Antitrust enforcement in digital markets requires a cautious approach, yet their fast-changing nature also calls for quick responses. In certain circumstances, interim measures may serve as tools to enable some level of preliminary intervention before it is too late, but could be especially challenging when imposed on digital platforms without in-depth knowledge on the investigated conduct or even on the market itself. This paper discusses the role of interim measures applied to digital platform exclusivity cases in the context of the Brazilian recent experience with the iFood and Gympass probes, from which one can draw important lessons to similar investigations in the future. To avoid the risk of overenforcement in the imposition of interim measures, we argue that a competition authority should be able to monitor their effects and amend or adjust its decisions overtime to ensure that they are specifically tailored to inhibit the occurrence or aggravation of competitive harm that may not be restored – otherwise antitrust enforcement may inhibit innovation and disruption in digital platforms, to the expense of consumers.
By João Felipe Achcar de Azambuja[1]
I. INTRODUCTION
Like many other antitrust authorities worldwide, the Brazilian competition agency (“CADE”) has been devoting increasing attention to vertical agreements in digital markets, including, inter alia, exclusivity clauses practiced by online platforms.
Antitrust enforcement in digital markets requires a cautious approach to avoid type 1 errors (overenforcement) that could discourage innovation and other benefits to consumers. Such a careful assessment to balance the pro and anticompetitive effects of a given practice in a digital environment may lead to longer fact-finding investigations and a thorough review of market conditions and the parties’ activities, business models and incentives. However, the fast-changing nature of digital markets also calls for quick responses, or else antitrust enforcement may be rendered ineffective, leading to type 2 errors (underenforcement), resulting in market tipping or other distortions.
As suggested by the OECD,[2] in certain circumstances, interim measures may be interesting enforcement tools to deal with this dilemma, to the extent that they enable some level of intermediate intervention before the conclusion of an antitrust probe. Nonetheless, the design of an interim measure in the context of digital markets may be especially challenging, as it naturally demands that the authority makes relevant decisions without in-depth knowledge on the investigated conduct or even on the market itself.
Under the Brazilian Competition Law, CADE may rely on interim measures to cease potentially anticompetitive conducts while further investigation takes place. By doing so, the authority shall take into account: (i) whether the alleged conduct may possibly be deemed illegal (fumus boni iuris); and (ii) whether the perpetuation of the conduct may entail significant competitive harm if no immediate action is taken by CADE (periculum in mora). CADE also considers whether the intervention itself may generate irreversible impact to the investigated party or to the market (reverse periculum in mora).
In the last two years, CADE granted its first interim measures in antitrust probes related to digital platforms: one against iFood, a platform for intermediation of online food delivery orders; and another against Gympass, an aggregator of fitness centers. Both cases involve allegedly anticompetitive platform exclusivity agreements and other vertical restrictions, signaling a trend of preliminary antitrust intervention in Brazil in early stages of investigations on vertical agreements involving dominant online platforms.
II. iFood CASE
iFood is a platform for intermediation of online food delivery orders headquartered in Brazil. Founded in 2011, the company is a pioneer in the online food delivery market in Brazil and a market leader ever since.
In February 2021, CADE launched an antitrust probe against iFood following complaints filed by Abrasel, a Brazilian association of bars and restaurants; and Rappi, a competing platform for intermediation of online food delivery orders.[3] The complainants argued that iFood abused its market power in the Brazilian online food delivery market by imposing exclusivity obligations to a wide range of partner restaurants and engaged in discriminatory practices favoring exclusive over non-exclusive partners. According to the complainants, the practice entailed market foreclosure, increased entry barriers to competing platforms by preventing multihoming and increased switching costs due to large fines for the termination of the exclusivity arrangements.
In light of these alleged anticompetitive practices, the complainants requested the imposition of interim measures determining the immediate suspension of all exclusivity obligations on restaurants in its platform, and that iFood should be prohibited to take any discriminatory action against partner restaurants that were not subject to exclusivity clauses.
In the course of the investigation, CADE’s Superintendence-General (“GS”) found that iFood enjoyed a dominant position with a significant market share (up to 86 percent, according to data provided by Abrasel). Based on the public case files, only a limited portion of its partner restaurants were subject to exclusivity arrangements, although the exact figures were kept confidential. Nonetheless, the GS considered that the practice potentially entailed market foreclosure, especially given that iFood’s exclusive coverage included strategic restaurants and food chains that could increase the attractiveness of rival platforms and potential entrants, should they be able to enlist them on their platforms as well. In addition to that, the duration of the exclusivity clauses (which was kept confidential), prior notice obligations and the establishment of contractual fines for exclusivity breach increased switching costs that could hinder competitors’ growth. According to the GS, such restrictions to competition could be particularly harmful given that iFood, a first mover and dominant player in the Brazilian online food delivery market, was already better positioned to enjoy the rapid expansion of the market than its rivals. In fact, the competing platform Uber Eats exited the market in the course of GS’s investigation, despite market growth.
In contrast, iFood argued that the interim measure requested by the plaintiffs would entail unpredictable consequences given the uncertain and fast-changing dynamic of the Brazilian online food delivery market. According to iFood, strategic platform exclusivity provided incentives to invest in the company’s partners while avoiding free riding behavior. Without exclusive obligations, iFood argued that it would not have the operational and financial capacity to protect its investments in benefits granted to restaurants, leading to a natural decrease in spendings with partners enlisted in its platform. Competing platforms, in their turn, would be able to employ less efficient commercial conditions to capture exclusive partnerships as a result of the proposed preliminary intervention.
In March 2021, CADE’s GS adopted an interim measure partially granting the plaintiffs’ requests, by determining that the platform would be allowed to maintain all exclusivity agreements in place (which could be renewed for subsequent periods of up to one year each), but should be prohibited to execute new exclusivity agreements with any other partner restaurants. As such, it restrained the expansion of the practice, freezing new contracts, without interfering in the alleged incentives for iFood’s investments in existing contracts. This interim decision was not appealed before CADE´s tribunal and remains in force, as the investigation continues.
III. GYMPASS CASE
Gympass is a digital platform that operates as an aggregator of fitness centers, offering subscriptions to corporate customers interested in providing their employees access to partner gyms. Founded in 2012 and headquartered in Brazil, the platform is a pioneer and market leader in the Brazilian market for aggregators of fitness centers.
In September 2020, CADE launched an antitrust probe against Gympass following a complaint filed by TotalPass, a competing aggregator of fitness centers that owns a leading fitness chain in Brazil.[4] TotalPass argued that Gympass distorted competition by (i) imposing exclusivity obligations to all partner fitness centers and to a portion of its corporate customers, inhibiting multihoming in the two sides of its platform; and (ii) establishing most favored nation (“MFN”) clauses with some partner gyms, preventing them from charging different prices in other sales channels. As such, TotalPass requested the imposition of an interim measure determining the immediate suspension of all default exclusivity clauses imposed by Gympass to fitness centers and corporate customers, as well as the suspension of all MFN obligations in relation to the prices offered by gyms outside the platform.
Similarly to the iFood Case, the GS assessed the extension of market foreclosure imposed by Gympass and the duration and renewal conditions of exclusivity clauses vis-à-vis their economic justification. CADE’s GS found that Gympass held a dominant position and enlisted up to approx. 80-90 percent of the gyms in Brazil, applying exclusivity obligations by default for indefinite term. In GS’s view, this would lead to extensive market foreclosure, as only a small portion (less than 20 percent) of gyms would be available to be enlisted in competing platforms. Considering that the economies of scale of this business model depend on vast cross-sided network effects and the attractiveness of aggregators of fitness centers depend, to some extent, on a broad geographic coverage available on the platform, this could render market entry and competitors’ growth unfeasible.
Gympass argued that the economic rationale underlying the exclusivity contracts related to enhancing user experience and avoiding free rider behavior in its investments on gyms. In a preliminary assessment, the GS understood that there was no evidence of direct investments in each partner gym to justify platform exclusivity and that the foreclosure of all gyms enlisted was not justifiable by diffuse investments made by the platform.
In December 2021, the GS partially agreed with the plaintiff’s request and granted an interim measure inspired in the iFood Case, pursuant to which Gympass should be allowed to maintain the agreements in place or in final stages of negotiation, but should be prohibited to execute new exclusivity contracts with fitness centers and enforce quarantine obligations. Also, all MFN clauses related to prices outside the platform should be suspended. According to the GS, this was a transitory intervention while negotiations of remedies to settle the case were already in course.
TotalPass appealed this decision to CADE’s Tribunal.[5] In February 2021 the Tribunal overruled the GS’s decision and, by a majority vote, significantly broadened the interim measure to limit the platform’s exclusive coverage to cases in which Gympass injected direct investments in the gym’s infrastructure and other capital goods, arguing that the GS’s decision would endorse the existing substantial level of market foreclosure indefinitely and there was no legitimate economic reason for imposition of platform exclusivity contracts with all gyms in its portfolio.
IV. RATIONALE OF INTERIM MEASURES AGAINST EACH EXCLUSIVITY ARRANGEMENT BY DIGITAL PLATFORMS
The interim decisions issued by CADE in the iFood and Gympass cases establish a similar theory of harm, pursuant to which the market foreclosure entailed by exclusivity clauses in one side of the platform was found to potentially harm competition among platforms, increase entry barriers and raise rivals’ costs.
On both cases, the preliminary assessment of competitive harm considered specific features of the online platform environment that could aggravate these negative effects. iFood and Gympass were first movers in their respective markets and secured a wide range of business partners (restaurants and gyms, respectively) on one end and consumers on the other, enjoying significant cross-sided network effects in markets that favor market tipping. Should the incumbents be free to expand their exclusive footprint, GS understood that they would substantially limit the scale and attractiveness of competing platforms and potential entrants.
It should be noted, however, that iFood and Gympass presented very different footprints of exclusivity agreements. iFood had a market share above 80 percent in some accounts, but, according to the public case files, only a reduced portion of its portfolio was subject to exclusivity clauses. Gympass, in its turn, held a market share of approx. 80-90 percent and allegedly made use of exclusivity clauses by default in all contracts with gyms and fitness centers.
In addition, different foreclosure concerns arose given specific market characteristics in each investigation. Upon assessing the obstacles imposed on competitors and entrants due to the exclusivity arrangements, CADE found that competing online food delivery platforms needed access to key must-have restaurants and food chains to provide an attractive service to end consumers, whereas this was not the case for aggregators of fitness centers. Rather, the attractiveness of a fitness platform to its corporate consumers (and their employees) is directly related to having a broad geographic coverage of gyms in each city or neighborhood in which it is active. Hence, competing platforms may need access to a larger portion of the market to succeed when compared with online food delivery platforms, which, in their turn, need access to attractive partners that increase interest of users on the other side of the platform, but not necessarily a significant portion of the market.
The iFood and Gympass cases also entailed distinct considerations on switching costs. According to CADE’s GS, iFood imposed prior notice obligations and large fines to partner restaurants for early termination, whereas Gympass did not impose fines on gyms that would exit their portfolio, but subjected them to a quarantine period in which they would be prevented from joining rival platforms. In the context of preliminary intervention, however, the GS found that the two investigated conducts were deemed to artificially increase switching costs to the platforms’ partners in a context in which, absent platform exclusivity, multihoming would have been possible and potentially desirable for trading partners and customers.
Despite relevant different features found in the iFood and Gympass cases, CADE’s GS granted rather similar interim measures with respect to platform exclusivity arrangements, allowing the maintenance of all exclusivity agreements in place and prohibiting new agreements of this kind. In practice, the GS froze the expansion of the exclusivity coverage until the final ruling of the matters, averting the aggravation of the potential effects of the alleged infringement.
The fact that GS avoided inflicting significant changes in the business models of the defendants indicates that it may feel compelled to take action in early stages of investigations involving digital markets, but tends to adopt a cautious approach in the context of preliminary antitrust enforcement. It also demonstrates that, to some extent, the GS acknowledged that the exclusivity business models practiced by iFood and Gympass could potentially lead to some procompetitive benefits in the protection of the platforms’ investments against free riding effects.
In the Gympass Case, while overruling the GS’s decision, CADE’s Tribunal considered that the conduct was potentially harmful to competition and Gympass’s widespread exclusivity partnerships should be prohibited, absent a valid economic rationale for the practice. The interim measure that prevailed conditioned exclusivity clauses to specific investment in the fitness centers, concluding that platform exclusivity could only be used to protect these investments from free riding behavior and nothing else. By suspending a large part of the practice and requiring justification for any exclusivity clause, CADE’s Tribunal indicated that it may be willing to inflict more significant changes to a digital platform’s business model than the GS.
Although Gympass made use of broader exclusivity arrangements than iFood, which could justify a stricter restriction, one may argue whether CADE’s Tribunal strict approach was really necessary under the terms of an interim measure, as it disrupted a large number of contracts without any detailed analysis and, according to the GS, settlement negotiations were already in course.
Another point that deserves a closer look is the requirement of specific investments in concrete assets to justify exclusivity clauses. This seems to disregard other potential investments (cash investments, publicity etc.) in long term partnerships with gyms and generates incentives towards only a single type of direct investments out of the multiple potential investments that a platform could choose from to promote its ecosystem and/or trading partners. This can be problematic, as the antitrust authority is rarely in a suitable position to decide which investment options of a company are optimal to enhance a procompetitive environment – especially in the context of a preliminary decision, even more so when dealing with digital platform markets.
V. CONCLUSION: TRENDS AND LESSONS FROM THE BRAZILIAN EXPERIENCE
The precedents involving iFood and Gympass show that CADE may be more inclined to enforce interim measures in antitrust probes involving vertical agreements entered into by digital platforms. Even though these are only two decisions, they point to similar concerns about digital markets.
This enforcement trend implies that the authority acknowledges the dynamic nature of digital markets and that quick reactions may be required to preserve competition, especially when concerns with market tipping and consolidation of first mover advantages are on the table. Indeed, interim measures can be important enforcement tools in platform markets with digital features, although the actual scope and intensity of such preliminary intervention on a specific case can be subject to debate. Moreover, the assessment of the suitability of an interim measure also requires additional thought on the use of scarce resources to achieve optimal enforcement levels in an environment of dynamic nature in which uncertainty is ubiquitous.
Given the unpredictability factor in preliminary antitrust intervention in early stages of investigations, there is always some risk of overenforcement. As argued above, the severe restrictions imposed by CADE’s Tribunal in the Gympass Case may be an example of this, as the interim measure that prevailed narrowed the acceptable protection through platform exclusivity arrangements to only one out of multiple investment options and other potential benefits that could be generated by the aggregator of fitness centers.[6]
With the risk of overenforcement in mind, CADE should be able to monitor the actual effects of the interim measures imposed and amend or adjust its preliminary decisions overtime[7] to ensure that they are specifically tailored to inhibit the occurrence or aggravation of actual or potential competitive harm that may not be restored or that would render the outcome of the investigation ineffective. Otherwise, harsh interim measures may negatively affect the development of new markets and new market players and inhibit innovation and disruption as a whole before a final and reasoned ruling, to the expense of consumers. CADE will likely be incited to address these challenges more often, as well as all other competition agencies worldwide.
[1] Associate at Machado, Meyer, Sendacz e Opice Advogados, in São Paulo, Brazil. Master’s student at FGV Direito SP, with support of the Mario Henrique Simonsen scholarship. Email: joao.achcar.azambuja@gmail.com.
[2] OECD, Interim Measures in Antitrust Investigations (2021), https://www.oecd.org/daf/competition/interim-measures-in-antitrust-investigations-2022.pdf.
[3] Administrative Inquiry No. 08700.004588/2020-47 (Plaintiffs: Associação Nacional de Restaurantes – ABRASEL and Rappi Brasil Intermediação de Negócios Ltda.; Defendant: iFood.com Agência de Restaurantes Online S.A.).
[4] Administrative Inquiry No. 08700.004136/2020-65 (Plaintiff: Total Pass Participações Ltda.; Defendant: GPBR Participações Ltda.).
[5] Voluntary Appeal No. 08700.007228/2021-88 (Plaintiff: Total Pass Participações Ltda.; Defendant: GPBR Participações Ltda.).
[6] Gympass settled the case shortly before the publishing of this paper in September 2022. In fact, the restrictions on exclusivity clauses under the settlement agreement entered into by Gympass were quite less strict than under the interim measure adopted by CADE’s Tribunal.
[7] Filippo Lancieri & Caio Mario S. Pereira Neto, Design Remedies for Digital Markets: The Interplay Between Antitrust and Regulation, Journal of Competition Law and Economics (2021) at 30.
