As reviewed in last week’s email/posts, the U.S. Supreme Court recently found that American Express’s ‘anti-steering’ rules did not violate U.S. antitrust law (in a decision reviewed here). In its judgment, the Supreme Court addressed a variety of topics essential to antitrust analysis – market definition, two-sided markets, harm through price effects and output effects, cross-market efficiencies and ancillary restraints – in ways which are at odds with the European approach. This paper, available here, seeks to compare the EU and US approaches in this respect.


It is structured as follows:

Section three contains a comparison of the AmEx majority and dissenting opinions. In the interest of clarity, I will review it here, instead of following the paper’s structure.

In Ohio v American Express, the majority held that only one market should be defined in two-sided transaction markets. Because there is a single relevant market, cognisable harm must refer to net harm across merchants and cardholders. Even demonstrating that the benefits granted to one side of the market do not fully offset the price increase to the other side does not suffice to show harm. Instead, establishing harm requires evidence of a price increase causing a reduction of output.

The minority’s dissenting opinion sets out a comprehensive rebuttal of this approach. First, the majority and the dissenting opinions disagree on the facts on the record. Second, the minority’s opinion argues that requiring proof of output reduction is too strict. Third, the dissent argues that the majority’s definition of a two-sided situation is not clear, and that the legal test laid down in the majority opinion will be difficult to apply in practice. The author finds support for this conclusion in a number of subsequent US cases where AmEx has been invoked to define the relevant market. Fourth, the dissent points out that there is no requirement for the plaintiff to set out net harm across the two sides of the market, since this collapses the various stages of the rule of reason. Fifth, the majority opinion does not sufficiently account for the fact that AmEx’s anti-steering rules removed its competitors’ incentives to lower their own payment processing fees.

A first section contrasts the EU and US overall approach to steering clauses.

In the U.S., anti-steering rules are prohibited with regard to debit cards by the 2010 Dodd-Frank Act, according to which merchants are legally entitled to offer discounts or other in-kind incentives to encourage their customers to use debit cards, but not as regards credit cards. By contrast, the EU’s MIF Regulation prohibits anti-steering rules with regard to both debit cards and credit cards. Even before 2015, the EU’s PSD1 Directive already prohibited some forms of anti-steering rules, although that prohibition was less far reaching than the prohibition adopted later in the MIF Regulation.

In the eyes of the European legislator, ‘steering’ exposes ‘more expensive means of payment’ to more competition, while anti-steering policies cause all consumers to subsidise the wealthier users of high-cost cards. This rationale for prohibiting anti-steering clauses in the EU extends to the US AmEx case. U.S. Federal Reserve economists have actually quantified this effect for U.S. households, with the wealth transfer estimated at an annual $1,282 transfer from the average cash payer to the average card payer’, and $438 from average low-income households to average high-income households. Interestingly, Justice Breyer’s dissenting opinion notes this point several times.

A second section contrasts the EU and US generic approaches to market definition in payment card systems.

In a horizontal ‘rule of reason’ case, evidence of harm obviates the need to define a relevant market in the US. A market must always be defined in vertical cases, even when there is already evidence of harm. According to the majority’s Opinion, this stricter standard falling on plaintiffs in vertical cases is justified because vertical cases ‘often pose no risk to competition’.

The EU also recognises that ‘vertical agreements are, by their nature, often less damaging to competition than horizontal agreements’. Nonetheless, market definition is only required ‘where it is impossible, without such a  definition, to determine whether the agreement or concerted practice at issue is liable to affect trade between Member States and has as its object or effect the prevention, restriction or distortion of competition’ (Volkswagen).

Going back to the third section, the paper looks at how competitive harm is approached in each side of the Atlantic.

The majority opinion notes that the alleged price increases caused by AmEx’s anti-steering provisions are not proven because ‘[w]here … output is expanding at the same time prices are increasing, rising prices are equally consistent with growing product demand’. According to the author, this may be true in general but is a problematic assertion in this case. In AmEx, people who choose to engage in card transactions (cardholders) are not the same as those who pay processing fees (stores). Since fees can increase without cardholders feeling the effects of that increase, the link between price and output is broken. Actually, ‘steering’ is a mechanism that can re-establishes a link between those who choose to engage in card transactions (cardholders) and those who bear the direct price of lack of competition (stores). In that sense, steering produces better price signals, which should be a good idea in the eyes of any antitrust lawyer or economist.

AmEx’s anti-steering clause could be seen as a ‘most favoured-nation’ clause: it ensures that stores will not treat AmEx less favourably than other card schemes in the way they present possible means of payment to customers. From an EU antitrust law perspective, a ‘most-favoured-nation’ clause can raise the exact same antitrust concerns that the prohibition on steering raised in the US context: (a) dampening competition among current competitors and (b) discouraging firms from entering the market through a differentiation strategy. Given all this, and even in the absence of evidence of actual price increases or output reduction, EU antitrust law would probably be able to tackle this situation. It seems plausible that a contractual clause shielding a firm from transparent and accurate price signals hampers the competitive process to such an extent that it produces likely anticompetitive effects, or even anticompetitive effects that are ‘so likely’ as to trigger the ‘restriction by object’ label.

Another sub-section compares the antitrust treatment of two-sided markets.

Like AmEx, the EU Cartes Bancaires case recognised the two-sided nature of payment card markets. The ECJ, having found that there were “interactions” between the issuing and acquisition activities of a payment system and that those activities produced “indirect network effects”, concluded that it was necessary to take into consideration the economic or legal context in which coordination takes place, it being immaterial whether or not such an aspect relates to the relevant market. That must be the case, in particular when there are interactions between the two facets of a two-sided system.

Particularly as regards efficiencies, the EU approach would likely be different from the AmEx majority’s approach. First, under EU antitrust law the defendant has the burden of proving efficiencies, and that the efficiencies fully outweigh the harm. Unlike what was required in AmEx, there is no requirement on the claimant to prove the absence of efficiencies or to provide a net account of harm and efficiencies. Second, under EU antitrust law, there is limited scope for so-called ‘cross-market efficiencies’ – negative effects on consumers in one geographic or product market cannot normally be balanced by positive effects for consumers in another, unrelated geographic market or product market. This echoes the position of the minority in AmEx.

A last section is concerned with ancillary restraints.

Near the end of the majority’s Opinion, Justice Thomas adds that merchants steering cardholders away from AmEx at the point of sale makes a cardholder less likely to use AmEx’s card. This externality endangers the viability of the entire AmEx network, so an anti-steering provision is a necessary ancillary restraint. In the minority, Justice Breyer, notes that the District Court already rejected this possibility on the (uncontested) facts. The author notes that the assessment under EU antitrust law would probably be the same as Justice Breyer’s view.


As the author notes, the AmEx ruling seems to diverge significantly from how EU competition law that would apply in a similar situation. That assessment strikes me as fundamentally correct. Without expressing my own views on the decision – you can find them in my comment to the case –, I think it would not be unfair to say that the author views the EU approach favourably, is sympathetic to the minority’s approach and is rather critical of the majority’s Opinion in Amex.

Author Socials A weekly email with competition/antitrust updates. All opinions are mine

What do you think?

Note: Your email address will not be published