Contact Us

Disruptive Competition Project

655 15th St., NW

Suite 410

Washington, D.C. 20005

Phone: (202) 783-0070
Fax: (202) 783-0534

Contact Us

Please fill out this form and we will get in touch with you shortly.

The Amex Decision: Cataloguing the Discussions, Debates, and Disagreements

The Supreme Court decision in Ohio v. American Express (“Amex”), referenced in my latest post on multi-sided markets, has spawned much discussion in antitrust circles. While there are those that fully support the entire decision (e.g., Gus Hurwitz and Kristian Stout), or modestly support the decision (e.g., Randy Picker), there are also critics that believe the ruling represents a major step backward for antitrust enforcement. This has caused a vigorous debate on antitrust panels, blogs, and even in Op-Eds. This debate is wide-ranging and hits on many practical and philosophical parts of antitrust policy. This post attempts to break down the discussion and disagreements into discrete arguments that can, perhaps, help newcomers understand the debate.

The Supreme Court’s Amex ruling can be broken down into three propositions:

  1. Multi-sided markets must be considered as part of the competition analysis in an antitrust case involving certain two-sided markets.
  2. Multi-sided markets are to be considered, as a practical matter, by plaintiffs defining a relevant market that describes all relevant sides of a multi-sided market.
  3. Case specific analysis of two-sided market effects. The Supreme Court heavily weighted increases in output and other benefits on the cardholder side, ultimately concluding it outweighed evidence of harm in the record.

These three propositions deal with, in order, how we first approach the issue, how plaintiffs procedurally must meet their initial burden, and how judges are to consider the facts and analysis presented. Using these three propositions, it becomes easier to categorize the complaints that critics have of the Amex decision. Additionally, there is one final bucket that we can put these criticisms in — that the Supreme Court misapplied multi-sided market theories and that the government should have still won its case against Amex even using a multi-sided model.

Proposition 1: Considering the impact of multi-sided markets

Proposition 1 is the least controversial proposition of the Supreme Court ruling. After all, the economics of two-sided markets are fairly well developed and accepted by economists. Adjusting competition enforcement due to new economic thinking is a part of a shared goal among virtually all antitrust advocates to reduce error costs in either under- or over-enforcement. Most would agree that banning free internet services due to a failure to account for two-sided market dynamics would be an enforcement error that would lead to consumer harm (see my last post for more on this). However, there is still some meaningful debate concerning whether this part of the ruling, that multi-sided markets should be considered as part of the competition analysis, is correct.

  • One debate concerns the line the Supreme Court drew between multi-sided markets that can be treated like single-sided markets and those that must be analyzed as multi-sided markets. The opinion states that when the indirect network effects operate in only one direction, as in the newspaper industry where readers don’t care about the amount of advertising, then the market can be treated as a one-sided market. However, when both sides are needed to simultaneously participate for a transaction to occur, like in the credit card industry, then both sides of the market must be considered. This guidance is not clear enough to determine which markets are in and which are out of this new framework. There is substantial discussion of who will or should be covered by the Amex ruling.
  • A criticism that has sparked further debate is whether multi-sided market analysis is even necessary as part of an antitrust case. This argument generally states that factors unique to multi-sided markets can still be addressed in a case without forcing an analysis of the multi-sided market. These critics argue that multi-sided econometric analysis is not worth the expense in resources in developing the case, and that we can achieve the same enforcement error avoidance using traditional single-market analysis. Although, it appears that these critics would still permit some evidence of multi-sided factors that explain particular conduct, so this criticism might not be entirely against a Proposition 1 that says multi-sidedness can be considered in some form as a defense under certain conditions.

Proposition 2: Performing a multi-sided relevant market analysis

The criticisms in this proposition largely revolve around whether the decision will substantially increase the burden of bringing an antitrust case, and therefore deter enforcement. It is important to note that these issues can be resolved without rejecting the premise that relevant multi-sided market factors must be taken into account in proper antitrust analysis.

  • One debate is over whose burden it is to develop multi-sided market facts. Courts use a burden shifting framework in rule of reason antitrust cases. Plaintiffs must clear an initial bar of proof, after which the defense has the opportunity not just to rebut the evidence presented, but additionally offer important context to their actions. The plaintiffs then have the opportunity to rebut the evidence put forth by the defense. In some cases, this context provided by defendants can prove that the challenged activity isn’t anticompetitive but is in fact procompetitive or competition neutral. The Supreme Court’s Amex ruling states that developing multi-sided market facts is part of the plaintiff’s initial burden, because it is a part of pleading a relevant market (see my post on market definition). Critics argue that it should instead be part of the defendant’s rebuttal to present relevant factors of multi-sided market dynamics in their industry.
  • A related criticism is that the Supreme Court does not provide a good roadmap on how a plaintiff would plead a multi-sided relevant market and how this kind of relevant market might impact later analysis like market share. This is important, because proving a multi-sided relevant market could be a large or a minimal additional cost for plaintiffs in building their initial case depending on how it is required to be done. Critics argue that if this creates a large additional cost then it will deter enforcement based purely on the resources required to bring a case (although some argue that getting it right is more important). This is an issue that will likely need to be resolved in the application of Amex by lower courts.
  • Another debate centers around the fact that the Supreme Court appears to be requiring plaintiffs to prove an output effect in addition to the price effect. Some critics believe that this is a new burden for plaintiffs that is nonsensical and will make cases very difficult to win in practice.
  • Finally, there is debate around a footnote that requires defining markets in all vertical restraints cases. The criticism here is based on a potential additional burden to some plaintiffs to plead their case. Relevant markets are generally used to provide evidence of market power through the use of market shares. If a plaintiff has direct evidence of market power, then there is an argument that the plaintiff need not spend the resources necessary to define a market. The Amex decision’s footnote 7 requires the plaintiff to plead a relevant market in all vertical restraint cases, even those in which the plaintiff could potentially prove market power directly.

Proposition 3: Weighing competitive effects in a multi-sided market

There are a few debates centering around the analysis the Supreme Court ultimately used to come to its decision that anti-steering provisions do not violate antitrust law under its multi-sided analytical approach.

  • One such debate focuses on the criticism that the decision favors one business model over another. The argument is that the decision allows one company to set the form competition will take, a cardholder benefits focused model, and makes it difficult to compete using other models, like a low merchant fee focused model.
  • Another debate is about what the decision says about offsets, or balancing factors. In some areas of antitrust enforcement, like price fixing, a defendant can’t use any offsetting benefits to another set of customers to justify their behavior. However, the Supreme Court’s analysis in Amex relies heavily on the offsetting effect of increasing output and cardholder benefits to reach its conclusion. There remains some questions as to what role offsets play in antitrust analysis and what role they should play.

The problem critics have with Amex might not be multi-sided markets

There is a large thread of pushback against the Amex decision that doesn’t seem related to the Supreme Court bringing multi-sided market analysis into competition law. Indeed, some believe that the Supreme Court misapplied multi-sided economics and ignored evidence that was still persuasive in a multi-sided framework (See Tim Wu’s Op-Ed).

  • One line of argument is that the court improperly focused on the high prices to merchants. High prices to one side of a multi-sided market are potentially explained as procompetitive by economic thinking. However, critics point to the anti-steering agreement as the actual problematic behavior, with higher fees just being a symptom. Critics believe that under a proper analysis it is clear that these agreements restricted an important model of competition that is still valid under multi-sided models, which is a strategy of winning merchant loyalty through lower fees. These critics also believe that such a restriction invited tacit collusion among the credit card companies (Visa and MasterCard were a part of the original complaint but settled with the government).
  • Another issue critics point to is the argument that the Supreme Court ignored important evidence that non-Amex consumers were harmed by the anti-steering agreement. Amex provided better cardholder benefits to their own customers by charging higher merchant fees. The district court found that merchants passed on these fees to all consumers, not just Amex customers.

The greater conversation

These debates are likely to continue and some consensus on how to improve the Amex decision could emerge. It is likely that some scholarship is needed to at the very least guide lower courts in their application of the Amex decision, especially in regards to when to apply the decision and how to define a multi-sided relevant market in a way that is both accurate and does not overly burden plaintiffs.

It is also important to note that there is a part of this debate that can’t be broken down granularly: the criticism that the decision as a whole amounts to a substantial weakening of enforcement, although this too is being debated. Some experts argue that the decision is actually very narrow and will not be widely adopted by lower courts. Additionally, economists Evans and Schmalensee authored a paper arguing that integrating two-sided market analysis in antitrust law will improve enforcement in a way that cuts both ways, sometimes strengthening plaintiffs’ cases while at other times bolstering the argument that no harm has occurred.

One enduring question for me is if the loss was due to the way the case was constructed or due to the facts. In other words, could the government have scored an easy victory if they only had the foresight to structure their case the way the Supreme Court wanted or would the Supreme Court have viewed the behavior as lawful no matter how the government presented their case. The EU may provide some help here, as they have already dealt with multi-sided market cases and the EU’s MasterCard decision shows that a government plaintiff can prevail under a multi-sided market analysis requirement.


Some, if not all of society’s most useful innovations are the byproduct of competition. In fact, although it may sound counterintuitive, innovation often flourishes when an incumbent is threatened by a new entrant because the threat of losing users to the competition drives product improvement. The Internet and the products and companies it has enabled are no exception; companies need to constantly stay on their toes, as the next startup is ready to knock them down with a better product.