In light of today’s Federal Communications Commission’s network neutrality proposals, I thought I’d look again at what I wrote in 2011 in Competition Law and Regulation of Technology Markets comparing EU and US telecoms regulation. Given that Chairman Wheeler has indicated that reclassification as Title II will not include unbundling requirements, the point I make below about the network neutrality debate in the US being conducted in behavioural rather than structural terms looks likely to remain true.
8.341 The US debate over whether cable or DSL service should be a telecoms or an internet service and the resulting regulatory consequences is necessary given the wording of the 1996 Telecoms Act, but seems metaphysical at best. The chokepoint, the point of the regulation existing in the first place, is the competitive problem caused by the cost of laying and maintaining local infrastructure. But the US regulation is not focused on local infrastructure, or issues of market power, but on distinctions between telecoms and information services, and the severability of one from the other. At the time of the passing of the 1996 Act, this distinction broadly made sense: the distinction between telecommunications and information services corresponded broadly to the distinction between the core activities of local infrastructure operators, and other services provided over that local infrastructure. The development of broadband services put the distinction under strain: although a simple movement of bits would seem to fit acceptably under the definition of a telecommunications service, this is not the route taken by the Federal Communications Commission.
8.342 Although there is more discontent in the US about the overall state of the telecoms market, and of the structure of telecoms regulation, there as yet appears no appetite for either a reversal of the 2002–2005 deregulation of broadband or a re-examination of the basic principles of the 1996 Telecoms Act. Discontent has found more of an outlet through the network neutrality debate.
8.343 Network neutrality The net neutrality debate continues many of the same concerns about the consequences of natural monopoly infrastructure providers, and it is likely no coincidence that calls for net neutrality regulation first began as the US FCC began to move away from network access regulation.
8.344 Terminating markets and net neutrality Conceptually, network neutrality is similar to an after-market type problem in antitrust (itself a controversial issue), and the terminating market problem in telecoms. However, one of the difficulties with this comparison is that the network neutrality debate covers a range of factual circumstances, some of which look like the termination problem, others which do not.
8.345 In telecoms call termination, customer A wants to call customer B, and customer C is not substitutable. Looking at the two examples above, this looks like the first, the blocking of VoIP services. This is not necessarily only a problem for voice services—there are a range of other any-to-any services on the internet—but the incentives are perhaps clearest here.
8.346 The second example cited is not a termination problem. A web content provider such as an online bookstore wants to access customers, but is, absent more, neutral as to whether it is customer B or customer C that is reached. This is not an issue of every terminator being a monopolist in respect of its own customers, but only that if one terminator gets market power on the overall customer market, that may lead to foreclosure of non-preferred bookstores.
8.347 In some cases, however, net neutrality is looking to cover a terminating monopoly problem, access to a particular customer is required: email, instant messaging, voice over IP; all are examples of any-to-any communications over the internet. There is an any-to-any imperative in communications, that, say, the provision of a search engine to potential customers, or the sale of books online, lacks. It is the difference between an economy of scale (more customers may reduce average costs), and a network effect (more customers may increase demand—or more pertinently, the loss of even a few customers might destroy it).
8.348 The internet as an any-to-any design The more general argument in favour of network neutrality is that the internet was originally designed as an any-to-any network, with the network being blind to the content, only ensuring this end-to-end nature. Network neutrality therefore, is an attempt to maintain this design decision in an environment where, some fear, individual networks may try to interfere with it, to the detriment of the network as a whole. Rather than being a concern flowing from the structure of the market (as in a classic termination case), this is a concern flowing from the intended design of the internet which brings benefits that proponents wish to preserve.
8.349 Behavioural and structural remedies It is a commonplace in antitrust discussions that structural remedies are preferred over behavioural ones. A typical structural remedy would be a divestiture of an overlapping activity in a merger, but structural remedies are not limited to simple divestiture. Although there is no generally accepted definition of a structural remedy, it is generally conceived as one which changes the structure of the market, in such a way that the market will operate to avoid the harm to consumer welfare that was being feared. This contrasts with a behavioural remedy whereby the—presumably dominant—company in effect promises to behave in a particular way, and remains subject to regulatory oversight of some form to ensure it keeps its promise.
8.350 The net neutrality debate in the US is interesting, because it is being conducted almost entirely in behavioural, not structural, terms. If the focus were instead to shift to a search for a structural solution, the likely obvious answer would be to re-open the network access debate that the FCC closed in 2005. Although there are some calls for such an approach, they are relatively muted.
8.351 Access and interconnection Just as US network regulation does not appear to have learned substantially from antitrust experience, so antitrust—in both the EU and the US—could usefully learn, in at least one point, from network regulation. One factor common to both EU and US network regulation is the imposition of interconnection requirements on all operators, and the distinction both regimes draw between access and interconnection.
8.352 EU telecoms regulation will mandate access on a ladder of investment argument to several elements of a telecom’s operator’s operations if those elements are not commercially replicable given the state of market entry. US telecoms regulation would formerly mandate access on similar grounds, but now mandates access only to the natural monopoly element. Similarly EU telecoms regulation will mandate access to new infrastructure—such as local fibre—if further local fibre investment is not commercially feasible. Both EU and US telecoms regulation mandate interconnection between voice networks for all operators, dominant or not; US telecoms regulation will only price regulate incumbent local exchange interconnection (roughly on operators with market power in the primary market). EU telecoms regulation regards termination as a market where every operator that controls access to an end-user is a monopolist, and therefore price regulates all termination, both fixed and mobile.
8.353 EU antitrust will mandate access to a facility if access is essential for a new product, or more generally if access is necessary to maintain consumer welfare in the long term; although previous cases have always involved supply to competitors, this may not be a precondition. US antitrust will likely only mandate access to a facility if there has been previous voluntary supply in a non-regulated area, regardless of any specific investigation into consumer welfare. To the extent that there could be an obligation to supply in other circumstances, which is unlikely, it only extends to competitors.
8.354 EU antitrust law on refusal to supply is therefore more interventionist than US antitrust law, though it is important not to overstate the level of intervention, as cases remain relatively rare. EU antitrust on refusal to supply is broadly analogous to EU and US telecoms law on mandated access to facilities (even though the EU and the US implementation of the telecoms laws currently differ on which facilities should be accessible). US antitrust law on refusal to supply is notably less interventionist than both EU antitrust law and EU and US telecoms regulation (even leaving to one side Trinko’s disapplication of the law in regulated sectors).
8.355 Significantly, however, neither EU nor US antitrust distinguishes between access and interconnection as different types of refusals to deal, with significantly different consequences. Learning from network regulation, antitrust should distinguish between (1) refusals to supply access to a non-rivalrous interconnection in a network industry, where the primary effect of the access would be to increase the value of both networks, and (2) refusals to supply access to rivalrous infrastructure where the primary effect of the access would be to deprive the infrastructure owner of its ability to use that infrastructure for its own purposes. Either may be abusive—at least within the definition of abuse used in the EU—but the first should logically be more easily seen to be abusive under the antitrust rules, just as the first is more readily imposed in both the network regulation of the EU and the US.