Arcep speaks

Net neutrality : Act II - Intervention of Nicolas Curien, member of the executive board, for the DigiWorld Summit 2011, Montpellier (France) - 17 November 2011

Over the last five years, first in the United States and then in Europe and Asia, Net Neutrality has triggered a fierce debate among network operators, content providers, regulators, end-users and citizens. Act I of the play is now over. The ideological clash between libertarians demanding unconditional equivalence of all bytes of information, on the one hand, and reactionary telecom or cable companies demanding unfettered management of traffic on their networks, on the other hand, has receded in favour of a more reasoned discussion on some of the details of net neutrality: what are the minimum regulatory tools needed to ensure an acceptable level of net neutrality while giving network operators flexibility to innovate and manage their networks? Act I of the debate resembled a war of religion, each side rejecting out of hand the other's philosophy. Act II introduces a fragile era of tolerance and cohabitation, each protagonist still distrustful of the other, but admitting that the other side's point of view has to be taken into consideration. In Act II, market actors have understood that neutrality is not a binary topic and that the subject must be addressed in a collective and collaborative way for the sake of achieving economic and social efficiency.

The debate has progressively shifted and focused on several key issues that are essential to the design of a well-functioning neutrality: (i) traffic management, (ii) IP interconnection arrangements, (iii) transparency, (iv) price differentiation and markets for enhanced quality, and (v) the setting up a suitable regulatory framework. Such are the main stakes of "Net Neutrality: Act II" and the focus of the latest special issue of Communications & Strategies around which this seminar is organized.

(i) Traffic management, up to a certain degree, is necessary to ensure quality of service but it should not exceed what is strictly necessary to fulfil that objective. Regulators, such as the ARCEP in France, have proposed to apply proportionality and non-discrimination principles to determine when traffic management is acceptable, or on the contrary excessive. Application of those principles should deter anticompetitive behaviour, such as throttling the traffic from competitors in order to favour one's own managed services. And applying general principles seems far preferable to creating a white list or a black list of traffic management techniques. Such a list would rapidly become obsolete.

Furthermore, the application of technology neutral principles to traffic management should yield coherent results even in different network environments. Fixed and mobile networks have different constraints, but mobile networks should also abide by the same general principles. The application of those principles should take due account of different constraints such as spectrum scarcity for mobile networks; the latter are currently less neutral than fixed networks and the scarcity of spectrum might not be the only explanation for the unavailability of some Internet applications on smartphones, such as VoIP or instant messaging.

Last, but not least, operators should not consider traffic management as a substitute for investment in NGAs. The two must go together. Investment is the only efficient long-term solution to the reduction of congestion and thus to the preservation of neutrality, which leads to the important question of how operators will finance their NGA build-out.

(ii) The IP interconnection market affects the quality of data delivery to end users and thus can impact neutrality. This market is changing fast because of the explosive growth of data flows, especially video, and the emergence of strongly asymmetric traffic patterns between operators. Consequently, so-called free peering agreements are increasingly challenged and replaced by paid peering (especially with ISPs). In addition, content delivery networks (CDNs) - either independent CDNs or CDNs self-provided by large content providers or ISPs - have become indispensable actors in the wholesale market, particularly for distribution of video content. But CDNs raise new challenges regarding IP interconnection. While ISPs and CDNs have often peered in the past to avoid bearing transit costs, they may soon compete.

The geometry of IP interconnection - around both transit and peering - is thus moving fast, giving rise to conflicts between ISPs and upstream operators which may affect quality of service and neutrality in the downstream market. Alternative solutions to these conflicts arise, including the possible payment of a data termination rate. This suggestion might be seen as generalization of paid peering with ISPs.

Nevertheless, the IP interconnection market has so far been deemed fully competitive and has therefore remained unregulated and it should probably remain so. This does not mean, however, that the wholesale market for IP traffic exchange should not be monitored and occasionally subject to soft guidance by regulators.

(iii) Transparency is put forward by regulators around the world as an essential condition to ensure neutrality, in particular by regulators in Europe. The argument is that if the end users are fully aware of traffic management practices or other restrictions by their Internet access provider, the consumer will switch to another provider if he or she is unhappy. Although necessary, transparency is by no means sufficient to ensure neutrality. Even in a competitive market, switching costs may prove significant, and even well-informed consumers will remain to some extent captive of their Internet access provider.

Moreover, where there are market rigidities due to switching costs, ISPs may be tempted to use transparency as a convenient safe-harbor to legitimize all forms of traffic limitations. Consequently, transparency should not be understood only as a requirement that ISPs communicate their practices to "passive" consumers but also as promoting - possibly through regulatory intervention - the provision of toolkits that allow a proactive consumer, a "prosumer", to measure by himself the quality of his access. The participation of end-users in the regulation of neutrality appears as an important factor for its success.

(iv) Price and quality differentiation remains one of the trickiest issues in net neutrality, because the very concept of neutrality was first defined by economists as the combination of a "zero-pricing rule", stating that in the two-sided market of Internet access the content provision side should not be billed, and a "non-discrimination rule", stating that ISPs should not differentiate the quality of service they offer to content providers and to end users. The evolution of the net neutrality debate shows, however, that this theoretical view should be seen as an ideal reference point rather than a normative prescription, and that these two "golden rules" may be at least partially relaxed without undermining the principle of neutrality.

Differentiation, i.e. efficient discrimination in the economic meaning of the term, remains consistent with neutrality as long as everyone, providers as well as users, may enjoy a basic access to the open Internet with a guaranteed quality without paying any mark-up. Some may choose better quality supplied at premium price and this differentiation should not be problematic. The difficulty here is to define what a basic access is, what a guaranteed quality is, and how it might be monitored and enforced through regulation or market dynamics. Were the operators in a position to set freely - upstream and downstream - their levels of quality and their prices, then the Internet content providers and users would face the serious risk of an eviction effect, the standard open Internet gradually vanishing to the benefit of premium and clustered Internets. On the other hand, price/quality differentiation may help operators to generate additional revenues to invest in NGA infrastructure, potentially benefiting all users. Numerous paid options are already being offered in the enterprise markets, with some QoS differentiation.

(v) Regulatory approaches to net neutrality vary from region to region and from country to country: regulators are more or less active in this field and they are also more or less successful in their attempts, as the repeated jurisdictional roadblocks encountered by the FCC have shown. Despite such variations, we see convergence among regulators regarding the goals of regulation: free and universal access to lawful content, no harmful traffic management, no anticompetitive price and quality discrimination, and transparency. These are the core principles - discussed above - that form the common base of all regulatory frameworks. The question raised now is how all these desirable objectives can best be reached? Is it through strong ex ante regulation or is it rather through market discipline with vigorous ex post enforcement of competition law?

The second alternative sounds both attractive and sensible to the ears of any economist believing that competition - even imperfect - still remains a useful guide towards efficiency and consumers' welfare maximization. But ex ante regulatory action will still stand as a useful complement to competition law, provided the regulation is performed in a participative and "maieutic" way rather than in a prescriptive top-down manner. The job of regulators is definitely not to impose on market players detailed rules on how to manage or finance their networks. For one thing, regulators lack the necessary information and would likely get many of the details wrong. By contrast, the regulators' job is to engage with all the stakeholders in multilateral discussions and negotiations, from which appropriate solutions will spontaneously emerge.

* * *

So what's next? What about "Net neutrality: Act III"? Is net neutrality a classical comedy or tragedy? In the optimistic case where the net neutrality story is a comedy, then stakeholders will all sing together in Act III their successful achievements in Act II. Network operators will have entered into profitable deals with one other and with content providers, significant investments will have been made in NGN and NGA infrastructure, severe IP congestion will have been avoided, prosumers and their lobbying groups will have contributed to the quality of their Internet access, to their liberty of expression, to their freedom of innovation on the Web and to the protection of their privacy. Net neutrality will have been dynamically preserved and enhanced through the responsible and collaborative acting of all parties, in a kind of Commedia dell'arte with more room being given to creative improvisation than to normative regulation.

But in the pessimistic and we hope unlikely case where Act II turns the net neutrality play into a tragedy, then Act III will take place in a world where underinvestment will have caused severe inroads to net neutrality, where a few vertically integrated Internet or telecom giants will have taken control over most aspects of the Internet, where captive and passive consumers will have been deprived of universal access to all content, where impotent regulators will have failed to give players the right incentives. At that point, very little could be done to restore efficiency…

Instead of such a tragic outcome, the editors of this issue "have a dream": in the continuation of Act I, where strong initial conflict has finally morphed into the recognition by stakeholders of several critical issues to be solved by negotiation, we may reasonably bet that the outcome of Act II will consist in some satisfactory compromises, and that Act III will constitute the happy end…

This text was first published as an introduction to Communications&Stratégies n°84, special edition on Net neutrality, co-edited with Winston Maxwell and Vincent Bonneau.

Linked documents