Essential vs Useful: Can Online Services be ‘Essential’ or are They Simply Very Useful?
This article explains how the ‘essential facilities’ concept in competition law can hardly be applied to online platforms. Online platforms can be very useful, but they are not ‘essential’ to other businesses.
As mentioned in James’ recent post on Europe’s policy debate around the concept of ‘platform neutrality’, online companies are increasingly becoming a discussion point for European regulators. Given the political nature of this process, it is more accurate to say that it’s primarily politicians who call for more regulation of online platforms. The best example is a common French-German letter signed by government representatives of both countries voicing concern about competition in the digital economy. The addressee of this letter is the new European political leadership gathered in the European Commission.
The interesting bit in this debate is how openly both governments mingle established concepts found in competition law with an ex ante regulatory agenda. Accordingly, the French-German letter talks about digital platforms becoming “essential intermediaries” or “gateways” for the provision of services and digital content. The two governments call on the Commission to organise a public consultation on the appropriate regulatory framework for “essential platforms”. The letter’s accompanying ‘non-paper’ goes into greater depth by pointing out that defining “essential” in congruence with the classic definition of an “essential facility” must be part of the debate.
The last point really is the crux of the matter. If one wants to define “essential” in congruence with the competition law concept of “essential facility”, one has to look at the Commission’s enforcement history and at the European Courts’ jurisprudence. Which online company could in theory meet this legal standard? Could the highly competitive and diversified online environment ever produce a digital essential facility?
First and foremost, the concept of an essential facility is principally used in vertical foreclosure cases. These have traditionally been refusal to supply cases in which a dominant company in an upstream market refuses to supply customers in a downstream market on which it also operates. Only after establishing that a company actually is dominant in an upstream market, European Courts employ an ‘indispensability test’. Under this test the key question is whether the dominant company’s product to which access is sought is indispensable to someone wishing to compete in the downstream market. (Note: EU Courts tend to use the word ‘indispensable’; ‘essential facility’ or ‘objectively necessary’ are alternative expressions used interchangeably in this piece).
It’s important to recall that the essential facilities doctrine is highly controversial among academics, economists and lawyers who largely agree that it should only be applied in extremely limited circumstances like to true natural monopolies. That is because forcing a dominant company to supply or deal with competitors could ultimately lead to less consumer welfare if it was to incentivise rivals to ‘free ride’ on the investment of the dominant company and hence disincentivise that company to create the facility in the first place. For that reason the Advocate General in his opinion in Oscar Bronner (discussed in greater detail below) concluded that “by retaining a facility for its own use a dominant undertaking retains an advantage over a competitor cannot justify requiring access to it” (para. 57).
For this reason the bar for a finding of indispensability has been set extremely high by the Court of Justice of the EU (CJEU). In Oscar Bronner a smaller Austrian publisher of a daily newspaper sought access to the only nationwide, highly sophisticated home-delivery distribution system run by Mediaprint, a much larger competitor. Mediaprint declined access leading to a legal battle up to the CJEU. The Court did not find an abuse of dominant position in Mediaprint’s behaviour because it did not consider the distribution network indispensable to Oscar Bronner’s business. Oscar Bronner could not demonstrate that the creation of an alternative distribution system is not a realistic alternative. It was not enough to argue that establishing a second system is not economically viable by reason of the small circulation of Bronner’s newspaper. This point is key: including Oscar Bronner’s newspaper in Mediaprint’s distribution network would have been useful or convenient, but under European law it was certainly not essential for Bronner to continue his business.
It is therefore of no surprise that in a lot of cases essential facilities were identified in relation to physical infrastructure like ports, rail networks and gas pipelines. Duplication of this kind of infrastructure is economically not viable and sometimes even physically impossible.
Crucially, even if an undertaking refuses to supply an indispensable input, this in itself does not amount to abuse under competition law. In Microsoft v Commission the General Court explained that the refusal to grant access to the essential facility must be liable, or likely, to eliminate effective competition in the downstream market. This point is also made in the Commission’s Guidance on Article 102 Enforcement Priorities (see para. 85). Accordingly, a true essential facility, i.e. a facility to which access must be granted (otherwise there would be an abuse under the law), has to pass both requirements: the indispensability test and it must be shown that denial of access is likely to eliminate all effective competition.
This careful and restrictive approach towards essential facilities makes sense. After all, even the owner of an essential facility is free to choose trading partners as long as effective competition is not foreclosed in downstream markets. This must signal caution to policymakers. The fact is that the concept is not easily transferable to the digital world. While the French and German governments do not specify which ‘digital gatekeepers’ they mean, it is fairly easy to predict that some of the most popular consumer brands would be targeted. Let’s discuss some hypothetical examples assuming that these companies have a dominant position in a given market for the purpose of this exercise — something which can never be taken for granted since every Article 102 TFEU enforcement action requires a deep and thorough market analysis.
Let’s start with Facebook. If one was to claim that Facebook constitutes an essential facility in the upstream market for social networking (which is ridiculous given the myriad of alternatives like YouTube, Twitter, WeChat, etc.), the question becomes in which downstream market(s) competition could be foreclosed because companies don’t have access to Facebook. Could Facebook for example foreclose competition in the market for social advertising and be forced to grant access to every social advertiser to maintain effective competition in social advertising? Obviously, the question is close to nonsensical given the inherently open nature of Facebook’s platform for all — users, corporates and advertisers. But even if one follows through the argument, Facebook would not even be forced to ‘supply’ every advertiser. That is because no one could claim that effective competition in social advertising is eliminated. Advertisers have alternatives like Twitter, Google, LinkedIn or significant country-based social networks like Nasza Klasa in Poland. For this reason Facebook could obviously not be considered to be indispensable to social advertisers in the first place.
What about Amazon? Amazon is clearly very useful for the distribution of e-books given its platform and the popularity of the Kindle e-book reader; but it is clearly not essential. Still, I wouldn’t put it past someone arguing that in order to compete in the downstream market for e-book sales, publishers need to have access to the Amazon platform which serves as a digital distribution system. This scenario would essentially be the digital, 21st century version of the Oscar Bronner case. There is no reason to assume that a competition authority or a court would decide the case differently. Amazon is not indispensable to any publisher because there is nothing that prevents publishers from building their own webpage and even the necessary hardware for readers of their content. Being on Amazon is surely useful, but not indispensable for selling e-books. Just like in the Facebook example, this discussion is very far away from reality because online distributors of content like Amazon have a very high incentive to not exclude anyone since they stand in competition with other platforms. Apple and Google are two high-profile examples exerting competitive pressure on Amazon’s e-book business — competitors that surely need to be taken seriously.
Google, currently facing a competition investigation, is another interesting example. Probably because of the ongoing investigation, no other online company has been subject to more forceful claims of being an essential facility or the “gateway” to the Internet. Leaving aside all other aspects in the ongoing investigation, let’s see in how far Google could be considered “essential” under competition law terms. Unsurprisingly, this doctrine runs into its limitations very quickly. If one was to define Google’s generic online (desktop) search engine as the indispensable facility, a legal requirement to have access is difficult to substantiate because no lawful website is excluded from the display of search results. Could one, more narrowly, define the top spots on the search listing on which Google increasingly competes with its own ‘vertical’ search services as an essential facility?
In light of European jurisprudence one would have to ask whether the top positions on Google’s search listing are indispensable for companies to carry on business in downstream markets like specific vertical search services. Secondly, would not having access to the top search results eliminate all effective competition in a related market? In the online world this can hardly be the case. Any online company has plenty of alternatives to attract customers including own webpages, mobile apps, social networks and even traditional TV and banner advertising. There are also other generic search engine providers that help in driving traffic. Being prominently displayed in Google’s search results is useful for everyone, but it’s not indispensable to any business in the meaning of Oscar Bronner. Also consumers have plenty of ways to get to the information they are looking for. It is worth noting that Bo Vesterdorf, former President of the European Court of First Instance (now the General Court), recently explained that under these circumstances EU competition law would not even place any obligation on a dominant search engine to deal with competitors in a particular way or to not favour its own services.
Let’s take a look at another example involving a (currently) less prominent European online platform. Delivery Hero was founded in Berlin four years ago and is now present on 10 European national markets. In Delivery Hero’s own words “visiting our [online] platforms is the easiest way to your favorite food”. Berlin-based e-commerce group Rocket Internet has just bought a 30% stake in the company valued at €496 million. Imagine Delivery Hero was to become the ‘dominant’, most popular platform for online food ordering. Could it be found to constitute an essential facility for restaurants competing in the downstream market of home delivery services in meals? Would Delivery Hero have to deal with all restaurants providing home delivery services? The answer must be a clear ‘no’ in light of established EU law. Delivery Hero’s hypothetical refusal to deal would neither eliminate all effective competition in home delivery services for meals nor would their platform be indispensable for the provision of these services. Again, it would be quite useful but not essential given there are alternatives. Remember the (annoying) flyer you had in your mailbox yesterday praising a given restaurant’s food and its amazing, below-30-minutes-waiting-time home delivery service?
All of these examples illustrate that in the digital environment one must be very careful to not equate ‘popularity’ with ‘indispensability’. Claims that the services of popular (and successful) online companies are ‘essential’ to the economic activity of others would certainly not pass the European Commission’s and the Courts’ standard for a finding of essential facility, even less so for establishing abuse under competition law. There are very good reasons for such a strict approach limiting the essential facility concept to exceptional circumstances. After all, competition policy is to encourage all companies to compete on the merits and to innovate, including to come up with popular platforms in the first place. This must caution policymakers to not hollow out established competition law enforcement with a short-sighted regulatory agenda. Anti-competitive refusal to deal scenarios have simply not materialized.
We have already described the problems with defining a ‘digital platform’ in a world in which the whole economy is becoming digital — imagine the problems with defining an ‘essential digital platform’. More importantly, do we really want policymakers to make that call for one of the most vibrant and innovative sectors of the economy?