Damned if you don’t, damned if you do?

What are the implications of the CFI decision to annul the Commission clearance of Sony BMG?

On 13 July the CFI delivered its verdict on an appeal by the Independent Music Publishers and Labels Association (Impala) against the decision by the European Commission to allow a merger combining the recorded music businesses of Sony and Bertelsmann Media Group (BMG) without imposing any conditions.1  Not only is this the first case where the Commission saw a merger clearance overturned on appeal, but following the Airtours verdict of 20022, it is another milestone decision in which the Commission’s analysis of joint dominance came under close scrutiny and was found to be wanting.

However, where in Airtours the Commission was criticised for being too ready to find joint dominance, the CFI’s findings in Impala were converse: the Commission had failed to show, to the requisite standard, that a position of joint dominance on the market for recorded music did not exist prior to the proposed concentration, or that the merger would not create such a position of joint dominance.


The proposed merger of Sony’s and BMG’s worldwide recorded music businesses (excluding Sony’s activities in Japan) was notified to the Commission in January 2004.  The newly created company (or companies) would operate under the name Sony BMG, and would be active in the discovery and development of artists, and the marketing and sale of recorded music, but would not undertake other activities (such as music publishing).  The notified concentration would bring together two of the five major recorded music businesses (which together accounted for around 80% of the EEA market), and would thereby reduce the number of so-called majors from five to four (namely Universal, EMI and Warner in addition to Sony BMG).

The Commission found that the notified concentration warranted further investigation, and in its Statement of Objections (SO), which was issued in May 2004, provisionally concluded that the proposed merger was incompatible with the common market as it would strengthen a position of joint dominance on the market for recorded music (as well as in the wholesale market for licences for online music).


Having considered the responses received to this SO, the Commission in July 2004 then declared the concentration to be compatible with the common market without imposing any conditions or undertakings on the merging parties.  It was this decision that Impala appealed in December 2004 because, amongst other reasons, in its view the Commission’s analysis that led it to reject the provisional finding of joint dominance in the SO was mistaken in fact and in law.

From an economic perspective, the key questions at stake clearly were what had caused the Commission to change its view on the existence of collective dominance, and whether the reasons given for this change of mind were strong and compelling.

The reasons for a finding of joint dominance in the SO

One of the main findings in the SO was that there had been substantial parallelism in prices for recorded music, both in terms of published prices to dealers (PPDs), and net prices (which were the result of applying various types of discounts to PPDs).  The Commission also noted that the responses to changing market conditions such as a fall in demand for recorded music, and the emergence of alternative distribution forms, was indicative of parallel behaviour.

In order to establish whether the observed parallelism was likely to be the result of tacit collusion (which would give rise to joint dominance of the majors), and whether the concentration would allow such tacit collusion to take place or make it more likely, the Commission considered a number of factors that are relevant in assessing whether a particular market is susceptible to tacit collusion3, including product homogeneity, transparency and the scope for retaliation against firms who might deviate from a collusive outcome.

The Commission considered that – despite the underlying heterogeneity of individual titles – recorded music was a fairly homogeneous product and that prices were transparent as PPDs were published in the majors’ catalogues.  Because a small number of titles (which would be easily identifiable given the publication of weekly charts based on sales) accounted for the large majority of sales, it would be sufficient to monitor a few price points in order to establish deviation from tacitly collusive behaviour.  Transaction net prices were closely linked to PPDs, suggesting that invoice discounts were broadly stable over time and did not show significant divergence owing to considerable transparency confirmed by customers.

The Commission also found that the majors were interacting in multiple markets, and that there were structural links, e.g. because of licensing and distribution agreements, distribution of joint ventures and collaboration on compilations, which bring together top titles of different majors on a single album.  The last point was also considered to be relevant with regard to the scope for retaliation: if one of the majors were to deviate from the collusive outcome, excluding that major’s titles from future compilations would provide the opportunity for punishing such deviation.

Overall, the Commission found that the substantial parallelism it observed in net prices could not be explained by competition, and that there were various features of the market that were conducive to tacit collusion.

Why the Commission did not find joint dominance in the decision

Whereas the Commission, in the SO, appeared to be of the view that the recorded music market was very prone to tacit collusion, and that the evidence indeed indicated that the five majors enjoyed a position of joint dominance, which would be strengthened by the concentration, its assessment of the market and the impact of the joint venture in the final decision was remarkably different.  Despite the finding of price parallelism in both PPDs and net average prices in the major markets, and price parallelism in PPDs in smaller markets (where net average prices were not examined), the Commission concluded that there was insufficient evidence to indicate that this might be the result of tacit collusion.  It therefore rejected the provisional finding that the five majors enjoyed a position of joint dominance, which would be strengthened as a result of the concentration.  It also rejected the view that the market was prone to tacit collusion, and that therefore the proposed joint venture might result in the creation of joint dominance.

The main reasons for this fundamental change in the Commission’s findings were that:

  • discounts, and in particular campaign discounts, were not transparent, and it was therefore not possible to infer the wholesale price from observing the retail price of any particular title; and that
  • there was no evidence that retaliatory measures had been used.

Because it was not easy to establish the level of discounts (and in particular campaign discounts) offered by a major from the retail prices, which could easily be observed, the Commission thought that PPDs – which were highly transparent – could not be used in order to co-ordinate pricing, and thus that one of the main conditions for tacit collusion would not hold.  Because of this, the Commission found that there was insufficient evidence to establish that the proposed transaction would lead to the strengthening of an existing collective dominant position, or that it would create such a position if none existed at present.

Why the Commission’s change of mind was challenged 

The stark difference between the provisional findings in the SO and those underpinning the decision was one of the main reasons for the challenge brought by Impala.  In particular, the applicants argued that the lack of transparency that the Commission found with regard to discounts was the result of focusing on the pricing of individual records rather than on the relevant question whether there were a few generally understood pricing principles, compliance with which was easy to monitor.

Indeed, the Commission had identified a number of pricing principles from its analysis of discounts (e.g. that campaign discounts vary with the size of the order, that they focus on customers with a reputation for selling a particular genre, or that they vary with the type of title or release).  The potentially large variation of discounts offered for particular titles and to particular customers, which might seem very complex to an outsider, was therefore the result of a handful of well-established pricing principles that would be easily understood by industry participants, who would equally easily be able to identify if one of the majors were not playing by these rules.  Thus, the most likely explanation for the observed alignment of net prices and PPDs (with regard to proper, volume-weighted averages rather than examples of individual titles) was tacit co-ordination amongst the majors on the basis of well-known pricing principles.4

Also, the applicants contended that the Commission’s focus on one retaliatory measure was too narrow, and that the fact that it had not been used did not suggest that it was ineffective – indeed, the most effective retaliatory measure is one that helps maintain co-ordination without having to be used.

The CFI decision

The CFI in its decision broadly supported the applicant’s view that the analysis supporting the Commission’s volte-face was insufficient and flawed. The Court noted that the decision contained contradictions and insufficient reasoning to support what the Court described as a “fundamental U-turn in the Commission’s position”.5  The Court also ruled that the evidence gathered was insufficient to support such a radical change of mind.

Thus, the Court “found that the theory that promotional discounts have the effect of reducing the transparency of the market to the point of preventing the existence of a collective dominant position was not supported by a statement of reasons of the requisite legal standard and was vitiated by a manifest error of assessment” and that “the Commission relied on the absence of evidence that retaliatory measures had been used in the past, whereas, according to case-law, the mere existence of effective deterrent mechanisms is sufficient, since where the companies comply with the common policy there is no need to have recourse to sanctions.”6

What about Airtours?

If the Airtours decision had created the impression that it would be safer to let a merger that might result in, or strengthen, a position of joint dominance go through unless the evidence in support of blocking it was overwhelming, this belief was thoroughly shattered by the CFI’s decision in Impala.  There is no comfort to be had from erring on the side of caution: as the CFI made clear, the Commission has to back up any decision it makes by a full and comprehensive analysis of the facts, and if it had (and had expressed) serious concerns about potential joint dominance, robust evidence would be needed to dispel such concerns.

This should not come as a big surprise.  As everybody who is making decisions under uncertainty knows, there is little point in focusing on avoiding type-I errors (taking an action even though it might be better not to do so) as this will only increase the likelihood of type-II errors (failing to take an action even if it were appropriate to do so).  The CFI decision only drove this point home to the Commission.  A rather different question is how the Commission should go about its business in a way that minimises the combined likelihood of type-I and type II errors.

What are the lessons?

The lessons from the CFI decision are thus both clear and uncomfortable: the evidence to support the decision to block or clear a merger needs to be of sufficient quality regardless of what the decision is.  Erring on the side of caution does not provide a safe haven.  In particular, significant changes in the Commission’s position from the SO to the final decision have to be supported by compelling evidence.

Notwithstanding the fact that the CFI found the Commission’s analysis to be flawed and insufficient in the specific case at hand, a comprehensive and detailed assessment of complex issues involved in a proposed concentration may often be difficult to achieve given that the timetable is tight, that the Commission has to rely mainly on information provided by the merging parties and interveners, and that the ability to leverage the experience of industry insiders by having one group comment on the arguments presented by the other is limited by confidentiality requirements.  Many of the arguments put forward by Impala in the appeal could have been advanced in the administrative procedure, had the interveners had sight of the response of the merging parties to the case put forward by the Commission in its SO.  This might have helped the Commission to probe the strength of its case for abandoning the provisional findings, and thus better to support its decision, whatever that decision might then have been.

The problem is particularly acute where the Commission proceeds from a strongly worded SO – which is perhaps intended to over-emphasise those issues that are of concern – to a clearance decision.  There, the time left for the Commission to change its mind is very limited.  This has a number of implications:

  • The obvious one is to call for more time for the Commission to understand market realities, or to increase its ability to involve interveners in assessing the strengths and weaknesses of the evidence submitted by the merging parties in response to an SO.  Both of these options have downsides – a more drawn-out procedure runs against the interest of merging parties to have a decision soon, and extending the involvement of interveners runs the risk of undermining confidentiality, which in turn affects the incentives of merging parties to put forward certain information.
  • Another implication is that the Commission should simply be more careful when preparing an SO.  Of course, any decision to clear a concentration following an SO that, by its very nature, proposes to block the concentration or impose significant undertakings on the merging parties involve a change in the Commission’s position, but the question is how fundamental this change can be.  The larger the change (in both substance and tone), the stronger the evidence that would be needed to support it.  Thus, rather than using the SO to probe the merits of the merging parties’ case by setting out the worst case scenario for competition in the strongest possible terms, it might be advisable to include only concerns for which there is already strong support.  As a result, SOs might have to become less aggressive in tone, and more balanced.  Although this might look like a desirable development, it may well make matters worse: the Commission would have to arrive at a fairly balanced and robust view by the time it issues an SO, which would further compress the timetable and create additional pressure.

Perhaps a better way of removing any tension between what is in the SO and what the final decision says would be to separate the role of prosecutor and jury.  If it were free to present the case against a proposed concentration (having decided that there are sufficient concerns to warrant further investigation) in the strongest possible form, with an independent agency make the final decision, the Commission could ignore concerns that it might be accused of having changed its mind, and having to present strong evidence to support this.  Calls for separating the two roles have been made in the context of Article 81 and Article 82 investigations – but it would seem that there are benefits from such a separation in the case of merger assessments, too.


Print This Post Print This Post


  1. 1. Judgment of the Court of First Instance of 13 July 2006 in Case T-464/04. []
  2. Airtours v Commission, Case T-342/99. []
  3. For a detailed discussion of these factors see M Ivaldi, B Jullien, P Rey, P Seabright and J Tirole, “The Economics of Tacit Collusion”, Report prepared for DG Competition, March 2003.  Following the CFI’s decision in Gencor v Commission (Case T-102/96), which upheld the Commission’s decision to block the merger between two platinum producers, and confirmed in Airtours, the notion of joint dominance is closely related to tacit collusion, and in particular whether a merger would make such tacit collusion more likely. []
  4. Indeed, at paragraph 254 of the Decision, the Court noted that even without considering whether or not the Commission had properly applied the conditions defined in Airtours in relation to the creation of a position of joint dominance, “a finding of a common policy over a long period, together with the presence of a series of other factors characteristic of a collective dominant position, might, in certain circumstances and in the absence of an alternative explanation, suffice to demonstrate the existence of a dominant position, as opposed to the creation of such a position.” []
  5. Paragraph 283 of the CFI decision. []
  6. CFI Press Release 60/06 of 13 July 2006. []

Comments are closed.