Indispensability and Article 102 TFEU: it is not about Bronner (and refusals), but about Van den Bergh Foods (and remedies)
As you know, the hearing in the Google Shopping case took place last month – speaking of which: huge thanks to Lewis Crofts for his quasi-live tweeting of what was going on (all that follows is what I gathered from his reporting).
The question of the applicable legal test was always going to be central in Google Shopping. Lewis’s tweets confirm this view. A lot of time was spent, it seems, on whether the Commission should have established that non-discriminatory display on the Google platform was indispensable within the meaning of Bronner (and, I would presume, IMS Health, which provides the most precise definition of the concept in the case law).
When I read about these matters, I am always surprised about two aspects of the discussions. One is that they are dominated by Bronner. A second one is the fixation with categorising conduct as a ‘refusal’ (as if it were the relevant factor to determine whether indispensability is an element of the legal test). I explored some of the themes in a recent paper, and tell myself it could be useful to revisit them in a post.
Bronner is just one of many cases on indispensability: it is not even the most important or relevant
Let me start with the first point. Bronner is just one of many cases addressing the indispensability condition: it was not the first and it will not be the last. It is not the most important either.
In the context of Google Shopping, it is not even the most relevant precedent: I have repeated many times (on the blog, at conferences and elsewhere) that the case is pretty much a re-run of Commercial Solvents and CBEM-Telemarketing.
As much as Commercial Solvents and CBEM-Telemarketing, the case is about a company changing its commercial practices to favour an affiliate at the expense of rivals. Indispensability was an element of the legal test in the two precedents (the latter judgment made an explicit reference to an ‘indispensable’ service).
Against this background, the fundamental question one should be asking, in my view, is why indispensability would not be an element of the legal test in Google Shopping, given that the self-preferencing has the same object and effect as the conduct considered in Commercial Solvents and CBEM-Telemarketing.
In other words: is there a good reason to depart from the legal test set out in the two closest precedents? Intriguingly, Lewis’s tweets suggest that this is not how the discussion took place.
It does not matter whether there is a refusal: what matters is the nature of the remedy
I have been able to gather from reporting-via-Twitter that a great deal of the discussion revolved around whether there had been a refusal to deal. I get the impression that this was perceived to be a crucial matter. As far as I can tell, the underlying idea is that, absent a refusal of some kind, indispensability is not an element of the legal test.
There is nothing in the case law that suggests this conclusion.
This is so, first and foremost, because there are cases where indispensability was required (including the two mentioned above) and which did not concern a Bronner-style refusal. Second, the question of whether there is a refusal can easily turn into a semantic discussion. Was the behaviour in CBEM-Telemarketing a refusal to deal? Maybe, or maybe not, depending on what one calls a refusal. How about Slovak Telekom? Is that a refusal or a strategy aimed at degrading the conditions of access? Finally, there are cases that involve a refusal and where indispensability is not required (including Slovak Telekom itself).
The criterion to determine whether indispensability is an element of the legal test was defined in Van den Bergh Foods. This criterion distils the essence of previous cases. According to this ruling, indispensability is required when intervention would require a firm to ‘transfer an asset or enter into agreements with persons with whom it has not chosen to contract’.
It is easy to illustrate this criterion by reference to Bronner. In that case, the defendant could have brought the infringement to an end in two main ways: (i) by giving access to its delivery network (that is, enter into an agreement with a firm with which it has not chosen to contract) or (ii) by transferring its assets to a third party (that is, a structural divestiture). To be sure, it could also have (iii) closed down its delivery division.
Accordingly, indispensability was found to be an element of the legal test (and the condition was found not to be fulfilled).
Apply the Van den Bergh Foods criterion to any vertical leveraging case and you will realise that, when intervention in a case necessitates one the three options mentioned above (enter into agreements with third parties, sell the firm’s upstream or downstream assets or close down its upstream or downstream activities), indispensability is a condition to establish an abuse.
This criterion helps one understand why indispensability is an element of the legal test in Commercial Solvents (the firm was asked to enter into an agreement with a third party on terms and conditions determined by the authority), Magill and IMS Health (in the last two, intervention forced the firms to change their business model and start licensing their intellectual property to third parties).
Conversely, the criterion is helpful to understand why indispensability is not required in ‘margin squeeze’ cases or in other constructive refusal scenarios. A mere cease-and-desist order was enough to bring the infringement to an end in cases like TeliaSonera and Slovak Telekom.
Whether or not there is a ‘refusal’ (however this tenuous concept is defined) does not come across as a relevant or decisive factor.
How Van den Bergh Foods is interpreted in Google Shopping, and why it is controversial
Interestingly (and reasonably), the Commission concludes in Google Shopping that the question of indispensability should be considered in light of the Van den Bergh Foods criterion.
The application of the criterion is even more interesting (and also controversial). Since the decision merely requires the firm to bring the infringement to an end (without specifying how), the Commission claimed, indispensability is not an element of the legal test.
As explained in my paper, I am not sure this comes across as the most reasonable interpretation of the Van den Bergh Foods criterion. It makes sense I go briefly over the reasons why.
In essence, this interpretation would give the Commission the discretion to decide when indispensability is an element of the legal test. The authority would have the freedom to choose when it imposes this threshold upon itself.
To illustrate the Commission’s approach and its implications, just consider Bronner (where indispensability was both required and not met).
According to the interpretation of the Van den Bergh Foods criterion advanced in Google Shopping, the Commission would be able circumvent the indispensability condition merely by requiring the defendant to bring its infringement to an end.
In other words: if, in a case identical to Bronner (same facts, same economic and legal context) the Commission left it for the firm to figure out how to comply with the decision, indispensability would disappear as an element of the legal test.
Since competition policy is implemented through law, not discretion, I struggle with this interpretation of Van den Bergh Foods .
The relevant question, as I understand the underlying principles, is not so much what the decision formally requires but what it entails in substance. What options does the firm have to comply with the decision? Are these options the same as in Bronner? If so, the Van den Bergh Foods criterion means indispensability is an element of the legal test. If there are other options (namely a negative obligation administered on a one-off basis), it is not.
Those who have read my paper knew I have nothing to disclose. Those who have not (no hard feelings) now do. I look forward to your comments!
_______________________________________________________________________
UPDATE: A few among you have reached out to explain that, in the hearing, some importance was given to whether there had been a ‘request’ for access, as if this factor had an impact on the applicable legal test (and whether indispensability is part of it).
I fail to see how the existence of a ‘request’ is a relevant factor. Just take a look at the facts of CBEM-Telemarketing (para 5 of the judgment). As soon as CBEM learnt about a firm’s decision to cease dealing with third parties, it brought an action before a court (and note that, because there was no ‘request’, there was no ‘refusal’ either). Still, indispensability was deemed to be an element of the legal test (para 26).
Thanks again for all the comments!
With regard to “refusal”, part of the answer may well be that common law (tort) lawyers have a strange obsession with the distinction between acts and omissions. A negligent omission case is much harder to win than a case about a negligent act. And sometimes that distinction seeps across the Channel, particularly when everybody hires English barristers to speak for them.
martinned
6 March 2020 at 5:34 pm
Thanks a million as usual, Martin!
What you point out is interesting. Curiously, in Google Shopping the Commission also distinguishes between an active and a passive act (a refusal apparently being an example of the latter). In my view (and yours too, I understand) this is a distinction without a difference (at least in competition law).
The reason behind the fixation on this distinction without a difference may well relate to the factor you mention. But something tells me that, just like the make up of the Court, it is likely to change soon.
Pablo Ibanez Colomo
6 March 2020 at 5:50 pm
Dear Pablo,
the article ends by saying you have “nothing to disclose”. Yet, maybe it is worth disclosing that the co-author of this very blog is representing Google before the General Court.
This would make it easier to understand the rather one-sided presentation of the facts and interpretation of the law that I find here, while claiming to be neutral (which I do not claim for myself).
In my professional and personal view, none of your main premises are correct. Let me just point out a few:
Firstly, you claim Commercial Solvents and CBEM-Telemarketing required indispensability. Yet, the reasoning of the Commercial Solvents judgment never mentions the word. The Court only examines whether the requesting party (instead of any third-party) could have switched (i) its production to other inputs (raw materials) or (ii) to producing other end-products (to still compete in the same market): “Speculations concerning the availability of such [other manufacturing] processes, of such other raw materials and/or other end-products do not alter the fact that the industrially tied buyer cannot switch to other suppliers of other materials, without changing the economic and industrial basis of his undertaking” (page 233). Identically, in CBEM-Telemarketing the court referred to whether a service was “indispensable for the activities of another undertaking” (para. 26) in contrast to “any” undertaking as it did in Bronner and IMS. Therefore, this test was much broader than the indispensability criterion defined in Bronner and IMS. While in Telemarketing it sufficed that the input was relevant for the individual, requesting party, only Bronner and (more clearly) IMS define indispensability as, ultimately, a natural monopoly. I explained this in my comments to your earlier post on the topic (and several publications dating back to 2004) – but you do not seem to be interested in this important distinction between a termination of supplies and a de-novo refusal to deal, even though this debate was at the heart of the entire IMS Health decision.
Secondly, citing Van den Berg Foods, you argue that indispensability shall be required if a firm is to “transfer an asset or enter into agreements with persons with whom it has not chosen to contract”. You then go on to claim that Commercial Solvents (and CBEM-Telemarketing) falls into this category. Yet, the Van den Bergh Foods definition does not fit: Commercial Solvents had been supplying … for many years. There was a termination of supplies. Thus, Commercial Solvents had already “entered into agreements” with Zoja and had “chosen to contract” them with them in terms of Van den Berg Foods. With a mere cease-and-desist order (= stop the discontinuation of supplies to Zoja), the abuse was remedied. This was the main difference to Bronner and IMS: here the dominant party had never supplied the requesting (potential) competitor before – and therefore a mere cease-and-desist may not have sufficed. But the remedy is irrelevant for the finding of an abuse. What matters is: if a dominant company has never voluntarily supplied inputs to downstream rivals, the legal threshold for an abuse needs to be higher as compared to a case where a dominant company has long supplied inputs to down-stream rivals but then suddenly stops doing so.
Thirdly, for those very reasons, it is also correct for the EC in Google Shopping to distinguish between an active and a passive act. A company that actively changes its business conduct to foreclose competitors (Commercial Solvents, Telemarketing, Google Shopping) does more than a company that merely insists on continuing to keep its assets to itself (Bronner, IMS).
Fourthly, while I agree that the mere fact whether there was a formal “request” may not matter for the category of a refusal to supply, any “refusal” pre-supposes a corresponding demand. There must be competitors that have an actual commercial demand for whatever is refused. Yet, in Google Shopping, no competitor ever wanted Product Universals/Shopping Units in the first place, or any access to these (let alone any demotions). CSSs have always been crawled, indexed and displayed in Google’s general search results pages (as generic and paid results). There was no “refusal”, there was merely a change of business conduct to foreclose them.
It is a shame the oral hearing did not last longer. Then we could have discussed this in greater detail.
Kind regards
Thomas
Thomas Höppner
11 March 2020 at 1:04 pm
Thanks again, Thomas! Your involvement is much appreciated.
If you have read the blog, you will know how seriously we take the issue of disclosure. You may also know about the strengthening of the rules on JECLAP, which I promoted. In the same vein: thanks for clarifying that you have an interest to declare.
It is indeed the case that I have nothing to disclose. There is more: if you check my profile on the blog, you will see that I have committed never to doing consulting, expert reports/witnesses, paid-for papers and similar stuff. As an academic, I believe I have a duty to society to preserve my independence.
If you read the blog, you will see that I have held the same views since 2010, when the investigation was announced. These views were held long before Alfonso became involved in these cases (e.g. my piece in the Common Market Law Review) and without ever having anything to disclose.
And I have to thank Alfonso: he accepted to stop blogging on Google and never to discuss with me any Google-related posts. He does not even get to see these posts prior to publication (nor is he warned about them).
On substance: you have already explained your view that indispensability has two meanings in the case law, and I have already shared my views on the idea that the Court would give two meanings to a single concept. But even assuming that your analysis is correct, a fascinating follow-up question is whether the decision establishes ‘indispensability-lite’ to the requisite legal standard.
Your second point is equally intriguing. The Van den Bergh Foods criterion is not so much about whether Commercial Solvents was dealing with third parties, but whether the remedy involves requiring the dominant firm to deal with parties with whom it has chosen not to deal (Commercial Solvents arises precisely because of the choice made by the firm not to supply third parties).
Pablo Ibanez Colomo
11 March 2020 at 5:21 pm
[…] Indispensability and Article 102 TFEU (Pablo Ibáñez Colomo) […]
Links of March 2020 – Concurrentialiste Review
7 September 2021 at 11:50 am