The treatment of fidelity rebates is one of the most difficult and controversial topics in EU competition law and US antitrust law. Unlike in the EU, where a number of fidelity rebates are deemed abusive without the need to engage in detailed economic analysis, in the US it is consensual that rebates should be subject to an effects-based analysis. Nonetheless, the legal assessment of fidelity rebates in the US remains controversial. Some courts have adopted an exclusive dealing framework, while others have used price-cost tests; others still have applied a mix of the two frameworks. This diversity of approaches has led to intense academic debate in US scholarship, which finds a parallel in debates regarding whether the appropriate approach to fidelity rebates should be based on predation or on a raising rivals’ cost (RRC) framework.
This paper, available here, compares the EU and US approaches to fidelity rebates, and seeks to draw lessons from the US experience and apply them in the European context, as follows:
The second section maps out the US economic debate on how to evaluate fidelity rebates.
One finds two camps in US academic scholarship when determining what the appropriate approach to fidelity rebates is under antitrust law.
One side argues that fidelity rebates are mainly pro-competitive discounting practices that should only be unlawful when a dominant company sets its prices below cost in order to exclude its competitors. Under this framework, the relevant benchmark is whether the price charged by the dominant undertaking covers the costs that an equally efficient competitor would incur when producing the same goods/services. If the discounted price is above the discounter’s own cost, such price could theoretically be matched by any rival that is as efficient as the discounter. As a result, an equally efficient competitor will not be excluded from the market; instead, a fidelity rebates scheme will lead to intense competition to the benefit of consumers. It has also been argued that fidelity rebates may generate pro-competitive effects and consumer benefits in ways that simple volume discounts cannot, and that the lack of a reliable safe harbour such as a price-cost test is likely to discourage discounters from engaging in discounting that might be beneficial to consumers. Other pro-competitive justifications for fidelity rebates include encouraging sales in markets with high fixed and short product lifetimes, where entry is unlikely unless firms can secure a certain share of the market demand; preventing free riding; and avoiding double marginalisation. According to this school of thought, it will be difficult to distinguish between pro-competitive and anticompetitive discounting practices without a price-cost test.
The second camp holds that fidelity rebates should be treated under the law of exclusive dealing. Such an approach argues that harm to competition and consumers can arise even with sales above cost when this has the effect of raising rival costs – particularly as regards access to necessary input or customers. A body of literature regarding raising rivals’ costs shows that above-cost fidelity rebates can potentially lead to anticompetitive foreclosure. In many industries, achieving a minimum efficient scale is necessary for a competitor or an entrant to realise economies of scale and compete effectively with the dominant company. In markets characterised by high fixed costs and constant demand, it may be difficult for a competitor to achieve minimum efficient scale by serving only part of the market. By adopting fidelity rebates, a dominant company can exclude its competitors from sufficient access to markets/clients, thereby depriving them of the opportunity to compete and to achieve a certain level of output/minimum efficient scale. In other scenarios, fidelity rebates may even preclude a competitor from achieving minimum viable scale, which is the minimum scale of output (breakeven output) that a company needs to achieve in order to cover its costs (average cost equals price). In such circumstances, the rebate will lead the competitor to exit the market. In short, fidelity rebates can exclude or marginalise competitors by raising their costs without sacrificing the dominant company’s ability to make a profit, which makes the price-cost test inappropriate. Another implication of this theory is that some discounting practices can be anticompetitive even if they do not completely exclude a rival from the market; instead, a practice can be anticompetitive because it limits an incumbent’s rivals ability to compete effectively and to impose competitive constraints on the incumbent’s conduct. It is recognised that this approach begs the question of when is it that rival costs are raised to a point where it becomes problematic. In practice, the relevant benchmark for whether a strategy of raising rival’s cost might result in competitive harm is whether that conduct forecloses a sufficient share of the market to deprive competitors of the possibility of achieving a minimum efficient scale, so as to remain able to impose competitive constraints.
The third section describes two US cases which provide examples of how different economic approaches to fidelity rebates are applied in practice.
A first case is ZF Meritor v. Eaton. In short, Eaton dominated the heavy-duty truck transmissions market from 1950 to 1989, when Meritor entered the market. Following this, Eaton entered into new long-term agreements (LTAs) with truck manufacturers that were unprecedented in terms of duration – at least five years – and market coverage. The LTAs included up-front payments, and a conditional rebate provision according to which a purchaser would receive rebates only if its purchases from Eaton reached a specified percentage of its requirements. Following these agreements being entered into, Meritor’s market share fell to the point where this company left the market. In its assessment, the Third Circuit found that the most significant issue was whether the antitrust analysis of rebates should be subject to the price-cost test applicable to predation, or to the ‘rule of reason’ approach applicable to exclusive dealing. The court chose the latter, arguing that a price-cost test is appropriate only in a case in which price is the clearly predominant mechanism of exclusion. In this particular case, exclusion occurred because Eaton was an unavoidable trading partner, and the cost of breaching the agreement for truck makers was very high given the size of the lost rebate. The majority therefore reached the conclusion that Eaton’s conduct amounted to an unlawful exclusive-dealing arrangement, which is typically unlawful where the market is highly concentrated, the defendant possesses significant market power and there is some element of coercion present. This position can be contrasted with that of the dissenting judge, who held that Meritor did not suffer antitrust injury because the LTAs provided discounted but above-cost prices, which any equally efficient competitor could match.
In Eisai v Sanofis, Sanofi had a market share of 80-90% in an anticoagulant drug, and Eisai had a drug with 4-8% of the market. Sanofi offered hospitals price discounts of 1% if the volume of Sanofi’s drug the hospitals purchased was below 75% of their total purchases, and increasingly higher discounts – ranging from 9% to 30% – if they increased their purchases above this threshold. The Third Court declined to characterise the conduct as one where pricing was the predominant mechanism for exclusion, and rejected the application of the price-cost test. While Sanofi’s prices had remained above cost, Sanofi offered discounts that bundled incontestable and contestable demand. It was this bundling, not the price, that served as an exclusionary tool. While the Court accepted the argument that the price was not the mechanism of exclusion, it also rejected the claim that Sanofi’s offer amounted to bundling, and ultimately Court dismissed Eisai’s claim.
Section four will compare the EU and US approaches to the use of price-cost tests as regards fidelity rebates.
In Intel, the European Commission chose to apply a modified price-cost test for the evaluation of the anticompetitive effects of fidelity rebates. I have covered what followed elsewhere, so suffice to say that the Commission’s decision was appealed and triggered an important discussion about how EU law should approach discounting practices by dominant companies. One of the questions that arose in this discussion was what role price-cost tests should play in such assessments. The General Court held that the price-cost test was irrelevant. It considered that the usefulness of the test is limited to pricing practices and, as such, not relevant for the evaluation of exclusivity rebates such as those at issue in Intel. In particular, a price-cost test could not capture the rebates’ anticompetitive nature, while foreclosure effects could arise even if an as efficient competitor could theoretically enter the market. This was in line with the Post Danmark II judgment, where the CJEU held that the as-effective-competitor test is neither legally required nor decisive for establishing an abuse. On appeal from the General Court’s decision in Intel, the CJEU similarly held that assessing the effect of fidelity rebates requires the evaluation of all the circumstances in order to assess the possible existence of a strategy aiming to exclude an as efficient competitor, but did not clarify when and whether a price-cost test may be required.
While coherent with past case law, it is here that European courts depart from the US Third Circuit, which made clear that that a price-cost test might be useful in cases in which price is the clearly predominant mechanism of exclusion. The US debate on raising rivals’ costs can also contribute to understanding developments in the EU. In Intel, the General Court observed that an anticompetitive effect occurs ‘not only where access to the market is made impossible for competitors, but also where that access is made more difficult’, thereby shifting the focus of analysis from the exclusion of rivals to their marginalisation. This is line with a raising rivals’ cost approach, according to which a dominant company would be in a position to successfully protect or enhance its market power f it can ensure that its rivals are not able to reach a large enough portion of the market to achieve minimum efficient scale. Intel’s rebate schemes fit very neatly with this theory. Because of the strong dominant position of Intel, its customers obtained a substantial part of the demand (the ‘non-contestable’ share) from it. Intel’s competitors could only compete for the contestable portion of demand, which was foreclosed by Intel’s rebate arrangements. Foreclosure in such circumstances naturally depends on market circumstances, but in the right cases even a small amount of foreclosure might deprive a competitor or an entrant from minimum efficient scale, thereby creating strategic barriers that might be enough to marginalise them. It follows that the EU courts’ approach is consistent with an analytical approach based upon a raising rivals’ cost theory of exclusion similar to the one adopted by the Third Circuit in the US.
Since Mira was my supervisee, and I provided feedback on earlier versions of this paper, I will be circumspect in my comments. Naturally, I enjoy the paper, which translates complex economic debates into a language that even an unsophisticated lawyer like myself can understand.
The reason why I review this piece here is that it focuses on how we should treat practices that can deploy different mechanisms to damage competition – in this case, predation and exclusivity leading to foreclosure –, and whether there is something we can learn from other jurisdictions. In this, the paper is different from many other papers that look at the Intel judgment through the prism of whether it creates different legal tests for the treatment of different debates, or what the CJEU meant when it required the General Court to evaluate Intel’s challenge of the as-effective competitor test pursued by the Commission – matters that Mira has addressed in a different paper. In short, if you want a different perspective on the debates regarding Intel and the correct treatment of rebates by dominant companies, I strongly recommend this article.