[Arnab is a student at Hidayatullah National Law University, Raipur. The following post was awarded the fifth-best entry in the IRCCL Blog Writing Competition 2020.]
The names of Google, Amazon, Facebook and Apple (or GAFA) have become ubiquitous on the files of competition authorities all around the world. With a towering aggregate market capitalization of almost $3.6 trillion, exceeding India’s nominal GDP, GAFA has been accused of repeated violation of competition laws in multiple states. While grappling with these organizations for their adherence with antitrust provisions, various authorities made a greater discovery; the tools possessed by them are not adequate to properly assess these digital platforms.
Digital platforms or multi-sided markets can be construed as firms that act as a medium for seller/service providers to meet the consumers and transact business, in a manner prescribed by the platforms; in addition to this, they are governed by indirect network effects. The most common examples of such platforms are web browsers which cater to the content providers (the websites), consumers (people browsing the websites through browsers) and advertisers (who appeal to the consumers); and application stores which cater to the content providers (application developers) and consumers (people using the apps through the store). These platforms are different from the conventional one-sided markets and by initiating proceedings against the former, the authorities have opened a Pandora’s Box.
This article strives to set out various factors that affect digital platforms, provide reasons for relatively difficult assessment of these platforms and propose various steps to combat such difficulties, all made with a reference to Competition Act 2002 (Act).
Need For a New Market Definition
It may be argued that the term 'multi-sided markets' is a misnomer because all markets consist of buyers and sellers and are, hence, multi-sided. The absence of a widely accepted definition of multi-sided market adds to the difficulty of pigeonholing these platforms as multi-sided. However, these markets can be differentiated and defined with the help of some common characteristics.
It is not possible for a single user to derive benefit from a platform. For example, the presence of a single user in a social media platform or a mobile network is redundant as the user needs others to join the network to derive benefit from it. This is termed as the same-side network effects. Similarly, in a social media platform, the advertisers or content creators gain more benefit with a greater number of users on board. This is termed as cross-platform network effects (as the users and advertisers/content creators are present on different sides of the market). Once a large number of users join the social media platforms, more content providers would want to join these platforms which would again result in more users joining the platform thereby forming a domino effect and affecting the strength of the platform.
Collection and use of big data
As a platform grows with increasing network effects, its interaction with users grows. This leads to a large influx of data continuously over time, and the aggregation of this data is known as big data. For example, Google collects information from other websites that use its advertising services and employs the aggregated data for personalised advertisements. An entity that possesses greater data can be more beneficial for users on the supply side as personalised advertisements have a higher chance of conversion to a sale. With an increase of users on the supply side, network effects will increase the users on the consumer side. Therefore, in the presence of network effects, large amounts of data can translate to an advantage to the platform.
The first step to assess an antitrust claim is to define the relevant market, which can be done with the help of various attributes provided in Section 19(7) of the Act. In one-sided markets, the relevant product market can be defined on the basis of demand substitutability. While this method can be extended to multi-sided markets, one may face the dilemma of defining a single relevant market encompassing markets for different users or multiple relevant markets for different users in each side of the platform; this can be solved by ascertaining whether the market is non-transaction or transaction.
A search engine can be cited as an example of a non-transaction platform, where there is interaction (in form of advertisements or accessing the compiled websites), but an absence of a direct transaction (in form of consideration) between the users. Here, although the advertisers benefit by the number of web surfers, the presence of advertisers does not encourage the web surfers to join the platform; the products, search and advertisement do not complement each other and therefore in these conditions, two markets can be defined. This scenario had come to play in the case of Matrimony.com Limited v. Google, where the Competition Commission of India (CCI) defined separate markets for online web search services and online search advertising in India.
Similarly, an online intermediary providing for different kinds of accommodation, listed by the respective owners, to the consumers through interaction can be cited as an example of a transaction platform. The present case is distinguished from the former as network effects will be present on both sides of the market. It is, however, important to note that in transactional platforms one has to ascertain the presence of constraints regarding price-setting among the users. If a user on one side is able to pass through the cost to the other side, the two-sidedness of the platform decreases. This theory was applied in the case of Booking.com, where the platform provided for booking along with price comparison and searching for accommodation. It was held that while booking accommodation was the primary service, the search and compare services were included to modify the platform through network effects; therefore, the services were defined as a single market: search, compare and book.
In a nutshell, a non-transaction market has separate market definitions while a transaction market generally has a single market definition. Once the type of the multi-sided market is ascertained, one can employ a mixed use of quantitative tests along with qualitative tests like surveys and questionnaires. This would help to gauge the extent of indirect network effects and its importance for consumer choices, providing a clearer picture.
Market Power for Multisided Platforms
Section 4 of the Act defines dominance as the ability to function in the market independently of prevailing competitive forces in the relevant market. Dominance is measured by the CCI through a non-exhaustive list of factors provided in Section 19(4) of the Act, which includes ascertaining market share, entry barriers, consumer dependence, and the structure of the market among others. Market share is a quantitative method to measure market power and one has to be cautious while extending this method to assess digital platforms. If the market share in only one side (Side A) of the platform is measured disregarding the other side (Side B), the platform may try to concentrate users on the other side (Side B) of the platform; network effects may come into play, eventually strengthening Side A of the market, and may make it difficult to offset these effects after a certain extent. Additionally, it is necessary to identify whether users can multi-home (use multiple platforms simultaneously). For example, if a distinct set of people use a specific food service aggregator without multi-homing, then the restaurants would find the platform more valuable; thus increasing its market power. In the case of Meru Travel Solution (Private) Limited v. ANI Technologies (Private) Limited, the CCI recognised these factors and took into account the market share of both the fleet (drivers) and the number of trips (consumer) over a period of two years along with taking note of the phenomena of multi-homing by the drivers and consumers.
In the above case, the CCI tried to plug in as many gaps as possible regarding an assessment of digital platforms for dominance. But, it placed an over-reliance on market share as a measure of market power and did not factor for entry barriers which is an extremely helpful tool to assess dominance. The General Court of the European Union in the case of Cisco System Inc. regarding internet-based communication service, which was a multi-sided and a nascent market like the cab aggregator, had stated that market share might not be the true measure of market power in a dynamic competition. Dynamic competition can be described as competition relying primarily on innovation and taking into account the effects of actions of incumbent firms on potential competitors. Barney (1991) laid down four attributes to assist identify entry barriers and in turn market power, they are:
Valuable resources: They should help the firm implement a strategy that could exploit the present opportunity and neutralize threats in the environment.
Rare resources: They should not be possessed by others so as to help them implement similar strategies.
Imperfectly imitable resources: A firm having rare and valuable resources can only sustain dominance is the resource is not imitable. The resource may be socially complex or dependent on unique historical conditions.
Substitutability: The degree of substitutability of the firm’s resources should be less.
The term 'resource' signifies a mix of human, physical and organisational capital; the scope of these attributes can be widened enough to accommodate contemporary issues like big data. As digital platforms thrive on innovation, an assessment compatible with dynamic competition should be given more importance to assess issues related to it instead of turning to the conventional static competition.
The woes of applying the antitrust provision to digital platforms do not subside after defining the relevant market and ascertaining the market power of an entity. The complexity of identifying violations committed in multi-sided markets are relatively more complex than those in single-sided markets. This assertion can be supported by the extensive market study undertaken by the CCI on e-commerce platforms, UK’s Competition and Market Authority report on digital competition, the European Union’s report Competition Policy for Digital Era to name a few. Some of the conflicts may be enumerated in the succeeding paragraphs.
The use of big data
The importance of big data for platforms can be measured by the fact that Walmart uses a supercomputer to process the massive amounts of data in a short time, helping them to make better decisions and more accurate predictions of the market conditions. The antitrust hearing conducted by the Congress in the US also highlighted complaints against Amazon that the platform uses data from third-party sellers to develop competing products, which often places the third-party sellers at a disadvantage. The importance of data was also a bone of contention in the merger between Microsoft and LinkedIn in the EU and the upcoming merger between Google and Fitbit. A simple solution would be to label data as an ‘essential facility,’ which can be defined as a resource without which an entity cannot provide service to the consumer, propounded in the case of Samsher Kataria. However, the Personal Data Protection Bill 2019, construed as the only safeguard for data privacy in India, is still in an inchoate stage and without proper safeguards, the implementation of the essential facility doctrine would be disastrous.
A recent trend of killer acquisition has been observed by competition authorities, where the acquiring firm discontinues the innovative strategies of the acquired firm in order to pre-empt competition. The acquisition of Whatsapp and Instagram by Facebook can also be identified as killer acquisitions, where Facebook acquired these firms in related markets, though continuing with their development, to pre-empt competition. In India, Section 5 of the Act deals with the threshold that defines which acquisition, merger or amalgamation should be notified as a combination. While Section 6 prevents any such acquisition that might have anticompetitive effects, most of the startups have a low turnover and their acquisition is not notified for review. Between 2008 and 2018, Google, Amazon and Facebook made 299 acquisitions but few of them cleared the first stage of the review due to their low turnover. The CCI may also suffer from this incapacity to take actions against killer acquisition as it follows a similar model, based on turnover, for notification of such acquisitions.
Platform competing with third-party sellers
Third party sellers can be at a significant disadvantage if the platform turns into a competitor as the former would have to compete with the facilitator itself masquerading as a rival firm. Section 4(2)(a)(i) of the Act prevents the platform from imposing unfair conditions or implementing a regime of self-preference.
The CCI held Google liable under this provision when it tried to distort the search rankings, implement self preferencing with regard to the search for commercial flights section and forbid its partners from using competing search engines. However, the authorities should be sensitive while providing relief to this kind of violation. In the case regarding Microsoft Windows Media Player (WMP), the EU had found Microsoft guilty of tying its media player to the operating software for personal computers. It was dominant in the latter market and was held guilty of trying to extend its dominance to the market of media players. This can be construed as a prima facie disadvantage to firms trying to provide media player software for personal computers. The EU, therefore, ordered Microsoft to bifurcate its products to an operating system with WMP and an operating system without WMP. However, when the different versions hit the market, users preferred the complete version of the operating system while the alternate version could put up only 0.005% of the total sales. While the remedy failed, one must wonder whether WMP is dominant in the present market; but, with the innovation of streaming platforms, which constitutes 80% of the revenue of the music industry, the viability of WMP comes into question. This episode posits a conundrum, whether the authorities should let the market forces decide the fate of such cases.
The US government has made ripples with its antitrust complaint against Google where it seeks to break up the internet giant. This directs our attention toward yesteryear’s famed breakup of AT&T due to its towering dominance and abuse thereof. Although the remedy was dubbed as successful, many scholars including the present commissioner of the FTC expressed their reservations against using structural remedies like divestiture. The reservations were mostly due to availability of alternates and uncertainty of the remedy.
Instead of providing structural remedies, authorities should focus more on remedies relating to specific cases. In cases related to network effects, efforts should be made to enable interoperability in order to facilitate multi-homing and prevent the concentration of users on a single platform. A time limit should be imposed on hearing cases on multisided platforms, as delay can cause a significant shift in dominance due to network effects. Such time limit can be complemented by the usage of temporary injunction as used under Order 39 of Code of Civil Procedure 1908, which can be accommodated as a step after the Director General submits his report under Section 26 of the Act. To prevent misuse of this power, injunctions should be used sparingly and according to the conditions laid down in Dalpat Kumar v. Pralhad Singh [AIR 1993 SC 276]. Platforms competing with third-party sellers should be governed by a strict ex post evaluation, so that innovation is not stifled and strong privacy laws should be implemented to prevent abuse of big data by firms. However, mergers should be governed by an ex ante evaluation so as to prevent improper speculations and unnecessary structural remedies afterwards. For non-notifiable mergers, an additional threshold relating to the transactional value of the acquisition can be introduced for better assessment. Lastly, market studies and surveys can be undertaken to provide a qualitative view as to the trends occurring in the market.
The CCI has not been oblivious to the developments taking place in the area of digital platforms as is evident by the market study undertaken by it. The judgements of CCI on multi-sided platforms also encompass a broader perspective to accommodate the economics related to multi-sided markets. But, with novel issues related to such markets springing up, CCI along with other competition authorities would have to span a long journey before grasping the intricacies of such markets.