England and Wales Court of Appeal (Civil Division) Decisions

IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE HIGH COURT OF JUSTICE
CHANCERY DIVISION
MR JUSTICE MORGAN

[2008] EWHC 1978 (Ch)

		Royal Courts of Justice Strand, London, WC2A 2LL
		28 July 2009

Lord Justice Lloyd:

Introduction

1. Horseracing is a substantial business in the UK. Racecourses and bookmakers are among the several essential elements in that business. They are interdependent: bookmakers draw in funds from punters but they need races, and therefore racecourses, without which this aspect of their turnover would not exist, and punters provide the most important source of outside income for racing. Betting on horseracing provides a substantial proportion of the turnover of British bookmakers.
2. The present litigation is between the interests of several major bookmakers, as Claimants, and those of about half of the racecourses in Great Britain, as Defendants. Not for the first time, recourse has been had to competition law in an attempt to affect the distribution of economic resources within the business of horseracing.
3. One element in the economic pattern, though not relevant to the present dispute, is that the betting industry pays substantial sums, which are passed to British racecourses, by way of the Betting Levy, administered by the Horserace Betting Levy Board. In 2000 the Government announced its desire to abolish the levy and opened a consultation as to the implications of that. In 2002 the proposal to abolish the levy was withdrawn, and the levy has been renewed year by year since then.
4. Betting takes place either at racecourses (on-course) or elsewhere (off-course), the latter largely at licensed betting offices (LBOs) of which there are some 8,700 in Great Britain, though nowadays it may also take place over the internet. Betting at LBOs is not limited to horseracing: greyhound racing, other sports, numbers games, and other events or contingencies provide suitable and popular subjects for betting, but the bulk of LBO turnover is on horseracing.
5. Until 1986 it was illegal to show live televised pictures of racing in an LBO. The law changed in that year. In 1987 such a service began to be provided for LBOs by Satellite Information Services Ltd (SIS). (At first it was supplied by its parent company, but nothing turns on the distinction for present purposes.) The parent company, Satellite Information Services (Holdings) Ltd (SISH) was founded by the then four leading bookmaking firms, in anticipation of the change in the law. At the time relevant to the present dispute it was owned as to about 49% by Ladbrokes PLC, William Hill Organization Ltd and the Horserace Totalisator Board and as to about 7.5% by Mr Fred Done. Directly or indirectly it acquired licences to broadcast to LBOs live coverage of horse races at courses in Britain and also in Ireland. (I will refer to the relevant rights as LBO media rights.) With that and other material it put together a live television service broadcast by satellite (called SIS FACTS) to which LBOs could subscribe. Until 2007 it had no competitor. It was therefore the sole buyer for the media rights which racecourses could offer, and the sole seller to LBOs of a broadcast service of this kind. Some races were and are broadcast live by terrestrial channels, including the BBC, and LBOs can show those races using the terrestrial service, but the SIS FACTS coverage was much more comprehensive.
6. In 2007 a rival service to SIS FACTS came into play, providing a similar broadcast service, known as Turf TV. This is provided by the First Defendant, Amalgamated Racing Ltd (AMRAC), which is a joint venture between the Fourth Defendant, Alphameric Gaming Ltd, and the Fifth Defendant, Racecourse Media Services Ltd (RMS), which itself is owned by 19 undertakings which run 31 out of the 59 racecourses in Britain. Thus, SIS FACTS is provided by interests aligned with major bookmakers, whereas Turf TV is provided by interests which include those representing a significant number of racecourses. Both of those services are regarded as essential to an LBO: they are "must-haves".
7. By the present proceedings, the principal bookmakers seek to undermine the Turf TV service, by showing it to be the product of a cartel illegal under the competition law of the European Union and of the UK. The racecourses object strongly, contending that it is absurd to try to use competition law in order to strike down a successful attempt to create competition in a situation where for 20 years there was none.
8. The proceedings came to trial before Mr Justice Morgan over some 6 weeks in 2008. He gave judgment in favour of the Defendants on the issues relevant to this appeal on 8 August 2008, and dismissed the claim: [2008] EWHC (Ch) 1978. He gave a later judgment on issues on a counterclaim ([2008] EWHC 2688 (Ch)); having done so he granted permission to appeal. I would like to pay tribute to the clarity of his long judgment, in the course of which, quite apart from having to consider the facts in more detail than we have had to, and to decide issues on the evidence, mainly concerned with that given by expert witnesses, he had to address several additional issues besides those which have been live on the appeal. Moreover, during the trial he had also to deal with a number of interlocutory points (including points as to the formulation of the Claimants' claim, on which he gave a preliminary judgment, [2008] EWHC 2503 (Ch)) and also with points on the counterclaim, as well as hearing argument on one point, under article 81(3) of the EU Treaty, on which it was not necessary for him, in the end, to deliver a judgment.
9. In grappling with the issues on the appeal, we have been assisted by clear and concentrated written and oral submissions from Counsel (who, in the case of the Appellants, were a new team who had not appeared before the judge), enabling us to conclude the hearing of the appeal in three rather full days.

The parties to the proceedings

10. The parties to the appeal are as follows. The Claimants included three of the largest four bookmakers in Great Britain: William Hill and Ladbrokes, each of which operates about 23% of British LBOs, and Done Bros, which trades as BetFred, with about 7% of British LBOs. However, Done Bros, although initially a party to the appeal, settled with the Respondents before the appeal hearing. Coral Racing Ltd (16%) had been a Claimant but settled with all parties before the trial. The other Claimant, BAGS, is a company limited by guarantee, established in 1967, which has as its members 22 of the UK's 680 off-course bookmakers. It was originally set up to encourage greyhound racing at times when horse racing might be cancelled because of bad weather, so as to provide an alternative betting opportunity. BAGS took a licence of LBO media rights from 49 racecourses, and granted a non-exclusive sub-licence of those rights to SIS.
11. The First Defendant is AMRAC, which is the joint venture company which provides the Turf TV service to LBOs. The Second Defendant, Racing UK Ltd (RUK), was set up to exploit the media rights of 30 racecourses, responsible between them for 54% of off-course LBO betting turnover in Britain. The two parties to the joint venture are the Fourth Defendant (of which the Third Defendant is the parent company) and the Fifth Defendant, RMS. The shares in RMS are held by the undertakings which run the 30 RUK courses, and by that which runs Ascot. The grant of LBO media rights for Ascot to AMRAC occurred in circumstances which the Claimants do not allege was in breach of competition law, so Ascot is not a Defendant.
12. The 6^th to 23^rd Defendants are the 18 undertakings which are responsible for the 30 RUK racecourses. Racecourse Investments Ltd (RIL) was once called the Racecourse Holdings Trust, and is a subsidiary of Jockey Club Racecourses Ltd. It operates 14 courses, of which all but one granted rights to AMRAC. (The other, Exeter, was previously an independent course, and licensed its LBO media rights to SIS at that time.) The other Defendants are the operators of the 17 other RUK racecourses.
13. SIS was also a party at trial, as a Third Party, and was represented and called its own expert and other evidence. It is not a party to the appeal.

British racecourses

14. The various British racecourses can be seen as forming different groups, not always in any formal way, and not always in a static formation. In all there are 59. (One additional course, Great Leighs, opened in 2008 but we were told it has since gone into administration.)
15. The RUK courses form the largest group, of which the 13 Jockey Club (or RIL) courses (i.e. all of them except Exeter, as explained above) is the largest sub-group.

i) The Jockey Club courses are Aintree, Carlisle, Cheltenham, Epsom, Exeter, Haydock Park, Huntingdon, Kempton Park, Market Rasen, Newmarket, Nottingham, Sandown Park, Warwick and Wincanton.

ii) Within the RUK group there are also the Large Independent courses, namely Ayr, Chester, Goodwood, Newbury and York.

iii) The other, smaller independent courses within the RUK group are Bangor-on-Dee, Beverley, Cartmel, Catterick, Hamilton Park, Ludlow, Musselburgh, Pontefract, Redcar, Salisbury, Thirsk and Wetherby.

16. Other groupings include the Northern racecourses, the Arena racecourses and the former GG Media courses.

i) The Northern racecourses are those operated by Northern Racing Ltd. There were 8 of them when they granted exclusive LBO media rights to BAGS: Bath, Brighton, Chepstow, Fontwell Park, Great Yarmouth, Hereford, Newcastle and Uttoxeter. Since then Sedgefield has become a Northern course; it had granted exclusive rights to SIS.

ii) The Arena courses are those operated by Arena Leisure plc: Doncaster, Folkestone, Lingfield Park, Royal Windsor, Southwell, Wolverhampton and Worcester.

iii) The former GG Media courses were originally Exeter, Fakenham, Hexham, Leicester, Perth, Sedgefield, Stratford, Taunton, Kelso and Towcester. Exeter is now a Jockey Club course; Sedgefield is a Northern course and Towcester has left the group as well. The group of 8 of these including Exeter but not Sedgefield or Towcester is sometimes called the ICAC courses.

17. The courses not included in any of the groups described above are:

i) Ascot, which is independent but followed the RUK courses in granting exclusive rights to AMRAC;

ii) Newton Abbot, Plumpton, Ripon and Towcester, each of which independently granted exclusive rights to SIS.

18. We were shown a chart giving a ranking of racecourses according to the percentage of off-course betting attributable to their respective races: the top three are Newmarket, Lingfield Park and Ascot. Of the first 14, all except Ascot, Lingfield Park, Wolverhampton, Doncaster and Newcastle are in the RUK group; as mentioned above, Ascot chose to accept AMRAC's offer, so that it is aligned with the RUK courses for present purposes. Evidently, the RUK courses are of substantial interest and importance to LBOs.

Article 81 of the EU Treaty

19. The Claimants' attack in the proceedings, as pursued on appeal, focuses on the agreements and arrangements by which the RUK courses got to the stage of granting, and did grant, exclusive LBO media rights to AMRAC. They argue that in several respects these arrangements infringe basic rules of EU competition law. I will describe the practical arrangements and the agreements later. First, I will set out the relevant basic provision of EU law, now in article 81 of the EU Treaty. (It used to be numbered article 85, which accounts for references in earlier cases.)
20. Article 81(1) is as follows:

"The following shall be prohibited as incompatible with the common market: all agreements between undertakings, decisions by associations of undertakings and concerted practices which may affect trade between Member States and which have as their object or effect the prevention, restriction or distortion of competition within the common market, and in particular those which:

a) directly or indirectly fix purchase or selling prices or any other trading conditions;

b) limit or control production, markets, technical development, or investment;

c) share markets or sources of supply;

d) apply dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive disadvantage;

e) make the conclusion of contracts subject to acceptance by the other parties of supplementary obligations which, by their nature or according to commercial usage, have no connection with the subject of such contracts."

21. It is mainly on that provision that the appeal turns. But it is necessary to consider also parts (2) and (3) of the article. Article 81(2) is as follows:

"Any agreements or decisions prohibited pursuant to this Article shall be automatically void."

That provision is relevant to one of the issues on the appeal.

22. Article 81(3) lies in the background to the appeal. The Defendants seek to rely on it if necessary; the Claimants argue that they would have to do so, but do not accept that they can. The judge heard argument on it but did not decide it, both because he did not need to on his reasoning, and in order to avoid further delay in giving judgment. It is as follows:

"The provisions of paragraph 1 may, however, may be declared inapplicable in the case of:

any agreement or category of agreements between undertakings;

any decision or category of decisions by associations of undertakings;

any concerted practice or category of concerted practices,

which contributes to improving the production or distribution of goods or to promoting technical or economic progress, while allowing consumers a fair share of the resulting benefit, and which does not:

a) impose on the undertakings concerned restrictions which are not indispensable to the attainment of these objectives;

b) afford such undertakings the possibility of eliminating competition in respect of a substantial part of the products in question."

23. Article 81 is of direct effect. Section 2 of the Competition Act 1998 mirrors its provisions, though limited to effects within the UK. No point arises as to territoriality, and it is common ground that there is no difference between the article and the national legislation, so reference need only be made to the article.
24. As appears from article 81(1), what is prohibited includes both agreements and concerted practices. The Appellants' challenge is primarily to the status of what are undoubtedly agreements. However, some of the arrangements (to use a neutral word) under examination are less formal than agreements, but would be sufficient to constitute concerted practices. It is unnecessary to differentiate between the two types of arrangement for present purposes.
25. It is also unnecessary for present purposes to differentiate between preventing, restricting and distorting competition. In general I will refer to restricting competition, rather than to all three prohibited types of conduct.

The issues on the appeal

26. The principal issues arising on the appeal can be summarised as follows.

i) First, the Claimants contend that the relevant arrangements had as their object the prevention, restriction or distortion of competition within the market. They rely on two aspects: first, that the RUK courses agreed upon minimum prices for the grant of LBO media rights to AMRAC; secondly, that they agreed on closed dealing with AMRAC, that is to say they locked SIS and BAGS out of the negotiations.

ii) Alternatively, they contend that the arrangements had as their effect the prevention, restriction or distortion of competition within the market. They rely on the two points mentioned above and also on a third, namely that the racecourses undertook collective instead of individual negotiation with AMRAC. At trial it was also contended that the grant of exclusive rights to AMRAC was in breach of the article, but that is no longer pursued.

iii) An issue relevant at least to the point about collective selling as an anti-competitive effect is whether the RUK courses were competitors with each other in the relevant respects. The judge held that they were, but this is challenged by a Respondent's Notice.

iv) The Defendants succeeded in defeating both the object and the effect challenges. One of their principal arguments can be referred to as objective necessity, namely that the arrangements were objectively necessary for the pro-competitive object of ensuring entry by a new undertaking into a market previously monopolised by one operator.

v) Lastly, if the appeal succeeds on other points, and subject to any determination of the issues under article 81(3), so that the arrangements between the RUK courses did infringe article 81, the Appellants contend that this of itself renders the licences granted by each undertaking to AMRAC void under article 81(2). That, too, is in dispute.

27. For the purposes of the proceedings, the economist expert witnesses were agreed, and the parties and the judge accepted this, that there are two relevant markets. The upstream market includes the acquisition of licences for media rights to British horseracing. The downstream market includes the supply of televised broadcasting of live British horseracing to British LBOs. British racecourses are sellers in the upstream market, and before 2007 they had only one possible buyer, BAGS/SIS, a monopsony. Now they have two: BAGS/SIS and AMRAC. LBOs are buyers in the downstream market. Before 2007 they had only one seller, SIS. Now they have two, SIS and AMRAC, but they need to buy both services. The broadcasters, SIS and AMRAC, will therefore compete in the upstream market, for the acquisition of the media rights from racecourses, in order to be able to offer a viable and attractive product to LBOs. At the heart of the dispute is, on the one hand, the Claimants' complaint that, at the relevant time, SIS/BAGS was not able to compete with AMRAC for the acquisition of the rights from the RUK courses, and, on the other hand, the Defendants' argument that it was essential to protect AMRAC at that stage in order that it should ever become viable and able to compete with SIS/BAGS.

Outline history leading to the AMRAC licences

28. The nature of the Claimants' attack on the Turf TV arrangements makes it necessary to examine the arrangements and agreements entered into, and how they were arrived at, under which Turf TV came to be launched. I will start with an outline of the previous history. When SIS FACTS was new, all of the then 59 racecourses, acting through the Racecourse Association (RCA), granted non-exclusive licences of the courses' LBO media rights to SISH for a term of 10 years from 1987. That agreement was renegotiated in 1993. The 1993 agreement was brought to an end by the RCA on 30 April 2002.
29. The RCA had been reviewing the exploitation of LBO media rights between 2000 and 2002. Ten of the racecourses decided to sell their media rights (including LBO media rights) to a new company called GG Media Ltd which, in turn, granted an exclusive licence to SISH for 5 years. The other 49 licensed their rights directly to BAGS, but BAGS then granted sublicences to SIS. These licences lasted until 2004 but were then renewed. For six courses the renewed arrangements would expire on 31 March 2007; for the other 49 the date would be 31 December 2007.
30. RUK was set up in 2003. In 2004 it started to provide a domestic television service covering racing on the Sky platform, broadcasting to domestic customers, not to LBOs. In 2005 it opened negotiations with Alphameric on the possibility of providing a service to LBOs, as a rival to SIS FACTS. During the spring of 2006 RUK made presentations to the RUK racecourses about the proposal to launch a rival service to SIS FACTS. In August 2006 RUK and Alphameric entered into an agreement for a period of exclusive negotiations and in September Alphameric made a stock market announcement disclosing exclusive negotiations with "a major UK based consolidator of horseracing pictures and data".
31. It was not difficult for the major bookmakers, or for SIS and BAGS, to deduce what entity was referred to. The significant bookmakers (whose respective attitudes to the new broadcast service, if it came to be launched, could make a vital difference to its economic viability) made their opposition to the new venture both public and very clear. SISH decided that SIS should acquire as much content as it could on an exclusive basis. Previously it had been sufficient for it to take non-exclusive licences, since there was no other party in a position to acquire rights that would impinge on the practical exclusivity enjoyed by SIS. Now it felt the need to protect its position by acquiring stronger rights. During the autumn of 2006 it did so: between them, SIS and BAGS obtained exclusive licences from Towcester in October, from the 8 ICAC courses, Sedgefield, and eight Northern courses in November, and by 6 December also the seven Arena courses (and, prospectively, Great Leighs). BAGS agreed to sublicense its rights to SIS. Thus by Christmas 2006 SIS had, or would soon have, the exclusive right to LBO media rights for 26 racecourses.
32. In the meantime, AMRAC was set up on 22 September 2006, jointly owned by RUK and Alphameric. RMS was established on 24 October 2006. On 31 October the 30 RUK courses granted RUK a non-exclusive licence as regards LBO media rights. RUK wrote on 27 November to the RUK racecourses to inform them of the proposed joint venture, offering to acquire their LBO media rights, and asking them to enter into a period of exclusive negotiation until the end of December 2006. At this time further presentations were being made to the RUK racecourses about the joint venture. The letters sent on 27 November were not the first that the courses had heard of this. Goodwood agreed to the request for a period of exclusive negotiation on the same day. By 14 December all thirty RUK courses had agreed to this. Originally it was to last until 31 December 2006, but it was later extended, by stages, to 15 January 2007 and then to 31 January 2007.
33. Before any formal request or arrangement was made for exclusivity of dealing, the RUK courses showed themselves unresponsive to approaches from BAGS. In October BAGS wrote to the six courses whose current non-exclusive licences were due to expire on 31 March 2007 to ask if they would discuss the renewal of the licence with BAGS. Only Ascot replied. In December BAGS wrote again to the five relevant RUK courses, of which, some time later, York replied, saying that it was in a period of exclusive negotiation for the time being.
34. The joint venture was set up by an agreement between AMRAC, Alphameric Gaming Ltd and RMS dated 1 December 2006, and on 4 December a draft business plan for AMRAC was prepared (with revisions later).
35. On 6 December, an agenda paper for a meeting of the York racecourse set out the prices that each RUK course would get if the LBO media rights were licensed to AMRAC as proposed. On 22 January 2007 a report to the board at Musselburgh recorded an expectation that the other RUK operators would grant licences to AMRAC on the terms set out in the report.
36. On 31 January 2007 AMRAC did acquire exclusive LBO media rights from five of the RUK racecourses, as from 1 April 2007, and from the other 25 as from 1 January 2008, by means of 18 separate agreements with the 18 operators of those courses. One of the conditions of the licence was that the course operator should sign the shareholders' agreement relating to RMS. As a consequence of that having been done, the shares held by RUK in RMS were transferred to the operators of the RUK courses, and the licences became unconditional.
37. The arrangements made up to then covered the LBO media rights to 56 out of the 60 courses. In March 2007 AMRAC obtained exclusive rights for Ascot, and SIS those for Newton Abbot; in May SIS obtained the rights for Ripon and in September also for Plumpton.
38. In April 2007 AMRAC started broadcasting the Turf TV service, with its coverage at that time limited to the six courses whose licences in favour of BAGS or SIS had expired on 31 March 2007. On 1 January 2008 the service was extended to all 31 courses.
39. Next I must describe the principal elements in the agreements or concerted practices which are alleged to infringe article 81.

The joint venture agreement

40. It is not said that the shareholder agreement between Alphameric Gaming, RMS and AMRAC dated 1 December 2006 is itself anti-competitive. One point which is relied on by the Claimants about this agreement is that completion of the agreement was conditional (by clause 2.3(a)) on at least 19 of the RUK courses (including the large independent courses) agreeing to make their non-exclusive LBO rights available to AMRAC, and their exclusive rights available as soon as they could do so on the expiry of the current licences. Thus, not only are the intended licences of the essence of this agreement, but the number of licences was not taken for granted, and was set at a minimum. As defined in the agreement there were 6 large independent courses, leaving 11 small independent courses and the 13 Jockey Club courses, the latter run by one undertaking. Since the 13 Jockey Club courses could be relied on to act together, the minimum of 19 would be satisfied if they and all the large independent courses granted the necessary licence. The number of 19 would also be satisfied by the Jockey Club courses and some of the small independent courses, but it was evidently regarded as essential that all the large independent courses be committed to AMRAC.
41. The Claimants do not allege that the joint venture by itself was objectionable. They argue that what put it in breach of competition law was the fact that the RUK courses together undertook closed and collective negotiation with the joint venture, which involved price-fixing. For those propositions they rely on the course of events, which I have summarised, and on the terms of the AMRAC licences.
42. I should mention that certain of the content of the relevant agreements, and some other aspects of the evidence, were and are regarded as commercially exceptionally sensitive, and were not to be disclosed to the opposite parties. To deal with this, the normal solution was used of a "confidentiality club", under which the exceptionally sensitive information (ESI) was identified and was disclosed to the opposing parties' lawyers but subject to undertakings not to disclose it further. The court has seen all the information. I do not refer to anything in this judgment which is in the ESI category, but from time to time that constraint means that I have to be less specific about certain factual matters than otherwise I would.

The AMRAC licences

43. The licences were to last from the effective date until 31 March 2013. The first provision in the licence on which the Claimants place reliance is clause 2.2:

"ARL undertakes that, without prejudice to clause 2.3 below, it will not before or during the Term contract with any British Racecourses who are Founder Members for the licensing of rights in the nature of the LBO Rights for any period during the Term except on terms which are identical to those set out in this licence and, if it seeks to offer any variation of such terms to any other Founder Member will obtain the prior written consent of the RMS Board (not to be unreasonably withheld or delayed) to such offer to any other Founder Member."

44. The Founder Members were the shareholders of RMS at the date when the licence became effective. It is not necessary to refer to clause 2.3.
45. Clause 5 dealt with payments under the licence. Some of the detail of this is in the ESI category. AMRAC agreed to pay to the relevant course a specified rate per race if the course had contributed its LBO rights on an exclusive basis, and a lower rate if on a non-exclusive basis, with a reduction in each case for a race with fewer than a stated number of runners, and with different provision for races shown on terrestrial television. In addition, as the course would be a shareholder in RMS, it was entitled to dividends in accordance with the Shareholders' Agreement relating to RMS. The courses' respective shareholdings in RMS reflect their relative value to AMRAC, so that courses with more fixtures, and therefore more races, receive more by way of dividends (as well as receiving a greater number of the fixed rate race-specific fees).
46. We were shown a number of other provisions of the licences, but none requires separate mention for present purposes.

The economic context

47. It is agreed that the arrangements under scrutiny must be considered in the economic context in which they were made. It is therefore appropriate to say something about that at this stage, in addition to what I have already mentioned.
48. The LBO media rights are of significant value to racecourses, because of the value to bookmakers of live broadcasts of races in the LBOs. The dominant presence of SIS in the downstream market of supplying such broadcasts to LBOs from 1987 for 20 years may seem remarkable, but it is consistent with the proposition, which is common ground, that there is a high cost to entry into that market. Although LBOs need as much live coverage as they can get, in order for such a broadcast service to be viable, it was necessary for the broadcaster to achieve a critical mass of coverage, and all the more so when seeking to establish itself against the incumbent service, SIS FACTS. It was also common ground that, once there was more than one broadcaster in the market, a degree of exclusivity was essential for the success of the rival broadcasters. The judge held that there was no prospect of a third broadcaster entering the market: see paragraph 462.
49. At the time of the arrangements which are now challenged, there was only one buyer in the upstream market (given that BAGS and SIS did not act in competition with each other) and only one supplier in the downstream market. Moreover, both BAGS and SIS were owned either wholly (as regards BAGS) or as to some 56% (SIS) by major interests active in the purchasing side of the downstream market, whose interest it was to keep down the prices payable in the downstream market, and therefore also the cost of acquisition in the upstream market.
50. The status quo, therefore, was one in which there was no effective competition as regards acquisition in the upstream market, or as regards supply in the downstream market.

The Appellants' case as to anti-competitive agreements

51. Mr Vajda Q.C. for the Appellants relies on the AMRAC licence primarily for its having fixed the minimum price for the benefit of the LBO media rights and having done so on terms such that all the courses participating in RMS and granting relevant licences to AMRAC would be paid at the same rate per race. He contends that the arrangements under which the licences were granted were not merely bilateral as between AMRAC and each undertaking running one or more relevant racecourses, but were also between the racecourse undertakings among themselves. Thus, they were not merely vertical arrangements between a supplier and a purchaser, but horizontal agreements between several suppliers as well, who, he contended, were competitors as between themselves. By the arrangements, he argued, the competing suppliers of LBO media rights agreed among themselves on the prices at which they would sell their rights, for the period of the licence, and as to the other relevant terms of the sale. In addition, he argued that this was the product of a prior anti-competitive arrangement by which they all agreed to negotiate exclusively with a single buyer, namely AMRAC, for a certain time, thereby excluding the rival buyer, BAGS/SIS, from a possible deal. It is clear that there was an agreed lock-out arrangement from a time early in December, which lasted until the end of December and was renewed successively to last until the end of January. It may be that in practice the lock-out started earlier than December on an informal basis, but nothing turns on that.
52. Thus, the Appellants' case derived from these dealings is that by the arrangements between the 18 undertakings which ran the RUK racecourses, among themselves, and with AMRAC, RUK and RMS, first of all, the courses agreed on a process of closed negotiation with a view to selling to a single possible buyer, namely AMRAC, and also on a process of collective negotiation with a view to such a sale, and that, in the course of the negotiations resulting from that, they agreed on fixed minimum prices for the sale of the relevant rights.
53. As a matter of fact, it seems to me that those contentions are justified. The RUK courses did exclude BAGS and SIS from the process of negotiation. As from what date this was the result of an agreement or concerted practice may not be entirely clear. The judge put it at late November or early December: paragraph 498. It was put on a formal basis by some date in the first half of December 2006 at the latest. The agreed lock-out continued until the end of January 2007, by which time the negotiations had come to fruition.
54. The judge said that it was not in dispute that there was a concerted practice as to the way in which negotiations with AMRAC were conducted: see paragraphs 424 and 487. As regards the results of the collective and closed negotiations, the terms of the AMRAC licences speak for themselves: the rates to be paid per race were to be the same under each licence, by virtue of clause 2.2, and the return by way of dividend was governed by the RMS Shareholders' Agreement which would bind the selling undertakings directly.
55. Mr Vajda criticised the judge for his comment at paragraph 87 that the AMRAC licences were only vertical agreements, between supplier and customer. He submitted that they were also horizontal, as between all the suppliers. This is not in itself a fair comment. What the judge said at paragraph 87 related to the individual licences, which are in themselves only vertical. However, when taken together, they form part of a horizontal agreement or concerted practice as between the racecourses as well. The judge held that this case was open to the Claimants on the pleadings (amplified by one amendment which he allowed at trial).

Were the RUK racecourses in competition with each other?

56. One issue relating to the effect of these dealings under competition law is whether, relevantly, the RUK racecourses were in competition with each other. In the context of considering the Appellant's challenge of restriction by effect, as regards closed selling, the judge held that they were at least potentially so, at paragraph 489, but this is challenged by a Respondent's Notice, seeking if necessary to uphold the judge's conclusion on this alternative ground. It is logical to consider this point first.
57. The Defendants accepted that, in many respects, racecourses in general, and the RUK courses in particular, do compete with each other, for example for owners, trainers, jockeys and thus horses, for sponsorship, for spectators, and for events other than racing. The issue is whether they are in competition for the sale of LBO media rights. Of course, until AMRAC came on the scene, the ability of racecourses to compete was very limited, since there was, in effect, only a single purchaser. The judge held that SIS and BAGS did not act in competition with each other: paragraph 111. Even the separate action taken by the GG Media courses had led to a grant to SIS.
58. The Defendants' case on this point is that two or more undertakings are only in competition if their products are substitutable for each other. If a buyer can regard the products available from A and from B as alternatives, either of which will supply his needs, then A and B are competitors, and they will need to attract the buyer's custom by whatever means are available and appropriate: reducing the price, improving product quality, or as the case may be. If, instead, A and B agree between themselves on the prices at which they will offer their products, or on other ways of reducing their exposure to competition, that is conduct of the kind at which article 81 is directed. If, on the other hand, A and B produce products which, in the relevant respects, are not substitutable one for the other, then an agreement between them does not affect competition. Thus, by way of illustration, Mr Vajda relied, in support of his argument on anticompetitive effect, on the undisputed fact that Turf TV and SIS FACTS are not substitutable or interchangeable in the downstream market: LBOs have to have both of them.
59. The Defendants' case as to the courses not being in competition for the sale of LBO media rights is based on two propositions. First, races held at British racecourses are deliberately scheduled by the British Horseracing Authority so as to take place at different times and not to coincide. Secondly, LBOs have an incentive to show live coverage of as many British races as possible and to screen a succession of races throughout the day in order to maximise their betting turnover. Dr Niels, the Defendants' expert witness, gave evidence of this which was not challenged. An illustration of this, he said, could be found in the experience of Plumpton racecourse, which was the last to commit itself as between SIS/BAGS and AMRAC. If LBO media rights of different racecourses were substitutable, Plumpton should have been in a weak position, since each of the two buyers would have had enough to satisfy their needs. In fact, Plumpton was able to obtain from SIS rates which show no indication of negotiating disadvantage on the part of the racecourse.
60. The judge dealt with this point at paragraphs 488 and 489, without referring to Dr Niels' evidence on this point (though he had referred to it at paragraph 484, including the point that there was no substitutability). His decision, that "racecourses are potentially in competition with each other in relation to the sale of their LBO media rights", at paragraph 489, was expressly based on a Commission decision known as UEFA. In full, its name and reference is Re: The Joint Selling of the Commercial Rights of the UEFA Champions League [2004] 4 CMLR 9. This was one of three Commission decisions about the joint selling of media rights to sporting events. As in the present case, there was an upstream market for the sale and acquisition of broadcasting rights, and a downstream market in which television broadcasters competed for advertising revenue and for subscribers. UEFA had a joint selling arrangement by which the commercial rights to broadcast the matches in the UEFA Champions League were handled on an exclusive basis through UEFA as a joint selling body, which prevented individual football clubs from marketing such rights on an individual basis. This precluded competition between the clubs, and between the clubs and UEFA, in supplying media rights to interested buyers in the upstream market. At paragraph 114 of the Commission's decision it said this:

"This means the third parties only have one single source of supply. Third-party commercial operators are therefore forced to purchase the relevant rights under the conditions jointly determined in the context of the invitation to bid, which is issued by the joint selling body. This means that the joint selling body restricts competition in the sense that it determines prices and all other trading conditions on behalf of all individual football clubs producing the UEFA Champions League content. In the absence of the joint selling agreement the football clubs would set such prices and conditions independently of one another and in competition with one another. The reduction in competition caused by the joint selling agreement therefore leads to uniform prices compared to a situation with individual selling."

61. The judge referred to the Commission's treatment of the question of competition between the clubs at paragraph 371, as follows:

"The Commission further considered the question of competition in the context of the special characteristics of sport. At paragraph 125, it noted UEFA's submission that football clubs were not truly independent competitors. At paragraph 128, the Commission held that UEFA and the football clubs were economic competitors in selling commercial rights (property rights and media rights) to football matches. If there were no joint selling arrangement then these parties would be selling their rights individually and in competition with one another. At paragraph 129, the Commission again stated that the clubs were economic competitors in relation to the sale of the relevant rights."

62. The judge's reference to paragraph 128 is a direct quotation of that paragraph, allowing for putting it in indirect rather than direct speech. The Commission decision does not disclose any consideration of whether the rights were substitutable. The position was also complicated by the question of ownership of the relevant rights, as between the home club, the visiting club and UEFA itself, though it was accepted that an agreement between the two clubs and UEFA relating to the particular match would not be caught by article 81, since such an agreement would be necessary to produce "one unit of output" namely the licence to broadcast the one match. That particular problem does not occur in the present case where the rights are clearly owned by the relevant racecourse.
63. On the other hand, one feature of the argument put forward by the Defendants in this case, as to the timing of races, was not available in relation to the UEFA case, since in the Champions League, until the semi-final stage, matches are played simultaneously (8 at a time in the group stage, 4 at a time in the second knock-out phase, and two at a time in the quarter finals).
64. The judge referred at paragraph 488 to the Commission's conclusion, which he had summarised at paragraph 371, quoted above, that the football clubs did compete with each other in selling their media rights. He went on to say this:

"In my judgment, that reasoning applies in this case. Although racecourses do not compete with each other in a number of respects and for a number of reasons, it is open to them to compete with each other when it comes to selling their LBO media rights."

65. The Defendants had also relied on a Commission Notice about Formula One motor racing. The judge discounted this at paragraph 489 saying:

"The specific point being addressed in those comments is not wholly clear and the reasoning is very brief. In my judgment, the reasoning in the UEFA case is clear and intelligible and applicable to this case. Accordingly, I hold that the racecourses are potentially in competition with each other in relation to the sale of their LBO media rights."

66. The possible relevance of the Formula One notice is that Formula One events do not compete with each other as they are not held, or therefore broadcast, at the same time. It seems to me that, given the limited status of a Commission Notice, and the very different nature of Formula One racing from that of British horse racing, it is not surprising that the judge should have gained little assistance from that document.
67. However, the point remains that the expert evidence of Dr Niels, which was not, as I understand it, challenged in cross-examination on this point, was to the effect that the respective courses' LBO media rights are not substitutes for each other, and therefore the courses are not in competition with each other for the sale of these rights, and the judge, though having referred to this evidence, did not say that he did not accept it, nor did he explain why he did not agree with the conclusion drawn.
68. Mr Vajda submitted that there was other relevant evidence besides that of Dr Niels which supported the judge's conclusion. Neither of the other experts dealt with this point. Moreover, quite apart from the fact that the judge did not refer to any other relevant evidence in support of his conclusion, on examination it seems to me that the evidence to which Mr Vajda referred us was either neutral or indicated the opposite. It cannot therefore be taken as even a silent justification for the judge's conclusion.
69. What Mr Vajda is entitled to point out is that different courses, negotiating separately, ended up with different terms. One example he gave was Ascot, which negotiated with AMRAC as well as with BAGS, and obtained a deal which was different in certain respects from that which the RUK courses had. That feature, however, as it seems to me, is a consequence of the difference between collective and individual negotiation. It does not by itself show that the respective suppliers' products are substitutable so that they are in competition with each other.
70. Mr Vajda submitted in reply that the very fact that the upstream market had been agreed to consist of the market for the sale and acquisition of racecourses' LBO media rights showed that all sellers within that market were in competition with each other. That was a point taken then for the first time, and not foreshadowed in his skeleton argument, even that part of it which dealt specifically with the issues raised by the Respondent's Notice. It had not been put to Dr Niels in cross-examination, nor covered in the Claimants' own expert evidence or that adduced on behalf of SIS, as Third Party. In those circumstances we did not allow Mr Vajda to take that point, and I need say no more about it.
71. It seems to me that the considerations addressed by the Commission in UEFA do not provide an answer to Dr Niels' observations in his evidence on this point, nor does the evidence identified by Mr Vajda in his submissions. I therefore do not understand why the judge rejected (without explanation) Dr Niels' evidence and Mr Roth Q.C.'s submissions for the Respondents on this aspect of the case. I would hold that Mr Roth is correct and that the racecourses do not compete with each other as regards the sale of LBO media rights.
72. Mr Vajda did, however, correctly point out that this argument, which in itself might appear to be fundamental to the whole case, was raised by the Respondent's Notice only in support of the judge's finding that article 81 was not infringed by the collective negotiation arrangements having the effect of restricting competition. It follows that it would not be right to treat it as an answer to the whole of the Appellants' case.

Arrangements with the object of restricting competition

73. Article 81 applies if the agreements in question have the object of restricting competition, or if they have that effect. If they have that object, it is unnecessary to consider what effect they have. It is therefore also unnecessary to embark on an examination of what would have been the position but for the agreement, in order to identify what is known as the "counterfactual". It is therefore appropriate to start with the Appellants' case based on the object of the agreements, which Mr Vajda identified as being the fixing of uniform minimum prices, as between the RUK racecourses, for the sale of their LBO media rights to AMRAC, together with the factor of agreeing to negotiate only with AMRAC.
74. Mr Vajda emphasised that, under both limbs of article 81(1), but particularly so in relation to the object limb, it is inappropriate and illegitimate to consider any possible pro-competitive effects of the agreements. That is properly the preserve of article 81(3), according to his argument.
75. He also submitted that the determination of the object of the agreement, for this purpose, has nothing to do with the subjective purpose or motive of the parties. He placed particular reliance on a decision of the ECJ delivered since the judgment in the present case: Competition Authority v Beef Industry Development Society Ltd Case C-209/07, [2009] 4 CMLR 310, known as BIDS.
76. In her Opinion in that case, Advocate-General Trstenjak said at paragraph 44:

"From its wording it is clear that the notion of restriction of competition by object refers primarily to the object of the agreement. The Community judicature has found an anti-competitive aim or tendency of an agreement to exist in particular where the necessary consequence of the agreement was the restriction of competition. In such a case in principle the parties may not argue that they did not intend any restriction of competition or that their agreement also pursued a different aim."

77. In turn, the court said at paragraph 21:

"In fact, to determine whether an agreement comes within the prohibition laid down in Article 81(1) EC, close regard must be paid to the wording of its provisions and to the objectives which it is intended to attain. In that regard, even supposing it to be established that the parties to an agreement acted without any subjective intention of restricting competition, but with the object of remedying the effects of a crisis in their sector, such considerations are irrelevant for the purposes of applying that provision. Indeed, an agreement may be regarded as having a restrictive object even if it does not have the restriction of competition as its sole aim but also pursues other legitimate objectives (General Motors v Commission, paragraph 64 and the case-law cited). It is only in connection with Article 81(3) EC that matters such as those relied upon by BIDS may, if appropriate, be taken into consideration for the purposes of obtaining an exemption from the prohibition laid down in Article 81(1) EC."

78. That was a case referred by the Supreme Court of Ireland, concerned with an agreement within the beef processing industry designed to relieve the effects of over-capacity within that industry. The arrangements were designed to encourage some within the industry to leave (the goers), to deal with their plant in such a way that it would not be used to increase capacity within Ireland, and not to re-enter the industry for 2 years; in return those who stayed (the stayers) were to make payments to the goers and were also subject to disincentives against increasing their own activities. The Competition Authority of Ireland took the view that these arrangements breached Irish competition law and brought proceedings accordingly. The High Court dismissed the claim on the basis that the agreement did not fall within article 81(1), though it also held that, if BIDS had had to rely on article 81(3), it would not have succeeded. The Supreme Court referred the case to the ECJ for a preliminary ruling.
79. The court's ruling was that the agreement did have an object which was prohibited by article 81(1). The essence of the decision can be seen from paragraphs 33 to 36:

"33. The BIDS arrangements are intended therefore, essentially, to enable several undertakings to implement a common policy which has as its object the encouragement of some of them to withdraw from the market and the reduction, as a consequence, of the overcapacity which affects their profitability by preventing them from achieving economies of scale.

34. That type of arrangement conflicts patently with the concept inherent in the EC Treaty provisions relating to competition, according to which each economic operator must determine independently the policy which it intends to adopt on the common market. Article 81(1) EC is intended to prohibit any form of coordination which deliberately substitutes practical cooperation between undertakings for the risks of competition.

35. In the context of competition, the undertakings which signed the BIDS arrangements would have, without such arrangements, no means of improving their profitability other than by intensifying their commercial rivalry or resorting to concentrations. With the BIDS arrangements it would be possible for them to avoid such a process and to share a large part of the costs involved in increasing the degree of market concentration as a result, in particular, of the levy of €2 per head processed by each of the stayers.

36. In addition, the means put in place to attain the objective of the BIDS arrangements include restrictions whose object is anti-competitive."

80. The reference in paragraph 36 is to the disincentive on stayers to increase their own production, by being subject to a substantially higher levy per head of cattle than the €2 otherwise payable, and the restrictions imposed on goers as regards dealing with their plant.
81. Thus, the arrangements at issue in BIDS were manifestly designed to reduce competition, by assisting the withdrawal of undertakings from the market. By contrast, the aim of the AMRAC arrangements was to increase competition by enabling the entry of an additional undertaking into the market, previously the preserve of a single undertaking which was not exposed to any effective competition.
82. Mr Vajda submitted that, regardless of this difference, BIDS showed that the object of the present arrangements was to be seen as restrictive of competition in two respects: it removed price competition between the RUK racecourses in respect of the sale of their LBO media rights, and it shielded AMRAC from competition by way of the closed selling. He pointed out that one important aspect of the object of the arrangements was to improve the profitability of the RUK racecourses, by those reductions in competition, and that by restricting competition between market participants and the proper function of the market of supplying consumers (here, LBOs) optimally with a product at the lowest possible price, the parties to these arrangements were offending against the most fundamental principles of European competition law.
83. I respect the ingenuity with which Mr Vajda put these submissions to the court, but I cannot accept them as correct, or as reflecting in any realistic way either the economic context in which the arrangements were made, or the object of the arrangements, ascertained objectively by reference to their nature and terms in the relevant context.
84. I note that, in a passage (paragraphs 50 to 58) dealing with assessing in the light of the relevant legal and economic context whether an agreement has as its object the restriction of competition, the Advocate General in her Opinion in BIDS, at paragraph 53, said this:

"The second category concerns cases in which an agreement is ambivalent in terms of its effects on competition. If the object of an agreement is to promote competition, for example by strengthening competition on a market, opening up a market or allowing a new competitor access to a market, the necessary restriction of the requirement of independence can, when matters are viewed as a whole, give way to the aim of promoting competition."

85. It seems to me that, even ignoring the issue as to whether the racecourses were truly competitors with each other as regards the sale of their LBO media rights, the object of the arrangements, by which AMRAC was set up and then came to have the benefit of the licences from the RUK racecourses, was to establish a second broadcaster which would be able to act as a rival to BAGS/SIS both in the upstream and in the downstream markets. To achieve that, it is agreed, the broadcaster would have to acquire LBO media rights for a minimum number of racecourses on an exclusive basis. It is not something that any racecourse could have achieved by itself. Given that the incumbent operator was dominated by the interests of the purchasers in the downstream market, given the high cost of entry, and given the very long period in which no other operator had shown any interest in entry to these markets, it seems to me that it was obviously necessary that the new entrant would have to be promoted by or in association with a number of racecourses, and that it would need to be protected, at the stage of its establishment, from competition from the incumbent, since otherwise it would never get off the ground.
86. The proposition that the arrangements were designed to improve the profitability of the racecourses is correct, but I cannot agree with the submission that this was to be done by restricting competition. On the contrary, it was to be done by introducing competition into the previously monopsonistic upstream market. Nor is increasing profitability objectionable in itself. It is, after all, the motive of most commercial activity.
87. On this aspect of the case I agree with the judge who gave his reasons for rejecting the case based on an object of restricting competition at paragraph 439. He said, in my view rightly: "The objective aim of the cooperation was to sponsor the entry of AMRAC into the market." That is supported by what the Advocate General said at paragraph 53 of BIDS quoted above.
88. The Appellants argue that the judge took into account, wrongly, the subjective intention and desire of the racecourses. I disagree. What the judge took into account in that respect was something that one could infer without knowing anything about the subjective views of the racecourses, namely that they had an interest in enabling competition to take place in the upstream market, where it was until then conspicuous by its absence.
89. Mr Roth drew our attention to a number of other cases in which regulatory authorities had treated the collective sale of rights relating to sporting events as infringing article 81 on the basis of the effect of the arrangements rather than their object. In particular he relied on proceedings in the Competition Appeal Tribunal, The Racecourse Association v OFT [2005] CAT 29, (RCA) being appeals against determinations by the OFT of infringements of article 81 by the collective sale of media rights on the part of 49 racecourses, via the Association. He showed us that the OFT had not treated the agreements in that case as having the object, but only the effect, of restricting competition. Mr Vajda fairly commented that an argument of that kind is of limited persuasive value, but nevertheless it seems to me to be interesting that the regulator should have approached the case on the basis of the effect of the agreement, and not by reference to its object, when it seems that much the same argument could have been made in that case as is made here on the basis of the object of the agreement.
90. Mr Roth also submitted that, where undertakings active in a market agree to set up a joint venture to do something which they cannot do by themselves, then as regards their respective dealings with the joint venture they are not competitors, and their agreements, whether horizontal or both horizontal and vertical, do not have the object or the effect of restricting competition. He accepted, of course, that if the joint venture's purpose is merely to do collectively that which the members can do by themselves, then for undertakings which are otherwise competitors to set such a body up would be likely to be anti-competitive. Two or more competing retailers cannot set up a single joint venture to sell products to the public on their joint behalf which otherwise they would sell separately and in competition with each other. But they can set up a separate undertaking to do something which individually they do not and cannot do, and can enter into agreements with the joint venture without offending article 81. He showed us Guidelines issued by the Commission in 2001 on the applicability of article 81 to horizontal cooperation agreements, where, at paragraph 24, this is said:

"Some categories of agreements do not fall under article 81(1) because of their very nature. This is normally true for cooperation that does not imply a coordination of the parties' competitive behaviours in the market, such as:

- cooperation between non-competitors

- cooperation between competing companies that cannot independently carry out the project or activity covered by the cooperation."

91. I see a good deal of force in that proposition, but I prefer not to decide this case on that basis.
92. At a more basic level, Mr Roth argued that there cannot be an agreement whose object (or for that matter whose effect) is to restrict competition if at the relevant time there is no competition to be restricted. That could be on the basis that the parties are not competitors; as already mentioned that is one aspect of his argument. But it could also be because, even if otherwise they would compete with each other, they do not and cannot do so because such competition (here, as sellers) depends on there being more than one buyer. Since there was only the one buyer, there was no competition to be restricted. That seems to me to be correct in principle, and correctly applied to the facts of this case. Mr Vajda submitted that article 81 applies not only to actual but also to potential competition. However, that comment attracts the same response: unless a second undertaking became active in the upstream market, there could be neither actual nor potential competition. Arrangements whose object was to enable such an undertaking to enter the market could not therefore be restrictive of competition that did not and could not exist at the time.
93. Mr Vajda made the correct point that the ways in which competition can be restricted, in breach of article 81, are by no means limited to those mentioned in article 81(1) itself. It does not seem to me that the judge said otherwise in his paragraph 324, and in any event what he said there did not, in my judgment, affect his reasoning at paragraphs 439 to 441.
94. I do not find it necessary or appropriate to address every aspect of the detailed submissions on anti-competitive object made on either side. Mr Vajda said in his skeleton argument that the judge's conclusion on object was plainly wrong. I disagree. It seems to me that it was plainly right.

Arrangements with the effect of restricting competition

95. I therefore turn to anti-competitive effect, as regards which Mr Vajda relied on the collective aspect of the negotiations between the racecourses and AMRAC, and on the closed aspect as well. He submitted that the effects of the restriction on competition were to be seen in the uniform prices obtained by the RUK racecourses for the sale of their LBO media rights, and in the higher prices obtained in the downstream market.
96. In this context it is relevant to refer (as the judge did) to O2 Germany GmbH v Commission Case T-328/03, [2006] 5 CMLR 258. In that case the Commission had held that an agreement between O2 and T-Mobile infringed article 81(1) but was saved under article 81(3), but O2 appealed and the Court of First Instance held that the decision should have been that the agreement did not infringe article 81(1). I quote paragraphs 68, 71 and 72 of the judgment of the CFI:

"68. Moreover, in a case such as this, where it is accepted that the agreement does not have as its object a restriction of competition, the effects of the agreement should be considered and for it to be caught by the prohibition it is necessary to find that those factors are present which show that competition has in fact been prevented or restricted or distorted to an appreciable extent. The competition in question must be understood within the actual context in which it would occur in the absence of the agreement in dispute; the interference with competition may in particular be doubted if the agreement seems really necessary for the penetration of a new area by an undertaking (Société Technique et Minière Case 56/65, [1966] ECR 235, at 249-250).

…

71. The examination required in the light of Article 81(1) EC consists essentially in taking account of the impact of the agreement on existing and potential competition (see, to that effect, Case C-234/89 Delimitis [1991] ECR I-935, paragraph 21) and the competition situation in the absence of the agreement (Société Technique et Minière at 249-250), those two factors being intrinsically linked.

72. The examination of competition in the absence of an agreement appears to be particularly necessary as regards markets undergoing liberalisation or emerging markets, as in the case of the 3G mobile communications market here at issue, where effective competition may be problematic owing, for example, to the presence of a dominant operator, the concentrated nature of the market structure or the existence of significant barriers to entry - factors referred to, in the present case, in the Decision."

97. The markets presently under consideration are not in general analogous to the emerging 3G mobile telecommunications market, but they do share the features of a dominant operator, and high costs of entry as a significant barrier to a new operator. The references in paragraph 68 to the need for arrangements to enable an undertaking to penetrate a new area, and to effective competition being problematic, in paragraph 72, seem to me to have relevance to the situation on the markets with which we are concerned, as they were in late 2006. Equally, the references in paragraphs 68 and 71 to considering the agreement in the light of the competition situation as it would be in the absence of the agreement in dispute are highly pertinent to the present case.
98. Mr Vajda relied heavily on the Commission's decision in UEFA in this part of his argument, as regards collective negotiation by the racecourses. I have already mentioned this at paragraph [60] above, and explained some of the content of the decision. The ruling was that the collective arrangements for selling rights to broadcast the matches in the later stages of the UEFA Champions League had the effect of restricting competition between clubs in the upstream market, by leading to uniform prices as compared with a situation of individual selling, and in the downstream market for broadcasters. Therefore article 81(1) was held to have been infringed, but the arrangements entered into were (with some exceptions) approved under article 81(3), having regard among other things to the fact that the arrangements made it possible to create a branded league product sold in packages via a single point of sale.
99. A similar concern lay behind the Commission's decision called Re FA Premier League, Case COMP/C-2/38.173 [2006] 5 CMLR 25. A single entity, owned by the clubs in the Premier League, had the exclusive right to negotiate media rights to matches played by the clubs, and offered them every three years by way of invitations to tender. The Commission expressed a preliminary view that these arrangements infringed article 81(1), in particular because the joint sales organisation could restrict output and create foreclosure problems on downstream markets, for example by the sale of large packages of rights. In response the League made successive proposals for changes to the arrangements, designed to reduce output restrictions and to enhance the scope for ex ante competition for rights, including a prohibition on any one purchaser (alone or in association with others) buying all the packages. Ultimately the Commission was satisfied that the varied arrangements were sufficient to meet its concerns expressed under article 81. Just as in the UEFA case, an important part of the Commission's concern was to ensure that existing operators in the market were not foreclosed from the possibility of acquiring relevant rights and using them as part of its broadcasting service.
100. In the present case, as matters stood in 2006, that could not have been a concern. There was only one undertaking active in the market. It had a monopoly of the downstream market. It was protected by the relatively high cost of entry, an effective barrier. It does not seem to me that the Commission's decisions about UEFA, the FA Premier League, or the similar decision about the German Bundesliga, carry the significance by analogy that Mr Vajda seeks to put on them. The relevant contrast is between the need for protection, in those cases, of the competitive position of existing broadcasters, on the one hand, and on the other hand the notion that the single existing broadcaster in the present case requires the same protection against the possibility that a rival might enter its market.
101. Mr Roth submitted that the decision of the Competition Appeal Tribunal in RCA (referred to at paragraph [89] above) is of more relevance. In that case, 49 racecourses (all except the 10 in the GG Media group) agreed to sell their media rights (other than LBO media rights) with a view to exploitation by way of a new interactive product to a single buyer, Attheraces plc (ATR). The OFT found that the arrangements which had been notified to it were in breach of the UK equivalent of article 81(1) and were not saved by the equivalent of article 81(3). An important part of the OFT's objection to the arrangements was the element of collective selling. On appeal from the OFT's ruling the CAT held that there had indeed been collective selling, but that it did not follow from that alone that article 81(1) had been infringed. The Tribunal referred to propositions to which I have already made reference, including a passage to similar effect to that from paragraph 68 of the decision in O2 already quoted. It also referred to the fact that there is unlikely to be an appreciable effect on competition where, for example, two competitors, each with a large market share, form a joint venture to develop a new product which neither has the resources to develop on its own (paragraph 152). Later it referred to the Commission's Horizontal Guidelines, paragraph 87, on a point which is relevant at a later stage of the present case:

"cooperation between firms which compete on markets closely related to the market directly concerned by the cooperation cannot be defined as restricting competition, if the cooperation is the only commercially justifiable possible way to enter a new market, to launch a new product or service or to carry out a specific project."

102. The Tribunal held that the arrangements were not in breach of article 81 because they were set up in order to exploit a new product which had never been marketed previously, and they were necessary for that purpose. In particular, the collective selling was found to be the only realistic way of achieving the necessary critical mass of relevant rights. Of course, in the present case the rights in question had been sold before, but up to that time there had been only one purchaser for them. The obstacles facing a new entrant to the existing market were substantial, and comparable to those facing ATR in seeking to establish for the first time a new downstream market. In each case there was a need for a critical mass of rights granted by racecourses. In the present case there was also time pressure on the racecourses and therefore also on the aspiring entrant to the market to acquire the necessary number of rights. It seems to me that Mr Roth is justified in seeking assistance from the RCA decision. The analogy is strong.
103. Coming to more specific points, Mr Roth pointed out that the sums payable to the racecourses as a result of granting licences to AMRAC were not in fact uniform, because, in addition to the rates per race payable under the licences, the racecourses also had their right to a dividend, as described briefly at paragraph [45] above, which resulted in the various racecourses receiving different overall amounts from their participation in the AMRAC operation. By contrast, under the previous arrangements with BAGS, there were only the rates per race, and those were uniform. He therefore submitted that the arrangements with AMRAC had reduced rather than increased the degree of uniformity of prices, as between the RUK racecourses. As for the point about higher prices in the downstream market, while it is clearly the case that LBOs are paying more now for the two services, Turf TV and SIS FACTS, than they did for the one service beforehand, he submitted that this comparison by itself is meaningless. What is needed is a comparison between the position as it is now, and the position as it would have been but for what is said to be the objectionable aspect of the agreement.
104. The judge held at paragraphs 493 and 495 that the Claimants had not shown that the collective nature of the negotiations had led to any actual or likely increase in the price payable by AMRAC to the RUK racecourses. He did not express a view as to the effect on prices in the downstream market, that is to say those payable by LBOs. However, in that case too the comparison has to be between the actual position and that which would have prevailed but for the collective negotiation. Mr Vajda might have had a valid point on this, but he did not make it, for good reason. This would be that, but for the collective negotiation, AMRAC would not have been in a position to bid for the LBO media rights at all. If that is right, however, the Respondents' argument based on objective necessity would gain force which Mr Vajda was not prepared to allow to it.
105. Quite apart from that point, the Claimant's expert Mr Biro said nothing about the effect of the collective negotiation on prices in the downstream market, and, we were told, no questions were put to any of the witnesses about this. Accordingly, it seems to me that this argument fails for lack of evidence on the relevant point, that is to say not the position before and after the AMRAC licences, but the position with and without collective negotiation.
106. It is in this context, moreover, that the point taken in the Respondent's Notice is specifically said to be relevant. On the basis of the conclusion I have expressed on that, at paragraph [71] above, the RUK racecourses were not competitors in this regard. On this point (even if, for pleading and forensic reasons, not on others) that is a complete answer to the Appellants' challenge to the arrangements.
107. As for the other aspect of the Appellants' case on this, the closed nature of the negotiation, Mr Vajda criticised the judge's conclusion which is expressed at paragraph 501:

"The primary consideration in this case was that the 18 operators of the 30 courses wanted to create a joint venture and to sponsor the entry of that joint venture into the market to provide competition for the purchase of their rights. The suggested alternative to that would be to sell those rights to the incumbent. That would undermine the joint venture and imperil its entry. It would not be logical to promote the joint venture and then to withhold from it the rights which it needed and wished to have to enter the market. There is not a true comparison between granting LBO media rights to BAGS and granting LBO media rights to AMRAC. A deal with BAGS would not result in the existence of competition in the market, would not result in the successful entry of the joint venture and would not result in participation in a successful joint venture. If the two deals are not alike, then in my judgment, it is not anti-competitive for an operator to decide which type of deal he prefers to pursue and then not to deal on an alternative and incompatible basis."

108. It is a fair, but inadequate, comment that the words "and wished" in the fourth sentence are irrelevant; subjective wish is not sufficient, but objective need is. Mr Vajda also picked on the words "an operator" in the last sentence, and submitted that this ignored the pernicious element of combination by different operators. That seems to me to be an over-literal criticism.
109. He further submitted that the judge had misstated the comparison. It was not between granting the rights to AMRAC, on the one hand, and granting them to BAGS, on the other, but between granting them to AMRAC, having excluded BAGS from the process, on the one hand, and offering them openly and equally to both BAGS and AMRAC to bid for them, on the other. There is a degree of apparent justification for that comment, but again it seems to me to be taking the judge's words too literally, and not in context. The judge's point is that the rights have to be sold to someone, and it will be either the incumbent or the new entrant to the market. The object of the arrangements, as I have said, was to enable a new undertaking to enter the market. That could not happen without the joint venture acquiring the rights from a sufficient number of racecourses. The judge held that, by December 2006, the sufficient number was as many as AMRAC could get, namely at least 30. It followed that all the RUK racecourses had to grant their LBO media rights to AMRAC, or none of them would achieve the object of the arrangements into which they had entered. In that light, the judge's comments about undermining the joint venture are entirely appropriate. In practical terms, closed selling was essential.
110. For that reason I would also reject this aspect of the Appellants' grounds of appeal.

Objective necessity

111. The arguments on this part of the case have been foreshadowed by some of what I have said already. The judge did not find it necessary to decide this point as such, but said that he would have rejected the challenge on collective selling on this ground as well: see his paragraph 496:

"It is not strictly necessary therefore to consider whether any restriction on competition, resulting from collective negotiation, was objectively necessary. If I had to decide that question I would unhesitatingly hold that it was objectively necessary. It must be remembered that AMRAC was a joint venture in which the operators of the racecourses were participating. The participants necessarily had to talk to each other and negotiate collectively in order to create the joint venture. It was up to them to decide between themselves as to the terms on which they should participate in the joint venture. They were entitled to decide that they would share the total price for the LBO rights on equivalent terms and that their different interests would be reflected in the dividends payable out of profits."

112. Mr Vajda argued that the scope of the principle of objective necessity is narrow, and is best explained in the judgment of the CFI in Métropole télévision (M6) v Commission Case T-112/99 [2001] ECR II-2459, in a passage headed "The concept of ancillary restriction" from paragraph 104 to paragraph 117. An ancillary restriction, in this context, means a restriction which is directly related and necessary to the implementation of a main operation (paragraph 104). To be directly related, it must be subordinate to the implementation of that operation and must have an evident link with it (paragraph 105). In order to decide whether it is necessary, two tests must be satisfied: first, whether the restriction is objectively necessary for the implementation of the main operation and, secondly, whether it is proportionate to it (paragraph 106). At paragraphs 107 to 109 the CFI said this:

"107. As regards the objective necessity of a restriction, it must be observed that inasmuch as, as has been shown in paragraph 72 et seq. above, the existence of a rule of reason in Community competition law cannot be upheld, it would be wrong, when classifying ancillary restrictions, to interpret the requirement for objective necessity as implying a need to weigh the pro and anti-competitive effects of an agreement. Such an analysis can take place only in the specific framework of Article 85(3) of the Treaty.

108. That approach is justified not merely so as to preserve the effectiveness of Article 85(3) of the Treaty, but also on grounds of consistency. As Article 85(1) of the Treaty does not require an analysis of the positive and negative effects on competition of a principal restriction, the same finding is necessary with regard to the analysis of accompanying restrictions.

109. Consequently, as the Commission has correctly asserted, examination of the objective necessity of a restriction in relation to the main operation cannot but be relatively abstract. It is not a question of analysing whether, in the light of the competitive situation on the relevant market, the restriction is indispensable to the commercial success of the main operation but of determining whether, in the specific context of the main operation, the restriction is necessary to implement that operation. If, without the restriction, the main operation is difficult or even impossible to implement, the restriction may be regarded as objectively necessary for its implementation."

113. At paragraph 111 the court gave (with approval) examples of determinations by the Commission under which a restriction was held to be within this principle where failing such a restriction, the operation in question "could not be implemented or could only be implemented under more uncertain conditions, at substantially higher cost, over an appreciably longer period or with considerably less probability of success".
114. If it is shown that a restriction is objectively necessary to implement a main operation, it is also necessary to decide whether its duration and its material and geographic scope do or do not exceed what is necessary to implement the operation. If the duration or the scope of the restriction does exceed what is necessary for that purpose, it cannot be justified as an ancillary restriction, and whether it is valid notwithstanding article 81 will depend on a separate assessment under article 81(3).
115. If a restriction is found to be directly related and necessary to achieving a main operation, the compatibility of the restriction with the competition rules will depend on that of the main operation.
116. In Métropole the object of the arrangements overall was to establish, by means of a joint venture, a new operator in the pay TV market which in France was very much dominated by one long-standing operator. The CFI held that the joint venture itself did not infringe article 81(1) because the parties to it were not at risk of co-ordination; in effect, it was not something in respect of which they were to be seen as competitors. It then had to consider three restrictions contained in the relevant agreements. One of these was a non-competition obligation under which the parties were obliged not to become involved in companies competing with the new venture. The Commission had held this to be justified as an ancillary restriction, but only for a period of three years, as being necessary during "the crucial launch phase", on these principles. It was "deemed pro-competitive in that it contributes to the creation of a new entrant on" the relevant market during that period. The CFI upheld the limitation to three years of the negative clearance given by the Commission to this restriction. The CFI also upheld the Commission's decision that two other restrictions were not justified, not being limited to that which was necessary and proportionate.
117. Thus, the object of the arrangements as a whole was not within article 81(1), for reasons similar to those mentioned in paragraph 87 of the Commission's horizontal guidelines, quoted at paragraph [101] above, and one of the detailed restrictions was held valid (for a limited initial period, covering the launch of the venture) as being necessary and proportionate to achieving the overall object. Mr Vajda submitted that the horizontal guidelines had been overtaken by the Métropole decision, and since then superseded by the Commission's Guidelines on the application of article 81(3), issued in April 2004. Those Guidelines do deal with this topic, and they make reference to Métropole, especially at paragraphs 28 to 31 on which Mr Vajda relied. However, it does not seem to me that what is there stated differs substantially from the previous Guidelines on this point, and the point made in the earlier Guidelines is supported by paragraph 53 of the Advocate General's opinion in BIDS (see paragraph [84] above) and also by later authority of the ECJ and the CAT to which I now turn.
118. The Métropole case was considered by the ECJ in a later case, Wouters, Case C-309/99, [2002] ECR I-1577. Both of them, and other relevant cases, were considered in turn by the CAT in RCA. The Tribunal said this at paragraph 167:

"We confess to some difficulty in reconciling the approach of the ECJ in Gøttrup-Klim and Wouters with that of the CFI in Métropole, but find it unnecessary to dwell on the explanation in Métropole as to the rationale that the CFI perceived as underlying cases such as Gøttrup-Klim and Wouters (the latter of course being decided after Métropole). We consider that these two decisions of the ECJ show that the assessment of whether or not a particular arrangement constitutes an infringement of Article 85(1) (now Article 81(1)), or therefore of the Chapter I prohibition, is a rather more flexible exercise than the CFI was perhaps willing to appreciate. It is not enough that the arrangement is apparently anti-competitive, as in Gøttrup-Klim and Wouters. What those cases show is that ostensibly restrictive arrangements which are necessary to achieve a proper commercial objective will not, or may not, constitute an anti-competitive infringement at all. Whether or not they will do so requires an objective analysis of the particular arrangement entered into by the parties, assessed by reference to their subjective "wants" and against the evidence of the particular market in which they made their arrangement. The task then is to consider whether the restrictive arrangement of which complaint is made is "necessary" to achieve the objective. The RCA appellants also submitted that the concept of "necessity" in this context is not an absolute one, but has an element of flexibility about it, for which they referred us to paragraph 109 in the Métropole case in which the course observed that "If, without the restriction, the main operation is difficult or even impossible to implement, the restriction may be regarded as objectively necessary for its implementation." We also accept this last submission: competition law is not an area of law in which there is much scope for absolute concepts or sharp edges."

119. If the creation of a new channel for pay TV in France was a legitimate objective for these purposes, so as to justify a non-competition obligation for the first three years, as in Métropole, and if the exploitation of a new market for interactive betting was in RCA, then it seems to me that to promote a new entrant into the markets hitherto occupied solely by BAGS/SIS in the present case also qualifies for the purpose, for reasons which I have discussed under the heading of restrictions by object. So the question is whether the aspects of the arrangements which are challenged by the Appellants, namely (now) closed and collective selling, and the fixing of prices, are properly regarded as necessary to achieve the object, and proportionate to that need. It is not now said that the grant of exclusive rights causes the arrangements to infringe article 81.
120. It was common ground, as already mentioned, that AMRAC required a critical mass of LBO media rights if it was to be viable. Even the Appellants' expert, Mr Biro, accepted that, insofar as exclusivity was necessary, it might have been necessary for the RUK racecourses to sell collectively. The collective negotiations, and preparation on the part of the racecourses for such negotiations, must have involved setting the consideration for the grant of the licence. It seems to me to be fanciful to suggest, in the present situation, that there could have been collective negotiation but without common terms as to the financial consideration for the licences resulting from that negotiation.
121. As for the aspect of closed selling, the judge explained his reasoning at paragraph 501 which I have quoted above, at paragraph [107]. In the passage after that quotation, I have explained why it seems to me that the judge was right to decide that closed selling was objectively necessary in order to achieve the object of the arrangements: see in particular paragraph [109]. The same reasoning leads me to the conclusion that this aspect of the arrangements was within the principle of ancillary restrictions discussed above. The analogy with the validity of the non-competition provision in Métropole for the initial launch period seems to me telling. I am satisfied that the judge was right to conclude that all three aspects of the restrictions were directly related and necessary to the implementation of the object of the arrangements as a whole, as I have identified it. They are subordinate to the implementation of that object, and they have an evident link with it. The restriction was both objectively necessary for the implementation of the main operation and proportionate to it. Without the restrictions, the main object would have been, at the very least, difficult, and in reality impossible, to implement.
122. In my judgment the case falls clearly within the scope of the observations of the Advocate General in her opinion in BIDS at paragraph 53, quoted above at paragraph [84], and within the principle of ancillary restrictions described by the CFI in Métropole.
123. Accordingly, for that reason as well as for the others given above, I would reject Mr Vajda's arguments on behalf of the Appellants, and hold that the judge was right to regard the various arrangements as not falling within article 81(1), largely for the reasons he gave, though I would uphold the Respondent's Notice on the one aspect of his reasoning which concerned whether the RUK racecourses competed with each other at the relevant time, even potentially, for the sale of LBO media rights. The position may well be different on the expiry of the current licences. At that stage, AMRAC will be present in the upstream as well as the downstream market as a viable undertaking, and will not need the sort of protection which was necessary to secure its entry into the market at the outset.

Are the AMRAC licences void by reason of article 81(2)?

124. This question does not arise, if I am right that the arrangements in question do not infringe article 81(1). The judge did not decide it, for the same reason. We do not know whether, if he had decided the article 81(1) point against the Defendants, and had found that the arrangements did not satisfy art 81(3), he would have held that the AMRAC licences were themselves void.
125. What he did say on the point was only in the context of his general discussion of the law, at paragraphs 409 and 410. There he said that the fact that the supplier under a vertical agreement is a party to a horizontal price fixing agreement which is void does not make the vertical agreement itself void. However, he also recognised that the offending arrangements might involve vertical as well as horizontal elements. As regards that possibility he said, at paragraph 410:

"If the facts of a particular case led to that conclusion then both the horizontal parties and the person taking the supply from one of them could be held to be party to an agreement and, if that agreement infringed Article 81(1), then that agreement is void under Article 81(2). In such a case, the consequence would be that the party taking a supply from the horizontal parties would not acquire rights under the offending agreement (or the offending part of it)."

126. It seems to me that there is force in Mr Vajda's submission that the AMRAC licences, though in terms vertical agreements only, are in fact part of a wider agreement or concerted practice which is both horizontal (as between the racecourses) and vertical (as between all of them and AMRAC and RMS): see paragraph [55] above. This is not like the sale of a motor car to a retail purchaser by a car dealer which has entered into a horizontal agreement in breach of article 81 with other dealers, or with dealers and one or more manufacturers. If there was a breach of article 81, the AMRAC licences were at the heart of it.
127. I propose to follow the judge's course and not to decide what the result would have been on this point if I had held that the arrangements between the parties did infringe article 81(1). I do so for three main reasons. First, it is unnecessary, and therefore hypothetical. Secondly, if I were wrong about article 81(1) it would be necessary for the case to be remitted to a further determination at first instance of the issues arising under article 81(3). The judge did not decide the article 81(2) issue against the Appellants, so it is not the subject of a subsidiary binding ruling which could only be challenged on this appeal. Thirdly, despite Mr Vajda's submissions, the essence of which I have identified very briefly, the matter is strongly contested on the part of Mr Roth. It seems to me that, since the issue (if it ever matters) can be left to be argued about and decided on another occasion, it is better not to lengthen this judgment still further, or to delay its delivery, by expressing my view on what is at present only a moot point.
128. For the reasons given above on the other grounds, I would dismiss the appeal.

Lord Justice Moore-Bick

129. I agree.

Lord Justice Mummery

130. I also agree.