The Future of Loyalty Rebates after Intel

By Abigail Hopper, J.D. Candidate 2018, Columbia Law School

The European Court of Justice (ECJ) took a step toward clarifying the antitrust implications of loyalty rebates offered by dominant firms in its Intel decision released on September 6. The decision rejected the interpretation of the General Court (GC), which ruled that Intel’s loyalty rebates were illegal per se because they were conditioned on exclusivity. The ECJ clarified that although rebates conditioned on exclusivity were presumed to violate Article 102 TFEU, the presumption is just that – a presumption. A firm offering exclusivity rebates can rebut the presumption by showing evidence that the rebates in question are not capable of distorting competition.

Critics who had urged the ECJ to embrace a more “economic” approach to rebates offered by dominant firms welcomed the decision. However, the ruling leaves a number of details to be clarified in future cases. Although the “as-efficient competitor test” (a form of price-cost test) must now be considered in exclusivity rebate cases, it is unclear whether and under what circumstances prices above cost can still be anticompetitive. The court was also vague as to how likely anticompetitive foreclosure would need to be in order for the rebates to be “capable” of restricting competition.

Background to the Intel Case

The GC’s decision that the Intel rebates violated Article 102 TFEU was based on its characterization of the rebates as “exclusivity rebates.” Based on prior case law, the court outlined three categories of rebates: 1) quantity rebates, which are presumed legal; 2) exclusivity rebates, which are formally conditioned on the customer’s purchase of all or close to all their requirements from the dominant firm and presumed illegal; and 3) rebates that are not formally conditioned on exclusivity but have a “fidelity-building effect,” in which case a court must look at “all the circumstances” to determine whether the rebates have the capability to foreclose the market. Rebates that fall into the third category encourage loyalty by conditioning retroactive rebates on the purchase of a certain target amount over a reference period, or a certain percentage of requirements (at a percentage not amounting to exclusivity) over a reference period.

Intel had offered loyalty rebates on its x86 Computer Processing Units (x86 CPUs) to four of its major equipment-manufacturing customers (Dell, Lenovo, HP, and NEC), and made payments conditional on exclusivity to a fifth customer. The discounts were conditioned on the customer obtaining either all or close to all of its requirements for x86 CPUs from Intel, according to agreements Intel had entered into individually with each of the customers. Because of the condition of exclusivity or near-exclusivity, the rebates fell into the second category of “exclusivity rebates.” The GC ruled that because exclusivity rebates were presumed illegal, it did not need to consider Intel’s argument that the “as-efficient-competitor test” (AEC test) carried out by the Commission was incorrect, and that a correctly applied AEC test would have shown that the discounts were priced above cost and therefore not capable of restricting competition from equally competitive firms (paragraph 151 of the judgment).

The AEC test looks at the effective price a competitor would need to offer in order for a customer to be willing to forgo the rebate and shift the “contestable share” of its demand to the competitor. It compares this effective price with the relevant measure of cost for the dominant firm. Prices above cost would in theory not harm an equally efficient competitor that would be able to match the effective price. In the case of rebates, the AEC test requires determining a contestable and incontestable share of demand, and applying the entire rebate discount to the contestable share. This contestable share is in theory the only share a competitor would be able to compete for, and it is this contestable share that a dominant firm can effectively foreclose to potential competition by using a rebate to leverage its incontestable share. A form of the test is used in both the European Union (EU) and the United States (U.S.), and was embraced by the Commission in its 2009 Guidance Paper. The ECJ ruled that the test was relevant to determining whether standardized rebates based on a target amount of purchases were anticompetitive in its Post Danmark II decision (decided after the GC’s decision in Intel). However, the rebate offered by Post Danmark was not an exclusivity rebate, and would fall into the third category of rebates that must be examined under “all the circumstances.”

In his opinion on the Intel case, Advocate General (AG) Wahl urged the ECJ to combine the second and third categories of rebates and examine “all the circumstances,” including evidence of prices above cost, for exclusivity rebates as well. This was in line with many critiques that argued that holding exclusivity rebates to a higher standard ignored the underlying economic reality of the discounts, which did not necessarily harm competition and could have procompetitive benefits.

The ECJ Opinion and the Future of Loyalty Rebates

Although the ECJ in Intel didn’t explicitly abolish the form-based distinctions, the presumption rebuttal it provides for in paragraphs 138-139 brings the treatment of exclusivity rebates and rebates that fall into the third category closer together in practice. In the case of exclusivity rebates, defendants will generally be able to produce economic evidence indicating that their rebates are not capable of restricting competition from equally efficient competitors, which means that a court must look at the factors outlined by the ECJ to determine if the rebates are in fact capable of anticompetitive foreclosure—it cannot say the rebates are illegal solely because they are exclusivity rebates.

Going forward, the future of loyalty rebates in the EU depends on how much weight is given to a showing that the dominant firm priced above the relevant measure of cost. If above-cost rebates can be anticompetitive (as the ECJ found in Post Danmark II), the question of how likely anticompetitive foreclosure must be in order for the rebates to be “capable” of foreclosing competition becomes more important.

The AEC Test

The decision does not discuss how much weight should be given to a finding that the effective price is above cost. In the U.S., prices above cost are usually given decisive weight. Most U.S. courts view loyalty rebates as a form of predatory pricing and require a showing of below-cost prices as an initial hurdle for any claim alleging that rebates violate antitrust laws. The price-cost test is applied not as another piece of evidence tending to show exclusionary effect, but as a necessary element for a plaintiff to have a claim at all.

In the EU, it is unlikely that the AEC test will ever be viewed as more than a (persuasive) piece of evidence, which can be outweighed by other circumstances indicating that the rebates can foreclose the market to equally efficient competitors even if priced above cost. This is in line with the ECJ decision in Post Danmark II, which dealt with loyalty rebates that did not rise to the level of exclusivity. The court found Post Danmark’s rebates capable of distorting competition even when priced above cost. The court referred to the test as “one tool amongst others” for determining whether a rebate scheme was capable of causing anticompetitive foreclosure (paragraph 61). The case presented a unique situation, however, where the dominant firm enjoyed a statutory monopoly and held over 95% of the market (making it all but impossible for an equally efficient competitor to exist).

A case may not need to be as unique as Post Danmark II, though, for above-cost rebates to be anticompetitive. Even in the United States, some commentators, including FTC commissioner Joshua Wright, have argued that allowing all above-cost rebates ignores economic evidence that rebates priced above cost can foreclose equally efficient competitors. Wright also points out that a price-cost test does not necessarily mean more certainty in the case of loyalty rebates, where it is often difficult to determine the contestable and incontestable shares of demand in order to accurately apply the test (in addition to deciding on the correct measure of cost).

Not all courts in the U.S. treat loyalty rebates like predatory pricing, subject to only a price-cost test. Most notably, the Third Circuit does not require a showing of below-cost prices in certain cases of conditional rebates where significant non-price incentives were used to exclude competitors. For example, in ZF Meritor v. Eaton Corp, the court rejected a price-cost requirement where the defendant, a dominant supplier of truck transmissions, had the right to terminate contracts with two customers (manufacturers) if they failed to meet the required purchase share for the rebates, and customers feared that not entering into the contracts would risk cancellation, price increase, and refusal to supply. Given the dominant position of the defendant in the truck transmission market, the Third Circuit determined that this resulted in significant non-price incentives for the manufacturers to purchase all or close to all of their requirements from the defendant. In Eisai v. Sanofi, the Third Circuit extended the holding to claims of non-price incentives resulting from the bundling of contestable and incontestable demand, at least where the incontestable demand was based on the dominant firm’s unique FDA approval for a particular use of a drug it manufactured.

A court in the EU would likely be less reluctant to find above-cost rebates anticompetitive than its U.S. counterparts. This is in part due to the EU’s traditionally more aggressive position toward abuses of dominance. In the U.S., the influence of the Chicago School and neoclassical price theory has led to the conviction that abuses of monopolization will be rare and that markets tend to self-correct. This view, combined with the fear of administrative error by judges and enforcement agencies have led courts and commentators to advocate for simpler rules and erring on the side nonintervention. Courts have been particularly cautious when it comes to price-cutting practices, like rebates. The EU by contrast has had both greater faith in the ability of regulatory agencies and courts to identify truly anticompetitive conduct by dominant firms and less faith in the self-correcting ability of markets.

Perhaps more importantly, antitrust in the U.S. is far more reliant on private plaintiffs suing for damages. The price-cost test weeds out vexatious claims and discourages potential plaintiffs from filing them. Given the more aggressive stance of the EU towards dominant firm abuses and the absence of the private plaintiff problem, EU courts will likely be more willing to consider circumstances under which even above-cost rebates could harm competition.

Capability and Likelihood

The impact of the ECJ’s Intel decision will in part rest on how courts determine whether a dominant firm’s rebates are capable of anticompetitive foreclosure. The ECJ did not specify how likely the foreclosure would need to be in order for the rebates to be “capable” of foreclosing the market. Is a theoretical capability sufficient? Or is capability synonymous with likelihood? If courts view above-cost rebates as potentially capable of foreclosing competition, the meaning of capability as it relates to likelihood becomes more important.

AG Wahl addressed the issue in his opinion and came to the conclusion that capability was synonymous with likelihood and the two phrasings had been used interchangeably in past decisions. According to AG Wahl, for rebates to be capable of restricting competition, it is not sufficient for anticompetitive foreclosure to be a “mere possibility” or even “more likely than not” (paragraph 117). The rebates would need to pose a higher probability of anticompetitive harm in order to violate Article 102 TFEU, although actual effects would not be required. AG Wahl’s view requiring a high level of likelihood is not a universal one. AG Kokott, in her opinion in Post Danmark II, wrote that anticompetitive foreclosure only had to be “more likely” than not in order for loyalty rebates to be capable of restricting competition (paragraph 82). The Danmark case addressed rebates based on a target amount (rebates that fall into the third category).

Form-based distinctions may come back into play as well, if rebates based on exclusivity require a lower showing of capability or likelihood than those that fall into the GC’s third category of rebates not based on exclusivity. Although the ECJ did not repeat the GC’s outlining of the three categories, it did reaffirm that a special category for exclusivity rebates exists, quoting Hoffmann-La Roche (paragraph 137 of the Intel appeal judgment). Such a distinction would undermine the ECJ’s move away from formalistic categories.


The ECJ’s Intel judgment brings EU case law on exclusivity rebates offered by dominant firms closer in line with mainstream economic thinking by acknowledging that such rebates are not always anticompetitive. The capability analysis the ECJ requires would allow firms to argue that their rebates are not capable of anticompetitive foreclosure. Whether this clarification will give firms confidence in adopting rebates as procompetitive pricing strategies depends on how courts perform the capability analysis, including whether and under what circumstances above-cost rebates can be anticompetitive, and how likely the anticompetitive foreclosure must be. Courts and commentators will continue to struggle with how to accurately address both the potential anticompetitive threats and procompetitive benefits of exclusivity and other forms of loyalty rebates.

Featured Image Source: