The legal status of rebates after Post Danmark II: my presentation
I was delighted to give a presentation on Post Danmark II in the Lunch Talk organised by the GCLC in Brussels earlier today. My slides can be found here. I am really grateful to the organisers and to Denis Waelbroeck, who chaired the session, for the invitation. The discussion with Nicholas Khan (European Commission) was really interesting.
Several participants in the Lunch Talk were of the view that some passages in Post Danmark II are not clear. In my presentation, I chose to avoid discussing individual paragraphs and focused instead on the aspects of the ruling that I find to be clearer and less subject to interpretation. The key points I made can be summarised as follows:
- The days of Michelin II and British Airways are over:
- A two-step test: The single most important contribution of Post Danmark II relates to the factors to consider when establishing the abusive nature of rebate schemes that are not conditional upon exclusivity. In Michelin II and British Airways, the assessment of ‘all the circumstances’ revolved around whether the practices had a ‘loyalty-inducing’ (or ‘fidelity-building’) effect. In the world of Post Danmark II, this is no longer enough (at least as far as standardised rebate schemes are concerned). The ruling sets out a two-step test that makes it necessary to consider both (i) the nature and the operation of the rebate scheme and (ii) the features of the relevant market, the regulatory context and the extent of the dominant position (paras 29-30).
- ‘Loyalty-inducing’ rebates are not abusive by object: In Michelin II, the GC took the view that ‘loyalty-inducing’ rebates are abusive by object (para 241). Similarly, in British Airways, the GC held that the scheme at stake in the case served an exclusionary purpose (para 288). After Post Danmark II it is clear that these schemes are only caught by Article 102 TFEU insofar as they are likely to have an exclusionary effect on competition. Thus, they do not fall within the ‘by object’ category.
- Towards the analysis of effects: Michelin II and British Airways were controversial cases because the impact of the practices on competition was not considered in the analysis. The exclusionary effects of the schemes were simply assumed to derive from their ‘loyalty-inducing’ nature. By contrast, the analysis of effects is an integral element of the legal test set out in Post Danmark II. The factors considered by the Court include the (i) extent of the dominant position; (ii) the coverage of the practice; (iii) the regulatory context and (iv) whether the dominant firm is an unavoidable trading partner. I pointed out that there are clear similarities between these factors and those sketched by the Commission in its Guidance. The threshold of effects (i.e. likelihood) is also the same.
- Open questions: The most obvious question raised by Post Danmark II is whether the same analytical framework applies to target rebate schemes. It is difficult for me to think of a reason why it should not. The second question is whether the ‘by object’ category is appropriate for exclusive dealing and loyalty rebates. The exclusionary potential of these practices, when implemented by dominant firms, is undeniable. The issue is instead whether it is appropriate to simply assume the exclusionary effects of these practices if this is not sufficient under the test set out in Post Danmark II. After all, rebates (including standardised schemes) are problematic from a competition law perspective when they come close in their nature and operation to exclusivity obligations. Why, then, should they be treated differently?
Chillin'Competition 
I very much enjoyed your insightful presentation at the GCLC lunch meeting and particularly agree with your assessment on adoption of concepts from the Guidance Paper.
I missed two points in your presentation with regard to the question whether exclusivity rebates should be treated as by-object violations:
(I) You correctly pointed to the plausible pro-competitiveness of exclusive rebates. But why does it not suffice to refer the dominant firm to the efficiency defence for this? (Which is of course limited in that it mirrors Art. 101(3) TFEU.) Isn’t it your view that a by-object classification should be based on a generalized balancing of anticompetitive effects and pro-competitive efficiencies? If so, on what side does an exclusive rebate by a dominant firm come out? (Relatedly, in Art. 102 TFEU cases, should this generalized balancing already take account of the “other” conditions in Art. 101(3) TFEU, particularly whether competition can be eliminated in respect of a substantial part of the products in question?)
(II) What happened to the notion that antitrust rules should be designed so as to minimize the costs of over-enforcement, under-enforcement, and administration of the legal rules? While treating exclusive rebates as a per-object abuse may in some cases amount to a type I error, the rule is very easy to administer. Also, dominant companies can still invoke the efficiency defence and only have to justify the actual foreclosure (which could theoretically only be one negligible customer), which limits costs from type I errors.
I’d be happy to hear your thoughts on this.
Robert
3 November 2015 at 7:46 pm
Thank you for your kind words, Robert.
– I was discussing the arguments that could justify treating exclusivity rebates as by object violations. One possible argument for the prima facie prohibition of the practice is that it is not pro-competitive, in the sense that it can be presumed to serve an exclusionary purpose. The Court and the Commission now accept unambiguously that exclusive dealing and loyalty rebates can be pro-competitive. Therefore, the ‘by object’ qualification cannot be based on a presumption of anticompetitive intent. In this sense, this conduct is very different from cartels. This has nothing to do with balancing. The point is rather that, when our experience shows that a practice has ambivalent effects on competition (it can be good or bad depending on the context), it is wise to prohibit it only when it is likely to have negative effects on competition. This is exactly the logic underlying Post Danmark II and what, in the context of Article 101(1) TFEU, the Court unambiguously held in Cartes Bancaires. I imagine you agree with me that there was no balancing whatsoever in Cartes Bancaires or in Post Danmark I and II.
– A prima facie prohibition is very easy to administer. No doubt about it. The problem with this argument is that an effect is required for some potentially abusive practices. Why is a margin squeeze not by object? As someone said during the talk, selective price cuts could also be made by object. A selective price cut does not reflect any cost savings or any economies of scale. It can also be a very cost-effective tool to drive rivals out of the market. Why did the Court hold that this practice is not by object in Post Danmark I? Standardised rebate schemes are now ‘by effect’ abuses. Why, if they can work in every way like exclusivity obligations and have the same effects, as the Commission rightly explains in the Guidance? The ease of administration argument misses this fundamental point: it would only be a valid argument if all potentially abusive practices were treated as by object infringements. And do not forget that the same argument can be used a contrario: a prima facie legality rule would be even easier to administer than a prima facie prohibition rule (and one could also say, as you do in relation to the efficiency defence, that it would always be possible for an authority or a claimant to rebut this presumption by showing that the anticompetitive effects of the practice outweigh its pro-competitive aspects)
Pablo Ibanez Colomo
3 November 2015 at 11:11 pm
Thanks Pablo, let me ponder over your response. 🙂
Robert
7 November 2015 at 4:40 pm
Dear Pablo, do I understand correctly that when you refer to “target rebate schemes” in the last section of the post (open questions), you mean “individualized” target rebate schemes only (and not standardized target schemes)?
Almos
12 December 2016 at 5:38 pm
Hi Almos,
Thanks for your comment. You are right. This is exactly what I had in mind.
Pablo Ibanez Colomo
14 December 2016 at 6:46 pm