This paper, available here, argues for more vigorous antitrust enforcement against Most Favoured Nation (MFN) provisions in the platform context.

A MFN clause requires providers to refrain from offering their products or services at lower prices on other platforms. During the past two decades, antitrust enforcement against MFN provisions has grown, particularly in Europe. In contrast, there have been almost no enforcement actions against platform MFNs in the United States. The authors make a number of proposals to reverse this trend.


The article is structured as follows:

Part I shows how platform MFNs can harm competition and consumers, despite their potential competitive benefits.

The authors’ draw on the economics’ literature on the effects of MFNs generally, and platform MFNs in particular. Simple MFNs commit sellers not to discount selectively, which assures covered buyers that they will be charged the lowest price offered by the seller. At first blush, one might expect this provision to lead to lower prices for covered buyers. However, as explained throughout the economics literature, there are compelling theoretical reasons to expect equilibrium prices to rise due to MFNs, and empirical evidence supports this prediction. The authors conclude that platform MFNs generally harm competition, except in narrow circumstances in which freeriding concerns are especially strong.

Platform MFNs differ from simple MFNs. Platform MFNs are agreements between sellers and platforms about the prices that sellers will charge buyers who purchase through rival platforms. They are thus not agreements between sellers and buyers about the prices that sellers will charge other buyers, as with traditional MFNs. The two types of MFNs nonetheless raise similar competitive concerns.

Some anticompetitive problems created by MFNs are “collusive”— MFNs weaken price competition. The term “collusive” includes both coordinated conduct and unilateral accommodating conduct that softens competition. A MFN creates a strong financial incentive for the seller not to offer low prices because any discount must be offered to all covered buyers. That penalty makes discounts offered to buyers expensive. By making it costly for firms to offer selective (and, in some cases, confidential) discounts to their customers, a MFN may reduce those discounts, soften price competition, and lead to higher prices. A MFN can alternatively or additionally facilitate coordination, including tacit collusion, and thus lead to higher prices. Lastly, a MFN may create an exclusionary anticompetitive problem since it can raise the costs of current or potential competitors by negotiating lower prices from suppliers of critical inputs.

Although the economics literature indicates that platform MFNs often harm competition, platform MFNs can generate efficiencies under the right circumstances. Online platform MFNs in particular have been justified as protecting investment incentives by preventing freeriding (or “showrooming”). While the balance between harms and efficiencies may vary across markets, the authors are sceptical that this justification will routinely prevail. Freeriding can occur when a match would occur in a platform, but the consumer is able to contract directly with the other side of the platform (e.g. book a hotel room directly on the hotel’s website), thereby undermining the platform’s business model and incentive to provide matches.

However, there are two reasons to question whether freeriding would be so substantial as to make full-service OTAs unprofitable and induce their exit. The first is consumer transaction costs: some consumers will not be willing to identify hotels on one site (the full-service online travel agent), and then find and book them on other sites (a no-frills online travel agent or hotel site). Second, the hotels may need the services that the online travel agent provides, but find it costly to provide those services themselves. If the hotels pay the online travel agent for its services, consumers that search the online travel agent but purchase on the hotel site would not be freeriding. These possible limitations on the costs imposed by freeriding – combined with the incentives to increase prices and the prevent entry caused by MFNs – make the economic analysis of a platform market critical. In this respect, the leading analyses find that platform MFNs lead to higher platform fees, drive up retail prices and discourage entry by firms with lower-cost business models.

Part II reviews how and why platform MFNs have been treated differently in the U.S. and in Europe.

In Europe, National Competition Authorities (NCAs) and the European Commission have brought cases against platforms such as Amazon Marketplace, and travel booking sites Booking.com and Expedia. Cases challenging anticompetitive platform MFNs are less common in the United States. U.S. courts have contended with the potential anticompetitive effects of MFNs for decades, although not in the context of online platforms. Before the 1980s, the major U.S. decisions grappled primarily with the problem of whether one should infer an agreement to fix prices when competitors used MFNs unilaterally but in parallel, thereby facilitating coordination. More recent cases have focused on MFNs that facilitate higher prices by both preventing rival discounting and discouraging entry. The recent Apple e-books litigation presents the primary U.S. example of antitrust enforcement against a platform MFN. Platform MFNs were also the subject of two other recent U.S. cases: the American Express case (which went to the Supreme Court, in a judgment I reviewed here); and in a 2014 private challenge to price-fixing of hotel room rates by online travel agents which outcome did not ultimately address the lawfulness of MFN clauses.

The platform MFN cases brought by NCAs in Europe appear to be appropriate enforcement actions, with sound theoretical and empirical bases. It is hard to explain on economic grounds why such cases have not been brought in the United States. The enforcement discrepancy is likely due to the relatively greater interest that European enforcers have shown in pursuing exclusionary conduct, which rules are more contested in the United States than in Europe. This difference may be exacerbated by the way resources are allocated at U.S. antitrust agencies, which prioritise merger cases.

Finally, Part III argues that U.S. antitrust enforcers should follow the lead of their European colleagues in investigating platform MFNs. It also explains how a case against platform MFNs could be structured to fit within current U.S. law.

In the most straightforward platform exclusion case, the platform imposing the MFN on vendors is the dominant online retailer of a product or service. The MFN protects the platform’s market power from erosion by discouraging entrants who would compete on price. The entrant might, in the absence of a MFN clause, charge a lower commission than the dominant company to vendors that agree to pass through their savings by lowering the price consumers would pay. A monopolisation claim requires two elements to be established: (i) the platform has market power; (ii) the dominant online platform engaged in exclusionary conduct to achieve or maintain its monopoly power. In the MFN context, this is straightforward. A plaintiff would prove its case by showing that the dominant platform prevented a rival or entrant from cutting fees to sellers willing to lower prices for customers.

Defendants will most likely attempt to cast their MFNs as protecting platform investment incentives by eliminating freeriding. A defendant might seek to prove this assertion through evidence that, before introducing the MFN, it had been considering limiting its investments in platform improvements due to the freeriding threat. A plaintiff can rebut this argument with empirical evidence of its own, such as studies suggesting that even narrow MFNs are unnecessary to protect investment incentives by online travel agents. A plaintiff might further support a rebuttal with evidence that the consumer transaction costs of freeriding are high, or that the dominant platform charges a fee sufficient to compensate it for the services it provides to freeriders.

Covered platforms may collectively represent a sufficient share of the market to create anticompetitive effects if they adopt MFNs, even if all individual platforms are relatively small. However, it may be challenging to prove market power absent large market shares. Accordingly, enforcement against platform MFNs that exclude platform rivals may have to be pursued under Section 1 of the Sherman Act, which does not require proof of monopoly power. Section 1 requires the investigated parties to have entered into agreements in restraint of trade. This requirement should be satisfied so long as the MFNs are included in (vertical) contracts between platforms and sellers. If multiple platforms use MFNs but none is dominant, the practice could nonetheless exclude the entry of rival platforms that would compete over price, as long as the sum of the market shares of the covered platforms is sufficiently large to ensure that the MFNs would discourage sellers from marketing their products on those platforms.

Furthermore, platform MFNs can also harm competition through two coordinated mechanisms: (1) discouraging cheating on consensus terms among competing platforms, and (2) inducing parallel accommodating conduct that leads rival platforms to charge higher prices.



This paper is a very good introduction to the economics underpinning competition concerns with MFNs clauses, and to cases brought on both sides of the Atlantic. A slight surprise for me was that the paper makes a case for MFN enforcement in the context of platforms, but does not discuss how that enforcement might be different in this context when compared to traditional one-sided markets. This may be asking too much – the paper already covers a lot of ground – and is in any event the topic of the paper below.

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