On April 19, 2018, the European Court of Justice (“ECJ”) delivered a judgment dealing with the notions of abusive discrimination and competitive advantage under Article 102(2)(c) TFEU. The case concerned abusive price discrimination by a dominant company, between two customers competing in a downstream market (so-called non-exclusionary abusive discrimination). While ostensibly attempting to adapt the Intel approach to the case at hand, the judgment rather follows a different path.
In 2014, pay-TV services operator MEO complained to the Portuguese Competition Authority about a possible abuse of dominance. The abuse was allegedly carried out by copyright-fee collecting society GDA. GDA allegedly enjoyed a de facto monopoly in Portugal for managing copyright of artists and performers. GDA applied tariffs determined pursuant to an arbitral award to MEO. These tariffs were higher than those GDA charged to competing pay-TV provider NOS. MEO claimed that as a result of GDA’s pricing practices it suffered a competitive disadvantage vis-à-vis NOS.
The Authority found no evidence that the difference in copyright fees could affect MEO’s competitive position in the downstream markets where it competed with NOS. Therefore, the Authority rejected the complaint.
MEO appealed the decision. The Portuguese court specializing in competition matters remarked that discrimination “may, by its very nature, bring about a competitive disadvantage”. But in this case the amounts at stake were very low compared to MEO’s costs, revenues and profits in the downstream markets. So price discrimination did not look likely to undermine MEO’s competitive position. The judges decided to refer the case to the ECJ for a preliminary ruling.
By way of background, price discrimination between customers by a dominant company infringes Article 102(2)(c) TFEU, if it places least-favored customers at a “competitive disadvantage” with respect to favored customers. In previous cases (British Airways, Clearstream), the EU judges had provided little guidance as to the conditions under which price discrimination actually translates into a competitive disadvantage. One could have been forgiven for getting the impression that, under EU competition law, any long-lasting price discrimination among competing customers was abusive.
In essence, the Portuguese court asked the ECJ to clarify:
- whether price discrimination is always abusive under Article 102(2)(c) TFEU, as it always entails a financial disadvantage for the disfavored customer compared to favored customers;
- whether, to the contrary, price discrimination is not abusive when the resulting financial disadvantage is so small that disfavored customers can easily absorb it. They can thus effectively compete with favored customers;
- whether one can identify minimum thresholds below which price discrimination is presumed not to be abusive. Thresholds could be for instance a minimum percentage of the disfavored customer’s costs. In the alternative, of the average costs of standard offerings in the downstream markets where customers compete.
The MEO judgment started its answer by recalling the legal test in prior case law. The main objective of the prohibition of abusive discrimination is to prevent distortions of competition between competing customers in the downstream market. To establish an abuse, both (i) price discrimination and (ii) its ability to inflict a competitive disadvantage on the discriminated customer should be proved. To assess the competitive disadvantage one must look at all relevant circumstances on a case-by-case basis. There is no necessity to find actual signs of deterioration in the competitive position of the customer, as long as it is clear that the conduct is potentially able to cause that.
Where the MEO judgment provides a first original contribution to prior case law is in the clear restatement that not all “immediate disadvantage” inherent in price discrimination is a “competitive disadvantage” for the affected customers. In other words, it is always necessary to additionally prove the “competitive disadvantage”. Price discrimination cannot be considered per se abusive, just because it inevitably entails negative financial repercussions for the discriminated customers compared to their competitors.
According to the EU judges, it is true that no “appreciability (de minimis) threshold” can be set out to decide whether price discrimination is so serious as to be abusive. Nevertheless, the legal test requires enforcers to look at “all the relevant circumstances of the case”, to check whether the discrimination’s “effect on the costs, profits or any other relevant interest” of the competing customers is such as to affect the disfavored customer’s ability to compete in the downstream market.
To carry out this analysis, the ECJ suggests “by analogy” the use of a revised version of the foreclosure test, as spelled out in Intel (see ECJ Intel judgment of 2017, §§ 138-139). This analogy is the second most important original contribution of the MEO judgment.
According to the ECJ, enforcers should consider the following as important elements of the analysis: the supplier’s “dominant position”, the customers’ “negotiating power” as regards the “tariffs”, the tariffs’ “conditions” and “arrangements” including their “duration” and “amount”, and the possible existence of a “strategy” to eliminate the strongest customers, who are no less “efficient” than their “competitors” (§ 31).
Curiously enough, in adapting the Intel foreclosure test to non-exclusionary abusive discrimination, the ECJ has taken out from the test the passages which required enforcers to (i) define the relevant downstream market(s) where the customers compete with each other and where the price discrimination may have an effect and (ii) identify the market shares and the competitive position of the affected customers in the relevant market(s). These elements are normally necessary in a foreclosure analysis to identify distortions of competition.
In applying this legal test to the facts at hand, the ECJ further considered that:
- MEO and NOS were GDA’s main clients and had some negotiating power;
- an arbitration clause ensured that GDA’s prices would be determined by arbitral award, in case of disagreement among the parties. The tariffs applied to MEO had in fact been established by an arbitration decision;
- the alleged price discrimination had lasted approximately four years and it concerned a “relatively low percentage of the total costs borne by MEO”. The difference in prices between MEO and NOS was also limited, compared to MEO’s profits;
- there was no evidence, nor reason to presume, that GDA had an interest in excluding MEO from the downstream market, because GDA was not vertically integrated in that market.
Discrimination under Art. 102(2)(c) TFEU: A Few Comments
Discrimination not an abuse per se
There is inherent tension among the passages of the MEO judgment saying that price discrimination is not per se abusive, and those ruling out the use of appreciability thresholds.
The MEO judgment confirms that a quantitative analysis of the “effect on the costs, profits or any other relevant interest” may indeed be relevant to confirm the lawfulness of price discrimination.
However, it also says that even minimal price discriminations could have important competitive implications in a downstream market, depending on the specific circumstances of each individual case. This is why, according to the ECJ, it would make little sense to provide specific indications, e.g., in terms of cost percentage, even only as a first analytical screen to create a rebuttable presumption of absence of competitive disadvantage.
The resulting rule of thumb should still be that price discrimination accounting for a very little percentage of the disfavored customers’ costs or profits should not be considered abusive, absent exceptional circumstances.
Competitive disadvantage and foreclosure
In theory, competitive disadvantage and foreclosure could be construed as two completely different tests.
As happens in cases of unfair competition for infringement of sector regulation (which may require proof that the infringer has gained an unfair competitive advantage as a consequence of the infringement), a competitive advantage analysis could focus only on the relationship between favored and disfavored customers. If the discriminatory disadvantage concerns an important parameter of competition, such that the disfavored customer can no longer compete on a level playing field with the favored customer, one could actually say that there is a competitive disadvantage among the affected customers, regardless of the discriminated customer’s ability to compensate for it and survive (or even thrive) on the downstream market by leveraging on other resources.
This kind of analysis does not even begin to question whether the risk of exclusion of least-favored customers would have negative consequences for competition in the relevant market as a whole. This is because price discrimination objectively interfering with the business of one customer to favor other customers may be considered unfair in itself, regardless of its potential impact on competition in the downstream market. It puts an unnecessary burden on the least-favored customer for no appreciable economic reasons: a dominant supplier could obtain comparable revenues by applying non-discriminatory prices to all customers, instead of subsidizing discounts to some customers with higher prices to others.
On the other hand, the foreclosure analysis normally looks at the consequences of a given conduct for competition in the market, from the angle of a more economic approach. Among other things, a foreclosure analysis aims at establishing whether the conduct is capable of excluding disfavored customers from the market, whether it only concerns a marginal part of the affected market or whether it tends to exclude only inefficient competitors. Evidence of actual competition dynamics in the affected markets or in comparable markets can also play an important part in the assessment.
By making reference by analogy to Intel’s legal test, the MEO judgment seems to suggest that the analysis of competitive disadvantage should be similar to that of foreclosure. After all, if the objective of the prohibition on abusive discrimination is to prevent distortions of competition in the downstream market, it would make sense to carry out a foreclosure analysis and to prohibit only discrimination which is capable of bringing about such distortions. In a way, non-exclusionary abusive discrimination is actually an exclusionary (rather than exploitative) type of abuse: the main difference with classic exclusionary behavior is that the dominant company’s conduct is not aimed at excluding its own competitors in the upstream or downstream markets where it operates, but rather some customers to the benefit of others.
The application by analogy of a foreclosure test to the notion of competitive disadvantage would thus imply, for instance, that if the least favored competitor is a minor and inefficient player in the downstream market, such that its disappearance would not affect competition in the least, in principle there should be no abuse. In Post Danmark I and Intel, the ECJ clearly stated that Article 102 TFEU does not intend to protect “competitors less efficient than [the dominant company]” and that “not every exclusionary effect is necessarily detrimental to competition” (§§ 133-134). In MEO, the ECJ actually confirms that an exclusionary strategy by the dominant company against its customers would only be relevant if it is aimed at customers which are “at least as efficient as [their] competitors”.
However, the test in MEO is really a hybrid and several pieces of the puzzle are missing.
Firstly, it is not easy to explain why the ECJ has not expressly required the enforcers to define the relevant market in which the discriminated customers compete. Nor has it required them to assess the customers’ competitive position and market shares in that market. Those are necessary steps in a foreclosure analysis, as Intel has made clear. Without looking at the state of competition and at competitive dynamics in the affected market, it is extremely difficult to understand whether price discrimination is actually able to affect competition.
Furthermore, the customers’ negotiating power and the criteria which were followed to set the prices (e.g., through arbitration) may be relevant to establish whether there is price discrimination. Much less so for an analysis of the competitive disadvantage. Once it is clear that there is a difference in price which is not justified by objective circumstances, the analysis of the competitive disadvantage should focus on the objective impact of the price difference on the least-favored customer’s ability to compete.
Price discrimination may be justified by differences in the customers’ negotiating power or by an arbitral decision. But once it is established that there actually is discrimination, the same elements do not say much as to whether the least-favored customer (which does not have sufficient negotiating power or evidently did not effectively exercise it, if it ends up being discriminated against) is actually able to compete on a level playing field with the other customers.
Also, the analysis of the supplier’s strategy (which is also recalled as an important element in the MEO judgment, following the Intel reasoning) is rarely useful, if at all, when the supplier is not vertically integrated. Suppliers would normally have no interest in damaging the business activities of their customers, to the point that they would no longer be able to sustain purchases. In MEO, the ECJ confirms that there was no reason to presume such an exclusionary intent.
Moreover, once the Intel foreclosure test is applied out of context to a competitive disadvantage scenario, the as efficient competitor parameter loses significance, because it can no longer refer to the costs of the dominant company (or to a reasonable proxy for these costs).
The MEO judgment’s “as efficient as its competitors” test is difficult to translate in practice: should one simply compare the costs of the disfavored customers to those of the favored customers in the downstream market? Should one look at the relevant underlying costs of comparable offerings? Should one look instead at the costs of the most efficient competitor that is active in the downstream market, regardless of whether it is also a customer of the dominant supplier? Should one look only at the most efficient among the possible various customers? Or is the ECJ really referring to an average efficient customer or to an average efficient competitor on the downstream market? The MEO judgment leaves us with this cliffhanger and we can only look forward to the forthcoming episodes in the hope of finding an answer.
A Spartan sense of ethics: acceptable to kill the weak?
According to the MEO judgment, it would seem that a dominant company is entitled to carry out price discrimination strategies aimed at imposing worse contractual conditions on customers who are already more inefficient than their competitors.
This is a very ancient principle. Myth has it that under Spartan law it was ok to abandon unhealthy babies on Mount Taygetus. The abandonment aimed at purifying society of unnecessary burden. Likewise, if one takes the MEO judgment at face value, dominant companies could lawfully prey on the weakest customers and accelerate their demise in the downstream market, by taking advantage of their customers’ lower contractual power and by imposing extra-costs on them.
Of course, a dominant company which is not vertically integrated downstream would normally have no interest in doing that. However, it is difficult to justify the MEO approach, at least from an ethical point of view. It is unclear why dominant suppliers rather than normal competitive dynamics should be entitled to decide who should remain active in the downstream market. It is equally unclear why dominant companies should be more wary of extracting revenues from more efficient customers who could possibly afford to pay higher prices.
The MEO judgment deserves praise for restating, in line with prior case law, that not every price discrimination by a dominant supplier is abusive, if it does not give rise to a competitive disadvantage for the least favored customer.
The MEO judgment is also groundbreaking in its attempt to transpose the more economic approach of Intel by analogy to a different kind of abuse.
But the judgment raises more questions than it answers. It alters the Intel foreclosure/AEC test beyond recognition. The resulting legal test for abusive discrimination seems inspired by an unorthodox idea of fairness, and has little in common with the more economic approach.
To make sure you do not miss out on regular updates from the Kluwer Regulating for Globalization Blog, please subscribe here.