Debate in the competition law community has intensified around the issue of enforcement in digital markets in recent months with (yet another) significant antitrust fine being levied against Google,1)Amount of actual fine may vary the Bundeskartellamt’s Facebook decision, and (most recently) the European Commission’s Report on Competition policy for the digital era.2)http://ec.europa.eu/competition/publications/reports/kd0419345enn.pdf
The emergence of digital markets (and these recent events) have raised the question as to whether existing competition law rules are fit for purpose to tackle the generation of competition challenges created by the technology revolution. This post will argue that the decisional practice of enforcers indicates that the competition rules and principles developed are largely fit for purpose. The closer the case being investigated is to a traditional market dynamic, the easier it is to bring a robust enforcement action. However, technological innovations and the emergence of new business models (such as Google Ads) are influencing the choice of case brought by the enforcer, meaning the identification of the competitive harm may require new enforcement rules for new kinds of markets.
This post will assess these enforcement issues in light of the AdSense decision and the European Commission’s earlier Decisions (Google Shopping in 2017 and Google Android in 2018) and the conclusion of the German Bundeskartellamt’s Facebook investigation in February 2019. These cases illustrate (to varying degrees) how the traditional antitrust analysis fares in assessing different kinds of digital markets: including in terms of assessing market definition, the competitive effects, barriers to entry and efficiency.
How to define a digital market?
Of course, it is worth noting at the outset that digital markets reflect traditional offline markets in many cases, and can be defined more or less in the usual way. For example, the Commission’s investigation into the “Steam” video game distribution platform is clearly one involving a digital market. However, the competitive harm being assessed in that case relates to contractual export restrictions or partitioning the market along national lines, reminiscent of Consten and Grundig (or any of your other favourite classic “by object” cases), and clearly did not require any innovations in competition law.3)Well, except perhaps that the Steam case is helped by the coming into force of the new Regulation (EU) 2018/302 (Geo-blocking Regulation).
A distinction should be drawn between traditional-type online markets and the more nascent and nebulous business models of the likes of Google and Facebook, who operate in a space with very different dynamics. A finding of dominance in those nascent “markets” may be very difficult to establish by traditional competition law measures; particularly where the services are free and the costs of switching are zero or virtually zero, with the ownership of data being a key factor in market influence. However, some traditional features of dominance continue to be clearly abundant in these dynamic markets such as a “winner takes it all” dynamic, zero marginal costs, and the presence of strong network effects leading to economies of scale. This dichotomy raises the question of where the line is between a successful traditional model and the need for innovative tools for new markets.
For the purposes of the competition rules, markets are defined by the substitutability of the goods or services on offer.4)Paragraph 7 – Commission Notice on the definition of relevant market for the purposes of Community competition law (97/C 372/03) Although Google and Facebook can be neatly described as “a search engine” and “a social media website” respectively, both entities have evolved into offering an expanding and seemingly limitless number of ancillary services, some or all of which could be or are broadly substitutable with other applications and websites. The level of differentiation in the apps and services they provide makes it difficult to reconcile conduct which intuitively “feels” like a distortion of competition with appropriate markets, dominance and particular abuses.
For example, it is unclear the extent to which different social media websites are substitutable and “compete” with one another (other than looking at less informative functional criteria like a differentiation in features, user interface and, perhaps, by the different kinds of emojis they offer to users). Similarly, in the absence of an end price and with a marginal cost of zero, it is unclear how to assess any alleged exclusion of an “as efficient competitor” in such an environment.5)That test being most recently discussed in Intel (Case C-413/14).
Such problems in defining markets and finding dominance are compounded by the fact that typically complex antitrust investigations are broad in scope and thus protracted in duration, meaning that the technology or user behaviour will often have changed by the time a decision on the competitive harm has been reached. All of these issues are particularly problematic for competition enforcers who rely heavily on their market definition to support their conclusions on competitive harm.
Network effects as a barrier to entry
Despite the difficulty in defining markets and dominance, there is one clear feature of digital markets that has allowed companies like Google and Facebook to wield significant influence. These companies both enjoy significant inherent advantages enjoyed by economies of scale known as “network effects”. By way of illustration, the more users that social media websites like Facebook and Instagram have, the more valuable that service becomes to each new user, which creates a virtuous cycle of more users leading to increased utility for users, in turn leading to more new users and so on. The same principle applies with business models such as Google Android and Google AdSense, whose market position becomes consolidated the more users they have.
Network effects certainly have positive externalities (a better end-user experience), while also leading to rapid growth in terms of market power and influence and operate as a type of barrier to entry for prospective competitors. Networks effects can multiply rapidly in online markets and encourage the large tech companies such as Google and Facebook to build and consolidate their market positions. However, on the other hand, innovation in the market and changing customer preferences mean formerly influential firms can fall rapidly by the wayside. Just as Myspace fell as a social media giant (or Bebo for the benefit of our Irish readers aged 25-35), so too will Facebook.
Have the competition rules been successfully applied in the Google and Facebook cases?
The European Commission’s conclusions in all three of its Google cases (as in any abuse of dominance case) are contingent on Google’s dominance in certain markets coupled with an abuse. They suggest that Google has managed to leverage and consolidate its market power by engaging in anticompetitive practices in its related (to a greater or lesser degree) online platforms.
The Google AdSense and Google Android decisions are similar in that both relate to Google seeking to consolidate its position as the dominant online search engine with close parallels that can be drawn to traditional abuse of dominance cases.
Google AdSense, the most recent decision, relates to the internet’s foremost advertising brokering service which is owned by Google. AdSense is integrated free of charge by a vast array of websites seeking to monetise their content by earning revenue through engagement with advertising facilitated by AdSense. These include traditional text / hyperlink-style ads in addition to images, animated graphics, and even imbedded videos (for example Google subsidiary YouTube monetises its content by displaying video ads through AdSense). This is not too dissimilar to a newspaper selling its advertising space
The European Commission found that Google has abused its dominant position in this market by imposing certain restrictions on third party websites using the AdSense including forcing them to: exclusively use AdSense for advertising, reserve premium placement for Google search, and take a minimum number of AdSense adds. The Commission found no justification for the imposition of these restrictions other than to exclude Google’s competitors (such as Microsoft and Yahoo) from the market for online search advertising. The Commission considered that this resulted in less choice for users in selecting advertising services and higher prices ultimately being passed on to consumers.
Last year, Google was also fined a record €4.3bn for engaging in restrictive practices related to its Android operating system including mandatorily pre-installing Google’s proprietary apps on Android devices. This is a case that is very much reminiscent of the infamous Microsoft case taken by the European Commission in the early part of the 2000s, which punished Microsoft for pre-installing its proprietary Media Player and Internet Explorer programmes on its Windows operating system, thereby excluding the entry of competitors. This is a case which – ironically – directly led to the rise of Google Chrome as a competitor browser.
Notwithstanding all the discourse around competition rules being in need of updating to reflect digital markets and the publication of the Commission’s Report, the current rules are adequate in addressing these wrongs. By contrast, the Google Shopping and German Facebook cases follow a less well trodden path, and the infringement of competition law is harder to identify using traditional tools.
The Google Shopping case related to Google systematically giving prominent placement to its own comparison shopping service and demoting rival comparison shopping services in its search results. In this case, the causal link between Google’s dominance (in online searches) and the finding of abuse (in an ancillary online shopping comparison “market”) is questionable. It is particularly problematic given that the services being offered are “free” and access to competitors in the ancillary markets (e.g. eBay and Amazon) is just a click away, indicates a switching cost of virtually zero. This is all the more surprising in circumstances where the Google search results were clearly labelled as “advertisements”.
Although dominant firms are under a special obligation not to abuse their market position, the Google Shopping decision goes further and appears to punish the vertically integrated Google for favouring its own services (in the manner of a leading supermarket chain not being allowed to favour its own brand groceries). In addition, the causal link between Google favouring its own products and the distorting effects on competition is not fully established by the decision.
Simultaneously, the Facebook case involved the German Bundeskartellamt imposing significant restrictions on data sharing by Facebook and its related applications and functions. It found that Facebook gathered data outside the Facebook website, including through its WhatsApp and Instagram platforms. This included data gathered through unclicked “like” and “share” buttons on 3rd party websites visited by Facebook users to improve its targeted advertising, and users were not asked to consent to such practices.
Given the difficulty in defining digital markets, in these circumstances, measuring the market power of social media websites by the “number of users” and therefore “information held on users” is an attractive method to define the market. It is clear that the Bundeskartellamt measured Facebook’s market dominance in this way. It viewed Facebook as a data-based business model (think of the Cambridge Analytica files), which likely informed its non-traditional approach to defining the market in this case.
This decision has suffered significant criticism for stretching the competition law rules beyond their natural limits. Most notably (and somewhat bizarrely) the decision also declared Facebook to be “in violation of the European data protection rules” without any formal decision by the German data protection authority or any clear link between this finding and competitive harm. Although the European Commission’s 2014 Facebook / WhatsApp merger decision expressly excludes privacy-related concerns relating to the concentration of data from the scope of EU competition rules, there is perhaps space for strictly competition law concerns to be addressed where that market power stems from the aggregation of user data.
However, a key problem of using a violation of user privacy as the measure of competition harm rather than (for example) increases in price is that it is almost impossible to establish to calculate a “competitive” or “market rate” for acceptable privacy standards from a consumer’s perspective. In addition, any consumer harm suffered seems to stem from a breach of privacy law, rather than competition law.
In any event, the Google Shopping and Facebook cases arguably extend this special obligation of leading (dominant) online businesses and seem to place a significant onus on the leading online platforms (which are often vertically integrated). Any outcome of a competition case which requires a firm to facilitate its rivals to compete with it is a far-reaching one indeed.
There is an uncomfortable relationship between the kind of influence enjoyed by Google and Facebook and the competition law harm identified in the Google Shopping and Bundeskartellamt decisions. Perhaps the issue stems from problem is the choice of cases to take. Both Google Shopping and Facebook likely have ambivalent effects on competition.
What next for competition enforcement in digital markets?
This is the key question that the Commission’s Report on Competition policy for the digital era attempts to answer. However, the truth is the rules work very well in most cases, as demonstrated by Google Android and Google AdSense.
Of course, competition policy is at least as much of an art as it is a science, and reasonable competition experts can disagree on the degree of enforcement and which cases should be pursued. Even the weight attached to each factor in a competitive analysis: network effects, the role of data, and whether there is harm to competition may be the subject of disagreement.
However, the themes and trends in this selection of cases related to market definition, dominance and market dynamics (such as network effects, and the role of data) suggest a certain need for further reflection before any changes to the competition rules are made. These reoccurring factors need to be reconciled and rationalised so that clear enforcement parameters can be drawn for competition enforcers in digital markets to avoid the risk of dilution of enforcement in less clear cut cases.
It is clear that large tech companies, such as Google and Facebook have enhanced competition and innovation exponentially in the digital age, which has put them at a competitive advantage. Case law tells us that the existence of a competitive disadvantage to rivals does not amount to an anticompetitive effect.6)MEO C-525/16 On the other hand, it is important to ensure a balance is struck to allow innovative procompetitive conduct in a market by the leading company, and allow entry and a viable playing field for new entrants to also innovate in ancillary markets. It remains to be seen how EU regulators will maintain this balance.
References [ + ]
|1.||↑||Amount of actual fine may vary|
|3.||↑||Well, except perhaps that the Steam case is helped by the coming into force of the new Regulation (EU) 2018/302 (Geo-blocking Regulation).|
|4.||↑||Paragraph 7 – Commission Notice on the definition of relevant market for the purposes of Community competition law (97/C 372/03)|
|5.||↑||That test being most recently discussed in Intel (Case C-413/14).|