Article 3 of the Platform Guidelines makes it clear right from the beginning that the regulatory framework for the platform economy sector is the same as for other sectors.[5] As such, the Platform Guidelines is the tool to further operationalize the principles and approaches stipulated in the current antitrust legislation, i.e., the AML, in light of the emergent nature and complexity of issues involved in the platform economy sector.

Article 2 of the Platform Guidelines defines “Platform" as a mode of business organization whereby, through network information technology, interdependent participants from two or multiple sides are enabled to interact with each other under the rules and matching functions provided by a specific medium, thereby jointly creating value.

The Platform Guidelines points out that the term “Platform" refers specifically to “internet platforms", which excludes other types of multi-sided markets, such as brick-and-mortar shopping malls or newspapers. The Guidelines does not appear to distinguish direct network effects and indirect network effects when defining a platform, which is consistent with the definition of online platforms by the Organization for Economic Co-operation and Development (OECD) [6] and the online platforms as defined in the EU Digital Services Act.[7]

The structure of the Platform Guidelines generally follows that of the AML, starting with market definition, followed by provisions on monopoly agreements, abuse of dominance, merger control (concentration of undertakings), and administrative monopoly in the platform economy sector.

Market definition has always been the starting point of antitrust analysis in China’s enforcement and judicial decisions. While the preliminary draft of the Platform Guidelines introduced an exception in which the step of defining a market may be skipped if there is “sufficient direct evidence" that the conduct at issue is “persistent and has significant anticompetitive effect",[8].the final version,  in Article 4, has omitted such language and states that market definition is generally required as the starting point even though the role of market definition may vary case by case. This approach seems to allow a wider latitude for enforcers in requiring market definition in novel cases.

Given that the platform economy usually involves complex business sectors and evolving com- petition dynamics, the Platform Guidelines provides some additional guidance on defining relevant markets in the platform sector based on the general principles provided by the AML and the Guidelines of the Anti-monopoly Commission of the State Council on the Definition of a Relevant Market.[9]

Article 4 of the Platform Guidelines emphasizes that market definition for a multi-sided platform is subject to a case-by-case analysis, but the Guidelines does offer several approaches to defining a market. First, the relevant product market can be defined based on the unilateral products in the platform. Second, multiple related product markets can be defined separately, depending on the interrelationships and effects between the relevant product markets. Lastly, when the cross-platform network effect can impose sufficient competitive constraints on the platform operator, the relevant product market can be defined on the basis of the platform as a whole. While undefined in the Platform Guidelines, we understand that a platform may provide different services to in-platform players (providers of goods or services) and its users. If such services provided by the platform to in-platform players and users are largely independent from each other, a separate relevant market may be defined for the products or services on each side. If such services are closely interconnected, that is, the platform’s services to different sides are largely integrated as a whole, the relevant market shall be defined by such multi-sided services. In the United States, the Supreme Court defined a single multi-sided transaction market in Ohio v. American Express Co.[10] and held that the proper definition of “credit-card market" must include “both sides of the platform—merchants and cardholders."

Not surprisingly, the market definition for online platforms may be delineated differently even for the same kind of practice at issue, given different circumstances. For example, in the Company A’s case, SAMR held that the relevant market should be defined as an e-commerce platform service market, separate from the traditional brick-and-mortar stores. And such e-commerce platform service activities do not need to be further segmented based on different operation models (e.g., B2C, C2C, live-streaming e-commerce, vertical e-commerce, etc.).[11] In contrast, in the Company S’s case,[12] Shanghai Administration for Market Regulation (Shanghai AMR) defined the relevant market as an online food delivery platform with English service in Shanghai, which is a much narrower market compared to that defined in A’s case. We understand that such divergence in the market definition between the two cases may be due to: (1) the difference in the business models of the two companies, as A offers a comprehensive e-commerce service, while Company S only focuses on the food delivery services for English-speaking expats in China; (2) the difference in the authorities dealing with the cases, as A’s case was investigated and sanctioned by SAMR, while S’s case was investigated and sanctioned by Shanghai AMR; (3) the difference in the timing of the sanction decisions, as the sanction on Company S was made in December 2020, about two months before the promulgation of the Platform Guidelines in February 2021, while the sanction on A was made in April 2021; and, (4) the difference in analytical approaches––a more in-depth economic analysis (including a hypothetical monopolist test) was conducted in S’s case than in A’s case. Given the limited number of cases so far, the issue of market definition in platform antitrust cases remains in a state of flux.

As noted above, Article 4 of the Platform Guidelines reiterates that market definition is generally the starting point for law enforcement (including the law enforcement concerning monopoly agreements) in the platform economy sector. Based on our experience and observations, however, it appears that SAMR does not put as much weight on market definition in investigating typical monopoly agreements. Monopoly agreements (so-called hardcore behaviors), such as price fixing, market allocation, and resale price maintenance, are expressly prohibited under the AML. In addition, the Platform Guidelines also discusses certain special types of monopoly agreements in the context of platforms.

Concerted Conduct Through Data, Algorithms, and Platform Rules. Under SAMR’s current enforcement, a finding of monopoly agreements in principle requires the finding of written or verbal agreement or of joint decisions by and among the participants. Articles 5 and 6 of the Platform Guidelines go further by deeming concerted conduct organized and coordinated through information exchange facilitated by data, algorithms, or platform rules as monopoly agreements, even if in such case, no written or verbal agreement or decision is actually made.

In this regard, Article 9 of the Platform Guidelines also intends to lower the burden of proof for SAMR in enforcement actions against monopoly agreements in the platform sector, where direct evidence may be difficult to obtain. With such lightened burden of proof, while the firm may provide rebuttal evidence, a finding can be made by SAMR based on the multiple indirect evidence which is logically consistent and infers participant’s knowledge of the relevant situation.

Also, Article 5 of the Platform Guidelines points out that “parallel conduct such as price-following by relevant firms resulting from independent expression of intent" is excluded in finding “collusion in substance through data, algorithms, platform rules". Unfortunately, it remains to be clarified or tested whether these provisions cover tacit algorithmic collusion in the platform sector.[13]

The Platform Guidelines refines the methods used by SAMR and its local counterparts of finding market dominance and assessing certain common practices by platforms. Indeed, assessing market dominance appears to be the focal point in recent enforcement and private action cases. Article 11 of the Platform Guidelines sets out in detail specific factors for assessing dominance in terms of market share, competition landscape, the firm’s ability to control the market, the firm’s financial position and technical conditions, the extent of reliance of other firms on that firm, and entry barriers, among others. Further, not surprisingly, the Platform Guidelines recognizes a variety of digital metrics that will be considered, such as the number of active users, click-throughs, network effects, lock-in effect, user stickiness, scale economy of the platform, access to data, etc.

Abusive conduct. Under the AML, typical instances of abuse of dominance include unfair pricing, predatory pricing, refusal to deal, exclusive dealing, tying, and imposing other unreasonable trade terms and discriminatory treatment without justifiable causes. The Platform Guidelines elaborates on each of these abusive practices in the context of a platform, addressing controversial practices such as free-of-charge or subsidizing tactics, “pick-one-side", technical blockage, search downranking, traffic restriction, and data-based discrimination.

Platforms’ free-of-charge or subsidizing tactics. Since the early days of China’s digital economy, platform operators have used free products and services or other subsidizing tactics to attract users while monetizing through other means such as advertising and/or paid membership. SAMR has not imposed penalties on such subsidizing tactics under the AML so far. However, there are law enforcement precedents penalizing subsidizing tactics under the Price Law. For example, certain community group-buy platforms subsidize users in the form of lower prices and discount coupons, resulting in prices of groceries lower than the actual cost. On March 3, 2021, SAMR imposed fines totaling RMB 6.5 million (around USD 1 million) on five community group-buy  platforms for violation of the Price Law in China.[14]

The Price Law regulates the pricing of enterprises and is distinguished from the AML. SAMR does not need to consider whether an enterprise has a dominant market position when penalizing such subsidizing tactics under the Price Law. In a recent nationwide compliance campaign against platform firms, the issue of predatory below-cost pricing has also come to the fore. Correspondingly, Article 13 of the Platform Guidelines provides that in calculating costs, “the costs interrelationship between the relevant markets in the multi-sided platform shall be comprehensively considered in general", while providing some cover for the below-cost pricing practice if it  is adopted to develop other in-platform business, promote new products, attract new users, etc., for a reasonable period.[15]

Although undefined, we think cost interrelationship can be understood to include the interrelationships between the costs of the platform facing different users in a multi-sided market––for example, for a community group-buy platform, the cost interrelationship may include the cost relationships between procurement costs for grocery sellers, transaction costs for consumers, and transportation costs for dispatchers. One economic perspective is that in a multi-sided market, pricing on one side of the market depends not only on the demand and cost on that side but also on how one side’s participation affects the other side’s participation and the profit that is extracted from users’ participation. A platform might charge a price below the cost on one side if those users have a large price elasticity and their participation attracts a large number of participants on the other side who are relatively price inelastic (and hence resulting in high mark-ups).[16] It is worth pointing out that below-cost pricing is regulated under the AML, and the Platform Guidelines reiterates that such abusive conduct by platforms is subject to antitrust scrutiny.

Platform   as   an “essential  facility". Once a firm is determined to be dominant, common platform operator’s practices, such as imposing restrictions and barriers (e.g., incompatibility, blocking, etc.) through platform rules, algorithms, or technology and traffic allocation against a counterparty (including a competitor) may be considered an unlawful refusal to deal, in particular if the relevant platform is deemed an “essential facility". Article 14 of the Platform Guidelines provides several factors to consider in finding the existence of an “essential facility", including the data owned by the platform, substitutability of other platforms, availability of potential competing platforms, the feasibility of developing a competitive platform, the degree of reliance on it by users or trading partners on the platform, and the possible impact on the platform operator by opening such platform to other (competing) trading partners. Nevertheless, such refusal to deal may be justified if, among others, transaction security will be compromised due to factors attributable to the users or trading partners, or if the refusal to deal will be unduly detrimental to the interests of the firm.[17]

The essential facilities doctrine originated from U.S. court cases is generally recognized. Most of the U.S. modern essential facilities cases rely on the four-prong test: (1) control of the essential facility by a monopolist; (2) a competitor’s inability practically or reasonably to duplicate the essential facility; (3) the denial of the use of the facility to a competitor; and (4) the feasibility of providing the facility.[18] Such four-prong test is also used in determining an essential facility in the Company H Metals case (as further analyzed below).[19] The Platform Guidelines also discusses incorporating certain platform-specific considerations, such as the possession of the relevant data by the platform, in finding whether an internet platform would constitute an essential facility.

In a recent antitrust case in China, Company B (which owns many popular apps) brought a lawsuit against Company T, alleging that T’s messaging services apps do not allow links to content on B’s App (“D"). Whether T’s digital platform constitutes an essential facility is one of the key issues in this case. If the answer is yes, T could be asked to open its platform to App D, allowing users to share information from D on T’s Apps. Some observers are skeptical about the application of the essential facilities doctrine in the platform sector because the doctrine originates from traditional sectors (such as telecommunication infrastructure). In this case, users can freely download and use D without downloading and accessing T’s Apps.[20] As this case is still at an early stage, and due to the lack of public information, it remains to be seen whether and how the court would apply the essential facilities doctrine with reference to the Platform Guidelines.

Notably, a Chinese court first applied the essential facilities doctrine in a recent case that involves novel antitrust and IP issues but does not involve platforms. In that case, a local Chinese court agreed with the plaintiff and found that H company’s patents were commercially essential and therefore constituted an “essential facility" within the meaning of antitrust law. The court went on to hold that H’s refusal to license without objective justification was anticompetitive and awarded damages of RMB 4.9 million (approximately USD 756,800) to the plaintiff. Significantly, the court in effect required H to license its patents to the plaintiff.[21] This case is currently on appeal.

Certain Platform tactics as exclusive dealing. Article 15 of the Platform Guidelines provides that “pick-one-side" is a typical exclusive dealing tactic and another focus of recent law enforcement. “Pick-one-side" refers to a tactic adopted by some platforms preventing in-platform players from cooperating or working with other competing platforms by a variety of “carrot and stick" measures, such as user data-based price discrimination, search downranking, traffic restriction, technical blockage, and subsidizing. Such tactics could lead to exclusionary effects and may be considered abuse of dominance. These practices, however, may be justified if the platform operator can show that such restrictions are necessary to protect intellectual property rights, trade secrets, or data security, for specific investments made for the transaction, to maintain a reasonable business model, etc.[22]

On April 10, 2021, SAMR imposed a fine on Company A  for its “pick-one-side" practice on its e-commerce platform; on April 12, Shanghai AMR announced that it had fined Company S for its “pick-one-side" on its online food delivery platform11; on April 26, SAMR announced a formal probe into Company M for “pick-one-side" and closed its investigation on October 8 by imposing a monetary penalty of RMB 3,444,439,866 (3% of M’s Chinese revenue in 2020)[23]; in a recent nationwide compliance campaign against 34 platform enterprises,[24] “pick-one-side" policy is also at the top of the potentially problematic practices identified by China’s enforcement authorities. As noted in the Platform Guidelines, and confirmed in relevant cases, the “pick-one-side" exclusive arrangements, including both written and oral arrangements, may trigger incentive and punitive measures. Further, the enforcement authorities will also consider the extent to which there is day-to-day monitoring and enforcement of such exclusive arrangements by the companies. Written contracts, e-mails, screenshots of WeChat (a primary messaging application in China), etc., may all be admitted as evidence against the firm concerned.

On the private action front, Company J (a leading e-commerce operator) brought a lawsuit against A.com (owned by A) in China, arguing that the latter has been prohibiting some of its merchants from participating in competitors’ sales activities since 2015. The case is still at the trial stage. We believe that the A case (especially regarding the determination of its market dominance and its adoption of “pick-one-side" policy) will have implications for this lawsuit.

Discrimination through Profiling.  Dominant online platform operators are not allowed to treat similarly situated users differently by manipulating big data and algorithms under Article 17 of the Platform Guidelines. In determining whether end-users are similarly situated, businesses may consider their differences in transaction security, transaction cost, credit status, transaction duration, etc., but not factors such as user privacy-related information, transaction history, individual preferences, and consumption habits. The relevant differentiation may be justified, however, if it conforms to proper trading customs and industry practices and is done for promotional purposes (e.g., for new users) within a reasonable period, among other potential justifications.[25]

In recent years, some Chinese online platform operators, especially travel service providers, collect and analyze information such as their customers’ consumption habits and preferences in order to implement price discrimination. There have been increasing public concerns regarding this practice, as China tries to tackle such big data- and algorithm-driven price discrimination. The Interim Provisions on the Management of Online Tourism Operation Services, which took effect on October 1, 2020, prohibits travel service providers from setting unfair trade conditions by abusing big data analysis or other technical means.[26]

We are not aware of any published antitrust investigation concerning discrimination through digital profiling. However, in the recent national compliance campaign against 34 platform providers discussed above, some platform providers have made compliance commitments in this area and promised to fine tune their algorithms and the use of big data to stop price discrimination through profiling.[27]

The Platform Guidelines clarifies several important questions regarding M&A transactions in the platform sector. This may be considered a response to the criticism that ex ante regulation has been too lenient. The main loopholes that need to be addressed are variable interest entity (VIE)-structured deals and below-threshold “killer acquisitions". The Platform Guidelines makes it clear that both are to be subject to vigilant antitrust scrutiny.

A VIE structure, which is a type of “contractual control", refers to a corporate structure where the actual controlling party does not own shares of the operating entity (i.e., VIE), but gains de facto control of business operations and financial consolidation of such an entity through a series of agreements (i.e., VIE agreements). In a typical VIE structure, the controlling entity is usually incorporated overseas for financing and tax planning purposes, while the controlled operating entity is a domestic company that can carry out business in the sector which is not yet fully open to foreign investors. Such a structure is believed to have been first introduced in the overseas IPO of Sina.com (also known as the “Sina Model") and later widely adopted in many emerging sectors, particularly in internet-related businesses.

VIE-structured deal expressly covered. For years, uncertainties have clouded the prospect of docketing and clearing a VIE-structured transaction under the AML. Although there exist no legal grounds for exempting such a transaction, in practice, parties often adopt various tactics to keep an otherwise notifiable transaction involving a VIE structure off the Chinese authorities’ merger review radar. This has become increasingly risky in light of recent regulatory developments.[28]

In July 2020, SAMR unconditionally cleared a joint venture between Company Y and Company G involving a VIE structure,[29] which is the first of its kind since the conditional clearance of M supermarket’s acquisition in 2012[30] and marks a milestone for future VIE filings. In December 2020, SAMR handed down failure-to-notify decisions against affiliates of three leading platform players,[31] and in 2021, SAMR has so far sanctioned over 40 failure-to-notify cases involving more leading platform players. [32] On July 10, 2021, SAMR prohibited the merger of two livestreaming platforms in China, the first blocked case in the platform sector.[33] On July 24, 2021, SAMR imposed a fine on Company T, citing violations in its acquisition in 2016, and ordered T and its affiliates to relinquish exclusive music copyrights within 30 days and to terminate the requirements on copyright holders to grant the company more favorable terms than its competitors.[34] All the platform players mentioned above operate at least partially through VIE structures.

Article 18 of the Platform Guidelines expressly points out that a VIE-structured deal is notifiable if it crosses the merger filing threshold. In the three failure-to-notify cases discussed above, SAMR for the first time imposed the largest ever monetary penalty under the AML and closed two cases within 40 days from the time of docketing. (The average investigation period for failure-to-notify cases is over 200 days.) In addition to imposing these fines, SAMR also pointed out that these industry leading players’ failure to notify despite their profile in the industry, the size of the deals, and the sophistication of their professional legal advisors all constitute aggravating factors. While the fines (RMB 500,000, approximately USD75,000) in each case were not economically significant, they have sent a strong signal that SAMR has stepped up its antitrust enforcement in the digital sector. [35]

Considering that the draft amendment to the AML released on October 23, 2021 proposes to significantly increase the upper limit of the penalty against a failure-to-notify (up to 10 percent of a firm’s group-level turnover in the preceding year for a concentration which will or will likely have eliminative or restrictive effect on competition; and RMB 5,000,000 for a concentration which will not have eliminative or restrictive effect on competition; the fine may be increased up to fivefold where the illegal act has extremely serious circumstances, extremely adverse impact, and extremely grave consequences), the legal consequences and non-compliance exposure has become much more severe.[36] Especially for digital platform players involved in a VIE structure, even a fine based on a low single digit percentage could see billions of dollars at stake.

Below-threshold “killer acquisition". Article 19 of the Platform Guidelines makes clear that the following below-threshold acquisitions may face merger review scrutiny:

• Either party in the transaction is a start-up or an emerging platform.

• The turnover of the parties in a transaction is low due to their adoption of the low-price or zero-price business model.

• The relevant market is relatively concentrated, and the number of competitors in the market is small.

The Platform Guidelines doesn’t provide an express definition of a start-up or an emerging platform, and we are not aware of any precedent in China where a below-threshold deal has been investigated for failure to notify or which results in any remedies. But the mention of below-threshold mergers is noteworthy as other jurisdictions have started exploring the potential anticompetitive effects of so-called killer acquisitions. For example, the European Commission is encouraging member states to refer transactions to the Commission if they fall below the national turnover thresholds, aiming to review killer acquisitions that are currently not scrutinized.[37] In the U.S., the Federal Trade Commission stated that it would review already consummated mergers in an effort to take enforcement actions against killer acquisitions by larger firms seeking to squash emerging competitors.[38] It is also reported that the FTC has raised concerns about acquisitions of nascent competitors in platform industries because these markets are prone to growing quickly due to network effects and economies of scale. For example, in the Edgewell/Harry’s case, the parties ultimately abandoned their USD 1.37 billion transaction after the FTC voted 5-0 to challenge it. The FTC alleged that Edgewell’s acquisition of Harry’s, which was founded as a direct-to-consumer wet-shave brand, would eliminate an innovative and disruptive competitor that reduced prices.[39]

Merger Remedies. Article 21 of the Platform Guidelines reiterates the importance of structural and behavioral remedies for clearing transactions with substantive competition concerns.

Possible structural remedies specific to the platform sector include the divestiture of intangible assets such as IP, technologies, and data. Possible behavioral remedies include the “opening-up" of networks, data, or platforms, licensing of key technologies, terminating certain exclusive agreements, modifying platform rules or algorithms, and interoperability commitments. It is noteworthy that SAMR’s recent practice seems to be more open to behavioral remedies (e.g., no tying/bundling, non-discriminatory (FRAND), interoperability with third parties, open-source commitments, protection of information of third-parties, continuation of supply, fulfilling the obligations of existing sales practices and procedures, and training sessions for management staff and employees) in high-tech sectors, especially for addressing vertical competition concerns.[40] It remains to be seen whether and when specific remedies will be imposed on platform operators and what these remedies will be.

* Gray Yu,contrib- uted to this article.