any evidence from Three as to the impact on it of failing to obtain the 700 MHz spectrum it wanted. Three must have engaged in detailed and sophisticated financial modelling in order to put a value on the spectrum being auctioned and to formulate its bidding strategy. It must have formed a judgment as to the amount of 700 MHz spectrum that it needed and assessed the impact of failing to obtain such spectrum. No doubt, as Mr Kennelly observed in argument, much of that analysis would be commercially confidential but arrangements could have been made to put it in evidence before the High Court in a manner which would have protected that confidentiality. Instead, the Court was invited to proceed on the basis of general statements as to the importance of the availability of spectrum to operators such as Three. No doubt, it is important but that did not absolve Three from putting before the Court evidence as to the actual impact on its business in the event that it lost out on spectrum as a result of strategic bidding by Vodafone or eir. As will be evident from the discussion above, I do not share the Judge’s view that Three was not in a position to provide a reliable estimate of the loss that it would suffer in such a scenario.

 

141.           Finally, as Mr Kennelly properly accepted in argument, any harm that might be suffered by Three, including harm under this heading, was exclusively pecuniary in nature. Prima facie, such harm is reparable. As Case T-95/09 R United Phosphorus Ltd v Commission illustrates, that rule is not absolute. But Three has not identified any particular feature of this case that would justify a departure from that rule here. Case T-95/09 R United Phosphorus Ltd v Commission is relevant by way of contrast rather than comparison. The detailed evidence before the President in Case T-95/09 R United Phosphorus Ltd v Commission as to the financial and market position of the applicant and the impact on it and on the wider corporate group of the contested Commission decision is strikingly absent here.

 

142.          I have not overlooked the Judge’s view that the harm to Three caused by strategic bidding engaged by Vodafone and/or eir should be regarded as serious and irreparable because ex post facto it would be very difficult for Three to establish that such bidding had occurred (Judgment, pages 34-36). Again, it is notable that this finding was based not on any evidence put before the High Court by Three but on the evidence of Mr Maldoom on behalf of ComReg. The point being made by Mr Maldoom was that the claimed harm relied on by Three was very difficult to make out because of the difficulties in drawing any hard line between legitimate commercial bidding on the one hand and illegitimate (though not unlawful) strategic bidding on the other. In my view, that evidence (which does not appear to have been disputed) significantly undermines Three’s case that strategic bidding would probably cause serious and irreparable loss to it and the Judge’s finding to that effect. If such harm cannot be established ex post, how can it be established as a matter of probability ex ante ? Nor am I persuaded that there is any meaningful analogy with American Cyanamid or Campus Oil in this context. Apart from the fact that a different threshold of proof is applicable here, there was no uncertainty in American Cyanamid or Campus Oil as to what the wrongdoing alleged against the defendants was or how it could be proved. The only relevant uncertainty was in the calculation of loss. The uncertainty here is of a much more fundamental and far-reaching character.

 

143.          In the result, I do not consider that the evidence supports the Judge’s conclusion that, absent a stay, there was a probability of serious and irreparable harm to Three by reason of the likelihood that Vodafone and/or eir would engage in strategic bidding.

 

The Second Head of Harm Identified by the Judge - The Integrity of any Re-run Auction

 

144.          The second head of serious and irreparable harm identified by the Judge was that, in the event that the auction was permitted to proceed but subsequently had to be re-run as a result of Three succeeding in its appeal, the integrity of the process would be compromised by reason of the disclosure of the bidders’ respective positions in the first auction. The Judge considered that such harm would be “very serious” (Judgment, page 41) and it was a significant factor in his assessment of the balance of interest (Judgment, page 47).

 

145.          As the Judge explained in his Judgment, this alleged harm had been identified very briefly in Three’s written submissions and had also been adverted to (again in brief terms) by its expert, Mr Marsden, in his report of 8 June 2022. However, the Judge did not base his findings on Mr Marsden’s report but rather relied on an affidavit which had been sworn on Vodafone’s behalf by Andrew Corcoran, its Head of Regulation.

 

146.          A finding that the integrity of any re-run auction process, however designed, would inevitably be compromised is a significant one and, in my view, is one which would require clear and compelling evidence. In my view, the evidence here does not sustain that finding.

 

147.          The suggestion that any re-run of the auction process would be compromised was vigorously disputed by ComReg. ComReg’s position was set out in detail by Mr Merrigan (first affidavit, para 118(a) - (c)) and by Mr Maldoom (report of 20 June, 2022, paras 138-150). The Judge noted that conflict of evidence and noted that he could not resolve disputed issues of fact. It was, he observed, “impossible to say whether ComReg or Three or Vodafone is correct” (Judgment, page 41). But if that was so, it is difficult to understand how the Judge then proceeded to find that the alleged harm had been established as a matter of probability. That necessarily involved the Judge implicitly resolving the conflict of evidence by preferring the evidence of Mr Corcoran (though, in fact, Mr Corcoran did not go so far as to assert that such harm was probable) over the evidence of Mr Merrigan and Mr Maldoom.

 

148.          In light of the Supreme Court’s decision in RAS Medical Ltd v Royal College of Surgeons [2019] IESC 4, [2019] 1 IR 63, such an approach was not appropriate. The onus of establishing a probability of harm was on Three. The Court was not concerned with whether there was a risk of harm or an arguable case that harm would occur; nor, in contrast to the position in IBB Internet Services Ltd v Motorola Limited [2013] IESC 53, was the court concerned with whether there was “reason to believe” that harm might or would occur. If Three’s case depended on factual assertions which were disputed - as was the case here - and if the Court was not in a position to resolve that dispute (as it might have been if Three had sought to cross-examine Mr Merrigan and//or Mr Maldoom), then it followed that Three was not in a position to discharge the onus on it.

 

149.          That issue aside, the evidence of Mr Corcoran is expressed at a level of assertion and generality and it does not identify, in any concrete or specific way, what information bidders would learn in the first auction process that would have the effect of compromising any re-run. Nor does Mr Corcoran explain how the disclosure of such information within the first auction process would fatally compromise any subsequent process without, it seems, negatively impacting on the competitiveness and integrity of the first process. In this context, it will be recalled that the “price discovery” that occurs within the auction process is seen as positively contributing to efficient outcomes. Mr Corcoran does not address that apparent paradox.

 

150.           Another key issue that Mr Corcoran fails to address arises from the fact that the auction process would only be re-run if Three were to succeed in its appeal. For Three to succeed, the High Court would have to find that the auction rules were significantly flawed, whether by reason of the imposition of the Sub 1 GHz spectrum cap, ComReg’s choice of the CCA format or the combination/interaction of those two elements. Any such a finding would necessitate the re-design of the auction process by ComReg. Any re-run auction would inevitably proceed under materially different rules. Mr Corcoran failed to take account of that fundamental fact and so too, with respect, did the Judge.

 

151.          In argument, Mr Kennelly contended that, even so, the concerns articulated by Mr Corcoran remained valid. Any information about the commercial strategy of other bidders risked distorting the outcome of a subsequent auction, even if held some significant time after the initial process (as would inevitably be the case, given the need for ComReg to engage in a further consultation process before making any decision on how to proceed) and even if conducted under different rules. That is, in my view, wholly unpersuasive. Three’s entire case is that auction design critically affects auction dynamics and outcome. An auction without a Sub 1 GHz spectrum cap (or with a differently configured cap) would have significantly different dynamics to an auction with such a cap. An auction that uses a format other than CCA will have significantly dynamics to a CCA auction.

 

152.          In the scenario being discussed, ComReg would of course be aware of the fact that an auction process had already taken place. It would be aware of the price discovery that took place during that process. It would obviously have a compelling interest in ensuring the integrity and effectiveness of any re-run auction process. There is no evidence that it could not adopt an auction design capable of achieving that objective.

 

153.          It follows that I do not consider that the evidence supports the Judge’s conclusion that, absent a stay, there was a probability of serious and irreparable harm to Three under this head.

 

THE BALANCE OF INTERESTS

 

154.          The balance of interest does not strictly arise in light of my conclusion that Three has failed to establish urgency in the form of serious and irreparable harm. It was, however, fully argued in the appeal and I will address it for completeness.

 

155.          The Judge comprehensively identified, and carefully assessed, the factors weighing for and against the granting of a stay. At page 46 of his Judgment, he stated that, if the stay sought was likely to be substantial in duration, it would be very difficult to conclude that the interests favouring a stay outweighed the public interest in the implementation of the Decision. He then identified three factors which, in his view, lessened the weight to be given to the public interest in the immediate implementation of the Decision. The first was that any stay was likely to be of a relatively short duration. The second was that ComReg would be free to issue temporary licences for use of the 700 MHz and 2.1 GHz spectrum during the period of any stay, thus mitigating the effect of the stay to some extent. The third factor identified by the Judge was the fact that, if the appeal succeeded and the Decision was annulled, the ability to re-run the auction successfully was likely to be very seriously compromised (at pages 46-47).

 

156.          By the time ComReg’s appeal came before this Court, the question of the duration of the stay had a quite different complexion. As already explained, the evidence before this Court established a very significant risk that, if the stay ordered by the High Court was permitted to remain in place until the High Court gave its decision on Three’s appeal (on 31 January 2023) then, even if the appeal was unsuccessful, a further significant period of time would elapse before the auction could proceed to completion, with the Main Stage commencing only late in 2023 or in early 2024. If that had been the picture presented to the Judge, it seems very likely that, on that basis alone, he would have reached a different conclusion as to where the balance of interest lay.

 

157.          In any event, that is the position as it is presented to this Court. The prospect that the auction process could be delayed to that significant extent was a significant factor against the maintenance of the stay granted by the High Court.

 

158.          As to the second factor identified by the Judge - the possibility of a temporary licensing regime being put in place by ComReg - while that is no doubt a factor of some relevance, the weight that can properly be attached to it is, in my view, significantly diminished given that the period during which any temporary licensing regime would be in place was likely to be much longer than the “few months” envisaged by the Judge. As the Judge himself recognised, a temporary licensing regime was no substitute for the allocation of spectrum pursuant to the Decision, with all of the significant network investment and roll-out obligations it would entail for successful bidders.

 

159.          For the reasons already given, the third factor identified by the Judge is not one which arises at all in my view.

 

160.          The Decision here is one of enormous consequence for the State. The economic implications of delaying its implementation are very significant (these are set out in detail in the reports of Mr Maldoom and Mr Clinch). Any loss or damage to the economy or to society arising from such delay is, of course, irrecoverable. There is a very powerful interest in giving effect to the Decision and allowing it to be implemented. That is the default position generally under the Framework Regulations. But that default position applies with particular force in the particular circumstances, given the nature and scope of the decision at issue. In my view, the Court should be slow to stay a decision of this kind and should do so only where there are compelling countervailing considerations.

 

161.          There are no such compelling considerations here. Having regard to the stay granted by this Court, the effectiveness of Three’s appeal will not be undermined by permitting the Main Stage of the auction process to proceed. Should Three succeed in persuading the High Court that the Decision is flawed to the extent that it should be annulled, that will provide a complete and effective remedy for Three.

 

162.          The balance of justice is therefore decisively against continuing the stay granted by the High Court.

 

WOULD THE APPLICATION OF OKUNADE/WORD PERFECT PRODUCE A DIFFERENT OUTCOME?

 

163.          The application of the Zuckerfabrik test therefore dictates that ComReg’s appeal should be allowed. Would there be a different outcome if instead the Court applied the the principles set out in Okunade, Word Perfect and the other authorities referred to earlier in this judgment?

 

164.          The application of Okunade would not affect my assessment of the first head of harm identified by the Judge. Any risk of such harm is excluded by the stay granted by this Court and that is so regardless of whether the test is that in Zuckerfabrik or that in Okunade.

 

165.           As regards the second head of harm, it may be that, as a matter of Irish law, a less exacting approach applies in terms of establishing the likelihood of loss and damage than applies under Zuckerfabrik. Undoubtedly, a different approach applies to assessing the adequacy of damages in this context, as is illustrated by Word Perfect. But even if one were to apply a lower threshold to establishing the likelihood of loss and damage than is applicable under Zuckerfabrik, the evidence before the Court does not establish such a likelihood in my view. But, even if one assumes for the sake of this analysis that Three could establish a sufficient likelihood of loss and damage, for which damages would not be an adequate remedy because of the restrictive conditions applicable to the recovery of damages from a body such as ComReg (or from the State) in the circumstances here, it would not entitle it to a stay on the Decision. The Court would have to assess the balance of convenience (more appropriately referred to, perhaps, as the balance of justice). In that assessment, as Clarke J explained in Okunade, the public interest in giving effect to the Decision would have to be given significant weight, as would the economic and societal implications of any significant delay in implementing the Decision. Against that, there would be risk of commercial loss to Three, as well as the wider public interest in the integrity of any re-run of the auction process that might be triggered in the event that Three succeeds in its appeal. In my view, the balance of justice in that scenario would clearly be against continuing the stay granted by the High Court.

 

THE STAY GRANTED BY THIS COURT

 

166.          I should, before concluding, return to the considerations that led the Court to grant the stay it did following the conclusion of the appeal hearing.

 

167.          Following engagement between the Court and counsel at the start of the appeal hearing, it became evident that, in the event that the stay granted by the High Court was simply discharged, there was a real possibility that the remaining stages of the auction process could be concluded before the High Court was due to give judgment on Three’s appeal and that, by the time judgment was given, licences might actually have been awarded by ComReg. Counsel for ComReg indicated that her instructions were that, in the event that the High Court quashed the Decision and found that the auction had not been lawfully conducted, ComReg would re-run the auction. Counsel indicated that, if licences had already been awarded when the High Court gave judgment, ComReg would seek the withdrawal of such licences. [38]

 

168.          In the Court’s view, the prospect that spectrum might actually be awarded before the High Court gave judgment on Three’s appeal raised immediate and significant concerns that the effectiveness of that appeal might be undermined. If new legal rights and interests had been created by the time the High Court adjudicated on Three’s appeal, its capacity to provide an effective remedy to Three in the event that its appeal was successful might be significantly impaired. In the Court’s view, that should not be permitted to occur and the Court made its position clear to the parties.

 

169.          In response to the concerns expressed by the Court, ComReg through its counsel indicated that it would be prepared to run the award process up to the point of grant only (i.e. without making any grant). [39] The Court did not, however, take ComReg to have necessarily agreed to such an order being made and Counsel for Three made it very clear to the Court that his client’s fundamental position was that the stay granted by the High Court should remain in place. Nonetheless, without prejudice to their respective positions, the parties were able to agree the stage in the auction process that ComReg should be restrained from going beyond, in the event that the Court decided to make such order. That was very helpful and that agreed position was reflected in terms of the order made by this Court on 8 November 2022 restraining ComReg from notifying the Winning Bidders of their entitlement to apply for licences as provided for in paragraph 3.259 of the Information Memorandum or taking any subsequent steps set out in the Information Memorandum pending the determination of the proceedings in the High Court and/or further Order.

 

170.          A stay in such terms was, in the Court’s view, clearly appropriate to ensure the effectiveness of Three’s appeal. It will be recalled that the purpose of the jurisdiction to grant interim measures under Regulation 7(2) of the Framework Regulations is stated to be “securing the effectiveness of the hearing and determination of the appeal”. That was also the rationale for the Court of Justice’s decision in Factortame: see para 107 above. The same essential point is made in Zuckerfabrik, in which the Court of Justice indicates that the court adjudicating on an application for interim measures must consider “whether immediate enforcement of the measure which is the subject of the application for interim relief would be likely to result in irreversible damage to the applicant which could not be made good if the [measure] were to be declared invalid” (at para 29).

 

171.          Whether under EU law (Factortame/Zuckerfabrik) or under domestic law (Okunade/Dowling), the right to an effective remedy is rightly regarded as a fundamental element of the rule of law and the principle of legality. Where immediate enforcement of the contested measure would result in an irreversible alteration in the status quo, so that any order that may subsequently be made annulling the measure would be deprived of any concrete effect, the effectiveness of the law, and the right to an effective remedy, would be gravely undermined.

 

172.          In the very particular (and, arguably, exceptional) circumstances here, where in the absence of a stay, the impugned Decision could have produced new and potentially irreversible legal rights and relationships beyond the reach of the High Court adjudicating on Three’s appeal, the limited intervention of this Court was required to preserve the effectiveness of that appeal. 

 

 

CONCLUSION

 

173.          As I have explained, the Court has already given its decision on this appeal. The above sets out my reasons for that decision.

 

174.          On 15 December 2022, when this judgment was at an advanced stage of preparation, the Court was notified of the fact that Three had issued a motion in the High Court seeking leave to discontinue its appeal. The Court was subsequently made aware that, on 19 December 2022, the High Court had made an order permitting the discontinuance of the appeal, with the consent of ComReg. In these circumstances, the order made by this Court on 8 November 2022 must be discharged. Even so, the Court took the view that, having given a decision on ComReg’s appeal on the promise that it would give its reasons later, it should proceed to give those reasons.

 

Haughton and Allen JJ concur with this judgment