Digital markets have come into sharp focus for competition authorities worldwide over the last decade, but never has the spotlight been brighter than in 2019. Platforms operated by tech giants like Google, Apple, Facebook and Amazon (GAFA) have taken on a pervasive role in everyday life and reshaped the modern economy.
At the same time, public concern over these platforms’ apparently unassailable market power, function as gatekeepers to other businesses, and the impact this may have on customers has led to competition law scrutiny around the globe. But it is not just these pure tech businesses that have been in focus. As digital channels become critically important for conventional businesses, they too have faced increasing scrutiny of their digital activities from competition authorities.
The competition law considerations for digital markets are highly complex. As competition authorities have been called upon to fill the regulatory lacuna of the internet, they have grappled with how to best apply their existing tools to markets which are fast-moving, often multi-sided with powerful network effects, and where in many cases, services are offered for free. Following a period of increasing competition law harmonisation, particularly among EU member states, divergence is now increasingly evident as each national agency experiments with ways to address these common challenges in their own way, in the absence of an agreed way forward. Enforcers still aim for harmonisation and consistency but in many cases the reality is different.
While many investigations are still only in the early stages, this potential for significant divergence in methodology and outcomes creates unpredictability for business that may spill over into other sectors. It has sometimes felt as though competition law is groaning under the weight of the digital world. Here we have summarised five of the key themes out of recent enforcement and policy work from competition authorities’ around the world.
1. Digital dominance: novel theories of dominance and abusive conduct
A number of novel theories of harm have been used against dominant platforms, reflecting the unorthodox characteristics of these markets and the appetite of many authorities to take action against their publicly perceived market power. The limits of many existing theories of harm have also been stretched beyond their traditional application. However, there remains debate about the validity and consistency of these new theories of harm, and whether new iterations of existing theories of harm are valid. One of the most striking has been the much-publicised action by the German Federal Cartel Office against Facebook1 , which found in February 2019 that it had abused its dominant position in Germany by making the use of its social network conditional on the collection of user data from third party websites, which was viewed as exploitative of Facebook users. However, the relevant German appeals court recently overturned this2 , saying it had “serious doubts” about the legal basis for the alleged offence.
The EC has primarily focused its antitrust actions against large tech platforms on two novel approaches to the conventional “leveraging” theory of harm applied to other markets3 . The first approach scrutinises platforms that reserve the use of data collected from their users and third-party suppliers to leverage their entry or expansion into related markets, threatening the competitive position of these competing suppliers. This theory of harm may play a role in the investigation launched by the EC to establish whether competitively sensitive data collected by Amazon from independent sellers on its marketplace is used for the benefit of its retail business, and whether this in turn is in breach of EU competition rules4 . This investigation follows Amazon’s recent settlement with competition authorities in Germany and Austria over its business terms for online traders that use its marketplace platform5 .
The second approach examines conduct that results in “selfpreferencing” – discrimination by a platform in favour of vendors vertically related to itself, and to the detriment of third-party suppliers that compete on the platform. This is broadly familiar from historic EC investigations like those against Microsoft6. Selfpreferencing may feature in the EC’s rumoured probe following a complaint by Spotify earlier this summer against Apple’s policy of “penalising” users of rival content-streaming services. The Spotify complaint, along with an investigation by SAMR in China into potential exclusionary conduct by electronic platforms, also highlights a wider global trend of investigation and enforcement against digital platforms by competition authorities, in particular those that both host services and compete with them.
These theories are also having an unprecedented spill over effect in the US, where monopolisation claims face a higher burden of proof to show potential exclusion. Both the DOJ and the FTC have launched broad formal competition investigations into large platform providers. While these investigations are reportedly still in the early stages, it is clear that the authorities are considering a broad range of potentially abusive conduct. State attorney generals are also actively investigating digital platform providers to address a perceived enforcement void at the federal level, with 50 state attorneys-general investigating Google, whilst 47 of them are also probing Facebook. Congress has also taken an unusually active enforcement role through public hearings in the Senate and House subcommittees on the conduct of large platform operators, and calls for legislative changes to support enforcement.
2. In focus: heightened scrutiny of online sales channels
Beyond the GAFA platforms that are household names in the tech space, over recent years, regulators have also focused their attention on the online dealings of businesses that we would not think of as tech at all. Guess7 is just one of the most recent fashion brands (following Asics 8, Coty 9 , Adidas10 and others) to have been found to have breached competition law (and fined – a whopping €39.9m to be exact) for imposing restrictions on online sales, reflecting the critical commercial importance of internet channels to conventional retail companies. Similarly, in the prohibition of the proposed Sainsbury’s/Asda UK 11 supermarket merger earlier this year, one of the CMA’s theories of harm was that the merger would lead to a reduction in competition online, illustrating the reach of tech theories of harm to even traditional “brick and mortar” businesses.
While EU and US enforcement against restrictive agreements in the tech space is not new, many of the cases taken historically within the EU have essentially represented old theories of harm in new clothes (e.g. bans on passive sales as per existing vertical supply exemptions). However, more recent enforcement has seen the focus turn to previously unconsidered forms of potential infringement. One obvious example is the focus on most favoured nation clauses. While commonplace in wholesale contracts offline and generally not regarded as problematic in that context, when used by platforms at the retail level these clauses have attracted close scrutiny from competition authorities around the world12.
Competition authorities have also considered entirely new restraints that can only occur online – for example “keyword bidding restrictions” in respect of online search advertising. In its decision against Guess in December 2018, the EC found that Guess’ prohibitions on distributors bidding for Guess brand names on Google’s online search advertising service constituted an object restriction. The German BKartA brought similar proceedings against Asics in 2015 for imposing restrictions on distributors’ online bidding activities. In addition to these cases focusing on keyword bidding as a vertical restraint, it has also been the basis of horizontal concerns. The FTC recently ordered 1-800 Contacts, Inc.,13 the biggest retailer of contact lenses in the US, to cease and desist from agreeing with competitors not to bid for search links containing a competitor’s brand name, thereby underlining that trademark settlements do not benefit from a blanket immunity from competition law. Wider issues linked to e-commerce, such as the nature of agency in an online context, also figure highly in the EC’s review of VABER, with a new version expected in 2022.
More novel still, pricing algorithms used in online markets have spawned fresh competition enforcement challenges for regulators to grapple with. The impact of these algorithms can be dramatic: in one well-publicised example, competing pricing algorithms led to the listing of a biology text book for sale on Amazon for the staggering price of US$23.6m. According to a CMA report14, three theories of harm in relation to algorithmic tacit collusion have come to the fore, namely sellers using the same algorithm and data pool to determine prices (i.e. hub and spoke “collusion”); algorithms’ predictable reactions to events tipping off competitors (predictable agent); and algorithms independently reaching tacitly coordinated outcomes without any human intention of colluding (autonomous machine). However, while we still await the first infringement decision on algorithms, the ECJ found in Eturas et al v Lietuvos Respublikos konkurencijos taryba15 that travel agencies which used Eturas, an online booking platform, and: (a) had actual knowledge that the platform capped discounts on travel packages and (b) did not oppose such restrictions; could be deemed to have tacitly colluded. The difficulties in adapting the hub and spoke cartel theory to algorithmic collusion were grappled with by the Competition Commission of India, which ultimately dismissed a complaint against Ola and Uber16 on the basis that arrangements with their platform did not amount to a set of vertical agreements with the drivers.
3. Put on the Google Glass(es): closer scrutiny of acquisitions of disruptive technologies
Authorities are looking to change their assessment of digital mergers amidst increasing concern about underenforcement, particularly “killer acquisitions” of potentially disruptive technology or competitors – the alleged practice by big tech (and also pharma) of buying up small innovative players to stifle future competition17. Despite the inherent uncertainty in proving future harm in relation to a nascent market entrant, commentators argue that the level of market concentration and the dynamic nature of these markets requires more active enforcement. Enforcers have also been revisiting consummated mergers where they allegedly missed opportunities to robustly scrutinise GAFA acquisitions such as Facebook/WhatsApp and Facebook/Instagram.
The EU’s report on competition policy in the era of digitisation18 suggests changes to the approach for substantive assessment of mergers, so that where an acquisition in an adjacent market can be seen as part of a strategy against users defecting from a dominant platform’s ecosystem, the parties should bear the burden of showing that the adverse effects on competition are offset by pro-competitive efficiencies. In the UK, the CMAcommissioned Lear Report19 conducted a review of historical mergers to inform merger policy going forward and concluded that the CMA’s approach to tech mergers had been too lenient. Similarly, the UK Government-commissioned Furman Report20 recommended imposing a positive obligation on major tech companies to notify all intended acquisitions to the CMA, and that a new balance of harms approach that emphasises the scale of competitive impacts of a merger, not just their likelihood, should be used for the substantive assessment. Outside of Europe, the BRICS countries’ competition authorities have released a digital markets report21 studying how the authorities should review mergers, acquisitions and conducts in digital markets.
While it is uncertain where proposals for legislative reform to the test for tech mergers will end up, we have already seen a significant uptick in enforcement against mergers in the tech space (e.g. the CMA’s probe of Amazon’s investment in Deliveroo22 or the EC’s review of Apple’s acquisition of Shazam23 and Google’s proposed acquisition of Fitbit). What does seem a virtual certainty is increased and deeper scrutiny of tech mergers in 2020 and beyond. Procedural measures have also been implemented to ensure greater opportunity to review acquisitions of smaller potential competitors before closing. For example, Germany and Austria have already introduced transaction value thresholds so that they can take jurisdiction over acquisitions of low-turnover, high-potential targets.
4. With great data power, comes great responsibility
One key source of market power for certain platforms, such as Google and Facebook (and a major barrier to entry for challengers) is their access to a huge amount of personal and commercial data. Historically in Europe, concerns over data have been considered through the lens of abuse of dominance, in cases where access to data can be considered an “essential facility”. However, ECJ precedents – Bronner 24, IMS Health25 and Microsoft 26 – make clear that the bar for establishing that data is an “essential facility” is high for parties in downstream markets. According to the 2016 joint Competition Law and Data report by the French Competition Authority and the German Federal Cartel Office27, these requirements would only be met if it can be demonstrated that the data owned by an incumbent is truly unique and that there is no possible alternative for the competitor to obtain the data that it needs to perform its services.
Notwithstanding these hurdles, enforcement cases have been brought: as part of the Commission’s AdSense investigation28, it was alleged that Google inhibited online advertisers’ ability to easily transfer or port across data and content related to these advertising campaigns to other rival platforms in breach of Article 102 and in 2013, Google offered commitments to terminate conditions which prevented such transfers. More recently, in July 2019, the Commission announced an in-depth investigation into Amazon’s use of data received from its marketplace merchants and whether this might breach Article 102 and/or Article 10129. Beyond enforcement on the basis of abuse of dominance, there is increasing consensus among policymakers and regulatory authorities globally that frameworks should be developed to enable greater access to data, while at the same time ensuring the protection of privacy rights. However, a key challenge is creating structures that allow data to be accessed and shared in a meaningful way – making data “portable”. A number of specific measures including GDPR and the Regulation on Free Flow of Non Personal Data have already been taken within Europe to increase data portability and therefore increase switching rates.
Similar initiatives have been pursued by policymakers around the world. In October, a bipartisan group of US senators introduced a bill which would require social media platforms with more than 100 million users to make user data portable and their services interoperable with each other. This bill – the Augmenting Compatibility and Competition by Enabling Service Switching (Access) Act – the first of its kind, will allow users on these platforms to delegate the management of their online interactions to a custodial third-party agent. Similarly, in Japan, GAFA representatives have been meeting with lawmakers this month to discuss a bill intended to improve transparency in such areas as search results and data access, following a report by the Japan Fair Trade Commission which found that conduct by large platform operators could harm competition30.
With so many governments and authorities around the world focused on data and multiple tools available to address its competitive effects, managing compliance with duplicative and potentially divergent rules on handling of data is bound to be a key challenge for not just data driven platforms, such as Google and Facebook, but all businesses handling data, over the coming years.
5. Out with the old, in with the new: “rethinking the toolkit” and the spectre of sectoral regulation
In addition to their enforcement efforts, governments and competition authorities around the world have published a multiplicity of policy reports and retrospectives, reflecting on how competition law has evolved over the life of the internet to date, and what (if any) new tools competition authorities need to face the digital age.
Most competition authorities that have considered these issues in depth think that, while a major overhaul of competition rules is not required (as per the joint statement issued by competition authorities of G7 Member States31), a sector-specific application is required in digital markets. One key challenge has been the (lack of) speed with which authorities have been able to intervene in markets which are dynamic and fast-moving. In particular, the potential for markets to “tip” towards a “winner takes all”32 can mean that a typical competition investigation lasting several years is ill-adapted to the challenges of the digital age. In response to this, there is a consensus that authorities need to be able to act faster. The UK has issued new guidance on interim measures, while the EC recently imposed interim measures on Broadcom33.
Related to this, one suggestion under consideration at the European Commission level is to shift the burden of proof on big tech companies to prove their conduct is not anticompetitive. The proposal, arising out of the Competition policy for the digital era report prepared for the Commission34 released in April this year, and recently endorsed by Margarethe Vestager, is that in the context of highly concentrated markets with strong network effects and high barriers to entry, authorities should err on the side of disallowing potentially anticompetitive conduct, requiring the putatively dominant incumbent to establish its conduct is pro-competitive. This could apply to both mergers (and a similar suggestion in the merger context was in the UK’s Furman Report) as well as behavioural restraints, e.g. companies preventing users from accessing multiple apps and companies blocking access to their data for third-party applications looking to offer supplementary services.
The upheaval of the digital era along with broader macroeconomic and political trends has seen an even broader rethink of the aims and remit of competition policy altogether, which could have significant ramifications for application of competition law in the tech space. While most major competition authorities around the world have historically applied rigorous economic analysis to the cases before them, there is an increasing political appetite for competition authorities to take into account broader considerations. In the UK for example, the Chairman of the CMA has proposed reforms which would make consumer welfare – rather than competition itself – the touchstone for the CMA’s enforcement work35. Similarly, in the German FCO’s probe into Facebook36, the focus was not a harm to competition per se, but breach of data protection rules as a potential abuse of market power.
A common proposal across most reports, potentially bringing these threads together, is to create a specialist digital body or unit to develop expertise in digital markets, with the goal of overcoming the information asymmetry that has so far left competition authorities following in the wake of big tech companies. In the UK, the government has backed the Furman Report’s proposal to create a new digital markets unit with specific powers including to set a code of conduct for companies with “strategic market status”. More recently, the French Competition Authority announced it is in the process of setting up a digital unit for antitrust and merger cases. Meanwhile, in the US, the Stigler Report37 suggests creating a specialist agency for digital platforms that would apply different rules to companies with “bottleneck power”, and in the Southern Hemisphere, the ACCC’s final report38 recommends the establishment of a specialised unit within the ACCC that could “build on and develop expertise in digital markets and the use of algorithms.”
While there appears to be an emerging consensus that a sector-specific regulator could add value in the tech space, the establishment of different national regulators applying potentially differing standards to online platforms that operate globally will raise obvious challenges for companies. It remains to be seen whether sufficient international cooperation will be possible to establish a truly joined up system of global regulation.