Ecosystems
The 2023 Booking/eTraveli merger prohibition by the European Commission (EC) was the first EC merger prohibition concerning digital platforms. Booking offered primarily a hotel online travel agency (OTA), amongst other smaller offerings in flights and car rentals, while eTraveli offered a flight OTA. The concept of an ‘ecosystem’ that was used in the Commission’s theory of harm was unique. Though the concept of an ecosystem has been used in other cases, also outside of merger control, the term still lacks an accepted universal definition. Also unclear is the extent to which such a theory of harm differs from conventional theories of harm, what the standards for evidence under this theory of harm are, and how such theories of harm would affect enforcement.
In this context, we recently published an article defining ecosystems, and potential theories of harm concerning ecosystems. In this blog post we investigate how these theories might change the enforcement of merger control.
We suggest that an ecosystem be defined as “a group of (possibly independent) components, with a governance system jointly managing network effects, scope economies, scale economies, or learning effects, while significantly exploiting complementarities between components.” Further, we define two avenues for potential enforcement under merger control (conventionally known as theories of harm), tailored to markets with ecosystems.
This definition shows how ecosystems combine features of platforms and conglomerates (or corporations active in several industries). Concerns around platforms, stemming from network effects, scope economies, scale economies, and learning effects, are combined with conglomerate concerns stemming from complementarities and foreclosure (such as through tying unrelated products).
A unilateral effects theory of harm
First, we suggest a unilateral effects theory of harm for ecosystems, where a merger can weaken competitive constraints and cause consumer harm. Competitive constraints on an ecosystem are not only exerted on the individual components within their own product market, but also on the ecosystem as a whole. Ecosystems compete by offering complementarities between components, as opposed to components themselves. Competitive constraints can be exerted on the ecosystem by another ecosystem consisting of different components (and different relevant markets) that offer substitutable complementarities.
For example, two travel ecosystems need not consist of substitutable components to offer substitutable complementarities. Consumers may not consider flights and trains to be substitutable, but might substitute flights and hotels with trains, taxis, and flexible homestays. Thus, a train service can exert a competitive constraint on a flight service at the ecosystem layer. Through acquiring the train service, the flight service can internalise this competitive pressure, allowing it to potentially raise prices, or act unilaterally at the detriment of consumers.
This suggests that competition on markets with ecosystems occurs in two layers: competition at the product market layer between substitutes, and competition at the broader ecosystem layer between substitutable complementarities. Within this context, a merger might allow an acquiring ecosystem to potentially remove an important competitive constraint on the ecosystem at the ecosystem layer and can therefore cause significant harm to competition.
An entrenchment theory of harm
Second, we suggest an entrenchment theory of harm. Entrenchment allows a dominant undertaking to expand its dominance, make it more difficult for rivals to enter and compete, and in such a way lower the contestability of a particular market. This follows from “the dominance of the ecosystem, in part because the new services [added to the ecosystem by a merger] add value to the consumers for which they are complements and in part because they help retain other users for which they are partial substitutes.” This can be achieved through pro-competitive means, such as by innovating more efficient methods, or by anti-competitive means, such as by exploiting brand identity.
Entrenchment can be defensive, where a merger is with a complement that itself is not a threat but enables the acquirer to reduce the contestability of its market position, such as the Booking/eTraveli merger. Alternatively, it may be offensive, where the target is not on the market but is building capabilities, such as the Adobe/Figma merger. Both defensive and offensive entrenchment are more likely on markets with ecosystems. Complementarities increase the market power in the ecosystem layer because of an acquisition (based on increased capabilities post-merger), facilitating defensive entrenchment. The ecosystem layer out-of-market constraints facilitate offensive entrenchment. Entrenchment can be the result of efficiencies, or of mechanisms that deviate from “competition on the merits.” When entrenchment is the result of “competition on the merits”, it can lead to both harmful effects and beneficial effects, in the short-, medium-, and long-term. This implies that there is a trade-off between the gains and harms to competition.
Regarding entrenchment, there is the open question of the distribution of harms and benefits across different markets within an ecosystem because of a merger. For example, harm as a result of an ecosystem merger tends to be concentrated among consumers who prefer to mix-and-match. There are parallels here with the treatment of sustainability agreements under Article 101, where collective benefits may accrue outside relevant markets. Under this standard, benefits can be relevant to cancelling out harm when there is substantial overlap in consumers between the markets where harm is concentrated, and the markets where the benefits accrue.
Enforcement benchmarks
For these theories of harm to be enforceable in practice, they must be falsifiable. This implies the need for benchmarks.
A benchmark may refer to a broader replicability standard than that applied by the Commission in Booking/eTraveli. In the case, the Commission considered that the positions of Booking and eTraveli in their relevant markets could not be replicated by competitors. Given the role of complementarities across relevant markets in ecosystems, replicability at the ecosystem layer may also be relevant. Such a replicability standard would concern the replicability of the capabilities and complementarities present at the ecosystem layer.
Features of the design of the ecosystem could be used to establish potential aggravating factors, the presence of which implies that a merger is more likely to require investigation. For example, controlling access points can increase the likelihood of anti-competitive effects. Furthermore, the presence and prevalence of practices like the use of dark patterns and hypernudging can increase the potential of anti-competitive effects. Transactions that show these characteristics are thus indicators of higher risk and therefore need for investigation.
Finally, in cases where a definitive assessment is impossible or undesirable, it may be beneficial to enforce general principles of openness, such as interoperability, data sharing, and open standards. This principle has legislative precedent, having been applied in the Digital Markets Act, from which inspiration could be taken. Such an approach reduces the role of ‘competition on the merits’, which has limitations in the context of ecosystems, given the ambiguous effects of complementarities on consumer welfare. It further emphasises protecting an open and dynamic competitive process and preventing lock-in and entrenchment from occurring in the first place.
These benchmarks, amongst others, can provide the means for authorities and merging parties to falsify and thus effectively apply both theories of harm. The following section looks at how this can affect enforcement.
Enforcement
The aim with developing clear theories of harm for ecosystem mergers is to provide competition authorities with effective frameworks with which to investigate mergers involving an ecosystem, while also providing merging parties with legal certainty and a means of justification. Based on this, how might enforcement change in such cases?
These theories of harm and limiting principles would hopefully allow for effective intervention in complex edge cases involving ecosystems. Rather than having to form a new theory for each merger case involving an ecosystem, or combining horizontal, vertical, and conglomerate theories of harm, authorities will be able to apply one of the ecosystem theories of harm. For example, authorities would no longer need to use abstract out-of-market effects but could use ecosystem layer effects. Additionally, the theories of harm offer indications to authorities for which mergers might have ecosystem layer effects. This would reduce the risks of false negatives. Alternatively, having clear definitions of the concepts and the theories of harm would provide merging parties with legal certainty. The enforcement benchmarks would ensure that merging parties can show pro-competitive, or at least competitively neutral, effects of ecosystem mergers. This would reduce the risks of false postitives.
Conclusion
With our article and this blog post, we hope to contribute to the understanding of the concept of an ecosystem, and to the development of theories of harm that represent the unique features of markets with ecosystems. For these theories of harm to be enforceable and falsifiable, it is necessary to place certain limiting principles that shape the enforcement effort. With appropriate benchmarks, such as ecosystem layer replicability, ecosystem design and access points, and general principles of openness, we suggest that these two theories of harm offer an enforceable means for competition authorities to protect against harm to competition on markets with ecosystems, while also providing clarity and legal certainty to merging parties. These theories of harm could improve the enforcement of merger control in cases involving an ecosystem, by indicating to authorities which mergers might have harmful effects, and by clarifying the concepts and the concerns behind the theories. They would also, hopefully, provide clarity to merging parties as to the investigation and the falsifiability of the theories, based on benchmarks.
