In particular, even where a dominant firm’s infrastructure was first established using public funds, provided that infrastructure was later acquired under genuinely competitive conditions and the firm retains full autonomy over access to it, the Bronner test must be met for a refusal to provide access to constitute an abuse of a dominant position.
We examine the background to the judgment, its reasoning and its implications.
Bronner and the evolution of the essential facilities doctrine
The CJEU’s 1998 Bronner ruling set a deliberately high bar for competition law to intervene to compel dominant undertakings – via the application of the prohibition of abuse of a dominant position, under Article 102 of the Treaty on the Functioning of the European Union (TFEU) – to share their own infrastructure with competitors (such scenarios often being referred to as “essential facilities” cases).
In particular, such a refusal will only constitute an abuse where: (i) access to the infrastructure is indispensable to the competitor’s business (i.e. because there are no substitutes and it cannot be replicated); (ii) the refusal is likely to eliminate all competition in a downstream market; and (iii) there is no objective justification for the refusal. These conditions seek to strike a balance between promoting competition on the one hand, and freedom of contract and the right to property on the other – recognising the need to preserve firms’ incentives to invest and to prevent free riding that could harm consumers over the long term.
Refusals to deal with or provide access to competitors in neighbouring markets can arise in a number of different constellations. As various allegations of abuse of dominance in such scenarios have made their way up through the EU Courts in recent years, the CJEU has clarified the situations in which the strict criteria in Bronner do and do not apply. In the process, it has limited the ability of dominant firms to shield behind the Bronner doctrine when defending themselves from such allegations. For example:
in Deutsche Telekom v Commission, the CJEU confirmed that where the dominant undertaking is subject to a legislative or regulatory obligation to grant access to its infrastructure, the Bronner criteria cannot be relied upon to defend a refusal to grant such access;in Lithuanian Railways, the CJEU concluded Bronner didn’t apply to cases where infrastructure was deliberately disabled or destroyed by the dominant firm, or where the infrastructure was not developed and owned by it; andin Android Auto, the CJEU ruled that a failure to grant a request for interoperability with a digital platform that was designed with third-party use in mind could constitute an abuse, even where such interoperability was not indispensable for the party seeking access. As discussed in our previous analysis of Android Auto, this underscored the limited role of Bronner in digital markets.Background to the Lukoil judgment
Two companies belonging to the Lukoil group (Lukoil Burgas and Lukoil Bulgaria) were owners and operators of fuel transport and storage infrastructure established under state ownership that was later privatised or operated under concession. The Bulgarian state continues to hold a “golden share”, conferring it certain rights in Lukoil Burgas.
In 2023, the Bulgarian competition authority found that Lukoil had, between 1 January 2016 and 31 March 2021, breached the Bulgarian equivalent of Article 102 TFEU by refusing rivals access to key transport and storage infrastructure that was unique and impossible to replicate.
In reaching its conclusion, the Bulgarian competition authority considered Bronner to be inapplicable on several grounds, including: (i) the fact that the infrastructure was built with public funds; and (ii) the existence of a legal obligation to guarantee third parties access to at least 15% of the storage capacity of certain facilities.
Lukoil disagreed and argued that Bronner should apply, appealing the Bulgarian competition authority’s decision before the national court, which subsequently made a preliminary reference to the CJEU. The reference asked three questions as follows.
Issue 1: whether separate types of conduct (which, owing to the specificities of Bulgarian law, the Bulgarian authority had to distinguish between), carried out by different legal entities within the Lukoil group, could lawfully be treated as a single abuse or a single infringement under Article 102 TFEU, or whether they had to be assessed separately.Issue 2: whether Bronner applies when the facility/infrastructure in question was publicly funded and has since been privatised or subject to a service concession; and if so, to what extent the relevant investment and contractual details matter.Issue 3: how the principle of proportionality should be applied to protect the dominant firm’s legitimate interests, whilst also preserving competition.The CJEU’s answers
Issue 1: Collective versus separate assessment under Article 102 TFEU
The CJEU confirmed that, when considering whether Article 102 TFEU has been infringed, it is sufficient to establish that the undertaking’s overall strategy is abusive. There is no need to establish that the conditions of Article 102 are met in respect of each type of allegedly abusive conduct. Furthermore, the existence of separate legal personalities within an undertaking’s corporate group does not fragment an abuse where the conduct forms part of a single plan producing exclusionary effects.
Issues 2 and 3: Bronner’s applicability, in light of infrastructure’s ownership, control and public origins
The CJEU referred to two criteria that must be present for the Bronner doctrine to be relied upon by a dominant undertaking as a “shield” against allegations of abusive refusal to supply: (i) the infrastructure must have been developed by the firm for its own needs; and (ii) the firm must own the infrastructure.
(i): origin of the infrastructure
The CJEU clarified that Bronner is not displaced simply because infrastructure originated in the public sector or was constructed using public funds. Where a dominant undertaking acquires such infrastructure at a fair, market-conforming price, under an open, competitive procedure (as those concepts are generally understood in EU state aid law, as clarified by Advocate General Medina in her Opinion) the asset is economically akin to infrastructure developed by the undertaking for its own business. In those circumstances, the firm has made an investment with a view to deploying the facility to serve its own commercial strategy, and therefore should not be compelled to share that facility unless the Bronner conditions are satisfied.
(ii): ownership/control of the infrastructure
The CJEU interpreted Bronner’s “ownership” criterion in functional terms. What matters is whether the dominant undertaking enjoys full decision making autonomy over access to the infrastructure it operates – i.e. owner-like control in substance. If legislation, regulations or the terms of concession contracts limit the undertaking’s autonomy with regard to that infrastructure, or prohibit it from refusing access, the undertaking cannot be regarded as the “owner” for Bronner purposes. Conversely, where a firm is not the formal owner but enjoys exclusive rights that confer owner-like autonomy over access (i.e. without access-related strings attached), the infrastructure is treated as if it is owned for Bronner purposes (provided those exclusive rights or concession were acquired under a competitive procedure).
Where does this leave us?
In the way it articulated the above cumulative conditions, the CJEU in Lukoil evinced a relatively nuanced, substance-over-form approach to applying the Bronner criteria in cases involving formerly public assets. The following two points are particularly noteworthy.
First, the ruling can be seen as rowing back (or at least clarifying) the approach taken by the CJEU only recently in Lithuanian Railways. In that case, the CJEU noted that the Bronner criteria were intended to apply only where infrastructure is owned by the dominant undertaking, and was developed for the needs of its own business by means of its own investments. The CJEU therefore concluded that the General Court, in its first instance judgment, had not erred in concluding that the criteria did not apply in situations where infrastructure was (as in that case) financed using public funds and not actually owned by the dominant firm. There was some uncertainty as to whether this should be interpreted (as the Bulgarian competition authority did) as excluding all publicly funded infrastructure from the scope of the Bronner criteria. Following Lukoil, however, even where infrastructure was originally developed under public ownership, and/or where the dominant firm is not its outright legal owner, further factual enquiries are needed before the application of the Bronner criteria can be ruled out. This is an important clarification given many essential facilities cases involve infrastructure established by state-owned enterprises that have long since been privatised.Second, the CJEU in Lukoil did not expressly confirm whether the limitations and obligations to which Lukoil was subject were enough to take the present case outside the scope of application of the Bronner criteria. The CJEU’s repeated references to the need for the dominant undertaking to “fully” control access suggest that they were so. But in her opinion, Advocate-General Medina noted that it was for the referring court to consider whether, among others, the allocation of a “golden share” and the imposition of an obligation to allow third parties to use a certain proportion of Lukoil’s storage infrastructure limited Lukoil’s autonomy sufficiently such that the Bronner shield was unavailable. Where the relevant boundary lies could perhaps be tested in future cases.
In sum, private operators of infrastructure established by former state monopolies – and those wishing to use that infrastructure to compete with them – would do well to pay close attention to this ruling.