Brussels à jour

Night Train to…Nowhere

The appeal filed against the General Court's judgment in the Lithuanian Railways case gave Advocate General Rantos the opportunity to analyse the so-called "Bronner criteria" for the classification of a refusal to supply by a dominant undertaking as an abusive practice. While the Court of Justice's judgment is still pending, stay with us, this one is a real page-turner.

When it removed a 19km stretch of railway track to the border with Latvia, little did LG (the Lithuanian national railway company) know that it would leave its mark on EU competition law. What came to be known as the Lithuanian Railways case is now testing the boundaries of the essential facilities doctrine, established in Bronner,[1] against a set of facts which can occur more frequently than one would think. We are taking a look at the history of the case through the lens of Advocate General Rantos.[2]

Facts and Figures

The facts of the case gravitate around the trendy world of trains. More precisely, LG concluded an agreement in 1999 with Orlen, a company specializing in refining crude oil and distributing refined oil products. The object of the agreement was the transport of the vast majority of the oil products resulting from Orlen's refinery in Lithuania to the Lithuanian sea ports, while the remainder was transported also by rail to and through Latvia for internal consumption on the Baltics' market. For the latter category, LG concluded a contract with LDZ (the Latvian national railway company), which also included the Lithuanian sector, under a subcontracting agreement with LG.

Following a dispute regarding rates between LG and Orlen, the latter considered contracting directly with LDZ for rail transport services of the internal consumption products, as well as switching its seaborne export business from the Lithuanian seaports to the Latvian seaports. Orlen made its intentions known to LG, which resulted in arbitration proceedings against Orlen, as well as LG's unilateral termination of the 1999 agreement. In parallel, Orlen requested a quotation from LDZ along the lines of its new transport plans. A few weeks later, LG, after identifying a defect in the rail track to Latvia caused, according to technical reports, by physical deterioration of rail elements, undertook the complete removal of the track. Orlen continued negotiations with LDZ, during which it concluded a new agreement with LG. However, discussions with LDZ were discontinued when it became apparent that LG had no intention to repair the track.

Orlen filed a formal complaint with the Commission, which concluded in its subsequent decision that LG had abused its dominant position as railway infrastructure manager in Lithuania by removing the track. The Commission noted that in so doing, LG had excluded the competing operator from Latvia.

LG challenged the Commission decision before the General Court, arguing – in essence – that the Commission should have treated LG's conduct as a refusal of access to an essential facility (the railway system) and, accordingly, should have applied the essential facilities doctrine as established in Bronner. The reason behind LG's insistence on having Bronner (and the essential facilities doctrine) applied is the high evidentiary standard which it imposes, as compared to other forms of abuses. If LG could frame the case in those terms, it would have an easier legal defense than if the case was judged as a straightforward competition abuse. In particular, LG argued that the facts did not meet the indispensability prong of the Bronner test because there was an alternative route to Latvia (though longer than the dismantled track). The General Court dismissed all pleas, but lowered the amount of the fine. Subsequently, LG filed an appeal before the Court of Justice, which is still pending. However, the (non-binding) opinion of Advocate General Rantos was published on 7 July, shedding light on the core issues of the case.

Bronner's Cube

One of the main takeaways of Rantos's opinion is the fact that the Bronner test is here to stay, if there was ever any doubt about it after cases like Google Shopping.[3] In short, the test says that a dominant undertaking is obliged to grant access to its own essential infrastructure only if the following cumulative conditions are met: (i) its refusal is likely to eliminate all competition on the relevant market, (ii) its refusal cannot be objectively justified and (iii) the access is in itself indispensable for carrying on the person's business.

Rantos spells it in capital letters that the test is still relevant nowadays. Rantos specifically recalls and reaffirms the rationale for applying a particularly strict test and high evidentiary burden where the (allegedly) abusive conduct concerns that refusal of access to an essential facility. In particular, Rantos's opinion reminds us that, as a starting point, undertakings – whether dominant or not – have a right contract and dispose of their property as they deem fit and that the Bronner test serves to protect that very (constitutional) right. Second, Bronner seeks to promote competition in the long term, in the interest of consumers, by allowing a company to reserve for its own use the facilities that it has developed and thus preserving that company's incentives to invest in the construction of these facilities in the first place.

The Bronner Junction – First Stop: Refusal of Access

Rantos goes to great lengths in his explanation of the Bronner criteria, in order to shed more light on the nuances of a test, in particular in terms of types of conduct that it applies to ("refusal of access") and the limitations that follow from the rationale underlying the Bronner test ("objectives of the Bronner criteria").

First, for Bronner to apply, the conduct in question has to amount to a refusal to grant access to essential facilities. That refusal can be explicit or implied ("constructive" or "de facto"). The issue that Rantos considers in this context is in which circumstances conduct that does not amount to an explicit refusal of access but may raise issues common to such a refusal (i.e. which may have a similar effect) qualifies as a (implicit or constructive) refusal of access for purposes of the Bronner test, as opposed to an (other) independent form of abuse distinct from a refusal of access that should not benefit from the higher evidentiary burden established by Bronner. Rantos concludes that the removal of the 19km of rail tracks fell into the latter category. Rantos sums up the main differences between a genuine refusal to grant access and other types of Article 102 TFEU which may have the same exclusionary effect, namely:

  • In cases of a genuine refusal to supply, the dominant undertaking reserves its exclusive right of access and does not suffer loss of infrastructure, while in the current case by destroying the infrastructure it intended to harm its competitors. As such, according to Rantos, LG's conduct only motivated by the willingness to harm competition;
  • LG's conduct followed a pure logic of predation, According to Rantos, the only rational explanation for LG's conduct was the monopoly reward that it would be able to reap by excluding LDZ from the market (similar to predatory pricing cases). That logic is different from a genuine refusal to grant access under Bronner;
  • The very logic of Bronner is based on maintaining infrastructure, by means of preserving the owner's incentives to invest in that infrastructure. LG's destruction of the rail tracks is inconsistent with that logic;
  • The consequence of finding a Bronner type of abuse has the consequence of requiring the dominant undertaking to conclude a contract (to grant access), while in the Lithuanian Railways case, the Commission merely ordered LG to bring the infringement to an end, without ordering the conclusion of a contract.

The Bronner Junction – Second Stop: Objectives of the Bronner Criteria

Second, even if the conduct amounts to a genuine refusal to grant access (explicit or implicit), Bronner will only apply if the rationale justifying the "preferential treatment" under the Bronner test are present. LG had argued that Bronner did not establish such an implied requirement. However, according to Rantos the Bronner test should generally apply only:

  • In cases where the infrastructure to which access is sought is owned by the dominant undertaking and has been financed with its own investment. In the case at hand, the railway tracks were not LG's property and LG had not undertaken any investments in the construction and development of the Lithuanian railway structure (which was financed by public funding);
  • In the absence of a regulatory obligation imposed on the undertaking in order to grant access to its infrastructure, which was not the case here, since LG was obligated, inter alia, to grant access to the public railway infrastructure in its capacity of infrastructure manager.

Next Station is…

We will certainly keep an eye on the Court of Justice's upcoming judgment. Together with Google Shopping, it paves the way for new categories of discrimination under Article 102 TFEU, especially in the essential facilities space.

In the meantime, have a good rentrée and don't forget to follow us on LinkedIn, for your favorite EU competition law topics!

[1]           Case C‑7/97, Oscar Bronner, 26 November 1998, ECLI:EU:C:1998:569.

[2]           Opinion of Advocate General Rantos delivered on 7 July 2022 in case C-42/21 P, Lietuvos geležinkeliai v Commission, ECLI:EU:C:2022:537.

[3]           Case T-612/17, Google and Alphabet v Commission, 10 November 2021, ECLI:EU:T:2021:763.


» DOWNLOAD Newsletter (PDF)


Don’t miss any edition of our Brussels à Jour Newsletter.
You can simply follow the hashtag #Brusselsajour on LinkedIn to make sure you receive our updates in your feed.