Winston Maxwell and Nicolas Curien explore whether competition law can provide help in dealing with net neutrality abuses, such as the situation where an ISP favourably treated certain content.

An important question in the European Net neutrality debate is whether existing competition law tools can deal with possible net neutrality abuses, and in particular certain forms of anti-competitive discrimination by an ISP in the treatment of content from upstream content or application providers

To answer this question, let us take a hypothetical example of an Internet Access Provider (IAP) that discriminates by providing a better quality of service to certain content providers -- perhaps to its own affiliated companies -- than to others. In our fictional case study an IAP ensures that its own video sharing service has a better quality of service than an unaffiliated video sharing service. During times of congestion the IAP's own video content service is easier to access than the unaffiliated site, whereas during times when the network is not congested, the difference in quality of service would not be noticeable by the end-user, best efforts being in that case as good as the premium service. This sort of discrimination in quality of service (QoS) has not so far occurred on the market, but it is one of the forms of abuse feared by net neutrality advocates.

Would this discrimination constitute an abuse of dominance?

To prove an abuse of dominance, two elements are necessary: a dominant position, and an abuse. In our example, the abuse will be relatively easy to show, because discriminating in favour of one's own affiliated entity at the expense of unaffiliated service providers is a classic form of abuse, well known in the telecommunications industry. The abuse would exist only if the IAP applied discriminatory terms to similarly situated entities. In other words, if the IAP offered the premium service to its own affiliated site and to other unaffiliated sites on a non-discriminatory basis, there would not be a manifest abuse, absent other factors.

Is there dominance? Let us assume in our example that the discriminatory practices of the IAP constitute an abuse. It would then be necessary to determine whether the IAP holds a dominant position. If it does not, the unilateral discriminatory practices would not violate competition law, at least under the "abuse of dominance" angle. In order to determine whether a given IAP holds a dominant position, we first have to define the relevant market. There are several possible markets to choose from.

  • If we look first at the retail market for Internet access, we will find that in most European countries no single IAP will hold a dominant position. In most European markets, alternative operators have been able to enter the market thanks to unbundled access to the local loop (ULL) and bitstream access. These alternative operators have in turn offered competing Internet access products on the retail market. Consequently on the retail market it may be difficult to show that a single IAP holds a dominant position.
  • If we now look at the wholesale market, the relevant market would be the market for exchange of Internet traffic. As we shall examine now, there are several ways to look at this market.

Do IAPs control a termination bottleneck? Some have argued that the IAP controls a form of termination bottleneck, similar to the bottleneck that exists for the termination of voice calls, each telephone operator being deemed to hold a dominant position over termination of voice calls to subscribers on its own network. The reason is that an operator (the "calling party's network") with a call that needs to be completed to a given telephone number has no choice but to address itself to the network (the "called party's network") controlling that number and subscriber. There is no physical substitute for the services of the called party's network, which thus controls a form of "essential facility." Moreover, the terminating network can set whatever price it wants for the service, because it is under no competitive pressure from its own retail customers to set reasonable prices at the wholesale level. The retail customer receiving the call does not know or have any reason to care about the wholesale price charged by his operator to other operators to terminate the call. The entire charge for the call is borne by the calling party's network, who generally passes the cost on to the calling party. Because the person bearing the cost is not a customer of the called party's network, the latter has no incentive to set reasonable termination rates. On the contrary, by setting high rates on the wholesale market, it can subsidize low retail tariffs for its own customers. Because of the called party's network dominant position, NRAs throughout Europe have imposed price caps on voice termination services, using the methodology for asymmetric regulation referred to above based on "SMP" findings.

Could it be argued that an IAP holds a similar form of monopoly for the termination of Internet traffic towards its own retail Internet customers? Like telephone calls, Internet content has to be routed through the IAP's access network in order to reach the retail Internet customer connected to the IAP's network. The IAP holds a form of physical bottleneck similar to the termination bottleneck of telephone operators. However there is an important difference between a termination of telephone calls and the termination of Internet traffic.

  • In the case of telephone calls, the person receiving the call is insensitive to the prices or other terms imposed by his or her operator on the wholesale market. If the telephone operator increases its prices for the wholesale termination of voice calls, fewer people may eventually call the operator's customer, but the customer may not even be aware of this because the customer does not know about calls he or she never receives.
  • For Internet traffic, the situation is different because in each case the retail Internet user places the call, i.e. he or she makes a request for certain content. If the content is not immediately available or is available with a degraded quality, this would be immediately visible to the customer having made the request.

Consequently, unlike the situation for voice telephony, the terminating operator for Internet traffic is constrained by his own retail customers in how he behaves on the wholesale market. The IAP will not therefore be able to block or degrade traffic without this immediately being known by his own customers and perceived as a degradation of service quality. Thus the IAP, unlike the telephone operator, cannot behave on the wholesale market independently of his competitors and of customers. A finding of market dominance on this market therefore is not obvious.

Since it appears difficult to find dominance based on the terminating network bottleneck theory, there may be other approaches, particularly if competition on the retail market is weak. In that case, it might be possible to project a strong retail market share (50% or more) onto the wholesale market and find dominance through that method. However, as noted above, in many European markets, retail competition is robust, making this method also uncertain.

Discrimination resulting from vertical agreements

The second angle under which we can examine possible problems of net neutrality is as a form of anti-competitive vertical agreement. To illustrate this path of analysis, let us take the fictional example of an agreement between an IAP and a popular video sharing site, pursuant to which the IAP and the video sharing site agree to develop together a premium service permitting subscribers of the IAP to share and watch 3D videos with an enhanced quality of service. In our hypothetical case study, the agreement is exclusive, which means that the IAP commits not to make the enhanced 3D quality of service available to another competing video sharing platform, and likewise the video sharing platform commits not to enter into similar agreements with other competing IAPs. The purported reason for the exclusivity is to protect the initial investment and advertising that will accompany the launch of the new 3D service.

Vertical agreements require a case by case analysis. There is clearly a discrimination in that the IAP has agreed not to grant the enhanced quality of service to another video sharing platform. For the purposes of our analysis we will assume that the IAP does not hold a dominant position for the reasons mentioned in the preceding section. The question therefore is whether the agreement is anticompetitive under Article 101# TFEU. There does not seem to be any reason why it would be considered illegal per se. Indeed, a vertical exclusivity of this kind appears similar to the case of an IAP entering into an exclusivity agreement for certain premium content, or an exclusivity agreement regarding a certain kind of smartphone. Nevertheless, the exclusivity and the discrimination that it creates carry significant competitive risks. Even if the IAP does not occupy a dominant position, the vertical agreement regarding the distribution of 3D videos would restrict competition in two ways.

  • First, any competing video sharing site would suffer from a significant quality disadvantage for a large proportion of potential users, i.e. all those Internet users who are customers of the IAP having entered into the exclusivity agreement.
  • Second, the exclusivity would restrict the competition between IAPs, particularly if the video sharing site having entered into the agreement is a popular site. In that case, customers may choose the IAP having signed the exclusivity not because it offers better broadband access services at better prices, but simply because it gives access to the popular 3D video sharing platform whereas the other IAPs do not.

Consequently, there are significant restrictions of competition both at the upstream level among video sharing platforms and at the downstream level among IAPs. The exclusivity agreement may however have positive effects that offset these competitive restrictions. One such positive effect would be that the agreement permitted the parties to make the investment both in technology and in advertising necessary to launch a truly new product. If enhanced video sharing for 3D films is truly a new product that may not have been brought to market as quickly in the absence of the exclusivity agreement, then the benefits to consumers may outweigh the restrictions to competition. Such a balancing can only be effected on a case by case basis. When analyzing vertical exclusive agreements between IAPs and content providers, the French competition authority indicated that the exclusivities would be acceptable if they are for a short duration in time and are limited to new innovative forms of distribution. In other situations, the competition authority suggested that an exclusive agreement between an IAP and a content provider would be problematic.

This article is an excerpt of a longer article published in the Revue “Concurrences”.