US regulators’ swipe at Amex declined by Supreme Court

Published On 29/06/2018 | By Mark Giuseppini | Consumer protection, Enforcement, Litigation

On 25 June 2018, the Supreme Court of the United States dismissed an appeal brought by the US Government and a number of US States against credit-card provider American Express (Amex) for alleged anti-competitive conduct in violation of section 1 of the Sherman Antitrust Act (Act).

The Ohio v American Express Co case deals heavily with the concept of ‘two-sided platforms’ – defined by Thomas J as platforms offering different products or services to two different groups, both groups depending on the platform to intermediate between them. One of the key issues explored by the Court was whether, when examining two-sided platforms for anticompetitive effects, it is necessary to consider the effect of the conduct on both sides of the platform, or only one.

As two-sided platforms become increasingly prevalent (particularly in the e-commerce space), the Court’s judgment helps to clarify that for US antitrust purposes, a market may include both sides of such a platform where the platform exhibits pronounced indirect network effects (further discussed below). This decision is likely to be welcomed by industry participants, and may also influence the manner in which courts in other common law jurisdictions (such as Australia) approach market definition in relation to two-sided platforms.

Factual background

Amex competed with Visa, MasterCard and Discover in the US credit-card market.  Its merchant fees were higher than those of its competitors, which Amex justified on the basis that it provided greater value to merchants by drawing in wealthier cardholders with a higher average spend. To encourage this greater cardholder spending (and thus increase the revenue that it earned from merchant fees), Amex provided more generous rewards to its cardholders than its competitors.

Because of Amex’s higher merchant fees, it was in merchants’ interests to encourage consumers to use payment methods with a cheaper cost of acceptance (e.g. Visa or MasterCard) – for example, by advising them that ‘it costs us more to accept Amex’, or by offering free shipping if customers pay with another card. To combat this practice, Amex included ‘antisteering provisions’ in its merchant contracts, which effectively prohibited merchants from dissuading customers from using Amex cards.

At issue was whether the inclusion of these provisions was anticompetitive such that it constituted an ‘unreasonable restraint of trade’ in violation of the Act. To establish their case, the plaintiffs were required to show that Amex’s term violated the ‘rule of reason’; that is, that the restraint had a substantial anticompetitive effect that harmed consumers in the relevant market. The key question in this case was how the relevant market should be defined, in light of the Amex credit-card network’s status as a two-sided platform.

Procedural history

District Court

At first instance, the District Court considered that the credit-card market consisted of two separate markets: the market for card issuance (cardholder market), and the network services market (merchant market). The Court held that despite the two-sided nature of the platform, the relevant market was the merchant market. Because the District Court found that the antisteering provisions eliminated any advantage that lower merchant fees might produce (due to merchants’ inability to steer customers to a lower priced card), Amex was held to have breached the Act.

Court of Appeals

On appeal, the Court of Appeals for the Second Circuit held that the District Court erred in its market definition analysis by excluding the cardholder market. It was necessary to include this market, according to the Court of Appeals, because the cardholder and merchant markets were interdependent – i.e. the antisteering provisions affected competition for cardholders, as well as merchants. This was due to the fact that if Amex increased its price in the merchant market, it would lose merchants. And this loss of merchants would have “feedback effects” in the cardholder market – namely, it would encourage cardholders to switch to alternative forms of payment.

Supreme Court decision

In a 5-4 decision, the Court upheld the judgment of the Court of Appeals. As emphasised by Thomas J, because the two-sided transaction platform under consideration in this case exhibited “pronounced indirect network effects and interconnected pricing and demand”, it was better understood as supplying a single product: transactions.

Proving anticompetitive effect therefore required the plaintiffs to prove that the antisteering provisions increased the cost of transactions above a competitive level, or reduced the number of transactions. According to the majority, the plaintiffs failed to prove this, as no evidence was led showing that the price of credit-card transactions was higher than would be expected in a competitive market. On the contrary, while the provisions were in place, the credit-card market had experienced “expanding output and improved quality”.

Furthermore, the Supreme Court noted that Visa and MasterCard’s merchant fees had increased even at merchant locations where Amex was not accepted (and thus its antisteering provisions did not apply). This suggested that the cause of increased merchant fees was increased competition for cardholders, rather than Amex’s antisteering efforts.

The Court concluded by suggesting that the antisteering provisions actually had the positive effect of promoting interbrand competition, as they prevented retailers from ‘free riding’ on Amex’s investments by steering customers away at the point of sale.

In a strident and comprehensive dissenting opinion, Breyer J (joined by Ginsburg, Sotomayor and Kagan JJ) held that contrary to the opinion of the majority, the antisteering provisions did in fact have “seriously anticompetitive effects”. This was for a number of reasons, but the crux of his Honour’s reasoning on the market definition issue was that only the network services market was relevant for the purpose of the antitrust analysis, as the relevant provisions appear only in Amex’s contracts with merchants. In addition, Breyer J took issue with the majority’s definition of two-sided platforms, criticising it as overly broad.


The case is significant because it is the first time the US Supreme Court has considered two-sided platforms in detail. Unfortunately, the Court did not take this opportunity to provide a detailed definition of what will constitute a two-sided platform for antitrust purposes.

In relation to when it will be appropriate to include both sides of a two-sided platform when defining a market, the Court stated that it will not be necessary to do so “when the impacts of indirect network effects and relative pricing in that market are minor”. In future US cases concerning two-sided platforms, it appears that it will therefore be necessary for courts to determine the extent of indirect network effects within the platform before reaching a conclusion on market definition. This adds an additional layer of complexity to what is already a relatively intricate field of law.

While two-sided platforms have previously been considered in Australian case law (including by the High Court in ACCC v Flight Centre [2016] HCA 49 (concerning travel agents), and the Full Federal Court in ACCC v ANZ [2015] FCAFC 103 (concerning mortgage brokers)), Australian courts have not yet addressed the question of when (if ever) it will be appropriate to include both sides of a two-sided platform in the relevant market definition. It will be interesting to see if and how the Australian courts resolve this issue over the coming years.

Image credit: Flickr – Hamed Al-Raisi / CreativeCommons 2.0 / remixed to B&W and resized

Like this post? Share it... Email this to someone
Print this page
Share on LinkedIn
Share on Facebook
Tweet about this on Twitter

About The Author

is a Solicitor in the Melbourne office of King & Wood Mallesons.

Leave a Reply

Your email address will not be published.

3 × 3 =