How Europe formulates internet policy

The internet has become so crucial to our knowledge-based economy that its preservation and operation is now an issue through which we see our political, social and economic freedoms. We no longer see ourselves as truly free unless we have both the ability to access the internet and the knowledge that the content circulating on it is not restricted. But the internet as we know and enjoy today is a result of historical circumstances, technological and social choices and policy making. We should have no reason to assume that it will remain unchanged unless further policy choices are taken to preserve it. This firm belief in the internet as the ultimate area of freedom has had its roots in the 1990s where laws in cyberspace were uncertain and chaotic and, the belief that the internet could or should not be regulated, widespread. But the internet today is a very different space – more regulated, more complex, more confusing and more important.

The EU is today seen as a ‘digital laggard’ which relies on divergent national regulation and whose digital policies lack coherence.1 In this article we will attempt to see how telecommunications policy and its dependance on the ‘Single Market’ framework contributed to that picture. We will provide a brief overview of EU policy making in the area of telecommunications, attempt to explain the Single Market background of EU internet regulation and look at present telecommunications policy in its potential for the future. Our aim is to clarify the origins of EU telecommunications policy in order to explain its more recent transformations.

There is a temptation to look at internet regulation and policy as applying to one distinct reality. Since the internet resides in our mental image as a single phenomenon, nothing could be more logical than to think that regulating it means regulating one entity. There are two reasons why this may not be possible today.

The first has to do with internet architecture. The Internet has since its inception been split between three technological layers2. The first of these is the telecommunications or the carrier layer, which conveys the signal. This is the physical level through which the communication travels - the wires, the hardware and the infrastructure that enables the internet. The second, logical layer, is the software and protocols that enables the hardware to operate. The third is the content layer, which carries the substance. Each layer can and is regulated separately, with its own distinct legal regime3. What we today call telecommunications law covers the first and some of the second, while other regulatory disciplines (such as copyright, privacy or e-commerce laws) cover most of the second and the third. Moreover, internet regulation is to a large extent a result of legacy regulation inherited from separate activities dating to well before the emergence of the internet. The regulatory structure of the internet is also layered - largely split between one group of laws that regulate the carrier side of the internet and the other which regulates the rest. There is, however, significant evidence today of convergence between information technology services, telecommunication services and media services4. This coming together may force a rethink of the layered regulatory structure’s appropriateness for the internet.

The second reason is that a single coherent EU “internet” policy did not, until fairly recently, exist at the EU level. Instead of that, one could find a number of individual policies scattered in various documents on both the content and carrier aspects of the internet5. These are, without exception, part of the EU Single Market drive. The closest the EU ever came to a coherent vision is the Digital Agenda 2020, dating back to 20106. This diversity has significant consequences for thinking about internet policy making. Internet policy is normally not created from one centre and with one vision. It is at least a subfield of either mass media (print or broadcast) or utilities (telecoms, radio, cable and satellite) and possibly many more7. In fact, it is reasonable to doubt whether the internet has the potential of being subject to a single policy at all8. In the EU, this is true on a rather fundamental level – different policy making efforts have taken place at different directorates. These have until recently proceeded with little coordination, without an overarching vision and have frequently had conflicting aims.

In the following sections we will look at how one of the layered regulatory regimes - that of legacy telecommunications regulation - informed, by necessity, a significant part of internet regulation and how the Single Market agenda repeatedly coloured not only that very telecommunications regime, but general internet policy making of both the carrier and the content layers. We will then, in the final part, attempt to suggest how some of these limitations may be overcome in future internet policy making.

A brief history of EU telecommunications policy making

In Europe, telecommunications have historically belonged to national monopolies which had almost complete and universal control over both the equipment and the services provided on it9 In 1987, the European Commission realised that American customers had access to better and cheaper services and introduced an initiative to inject more competition into the European telecommunications sector. This was the 1987 Green Paper10. The authors had liberalisation in mind, hoping that increased competition would lead to a gradual breaking up of national monopolies. It called for the gradual but complete opening of the terminal market to competition, freedom of access to services from any connection point and “complete separation of regulatory and operational functions” (p.17). Directives soon followed which started the liberalisation process. First voice telephony in 1988 with the Terminal Equipment Liberalisation Directive11, then with services other than voice in the Liberalisation Services Directive12 and finally, with a directive liberalising the infrastructure13. The main purpose of the said directives was the removal of privileges granted to local operators. The legal basis chosen for the directives above was the Article 106 TFEU (ex Article 86 TEC), a hitherto unused legal basis which talked about services of general economic interest. The European Commission’s intervention would have forced derogations for such services to be limited to specific services only and to be made accessible to all without discrimination14. Several member states challenged the Commission’s authority to act on various grounds, including on a legal basis. They argued, essentially, that the Commission lacked competence to act in the manner in which it acted. The European Court of Justice (as it then was), however, confirmed the Commission’s authority to act, opening up avenues for more profound changes15.

Member states were not readily willing to abandon their monopolies and were by and large not supportive of the Commission’s initiative. The late 1980s and early 1990s saw an increasingly complex web of proposed and adopted directives, a tug of war between the Commission and the member states, of governmental and non-governmental actors. Although member states maintained their resistance, the balance of power slowly shifted from national monopolies to the Commission16. The full liberalisation came in 1998 with the obligation imposed on national governments to relax entry into their telecommunications markets. The 1998 package comprised two types of directives: liberalisation and harmonisation. The former had the task of eliminating national privileges. The latter wanted to secure the EU-wide implementation of telecommunications policy.

The post-1998 period saw the price of calls throughout the EU fall, the range of products increase, the efficiency improve and mobile and broadband growth stabilise17.

However, although overall efficiency improved, the results in the broadband market were less impressive with penetration remaining low and price reductions remaining relatively insignificant. The next reform, the 2002 package, had less ambitious goals than the Green Paper or the 1998 reform, but it sought to address, among other things, the broadband problems outlined above. More significantly, however, it sought to adapt the then existing framework, where voice, data or audiovisual signals were all transmitted through different “pipes”, to the changing needs of the modern internet - where all three use only one carrier. This convergence was to be brought into the multitude of directives adopted under the liberalisation and harmonisation initiatives of the 1990s. Coherence was also needed since the above-mentioned harmonisation directives were enacted by the Parliament and the Council and had Article 114 TFEU as a legal basis. Their main aim was to improve the Single Market in the telecommunications area. In other words, they sought to prevent states from putting obstacles to foreign telecoms services. Liberalisation measures, on the other hand, were based on competition objectives. Put differently, they sought to improve competition between private actors. Due to this difference in aim, they had different legal bases and had (for procedural reasons) the Commission as the initiator.

The 2009 package is the latest and currently valid set of telecommunication measures18. It included, among other things, new consumer rights, measures to increase broadband access and the establishment of a new European authority. Among its most relevant contributions are convergence measures between fixed and mobile as well as convergence between telecoms, broadcasting and information technologies.  In the 2009 package, transport is treated in the same manner, irrespective of which sector it comes from. Content had been completely separated – the package not applying to content at all.

The three significant policy packages outlined above – the post-1987, the 1998 and the 2002, had a relatively vague notion of the direction in which the internet (from the carrier perspective rather than the content side19) should develop. Although it is true that increased competition and cheaper and faster internet for everyone remained as themes in these packages, a more coherent vision for the internet was absent. In addition, the perception (shared by the companies and the public alike) was prevalent that Europe was still a patchwork of regulatory regimes and that a true Digital Single Market was absent20.

This lack of a coherent policy was apparent through two controversial measures, both subject to intense lobbying and both suffering numerous drafting changes as a result of this lobbying21.

The first controversy arose out of the desire of network operators to introduce traffic management on the internet, thus increasing control over data flow and opening the possibilities for differentiated pricing. The problem, today commonly framed as ‘net neutrality’, arises out of the fact that the internet was originally built ‘neutral’, without discrimination between different data packages passing on it based on user, content, platform, etc. The traffic management proposed here would include managing access speed and throttling bandwidth-hungry services, thus eroding neutrality. The telecommunications operators were openly against net neutrality. The 2002 Universal Services Directive was eventually amended to include obligation to inform end users of any limitations imposed on access or distribution possibilities in Articles 20 and 21 but also minimum quality of service guarantee in Article 22. At the same time, the EU Digital Agenda committed to maintaining an open and neutral character of the internet.These provisions fall short of providing full net neutrality protection while maintaining a semblance of basic protection. The targeting by lobby groups resulted in a ‘language and structure’ that suited both agendas and which enabled operators to impose restrictions22.

The second controversy relates to the content regulation pushed through a carrier regulatory framework. It was an attempt by rightholders to introduce the so-called graduated response to perceived copyright violations. This would include the so-called ‘three strikes’ system, which would cut off access for repeated copyright violators after a series of warnings had been issued. The proposal was met with resistance by Parliament, which attempted to introduce (in the form of Amendment 138[xxii]) an obligation that no restriction to “fundamental rights and freedoms of end-users” could be imposed “without a prior ruling by the judicial authorities.” The final solution was a compromise. It did allow end-users’ access to be restricted but under strictly controlled conditions. The so-called ‘freedom provision’ (contained in Article 1(3)a of the new 2002 Framework Directive) introduced protection against graduated response, included proportionality, procedural safeguards (including judicial protection) and respect of the right to privacy23.

This apparent confusion in two critical issues – one affecting the intellectual property model for the digital world and the other network openness – is not accidental. Neither is the attempt to address a copyright enforcement (a content layer issue) through what is essentially a telecommunications (carrier layer) law. Rather than that, it is a symptom of a more serious lack of vision and vulnerability to pressures from legacy networks and content providers. Furthermore, as Comcast’s dispute with BitTorrent demonstrates24, the convergence of the content and carrier layers has blurred interests and confused the lawmakers. Large internet service provider (ISP) providers have an interest in throttling peer-to-peer traffic where they feel their own content is suffering as a result. Such providers push for restrictive telecommunications regulation (the carrier layer) in order to affect a content issue. In recent years, however, large network operators have resorted to practical solutions such as storing popular content closer to end-users, demonstrating the potential of self-regulation and cooperation25.

Telecommunications as part of the Digital Single Market

The most recent proposal for the reform comes in the form of Regulation for the EU Single Market on electronic communications (2013 Connected Continent Regulation)26. A casual look at the opening passages of the proposal shows the Commission’s desire at simplifying the rules for telecom operators, removing roaming and international call premiums and introducing legal protection for the “open internet (net neutrality).” The core of the Commission’s effort, however, goes deeper and is revealed in the three “outstanding integration challenges” which the present telecommunications regime still needs to address27. The first is to remove “unnecessary obstacles in the authorisation regime and in the rules applying to service provision” thus enabling multi-country operation. This effectively removes the previous regime of national authorisations replacing it with a single authorisation. The second is to improve the situation for “accessing essential inputs” regarding “predictable assignment conditions and coordinated timeframes to access spectrum for wireless broadband across the EU.” In plain terms, this is an attempt at spectrum policy harmonisation. It aims to improve coordination but could slow things down as countries that have rolled out 4G networks faster are encouraged to slow down while waiting for the slower ones to catch up. The third aim is to increase consumer protection and  “common commercial conditions in this respect, including the persistent problems of mobile roaming surcharges and of access to the open internet.” This is a package of consumer-protection measures that aims at improving general conditions, operator changes, roaming and similar.

One of the declared aims of the Proposal was to protect network neutrality. In order to do so, discriminatory blocking and throttling is prohibited and traffic management measures need to be “non-discriminatory, proportionate and transparent28.” The Proposal allows operators to directly differentiate their offers by speed and provide ‘enhanced’ quality of service. The neutrality proclaimed in paragraph 5 of Article 23 is directly cancelled out by what is said in paragraph 2. Paragraph 5 essentially provides that services cannot be blocked, slowed down or degraded based on content, application or services. Paragraph 2, however, allows end users to conclude contracts with the providers for the specialised services with the “enhanced quality”. In order to deliver these, “providers of electronic communications” shall be free to enter into contracts with “providers of content, applications and services.” This would enable them to “transmit the related data volumes or traffic as specialised services with a defined quality of service or dedicated capacity.” The ISPs as carriers are entitled to enter into contracts with content providers and request higher fees for more demanding contents and applications. The provision of these ‘specialised services’ must not impair the operation of the ‘general’ internet.

Effectively, this is the end of network neutrality, which, in simple terms, means precisely that there can be no difference between general and special services29. The explanation the Commission provided was that “such offers will enable telecom operators to generate additional revenue streams from [...], content providers as well as from consumers who are willing to pay for better or faster services. These revenues in turn, will enable operators to finance investments into network upgrades and expansion.”30 This is a variation of the ‘investment incentives’ argument which ISPs often use, claiming that net neutrality regulation (i.e. protecting net neutrality) will have a negative impact on their incentive to invest. The logic upon which this assumption rests has been called into question31.

In a press release on September 11, 201332, the Commission calls the Proposal the “most ambitious plan in 26 years of telecoms market reform.” This is a surprising statement, even taking into account the introduction of a single authorisation regime and the confusing attempt to “protect network neutrality.” Reducing and eventually eliminating mobile roaming charges, a substantially publicised move, although popular, is not critically improving the situation. One obvious step, introducing a single EU telecoms regulator, had not been taken although it has repeatedly been called for. In fact, a low-scale conflict arose when the Directorate General (DG) Competition called the current proposal unambitious33.

Is the Proposal really as ambitious as the Commission claims? Three policy options were originally considered. The first was regulatory coordination through review of the present instruments. This would involve a relatively insignificant intervention. The second was a set of targeted measures with improved EU coordination and this is the option eventually chosen in the Proposal. Option 3 was the same as option 2 in substance but suggested a different governance structure through a single EU telecommunications regulator. Different stakeholders supported different options during the drafting process34 – making it all too apparent that ‘Single Market’ has many meanings, depending on who’s talking. The incumbent operators wanted a far-reaching change in fixed and mobile networks and deregulation. Consumers and their organisations were focused on removing costs in roaming but also on the prevention of blocking and throttling of services. Larger-scale industry users wanted high-quality connectivity, more broadband coverage and less congested networks.

The first option would not have achieved more than a cosmetic change and would, as such, have been largely unnecessary. The third option would have gone a long way towards creating a true European Single Market in telecommunications. A number of actors including most member states, however, were sceptical or openly antagonistic towards the cornerstone of the third option - the introduction of a single EU telecommunications regulator. In such a climate, the second option seemed the logical way forward.

The content side of that Single Market rests on relatively stable grounds. The Electronic Commerce Directive35, together with the Copyright Directive36 and Data Protection Directive37 form the three cornerstones of internet regulation in the EU. Whereas the latter two only indirectly focus on the Single Market, the first concerns itself with it directly. These directives have provided a stable basis for provision of information society services in the European Union. In spite of this, Michel Barnier, the European Commissioner for the Internal Market and Services, called the Digital Single Market a “new frontier” in a 2013 speech38. He produced an unusually long wish list that remains unachieved as of 2013: the consumer ability to compare products from all over Europe, the ability to access digital entertainment from across Europe, easy access to “free, independent and multiple” sources of information, easy and convenient access to digital banking and other financial products, the ability of small and medium enterprises (SMEs) to go online, the ability of entrepreneurs to avail themselves of the internet’s many opportunities and, efficient public administration. The list ends with a statement that the Commission “has undertaken a major work programme to achieve the Digital Single Market.” This statement has to be seen in its wider context – it relates to the ‘content’ side of the internet as much as it relates to the ‘carrier’ one. But how is it possible that 20 odd years of intensive work on both the content and the carrier side of the problem resulted only in an “ambitious plan”, as the Commission points out in the quoted press release, or in a “new frontier”, as Commissioner Bernier would say? Was the second option that eventually became the 2013 Connected Continent proposal ambitious enough when put side by side with the declared goals?

The answer lies in the fragmentation and it has two aspects. The first has to do with the nature of the telecommunications fragmentation. Whereas it may be true that the single authorisation regime will greatly facilitate access to foreign markets and increase potential for multi-country operation, it will not remove financial, cultural, linguistic, political or other obstacles. Unlike American, European fragmentation is cultural. In other words, a Californian company wanting to operate in Texas may still be in a better position than a Portuguese company seeking access to the Polish market. This ‘inherent fragmentation’ present in the EU system acts as an impediment both at the carrier and at the content levels and therefore impedes equally.

The second has to do with the extent of fragmentation. Considerable fragmentation remains in both content and carrier arenas, in spite of all the efforts post 1987 - for the former, and post 1995 - for the latter. In terms of content, an achieved Single Market would mean an unimpeded ability to provide information society services across the EU. A cursory look at the Digital Agenda 2020 list of actions for barrier removal, however, reveals to what extent this still remains a problem. Difficult licensing conditions (Action 1), uncertainties in terms of orphan works (Action 2), copyright enforcement issues (Action 6) or convergence in the audiovisual world (Action 108) are some of the examples of the outstanding issues not regulated in any of the present content-related directives. In the telecommunications world, a fulfilled Single Market would mean complete freedom to provide telecommunication services of the kind that would exist within a member state. The fragmentation and the resistance on the national basis, the complexity and multitude of interests together with the lack of an all-encompassing regulatory framework under a single EU telecommunications authority testify to the state of affairs post 2009.

Europe, in other words, lacks a common market both in the telecommunications and in the content sectors. How is this situation affecting the future internet and what lessons might European policymakers learn from the past?

Connected continent: a way forward?

The internet today, in terms of performance at least, has been subject to very profound changes as it moved from networks to platforms, from telephone lines to broadband, from mobile phones to smart phones. The preceding sections have concentrated on the internet-independent origins of EU telecommunications law and the Single Market nature of the rest of internet regulation. We will attempt to make some generalised observations about EU telecommunications policy deficiencies in the Digital Single Market context that we believe arise from the aforesaid and, suggest possible policy prescriptions.

First, as of 2010, the EU does have at least an outline of a coherent internet policy. The Digital Agenda 2020 integrates the content and carrier sides. It shows awareness of the majority of problems facing policymakers wishing to tackle internet regulation. Its reference to trust, security, speed, innovation, literacy and inclusion demonstrates an alertness going beyond the commercial. Its basic premise, that fragmented digital markets contribute to rising cybercrime and low trust, lack of investment in networks or insufficient research and development, can also be accepted39. As a list of goals to be achieved, the Digital Agenda is an acceptable blueprint for policymakers. The problem, therefore, is not the fact that Brussels lacks general awareness of the most pressing problems on either the carrier or the content side. The problem is that the bulk of laws hitherto enacted are a result of internet policies, not an internet policy. The Digital Agenda reads like a lengthy wish-list but is vague on crucial concepts such as copyright reform or the future of ISPs. The Connected Continent proposal is not different as it essentially deals with three carrier-specific problems only. Guided by practical requirements rather than inherited theoretical and historical legal bias, the EU is a practical lawmaker but should not be afraid to continue the work began under the Digital Agenda 2020. That document is presently a strategy - a high level plan - informative but not binding. Its crucial ideas should be sharpened and turned into a framework which encompasses both the carrier and the content levels.

Second, a commonly repeated cliché regarding the internet is that it is changing rapidly. This assertion, however, acquires a new and more threatening dimension in the European Union. In spite of its readiness to deal with the problems of the cyberworld, the EU is a slow40 and cautious41 policymaker, often subject to resistance at the national level, prone to making compromises and lacking in transparency and democratic accountability. A careful observer will have noticed that basic liberalisation in the telecommunications sector took over one decade ago, or that the privacy (Data Protection Directive) or copyright (Copyright Directive) reforms are currently progressing very slowly. Such a regulator exposes itself to a perpetual delayed action. A rapidly changing internet combined with a slow-reacting compromise-prone political entity results in confusing policies. Instead of aiming at comprehensive coverage of a very large number of issues, the EU should keep writing technology-neutral laws and should stick to a minimum rather than full harmonisation (demonstrated recently in the 2005 Unfair Commercial Practices Directive or 2012 General Data Protection Regulation proposal). The EU is familiar with both approaches as some of its core internet directives, including the 2001 E-Commerce Directive, are both technology-neutral and minimum harmonisation measures. Full harmonisation, on the other hand, antagonises member states and increases resistance.

Third, European internet policy making is replete with the relatively non-assertive preambles42, usually appearing in a very similar form in most of the relevant internet directives. This language reflects, among other things, the lack of a proper empirical basis (although not necessarily the lack of a desire to acquire it), the usual approach being to call for a review of the instrument after a certain number of years (as is the case with Article 21 of the E-Commerce Directive). The situation has changed in recent proposals (e.g., Article 39 of the 2013 Connected Continent as well as calls for Copyright Directive Reform from 2008, 2010 and 2013) which are preceded by public calls and call for review reports to be submitted to the Parliament and the Council. Good policy requires good evidence on which to base that policy. In 2013, the Commission initiated the Global Internet Policy Observatory (GIPO) with the aim of creating “an online platform to improve knowledge of and participation of all stakeholders across the world in debates and decisions on internet policies.”43. This is an important initiative whose results ought to begin making a significant impact on EU internet lawmaking.

Fourth, the European Union’s habit of invoking the Single Market may be a relic of its origin as much as a testimony of that idea’s real importance. Its almost universal presence may now have run into the real boundaries imposed by the realities of cultural, political, social, economic and linguistic differences. In other words, there is a limit to what legal intervention can achieve. A multicultural, multi-linguistic perpetually dynamic multi-player setting tends towards maintenance of the differences, not their erasure. This produces new ‘barriers’ to trade. The Single Market paradigm may no longer be well suited for addressing these. The EU as a whole is less innovative than either the United States or Japan. The financing of technological and IT innovation and the related problems of public policy and efficiency ought to be high on the EU’s priority list. Knowledge creation, management and preservation are all at the very centre of a modern IT-based economy. Single Market is but one tool to achieve success in these areas. The fixation on the Digital Single Market may then have to be replaced with a more long-term-focused view on innovation and knowledge. Put differently, internet policies should not aim at achieving the Single Market as much as creating and sustaining an innovative knowledge-based society.  

Fifth, few of the European internet regulatory solutions are truly unique. Sometimes the limits are imposed by internationally binding documents, such as the TRIPS agreement or the Berne Convention in intellectual property matters. These put real limits to what the European Union can do. At other times, the historical role of the United States puts political limits on EU involvement. This is particularly true in internet governance44 and the domain name system. Auspiciously, there is readiness in Europe to experiment with internet lawmaking. Overall, EU internet law is a combination of traditional law, soft law, self-regulation, co-regulation and controlled chaos. This readiness to experiment and admit flexibility can be seen as an asset and enabling factor – introducing the needed flexibility/adaptability. The results have occasionally been noteworthy as evidenced by home country control, now present in many internet-related directives or comprehensive consumer protection. Europe needs original solutions to internet regulatory problems. Where Europe has an opportunity to come with such solutions, this advantage needs to be asserted.

Sixth, the modern internet will be the world of convergence of IT, media and telecommunications services45. This world requires original policy and a critical distance to legacy solutions. At present, it is not clear to this author that converging services necessarily require converging laws. In other words, it is premature to advise that the legal worlds of carrier and content directives begin merging. But, it is not premature to advise that future internet policy be formed with that convergence in mind. In that sense, reflexive laws, soft law and standardisation may provide a better solution to the challenges posed than hard laws.