Defining Relevant Product Market for Multi-sided Platforms

 

Our life nowadays revolves around many platforms. We interact with our acquaintances on social media platforms like Facebook, shop different merchandise from Amazon/Flipkart, order food from our favorite restaurants through Swiggy/Zomato, and book our hotels or flights from MakeMyTrip/Oyo, taxis from Uber/Ola, and there are many such examples. One of the unique characteristics of these platforms is that they are multi-sided, e.g., all of these platforms have two or more distinct user groups.

A multisided-sided platform (MSP) is one in which 1) two or more sets of agents interact through a mediator or platform, and 2) each set of agents' actions impact the outcomes of the other set of agents, generally through an externality (Rysman, 2009). Like in the case of a typical e-commerce side, an increase in the number of sellers will attract more buyers as they will get more options to choose from, and as the number of buyers increases, more sellers will come to the e-commerce platforms. This same equation is valid for advertisers and consumers on the same platform. Thus, such platforms create what economists refer to as "reciprocal positive externality between two distinct groups" (Bhattacharjea, 2018).

The advent of multi-sided markets has eliminated or lowered transaction costs in conventional marketplaces that they have tried to organize and frequently replace. On the one hand, they have frequently assisted in overcoming market inefficiencies without the need for govt interference, rendering many of the early grounds for regulatory interventions obsolete and unviable (Krishnan & Unni, 2019). On the other hand, some of them have been subjected to several antitrust cases around the world, and others are under the scanner. Competition Commission of India (CCI), in the last 3-4 years, investigated more than 30 cases concerning MSPs. Recently, the CCI concluded a market study (CCI, 2020) on the e-commerce platforms and raised concerns about various anti-competitive practices.

Most antitrust cases concerning digital platform markets involve allegations of abuse of dominant position. For Competition Authorities (C.A.s) to study any such cases, determining the relevant market is crucial. Market definition of multi-sided platforms is not the same as in the case of traditional one-sided businesses. The initial question that arises is how many markets to define. As competition analysis is empirical, the approach to market definition in cases of MSPs has been different in different jurisdictions. This article wants to analyze the approach of CCI in this regard.

The Rationale for the topic as an entry point

With the growing popularity of e-commerce, it is critical to resolving the issue of relevant market definition, primarily between e-commerce platforms and traditional platforms. The CCI must be well-equipped to carry out the findings of competition breaches. The relevant product market must be carefully defined for the CCI to adequately define and assess the borders within which the anti-competitive behavior occurred. For that to happen, it is crucial to understand the difficulties of assessing the definition of the relevant market in cases involving MSPs and examine if the existing tools and criteria are enough to regulate the anti-competitive behavior.

The Research Questions

1.     What are the distinct characteristics that distinguish MSPs from general one-sided markets?

2.     Are the existing tools for relevant market determination enough for the MSPs?

3.     What is a trend in CCI's approach to relevant market definition in cases pertaining to MSPs?

To answer these questions, I will conduct a brief literature review to understand the basic economics of the MSPs and then analyze the CCI's approach to defining relevant markets in some cases pertaining to MSPs in India.

Brief Literature Survey

Despite their economic relevance, multi-sided platform markets have only recently received attention from economists in the scholarly literature on antitrust. The older literature mainly focused on two-sided platform markets: traditional newspapers or payment cards industry and operating system antitrust cases involving Microsoft in the 1990s. Viscusi (2018, pp. 353-354) explained how network effects helped Microsoft create the near monopolist position in the operating systems market. The author explained the network effect with the below curve, where “there are two types of consumers, denoted L and H, and type H consumers value the good more. When a total of Q consumers use the product, V.H. (Q) denotes the value that a type H consumer attaches to the product, and V.L. (Q) denotes the value of a type L consumer.  As the product has network effects, V.H. (Q) and V.L. (Q) are both increasing with Q.”

 Figure 1- Network externalities in two-sided markets Source: (Viscusi, Harrington, & Sappington, 2018)

In the OECD report (2018) titled “Rethinking Antitrust Tools for Multi-Sided Platforms," the difficulties in defining relevant markets are elaborated in the international context. The report explained whether to define one or more than one market in antitrust cases involving MSPs and the constraints of traditional market-defining tools like the SSNIP.

The market study by the CCI (2020) on e-commerce is beneficial in understanding the stakeholder's view of the market dynamics in various MSPs in India.

Defining Relevant Market

The relevant market is the filter that defines the region of commerce within which competition authorities examine a firm's behavior. Market determination aims to find items that put a firm's products under competitive pressure and are interchangeable to the point where the businesses producing them may be considered rivals.

According to Section 2(r) of the Competition Act, 2002, “relevant market means the market which may be determined by the Commission concerning the relevant product market or the relevant geographic market or with reference to both the markets." The first step in determining dominance is to define the relevant product and geographic market. Sections 19(6) and 19(7) of the Act establish the parameters for establishing the relevant geographic and product markets. The Competition Act, 2002, like other competition laws across the world, focuses on "substitutability" as a criterion for determining the relevant market.

The "Small but Significant, Non-Transitory Increase in Price" test, sometimes known as the SSNIP test, is an essential technique for establishing substitutability. SSNIP assesses whether consumers of a given product will switch to another product in exchange for a minor but considerable price increase (roughly 5% to 10%).   If this is the case, the two items might be in the same market. This test is often referred to as the "Hypothetical Monopolist Test" since it determines whether "a relevant market is worth monopolizing" (Sharma, 2011).

A narrow market determination that includes fewer substitute goods increases the likelihood that a firm will be deemed dominant. Conversely, a broad market determination may include many substitutes, leading Competition Authorities (C.A.) to overlook instances of anti-competitive conduct.

This essay focuses on determining appropriate product markets in scenarios involving MSPs. While digital platform characteristics such as worldwide/national reach complicate defining local markets, there appears to be some uniformity in how C.A.s, notably the CCI, have tackled the matter (Koul & Prasad, 2020). In contrast, there is a substantial dispute and lack of standardization in defining important product markets.

Difficulties in determining Relevant Product Market for Multi-sided Platforms

Multi-sided marketplaces allow interactions between two or more user groups that produce value for at least one of the user groups. (Rysman, 2009) One distinguishing trait is that the user groups, while diverse, are frequently interconnected (Hagiu & Wright, 2015). Their interconnectedness originates from network effects, which occur when the value obtained by each user from a product grows as the number of individuals who use that product increases.

Network effects can be of two types:

1.    1. Direct network effects arise when the total number of users increases, and the product becomes more valuable to each user. For example, social media networks such as Facebook are more valuable to users if a bigger proportion of their contacts also use the platform.

2. When the value obtained by one customer group from a product or service is dependent on the presence of another, this is referred to as an indirect network effect. A cab aggregation platform is a classic example: the two sides of the market are consumers wishing to book taxis and drivers who want to give their services. The value of a cab aggregator to a user is determined by the number of drivers available, which leads to shorter wait times, reduced costs, and so on. Drivers will place a higher value on a platform with more users. As a result, the utility or value generated by either side of the market is dependent on the other.

Conventional competition analysis has seen a platform's network size as a major component of its value and market power. 

One of the first instances involving two-sided markets in India was the MCX Vs. NSE case, in which the CCI focused on the idea of network effects in its minority order (MCX Stock Exchange Ltd vs. National Stock Exchange of India Ltd., 2009). The CCI recognized that network industries vary from traditional markets to rely on network effects. However, as seen above, indirect network effects in multi-sided marketplaces tend to build ties between distinct user groups, influencing which products and services customers perceive as replacements. This directly influences demand-side substitutability, which is the benchmark used to establish the relevant market. As a result, analyzing indirect network impacts is becoming increasingly crucial in determining relevant markets.

Single market vs. Multiple markets: Platform Typology Approach

Focusing on platform nature or typology is recommended to facilitate the evaluation of how many markets should be defined. The OECD recommends distinguishing between transactional and non-transactional platforms (OECD, 2018).

Payment card systems, e-commerce marketplaces, and travel aggregators are examples of transactional platforms where indirect network effects flow from both sides. 

Non-transactional platforms, on the other hand, permit interactions between user groups that are not transactions. As a result, they display solely unilateral cross-side network effects. For example, increasing the number of users on a social media site will almost certainly raise its worth to advertisers. However, growth in the number of marketers does not necessarily harm the value gained by users.

Because of bidirectional indirect network effects, both sides of transactional platforms should be viewed as a single relevant market. Because distinct economic considerations and alternatives may exist for the multiple sides in non-transactional platforms, each side should represent a separate market (OECD, 2018).

Substitutability Determination:

a transition from Quantitative to Qualitative Analysis

 

Applying the SSNIP test to multi-sided marketplaces is complicated. The original SSNIP test ignores the influence of indirect network effects on pricing mechanisms. A price rise on one side of a multi-sided market might affect demand on the other. A raise in the commission rate charged to drivers by a cab aggregator, for example, may limit their motivation to give their services. Users would migrate to platforms with a broader roster if there were fewer drivers. Fewer consumers would reduce the number of drivers ready to provide their services on the site even further. As a result, a hypothetical monopolist must account for the impact of a big, non-transitory price rise on both sides of the market. OECD suggests that to arrive at an accurate determination in multi-sided marketplaces, the SSNIP test should be applied to both sides of the market (OECD, 2018).

Adding more issues, some of the platforms offer services to one side of the user at zero cost, like Facebook, and charge from the other side of the user. So, applying the SSNIP test to the users of Facebook would be a failure because a 5-10% increase in prices will still make the price for the user a zero. So, the SSNIP test should be applied to both sides of the users.

The CCI’s Approach to Multi-sided Platform Markets

Single market vs. multimarket approach

The CCI does not account for the interdependence between numerous user groups generated by indirect network effects in most situations involving multi-sided marketplaces. As a result, it has formed several unique and autonomous marketplaces while ignoring linkages between the various sides supplied by digital platforms. CCI often based its determination on which side had made the complaint in question. In FHRAI vs. MakeMyTrip case, it said, “Since the allegations in the present case are primarily concerning the hoteliers, the relevant market analysis needs to be carried out from the perspective of hoteliers” (Federation of Hotel & Restaurant Associations of India (FHRAI) and Ors. v. MakeMyTrip India Pvt. Ltd. and Ors., 2019). CCI overlooks the competitive restrictions imposed by either party in this scenario. The limits put on a site like MakeMyTrip by customers' behaviors against hoteliers were not thoroughly examined.

However, several of the CCI's recent rulings show a growing appreciation of multi-sided marketplaces and the impact of network effects. The relevant market was described in Lifestyle Equities CV Vs. Amazon as the "market for services supplied by online platforms for selling items in India. (Lifestyle Equities C.V. and Ors. v. Amazon Sellers Service Private Limited & Anr., 2020)." This is significant because it differentiates online platforms, which are often multi-sided markets, from other types of online shopping. It also expressly stated that network effects define multi-sided marketplaces and must be considered.

Assessment of Substitutability

The CCI was tasked in Harshita Chawla v WhatsApp to examine whether the 'bundling' of WhatsApp Pay, a UPI-enabled payments system, with WhatsApp amounted to an abuse of power. The CCI designated the primary relevant market as the “market for Over-The-Top (OTT) messaging applications using smartphones in India” to gauge dominance. Also, noting in its analysis that communication apps may be split into numerous distinct categories based on their unique features and functions. It opted not to depend too heavily on the unique properties of communication applications, instead of focusing on the “primary or most dominant feature of the app to categorize it into a particular relevant market” (Harshita Chawla v. Whatsapp and Facebook, 2020).

According to the CCI, in the case of Ashish Ahuja v. Snapdeal and others, the buyer prefers to examine discounts and shopping experience in both the online and offline markets before making the ultimate choice to purchase the product. A price rise in one sector may cause the consumer to migrate to another section. As a result, CCI concluded that physical and online markets are not distinct relevant markets since they differ primarily in their distribution channels (Ashish Ahuja v. Snapdeal and others, 2014). This may not be true because the customer may examine other platforms as substitutes but not offline marketplaces. So, in this case, a broader definition of the relevant market, that is, the online and offline both, does not assess the dominance of Snapdeal correctly.

In Meru vs. Uber case, the CCI separated radio taxis from other modes of transportation such as rickshaws and other private taxis based on capabilities such as time savings, predictability, and round-the-clock availability. Despite these operational differences, it was determined that customers in Kolkata significantly relied on yellow cabs for their everyday transportation needs, and so the relevant market should comprise the services provided by radio taxis and yellow cabs (Meru Travel Solutions Private Limited vs. Uber India Systems Private Limited and Ors, 2015). This illustrates a sophisticated approach: the CCI determined the relevant market based on customer behavior rather than characteristics or functioning.

A similar approach was used to approve the merger of Facebook and Jaadhu Holdings Pvt Ltd, with the CCI noting a gradual trend toward convergence in terms of functionality between user apps, implying that relying on minor differences in features may not be appropriate for relevant market determination (Chunduru, 2020). IN THESE CIRCUMSTANCES, the CCI's approach is appropriate for digital marketplaces, which are dynamic and complex, with new features and functionalities being introduced regularly.

Conclusion

Initially, the CCI's decisional practice did not wholly reflect the peculiarities of multi-sided marketplaces, but more recent rulings consider the distinctive aspects of digital platforms, such as indirect network effects. The CCI's adoption of a more sophisticated approach to the relevant market determination is unquestionably a step in the right direction. However, they need to be standardization in defining relevant market in the cases involving MSPs in India; for that to happen, the section 19(6) and 19(7) should not be an exclusive list, and CCI had to have the independence in assessing different features and functionalities and even consumer behavior when determining the relevant market in cases involving MSPs.

 

References

Ashish Ahuja v. Snapdeal and others, 17 of 2014 (CCI 2014). Retrieved from https://www.cci.gov.in/sites/default/files/172014.pdf

Bhattacharjea, A. (2018). Predatory Pricing in Platform Competition: Economic Theory and Indian Cases. doi:10.1007/978-981-13-1232-8_11

CCI. (2020). Market Study on E-commerce in India.

Chunduru, A. (2020). Facebook Assures CCI Its Deal With Jio Involves Exchange Of Only ‘Limited Data’. Retrieved from medianama: https://www.medianama.com/2020/10/223-cci-facebook-jaadhu-jio-platforms-limited-data/

Federation of Hotel & Restaurant Associations of India (FHRAI) and Ors. v. MakeMyTrip India Pvt. Ltd. and ors., 14 of 2019 (CCI 2019). Retrieved from http://cci.gov.in/sites/default/files/Interim_Order_14-of-2019and01-of-2020.pdf

Hagiu, A., & Wright, J. (2015). Multi-sided Platforms. Harvard Business School. Retrieved from https://www.hbs.edu/ris/Publication%20Files/15-037_cb5afe51-6150-4be9-ace2-39c6a8ace6d4.pdf

Harshita Chawla v. Whatsapp and Facebook, 15 of 2020 (CCI 2020). Retrieved from https://www.cci.gov.in/sites/default/files/15-of-2020.pdf

Koul, P. (., & Prasad, P. (2020). An Analysis of the Abuse of Dominant Position by the E-Commerce Retailers in India. Amity.edu, 55-61.

Krishnan, A., & Unni, V. (2019). Competition Regulation in Two-Sided Markets: The Indian Jurisprudence. National Conference on Economics of Competition Law , 1-35.

Lifestyle Equities C.V. and Ors. v. Amazon Sellers Service Private Limited & Anr., 09 of 2020 (CCI 2020). Retrieved from https://www.cci.gov.in/sites/default/files/09-of-2020.pdf

Mandrescu, D. (2018). Applying (EU) competition law to online platforms: Reflections on the definition of the relevant market(s). World Competition: Law and Economics Review. Retrieved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3271624

MCX Stock Exchange Ltd vs National Stock Exchange of India Ltd., 13/2009 (CCI 2009).

Meru Travel Solutions Private Limited vs. Uber India Systems Private Limited and Ors, 96 of 2015 (Competition Commission of India 2015). Retrieved from https://cuts-ccier.org/wp-content/uploads/2019/02/Edition-1-Meru-travel-solutions.pdf

OECD. (2018). Rethinking Antitrust Tools for Multi-Sided Platforms. OECD.

Rysman, M. (2009). The Economics of Two-Sided Markets. Journal of Economic Perspectives, 125-143.

Sharma, K. (2011). SSNIP Test: A Useful Tool, Not A Panacea. COMPETITION LAW REPORTS, 188-192.

Viscusi, W. K., Harrington, J. E., & Sappington, D. E. (2018). ECONOMICS OF REGULATION AND ANTITRUST. Masachuttes: Massachusetts Institute of Technology.

 

 

Table of Figures

Figure 1- Network externalities in two-sided markets Source: (Viscusi, Harrington, & Sappington, 2018) 3

Figure 2 Direct Network Effect Source: Author 5

Figure 3 Indirect Network Effects Source: Author 5