A New Regulatory Paradigm for European Electronic Communications: On the Fallacy of the ‘Less Regulation’ Rhetoric Alexandre de Streel*

DRAFT PAPER ITS-EUROPE REGIONAL CONFERENCE, BERLIN, SEPTEMBER 2004

Abstract: This paper addresses the manner in which electronic communications markets are selected for the main part of economic regulation applicable in the EU. It shows that the new 2003 regulatory framework is innovative with regard to its substantive as well as its institutional approaches. On substance, the generic selection criterion is now based on the relative efficiency of antitrust remedies compared to sector regulation to address market failures. Moreover, a safeguard clause is provided for emerging markets. On institutions, this criterion should be applied jointly by the Commission (DG Information Society and DG Competition) and the NRAs in co-operation with NCAs, under the control of National and European Courts. The paper submits that the reform may lead to over-regulation and legal uncertainty, increasing dramatically regulatory costs. To alleviate these dangers, the paper suggests to clarify the generic criterion and to rely on the ‘regulatory brakes’ more than it has been the case until now.

1. Introduction When the Federal Communications Commission was created in 1934 the US, it was supposed to be for a limited period of time. Yet after more than 70 years, the FCC is still there influencing very much the evolution of the American telecommunications industry. Similarly when the US Telecommunications Act was adopted in 1996, it was supposed to lead to less regulation. Yet, the number of pages in the official compendium of the FCC decisions and proceedings has nearly tripled since the passage of the Act, while the membership in the Federal Communications Bar Association increased by 73% between 1995 and 1998 and has remained essentially at that level1. Across the Atlantic, a new regulatory framework has just become applicable in July 2003 and is also deemed to lead to less regulation, and even in the long run to a complete phasing out of economic regulation to the benefit of the mere application of competition law2. Yet, I argue in this paper that the 2003 framework has changed the paradigm for regulation that may lead to more, and not less, public intervention. *

European University Institute, Florence. This paper reflects the law as of 1 August 2004 unless indicated otherwise. Thanks to Yves Blondeel and Christian Hocepied for very useful comments, any opinion or error remaining mines. 1 Sidak (2003:207). Similarly in Australia, the incumbent is reported to be the biggest consumer of legal services. 2 Communication of the Commission of 10 November 1999, The 1999 Communications Review, COM(1999) 623, hereafter 1999 Review Communication, at 49; Resolution of the European Parliament of 13 June 2000 on the 1999 Communications Review of the Commission A5-0145/2000, O.J. [2001] C 67/53, Point A; Statement

1

In effect, the main part of the economic regulation is imposed through the so-called ‘Significant Market Power’ regime. Under this regime, regulatory obligations are imposed in three steps: (1) Markets areas potentially to be regulated are selected and then defined; (2) Operators with SMP are designated on these selected areas; (3) regulatory obligations are imposed on the SMP operators. This regime was originally put in place by the ONP Directives in the nineties, but all its three steps have been radically reformed by the new 2003 European regulatory framework to ensure that regulation is more flexible and less pervasive. This paper focuses on the first step of the regime and more particularly on the selection of the markets. This step has been often overlooked in the literature as it is somewhat hidden in the Directives and less explicit than the two other steps. Yet, it is the most important one as it starts the whole process and contains the rationale for relying on sector regulation instead of leaving the market to the mere control of competition law. The paper shows that the 2003 framework introduces a new regulatory paradigm by basing market selection on the relative efficiency of sector regulation compared to antitrust to solve market power problems. On the one hand, this is to be welcomed as it makes the rationale for regulation more explicit than previously. On the other hand, this is to be feared as it may lead to an increase, and even a perpetuation, of regulation contrary to the political rhetoric of the European legislator. The paper is organised as follows. Section 2 explains the previous 1998 SMP regime and the need for its reform. Section 3 focuses on the new 2003 regime, detailing the substantive and procedural new approach, and the special cases of emerging markets. It also takes a critical look at the reform and makes some proposals for the drafting of the next Commission Recommendation on relevant markets. Finally, Section 4 briefly concludes.

2. The 1998 Paradigm 2.1. A regime for the early phase of the liberalisation The 1998 framework3 was designed for the early phase of liberalisation when market structures were relatively stable and homogeneous across Member States and when the scope of the regulation was relatively easy to determine. Indeed, there was a broad political consensus that regulation should focus on the retail and wholesale services of the previous monopolists because there were the key enablers for the development of competition and their infrastructure had been deployed under past legal protection. Thus, the SMP regime was straightforward: (1) Markets areas to be regulated were specified directly in the Directives by the European legislator (Council of Ministers and European Parliament); (2) The SMP threshold mainly equate to the crude assumption of 25% market shares4; (3) A full suite of five obligations had to be imposed on the designated SMP operators (transparency, nondiscrimination, accounting separation, compulsory access, and cost-orientation).

of Reasons of the Council Common Position 38/2001 of 17 September 2001 on the Framework Directive, O.J. [2001] C 337/51, para II.1. 3 See in general Garzaniti (2000), Scherer (1998), Walden and Angel (2001). Specifically on the evolution of the SMP regime: de Streel and Queck (2002). 4 Other factors could also have been taken into account like capacity to influence market conditions, turnover relative to market size, control of access to end-users, access to financial resources, experience in providing products and services, international links. These factors have sometimes been taken into account by NRAs like in Italia in the Omnitel decision of 1998 or in Sweden in the Vodafone/Tele2 decision of 2002.

2

In effect, regulation focused on some fixed-line narrowband services, and to a lesser extent, on some mobile 2G services. As regard to the fixed services, the 1998 directives identified three main areas: wholesale interconnection, retail voice telephony and retail leased lines5. They did not selected any fixed broadband services (like wholesale unbundled access to the local loop, wholesale bitstream access, or retail ADSL services) as they were developed when the directives were adopted. Yet there was a slight flexibility in the directives with the imprecise notion of ‘special network access’ which has been relied upon in some countries to regulate bitstream access6. In addition, the European Parliament and the Council adopted in December 2000 another legal instrument for the NRAs to regulate the local loop7. As regard to mobile 2G services, there was a political disagreement between the European Parliament in favour of regulation and the Council reluctant to intervene in such promising emerging markets. Indeed, the case for regulation was weaker than for the fixed services as the economies of scale and scope are much less important (and each Member State counts at least three mobile networks) and networks have been deployed under more competitive conditions. As compromise, the directives provided for a possible, albeit lighter, regulation for mobile services. Indeed, there was no possibility to intervene in retail markets and the market share to trigger cost-orientation on wholesale interconnection had to be calculated on a combined fixed and mobile interconnection market. Moreover, this lighter regulation was reluctantly and only slowly applied by the NRAs. Thus, the 1998 Framework had several characteristics. First, markets to be regulated were relatively limited and focus mainly on the wholesale and retail services offered by the previous monopolists. Indeed, there was a sense that regulation should be linked to the ‘original sin’ of operators having enjoyed past legal protection. Second, markets to be regulated were mainly directly selected by the Council and the European Parliament in their directives, with very little flexibility left to the implementing authorities (being the NRAs or the Commission). On the one hand, that made the regime very rigid, as new market selection would trigger the whole legislative process as illustrated by the local loop unbundling. On the other hand, this very rigidity was a powerful protection against over-regulation and unjustified market selection. Third, market areas were defined more according to their technical characteristics than according to the economic demand and supply side substitutions8.

2.2. The need for reform However as liberalisation was progressing and technological innovations were developing, the market structure became much more complex, with two consequences for the design of the SMP regime9. First, the rationale for regulation could not be any more simply linked to the 5

ONP-Voice telephony directive 98/10; ONP-leased lines directive 92/44 as modified by directive 97/51; ONPinterconnection directive 97/33 and Note of the Commission services/DG XIII of 1 March 1999 on the Determination of Organisations with Significant Market Power (SMP) for implementation of the ONP Directives, available at . 6 Like France in the LDCOM decision of 2003 or Ireland. See also the Communication from the Commission of 26 April 2000 on the unbundled access to the local loop, O.J. [2000] C 272/55. 7 Regulation 2887/2000 of the European Parliament and of the Council of 18 December 2000 on unbundled access to the local loop, O.J. [2000] L 336/4. 8 Explanatory Memorandum of Market Recommendation, section 2. 9 Some argue that the 1998 framework has been very slow to be implement, and in fact has never been fully applied. Many innovations developed in the run-up of the 2003 Framework (like bitstream access, cost

3

previous monopolist but should be more precise and sophisticated. As put by Larouche (2002:141): “As time goes by and the incumbent progressively loses its grip on the market, then the fundamental question begin to arise more seriously: were regulatory burdens imposed on the incumbent because of its size and market position or because of certain features of telecommunications which were only present in the incumbent since it represented for all intents and purpose the whole sector?”. Second, the determination of the regulated areas should be more flexible to keep pace with the rapid market evolution. In the run-up of the 99 Review, several proposals were made10. For instance some external consultants of the Commission (Squire-Sanders-Dempsey and Analysys, 1999:147) proposed to base access regulation on the concept of bottleneck. In practice, they identified interconnection (especially termination practices), access to networks or digital gateways, local loop, distributions and access to scarce resources. For the future, they also identified intellectual property rights, directory services, programming guides, control over interfaces/web navigators. This approach was interesting because of their pragmatism. They have been rejected by the European legislator because they were creating an unacceptable level of legal uncertainty as the very concept of bottleneck was not at defined neither in legal or economic terms11. Larouche (2000:368-398) proposed to focus regulation on cases of bottleneck and network effects. He argued that competition law could not deal with bottleneck problem unless being stretched outside (and un-legitimately) its traditional boundaries with the essential facilities doctrine. As he put: “when competition law deserts its well-established framework of analysis under Article 82 EC Treaty – by moving from market definition to market structuring, by replacing dominance with the vague notion of ‘essentiality’ and by abandoning the requirement of abusive behaviour, then the proper limits of competition law, a case-bound general regulatory framework, has been exceeded”. Similarly, competition law cannot deal properly with network effects as it may be necessary and efficient to impose compulsory access -supplier access, customer access and transactional access- even in cases where there is no single or joint dominance (like in case of national roaming). However seducing, this approach has not been followed for several reasons. First, it was based on a very narrow conception of the role of European competition law in electronic communications that did not correspond to the practice of the Commission as endorsed by the Community Courts12. Second, the definition of bottleneck and network effects was imprecise and the relationship between both concepts was not clarified.

3. The 2003 Paradigm 3.1. A regime for complex market structure

orientation for mobile termination, leased lines interconnection) could have been applied under the previous framework, hence it would have been better for the Commission to ensure a correct application of the 1998 framework instead of reform it. 10 See 1999 Review Communication, section 4.2.1. 11 For a tentative definition, see Vernon (2002:XXX). 12 See de Streel (2004:Section 1).

4

The 2003 Framework radically reformed each step of the SMP regime13. (1) In the first step, markets to be analysed are now defined in two sequences. The Commission adopts a Recommendation that defines, in accordance with the principles of competition law, the product and service markets within the electronic communications sector, the characteristics of which may be such as to justify the imposition of regulatory obligations. In practice, the Commission has to select the markets justifying ex ante regulation because of their structural problems and then, delineate the boundaries of these markets on the basis of antitrust methodologies. Taking account of this Recommendation on relevant markets and the Commission Guidelines on market analysis, the NRA then defines markets appropriate to national circumstances, in particular their geographical dimension within its territory, in accordance with the principles of competition law. (2) In the second step, the NRA analyses the defined markets to determine whether one or more operators enjoy SMP, which is equivalent to the dominant position as defined in antitrust14. (3) In the third step, the NRA imposes on the SMP operators the appropriate specific regulatory obligations to be chosen from a menu provided in the Directives (transparency, non-discrimination, accounting separation, compulsory access, and cost-orientation, as well as any other obligations in exceptional circumstances). In case of an SMP operator on a wholesale market, the regulator should rely on the menu of remedies provided in the Access Directive. In case of an SMP operator on a retail market, the conditions of regulation are much more stringent as it has to be shown in addition of dominance that an intervention on a wholesale market could not solve the identified retail problem. In such circumstances, the NRA should choose the appropriate remedies in the non-exhaustive list of the Universal Service Directive. The Commission review each NRA decision that affect the trade between Member States under the so-called Article 7 procedure, with a possibility to veto part of them. Indeed, it may oppose a market definition that would be different than one of the Recommendation on relevant markets and/or designation (or absence of designation) of SMP operator. Otherwise on remedies, the Commission may only issue a non-binding opinion. To manage this review, the Commission has set up two tasks forces (one in DG Competition and the other in DG Information Society) and work closely together for each case (NRA draft decision), and adopted a procedural Recommendation15. This regime is thus largely different from the 1998, as illustrated in Table 1 below. (1) For the first step, the selection of the markets is not any more done by the European legislator on the basis of unclear political criteria, but by the Commission and the NRAs on the basis of a more legal (although difficult to interpret) criterion related to the insufficiency of competition law. Moreover, these markets are not any more delineated on a technical basis, but on the economic SSNIP test used in antitrust cases. (2) For the second step, SMP does not any more equate to the crude test of 25% market shares, but to the much more comprehensive and economically oriented dominant position test. (3) For the third step, NRAs are not any more obliged to impose on the SMP operators the full suite of the five remedies provided in the directives, but should choose the most proportionate ones. In short, the 2003 framework introduces more flexibility and economic analysis in the regulation rationale. 13

See for a description: Bak (2003), Bavasso (2004), Cave (2004a), Cawley (2004), Garzaniti (2003:chapter 1), Krüger and Di Mauro (2003), Maxwell (2002), Nihoul and Rodford (2004:chapter 3), de Streel (2003b). For two critical views Larouche (2002), Veljanovski (2001). 14 Position of economic strength which gives the power to behave to an appreciable extent independently of competitors, customers and ultimately consumers: United Brands 27/76 [1978] ECR 207, para 65; Hoffman-La Roche, 85/76 [1979] ECR 461. 15 Commission Recommendation of 23 July 2003 on notifications, time limits and consultations provided for in Article 7 of Directive 2002/21/EC of the European Parliament and of the Council of 7 March 2002 on a common regulatory framework for electronic communications networks and services, O.J. [2003] L 190/13.

5

Table 1: Comparison between the 1998 and 2003 SMP regime 1998 Framework 1. Market Definition 1a. Selection

Directly in the directives on the basis of ‘political’ criteria By Council and European Parliament 4 main market areas 1b. Delineation (product On technical basis and geo dimensions) Presumption at 25% market share 2. SMP assessment Additional criteria may be used

3. Choice of remedies

Automatically imposed the full suite of five remedies (transparency, nondiscrimination, accounting separation, compulsory access, and costorientation)

2003 Framework

Not in the directive, but on the basis of insufficiency of competition law By Commission and NRAs 18 markets On economic basis (SSNIP test)16 Equivalent to dominant position (single or joint) Market shares important but not the sole criterion17 Imposed at least one remedy Choice should be based on the nature of market failure, justified to achieve the objective of the directives, proportionate, and incentive compatible18

Even though the most radical reform took place at the first step of the SMP regime (in particular the way markets are selected), it remains largely unnoticed by the literature to date19. That is not surprising because on the one hand, this step is largely hidden in the Framework Directive20 and has become apparent only in the Recommendation on relevant markets, and on the other hand, this selection process has not yet been fully applied as the markets to be selected in the first Recommendation were directly listed by the European 16

Directive 2002/21/EC of the European Parliament and of the Council of 7 March 2002 on a common regulatory framework for electronic communications networks and services, OJ [2002] L 108/33, hereafter the Framework Directive, Article 15; Commission Recommendation 2003/311 of 11 February 2003 on relevant product and service markets within the electronic communications sector susceptible to ex ante regulation in accordance with Directive 2002/21/EC of the European Parliament and of the Council on a common regulatory framework for electronic communications networks and services, OJ [2003] L 114/45, hereafter Recommendation on relevant markets; Commission Guidelines of 9 July 2002 on market analysis and the assessment of significant market power under the Community regulatory framework for electronic communications networks and services, OJ [2002] C 165/6, hereafter the Guidelines on market analysis, Section 2. 17 Framework Directive, Articles 14 and 16; Commission Guidelines on market analysis, Section 3; ERG Working Paper of May 2003 on the SMP concept for the new regulatory framework, ERG(03) 09rev2. 18 Framework Directive, Article 16; Directive 2002/19/EC of the European Parliament and of the Council of 7 March 2002 on access to, and interconnection of, electronic communications networks and services, OJ [2002] L 108/7, hereafter the Access Directive, Articles 8 to 13; Directive 2002/22/EC of the European Parliament and of the Council of 7 March 2002 on universal service and users' rights relating to electronic communications networks and services, OJ [2002] L 108/51, hereafter the Universal Service Directive, Articles 16 to 19; Guidelines on market analysis, Sections 4 to 6; ERG Common Position of 1 April 2004 on the approach to appropriate remedies in the new regulatory framework, ERG (03) 30rev1, hereafter the Common Position on remedies. 19 For instance, Bak (2003) or Bavasso (2004) do not mention the reform of the selection process. 20 Larouche (2003:XXX).

6

legislator in the Annex I of the Framework Directive. In other words, the full implication of the reform will only take place when the Recommendation will be revised at the end of 2005.

3.2. The Substantive Test This first step has been reformed on substantial as well as on institutional (or procedural) matters. On substance, the selection criterion is not very clear at a first sight. First, when selecting markets for regulation, authorities should respect the general objectives assigned to them by Article 8 of the Framework Directive stating that: 2. The national regulatory authorities shall promote competition in the provision of electronic communications networks, electronic communications services and associated facilities and services, by inter alia: (b) Ensuring that there is no distortion or restriction of competition in the electronic communications sector; (c) Encouraging efficient investment in infrastructure, and promoting innovation (…) 3. Contribute to the development of internal market, by inter alia: (c) Ensuring that, in similar circumstances, there is no discrimination in the treatment of undertakings providing electronic communications networks and services; (d) Co-operating with each other and with the Commission in a transparent manner to ensure the development of consistent regulatory practice and consistent application of the Directive and the Specific Directive. 4. Promote the interests of the citizens of the European Union (…)

More specifically on market selection, the Article 15(1) of the Framework Directive states that: The Recommendation shall identify (…) those product and service markets within the electronic communications sector, the characteristics of which may be such as to justify the imposition of regulatory obligations set out in the Specific Directives.

This provision is circular and does not mean much. It instructs the Commission to select for sector regulation the markets that need regulation. Recital 27 of the Framework Directive is a bit more explicit by stating in a generic way that: It is essential that ex ante regulatory obligations should only be imposed where (…) national and Community competition law remedies are not sufficient to address the (competitive) problem.

In addition, Recital 13 of the Access Directive provides that: The aim is to reduce ex ante sector specific rules progressively as competition in the market develops. However the procedures also takes account of transitional problems in the market such as those related to international roaming and the possibility of new bottlenecks arising as a result of technological development, which may require ex ante regulation, for example in the area of broadband access networks (…).

7

In other words, the Commission should select the markets that have characteristics such that competition law will not be sufficient to address problems created by the existence of hardcore market power, for instance in case of bottleneck. If this sufficiency criterion seem to follow common sense, it is difficult to apply in practice. Given its tools, competition law appears to be sufficient to deal with all problems created by every kind of market power21. Therefore, it is reasonable to go from a sufficiency criterion to a relative efficiency criterion comparing the respective strengths of antitrust vs. sector specific intervention and base market selection on such a generic efficiency criterion. Yet, this remains very difficult to apply in practice as it requires a double comparison. First, the efficiency of two sets of legal instruments (antitrust and sector regulation) should be compared. This is particularly complex in electronic communications as both instruments have converged over time and few divergences remain22. Second, the efficiency of two sorts of institutions (NCA and NRA) should also be compared. Again, this is complex, all the more so that in some Member States a single institution may apply both legal instruments23. This double comparison is rendered even more complicated as it is done at the European level when the Commission selects markets, although the relative efficiency of legal instruments or institutions may strongly vary across Member States. Despite all these difficulties, the generic relative efficiency criterion has been interpreted by the Commission in its Recommendation on relevant markets has covering three cumulative specific criteria. As Recital 9 of this Recommendation states24: In identifying markets in accordance with competition law principles, recourse should be had to the following three criteria. The first criterion is the presence of high and non-transitory entry barriers whether of structural, legal or regulatory nature. However, given the dynamic character and functioning of electronic communications markets, possibilities to overcome barriers within a relevant time horizon have also to be taken into consideration when carrying out a prospective analysis to identify the relevant market for possible ex ante regulation. Therefore, the second criterion admits only those markets the structure of which does not tend towards effective competition within the relevant time horizon. The third criterion is that application of competition law alone would not adequately address the market failures concerned.

Thus, the first criterion is static and relies on the presence of high and non-transitory barriers to entry. The barriers may be structural and result from original cost and demand economic conditions that create asymmetric conditions between incumbents and new entrants impeding or preventing market entry of the latter. The entry barriers may also be legal or regulatory and result from legislative, administrative or other state measures that have a direct effect on the conditions of entry. As noted by Cave (2004a:35), both types of barriers are non-strategic (i.e. not artificially manufactured by the firms) as it was considered that strategic barriers like excessive investment or reinforcement of network effects would require idiosyncratic and episodic intervention, which would be better done under competition law. 21

As European competition law may control exploitative abuses, even excessive prices of a natural monopolist may be controlled by antitrust authority, see Motta and de Streel (2003) with some examples of excessive pricing cases in telecommunications. 22 See de Streel (2004) for a demonstration of this convergence and an comparison between antitrust and regulation. For another comparison, see the references cited in footnote 166 of the latter paper, and in particular: Intven (2000:Table 5-1); Kerf and Geradin (2003); Laffont and Tirole (2000:section 7.2). 23 For instance, the British or the Greek regulators may apply sector regulation and competition law. 24 See also Recitals 10 to 16 of the Recommendation on the relevant markets, as explained by section 3.2 of the Explanatory Memorandum. The Commission developed the three specific criteria with the help of some external consultants whose report is public: Squire Sanders and Dempsey and WIK (2003).

8

The second criterion is dynamic and amounts to evaluating if the market does have the characteristics such that it will tend towards effective competition over the relevant time horizon considered. If it is the case, the market should not be selected. The application of this criterion involves examining the state of competition behind the entry barriers, taking account of the fact that even when a market is characterised by high entry barriers, other structural factors or market characteristics may mean that it tends towards effective competition. This is for instance the case in markets with a limited, but sufficient, number of undertakings behind the entry barriers having diverging cost structures and facing price-elastic market demand. Entry barriers may also become less relevant with regard to innovation-driven markets characterised by ongoing technological progress. In such cases, competitive constraints often come from the threat of innovation by potential competitors that are not currently in the market. The third criterion relies on the relative efficiency of competition law remedies alone to address the market failure identified according to the two first criteria compared to the use of complementary ex ante regulation. It is fulfilled when ex ante regulation would address more efficiently the market failure than antitrust. Such circumstances would for example include situations where the compliance requirements of intervention are extensive, where frequent and/or timely intervention is indispensable, or where creating legal certainty is of paramount concern. No exhaustive list of all the circumstances where antirust would be less efficient than sector regulation can be given. In practical application, NRAs should consult with their competition authorities and take account of that body’s opinion when deciding whether use of both complementary instruments is appropriate to deal with a specific issue25. The application of the third specific criterion may be problematic as it sounds like a repetition of the generic criterion of Recital 27 (relative efficiency of sector regulation) that all the three specific criteria were supposed to developed26. This critique is relevant and I suggest that the two first specific criteria are the most important and that once both of them are met, in most (if not all) the cases, the third specific criterion is fulfilled as well. All the three specific criteria should be applied cumulatively, which means that the absence of any of them implies that the relevant market would not be selected27. As put by Geradin and Sidak (2003:21), “the combined analysis of the three criteria suggests that markets characterised by the presence of high barriers to entry (first criterion), which is not compensated by a dynamic market structure (second condition) should generally be selected unless the situation can be dealt with adequately by competition law remedies (third criterion)”. Note the complex relationship between the different steps of the SMP regime. First, there is a link between the three specific selection criteria (step 1a) and the market delineation (step 1b). 25

The application of this criterion may change over time. For instance, if capacity of antitrust authorities is enhanced over time (because of the decentralisation of competition law for instance), the efficiency of antitrust remedies may be considered as enhanced as well. 26 It is reported that DG COMP was very much supporting the third specific criterion to alleviate calls that any sector fulfilling the two first specific criteria (like pharmaceuticals) should be subject to sector regulation. 27 The cumulative character of the specific criteria has been criticised by Garzaniti (2003:14) who argues that there might be a need in certain cases to select a relevant market for ex ante regulation, even where there are low entry barriers and strong dynamic market factors are present because the application of competition law had proved not to be sufficient to address those concerns. However, this author fails to give any concrete examples where although the two first specific criteria would not be met, there would be competitive problems not addressable by antitrust. I submit that there is none, hence Garzaniti’s critique is unfounded.

9

Indeed, the delineation of market boundaries and the fulfilment of selection criteria move together, according to the technological and market developments. More the markets are defined narrowly, more there is a probability that this narrow market would be selected for regulation28. At the limit, when the competitive conditions are such that each network constitutes a relevant market, then these very narrow markets should automatically be selected29. Second, there is a link between the two first selection criteria (step 1a) and the SMP designation (step 2) because both steps rely on many similar indicators as illustrated in Table 2 below.. Indeed, entry barriers and dynamic state behind these barriers also play a role in the dominance assessment30. In effect, the criteria aim to detect the sub-sets of dominance cases (hard-core market power) that justify regulation. Thus when a market is identified by the Commission in the Recommendation, there is a strong probability that SMP operators will be found the NRAs during their market analysis. Conversely, if SMP are not found on an identified market by many NRAs, there is strong probability that market will be removed for a subsequent revision of the Recommendation provided that the removal of existing obligations would not remove competition on that market31. However, the Framework Directive separates the three steps (maybe artificially) and assigns them to different institutions, hence each step should be done separately without pre-empting each other. Thus, the fact that a market is identified by the Commission in its Recommendation does not necessarily imply that the NRA is bound to designate one or more SMP operator on this market32. This is so because on the one hand, the Commission can not make a full and comprehensive analysis of each selected market and on the other hand, the Commission should select market for the whole Community without being able to take into account the characteristics of each of the 25 Member States. Table 2: Comparison between the criteria to select market and to assess dominance Selection criteria33 Criterion 1 Entry barriers Non-strategic, high, and nontransitory - Economic - Legal and regulatory

SMP – Single dominance34

Market shares Economies of scale/scope Barriers to expansion Control of infrastructure not easily duplicated Absence of potential competition

SMP- Joint dominance35

Similar market share High barriers to entry Absence of excess capacity Lack potential competition

28

Similarly, more the markets are defined narrowly, more there is a probability to find a dominant position (which is commonly known in antitrust practice). Thus step 1b will have a influence on step 1a and step 2, which is not surprising as step 1a and step 2 are closely linked together. 29 This is the case for the wholesale markets for fixed or mobile termination (markets 9 and 16 of the Recommendation). In opening a recent excessive prices antitrust case against Vodafone and O2 of the UK, the Commission suggested in its statement of objections that the wholesale international roaming may also be defined by each individual network for the past three years: IP/04/994 of 26 July 2004. However, this definition is not convincing and contradicts the broader market definition comprising all national networks given previously by the Commission in the Recommendation on relevant markets (market 17). 30 On the role of barriers to entry in the dominance determination, see: Bak (2003), Bishop and Walker (2002:XXX), Harbord and Hoehn (1994). 31 Recommendation on relevant markets, recital 16. 32 Recommendation on relevant markets, recital 20. 33 Recommendation on relevant markets, recitals 9 to 16. 34 Guidelines on market analysis, para 78. 35 Framework Directive, Annex II.

10

Criterion 2 Absence of characteristics that lead to effective competition - diverging cost structure - high elasticity of demand - technological progress

Similar cost structure Mature market Moderate growth on demand side Mature technology Technological advantage or superiority Low countervailing buying power Privileged access to financial resources Products/services diversification Vertical integration Developed distribution and sales network

Lack countervailing buying power Links between undertakings Retaliatory mechanisms Reduced scope for price competition

Third, there is a link between the third specific selection criterion (step 1a) and the choice of remedies (step 3). Indeed, the insufficiency of competition law remedies would probably be better assessed when remedies are chosen36. However, the Commission is bound to assess (maybe in a very abstract way) the relative efficiency of remedies during the first step and once a market is selected (and SMP is found), at least one sector remedy should be imposed without the possibility to argue at that stage that antitrust remedies would be sufficient37.

3.3. Institutional and Procedural Design The new SMP regime has also be reformed with regard to its institutional structure. The European Commission, in particular DG Information Society and DG Competition, starts the process by selecting markets, applying the three specific criteria at a European level and necessarily in a superficial way as it has only limited information on the characteristics of the market structure. Then the NRAs define (i.e. select and delineate) markets according to their national circumstances. They have three possibilities: to copy the list of the Recommendation, to add markets, or to remove markets38. The ERG and the Commission suggest that the NRA should not re-assess if the selection criteria are met in their own national territories when copying the list39. They should only do so when adding markets40 or removing differently defined markets41. However, I submit that NRAs should carefully assess the criteria in all 36

Nihoul and Rodford (2004:273) go even further arguing that: “It is not possible to determine whether preference must be given to general competition law at the moment when authorities are considering in abstracto what markets are to be defined for an assessment of competition”. 37 Guidelines on market analysis, para 114. Contra Hocepied (2002). 38 Note the difference between on the one hand removing (or not selecting) a market, and on the other hand analysing and deciding not to regulate a market. Both cases lead to the absence of regulation, but under a different procedure. If the first case leads to a limited analysis of the competitive conditions of market (only to the extent that the three selection criteria should be assessed), the second case leads to a full competitive analysis. 39 Common Position on remedies, p. 20. Note however, that the Portuguese NRA has re-assessed the three selection criteria for the markets of fixed call origination and fixed call termination: Commission Decision of 25 June 2004, PT/2004/60 and PT/2004/61, p. 2. 40 Like the NRA of the United-Kingdom which added the market of wholesale un-metered narrowband Internet termination services on the basis of the three selection criteria: Commission Decision of 29 September 2003, UK/2003/4 and UK/2003/5, where it noted (p. 4) that the three criteria test should in general be prospective in nature and not rely on data reflecting past market situation. 41 It has already be the case for the market for broadcasting transmission services to deliver broadcast content to end-users (market 18). The NRAs of Austria, Ireland and Finland have broken down this overall market in

11

cases, and even when they analyse a market already identified by the Commission. Indeed, if Article 15(1) of the Framework Directive and its associated Recital 27 is specifically addressed at the Commission when selecting a market, it seems logical that it should also apply to the NRAs when selecting. All the more so that NRAs are surely better informed to assess correctly this task given the link between the three selection criteria and the criteria used to determine the dominant position. When assessing the specific selection criteria (and in particular the third one), the NRA should co-operate closely with the NCA42. In addition, this assessment is controlled by the Commission when reviewing NRAs decision under the Article 7 procedure. However until now, this control focus all steps of SMP process, but the market selection. It is also controlled by the National Court (and ultimately the European Courts) in case of dispute43. I submit that the review of the Court should leave some discretion to the NRA as the assessment is complex in nature, but should not be solely marginal like for the question of purely economic nature (market definition, SMP assessment44) because the issue is the choice between two legal instruments (antitrust or sector regulation). The Court should also review carefully if the procedures have been respected, in particular the consultation of the NCA45. To sum-up, the market selection is a common task involving different services of the Commission, the NRAs and the NCAs, under the review of National and European Courts.

3.4. The special case of emerging markets As the 2003 framework eases the selection of market for regulation and aligns NRAs’ methodologies on antitrust working methods, some fear that the regulators would follow their antitrust counterparts and excessively intervene in emerging markets. Indeed, during the past ten years of merger control, the Commission and the NCAs have heavily intervened in ICT emerging markets. They have defined nascent markets and impose far-reaching access obligations to alleviate their foreclosure46. This interventionist approach has been justified to prevent a dangerous circle of self re-inforcing market power between related markets, whereby parties leverage their power from established markets to secure a dominant position on emerging markets and, in turn, leverage from the emerging market to strengthen their power on the established markets47. If this approach may be justified for merger control that is several narrower markets (distinguishing radio and TV, as well as the terrestrial, cable and satellite platform) and exclude some of these narrower markets on the basis of the three selection criteria: Commission Decision of 11 December 2003, AT/2003/18; Commission Decision of 2 March 2004, IE/2004/42; Commission Decision of 14 July 2004, FI/2004/76. 42 Framework directive, Article 16(1); Guidelines on market analysis, para 135-137; Explanatory Memorandum of the Recommendation on relevant markets, p. 12. 43 According to the Framework Directive, Article 4. 44 Guidelines on market analysis, para 22. 45 Note however that one drawback of large Court intervention in the SMP process is that there is much less harmonisation mechanism at the Court level than at the NRA level. Indeed, there is no ERG equivalent between the Courts of the Member States and the Commission may not veto a Court decision. One of the only coordination mechanism is the intervention of the European Courts that have precedence on their national counterparts, but that relatively slow device is not always efficient in fast evolving markets like those in the electronic communications sector. 46 For instance, the Commission defined a (not existent at the time of the decision) market for seamless panEuropean mobile services to corporate customers and imposed a three years compulsory access to the integrated mobile network of the merging parties: Decision of the Commission of 12 April 2000, Vodafone/Mannessmann, M. 1795. For a comprehensive review of merger cases in ICT, see Garzaniti (2003:chapter 8). 47 This vicious circle is particularly worrying in the ICT sector for three reasons. First, a lot of markets are only emerging and their development should not be controlled by a particular company. Second these markets are

12

one-shot, it is less justified in case of sector intervention that is continuous. Thus, Recital 27 of the Framework Directive provides that (… ) newly emerging markets, where de facto the market leader is likely to have substantial market share but should not be subjected to inappropriate obligations.

This Recital is of limited use as it does not define the notions of ‘emerging market’ or ‘appropriate obligations’, nor articulate a comprehensive rationale of the regulation of emerging markets. With these limitations, the Recital implies that sector regulation is not precluded on emerging market but should be apply more parsimoniously than in ordinary markets. To go further, it is useful to distinguish between three different cases depending on the service offered and the infrastructure used, as illustrated in Table 3 below.

Table 3: Different cases of emerging markets Non-emerging service Non-emerging infrastructure

Emerging infrastructure

Emerging service

Ordinary service based on legacy New service based on legacy network network e.g.: ADSL offer based over fixed copper e.g.: Voice over fixed copper pairs pair Case 2 Ordinary service based on new network New service based on new network e.g.: Voice over mobile 3G network Æ May not fulfil the market selection criteria Case 3 e.g.: Mobile broadband over mobile 3G networks

Case 1a Æ May fulfil the market selection criteria e.g.: Broadishband data over fixed FTH

Case 1b

The most extreme case is a new service provided over a newly deployed network, like mobile broadband data services provided over 3G networks that has been explicitly given by the ERG as an example of an emerging market48. In such circumstances, there is a strong case against regulation even if the first mover enjoys substantial market shares. Several reasons have been put forward (in a very disordornarily way) by the diverse European regulatory actors. First, there may not be hard-core market power justifying regulation and the high market shares of the first mover are contestable because there are low entry barriers49. Second, there is hardcore market power but regulators should forbear because the first mover should recoup its investment risk and investment incentive should be preserved50. Third, premature intervention

evolving very quickly and any anti-competitive behaviour could have rapid and irreversible effects. Third, most of the markets are characterised by network effects, that lead to path dependency with early developers (first mover advantage) becoming dominant by capturing new growth (bandwagon affect). See Monti (2000). 48 Common Position on remedies, p. 22. 49 That is implied by the Framework Directive of the European Parliament and the Council at Recital 27, and the Recommendation on relevant markets of the Commission, recital 15. 50 That is implied by the Common Position on remedies of the ERG at p. 90.

13

may constrain spontaneous market developments and unduly influence technological progress51. In fact, it is useful to distinguish between two sub-cases when a new service is offered on a new network52. First, the new service-new infrastructure does not and will not in the future meet the three specific selection criteria (case 1a above). For instance, that may be the case for mobile 3G access and origination. In such circumstances, there is no need to intervene because the market is emerging, and more importantly, because there is no hard-core market power that justifies regulation. Second, the new service-new infrastructure may in the future meet the three specific selection criteria (case 1b above). For instance, that may be the case of broadishband data services provided over FTH network. This situation is trickier because on the one hand there is hard-core market power, and on the other hand, investment incentives should be preserved53. This is particularly the case if there was a competition for the market (ex ante), even though there is no competition in the market (ex post). There, regulators face the classical conflict between static and dynamic considerations. At this stage, the position of the regulatory actors on this issue is not clear, although the Common Position on remedies implies that an emerging market meeting the three specific criteria may be regulated after the elapse of a sufficient amount a time54. The Australian Productivity Commission (2001:section 9.6) proposes two options. Regulators may adopt a radical approach and guarantee to the operator ‘access holidays’ for a certain period of time55, like an intellectual property right. They may also be a little more interventionist and impose ‘open access regulatory compacts’ that leave operator the freedom the set the level of prices but set a structure of prices such that the operator can not foreclose its competitors on related markets. In other words, the dominant operator is entitled to reap its monopoly rent on the dominated market, but can not leverage this position on related markets. In the US, FCC Commissioner Abernathy (2002) argues for the radical approach with her Nascent Services Doctrine. This doctrine holds that regulators should exercise restraint when faced with new technologies and services in order to facilitate the development of new products and services without the burden of anachronistic regulation, and in turn promote the goal of enhancing facilities-based competition. The doctrine has two applications. It applies to nascent technologies (corresponding to cases 1a and 1b above), which appear in the market without 51

That is implied by the Guidelines on market analysis of the Commission at para 32. Another sub-category may also be done between new infrastructures put in place by incumbents and by new entrants. I’m not sure such distinction would be relevant. 53 It may be argued that regulation will not impede the recoupment of investment risk (hence will not undermine future investment incentive) as any access regulation (and access price) should provide a premium for investment risk. However, the calculation of this premium is far from easy as regulators face difficulty in distinguishing the ex post rewards for risky investment from monopoly rents, hence there is a possibility that the premium will be set too low. On this point, see Australian Productivity Commission (2001:268). 54 In the draft Common Position on remedies of 21 November 2003 (at p. 72), the ERG noted that emerging markets may be regulated if meeting the three specific selection criteria after some time of no regulation allowing investor to make adequate return. However, in the final Common Position (at p. 22), this statement has been watered down, by simply noting that an assessment of the three specific selection criteria can not be done when a market is emerging and that this assessment can be undertaken only after the elapse of a sufficient amount of time. In any case, in the explanatory Memorandum of the Common Position (ERG (04) 14, p. 9), the ERG acknowledges that further work on emerging markets is warranted. Commenting on the 28 potential control points identified in Devoteam and Siticom (2003), one senior Commission official is reported to have said at the Next Generation debate at the ETP in 2003: “regulators will have to ask whether there are high barriers to entry such as first mover advantage or non-replicability of legacy network elements. If so, there will be a need to keep looking at these control points but they will be regulated only if the three criteria of the Recommendation on relevant markets are met”. 55 This suggestion has been taken over by Cave (2003:14) in case of some fixed broadband services. 52

14

any clear sense of the services they will ultimately support or the markets in which they will ultimately compete, for instance ultra-wideband. It also applies to nascent platforms (corresponding to case 3 above) which are new competitors to incumbents in already-defined markets, for instance satellite broadband which are beginning to compete with more established cable modem and DSL providers. A different -and less extreme- case of emerging market is a new service provided on a legacy infrastructure possibly upgraded with minor new investment, like ADSL services provided over the fixed local loop (case 2 above). In such situation, the position of the European regulators is relatively clear56. On the one hand, the retail market is emerging and should not be regulated57. On the other hand, the wholesale market can not be considered as an emerging one (even though some upgrading investment may have been done), hence it may be regulated if meeting the three specific selection criteria. Incumbents and new entrants should be put on an equal footing to enter the emerging market, and incumbents can not leverage their power on the legacy network to foreclose entry on the new market. The last case of emerging market is an ordinary service provided on a newly deployed infrastructure, like voice over mobile 3G network. This case is difficult because there are two conflicting arguments. On the one hand, there is new investment to protect, hence a case to forbear from regulation. This is the position of the Nascent Service Doctrine of FCC Commissioner Abernathy mentioned above. On the other hand, market are defined according the services provided and not (any more) according to the underlying technologies58, hence it is not possible to separate say a market for 2G voice and 3G voice59. Thus, the 2003 Framework requires symmetry in regulation of infrastructures providing similar services. This symmetry does not necessarily imply more regulation as it can go both ways: either the regulator extends the existing regulation to the new infrastructure (albeit there may still be a differentiation in the choice of the remedies imposed), or it removes regulation from the old network. Indeed, the very fact that a new infrastructure is deployed and is economically viable60 may be an indication that entry barriers are low, hence sector regulation is not justified any more.

56

Guidelines on market analysis, footnote 92 stating that the leveraging provision of Article 14(3) of the Framework Directive is not intended to apply in relation to market power leveraged from a ‘regulated’ market into an emerging ‘not-regulated’ market and that in such case, any abusive conduct in the ‘emerging’ market would normally be dealt with under Article 82 EC; and Common Position on remedies, p. 90. 57 Note however that there may be a need to intervene on the emerging retail market if the wholesale nonmerging market can not be efficiently regulated. For instance in France, the Conseil de la concurrence suspended the retail ADSL offer of France Télécom group until a satisfactorily wholesale offer was made to ADSL competitors, see XXX, and Bourreau (2003). 58 As I explained elsewhere, market definition should primarily be based on the needs of the end users, and not necessarily on the technology used but there are links between both (de Streel, 2003b:498). 59 Commenting on Oftel notification regarding the UK mobile access and call origination market, the Commission noted that “market definition should be technologically neutral, i.e. based on the nature of the products or services provided, not on the technological platform used to provide them. 3G voice and SMS services offered at present or in the near future are, from a demand-side perspective, not distinguishable from their 2G equivalents, and appear to be part of the same relevant product market”: Decision of the Commission of 29 August 2003, UK/2003/1, p. 3. The Swedish NRA has taken the point when defined the mobile voice call termination market as comprising both 2G and 3G: Commission Decision of 9 June 2004, SE/2004/52. 60 Indeed, if a new operator deploy additional infrastructure and then rapidly goes bankrupt, it cannot be safely assume that entry barriers are low.

15

To conclude, the issue of emerging market61 is very complex and deserves rapidly further clarification from the European regulatory actors because if regulation may be bad for new investment, uncertain regulation is even worse. At this stage, I think everyone will agree that (1) regulation designed for legacy networks should not be transposed telle quelle to newly deployed networks and services to the risk of severely undermining investment incentives, and (2) there is a need to distinguish between emerging services provided on legacy infrastructure and those provided on new infrastructure. This latter case may prove controversial in the future. I would suggest to follow the approach advocated by the Australian Productivity Commission with access holidays and possible open access regulatory compacts.

4. Risks and Future of the 2003 Regime 4.1. Appraisal of the new regime As explained, the new regime is innovative with regards to its substantive as well as its institutional approach. On substance, I think that the three specific selection criteria aimed to detect the ‘hard-core’ market power62 that justifies regulation mainly boil down to two alternative structural market problems where economic theory shows that the market left alone would not achieve an overall long term efficiency and that public intervention would be needed on an on-going basis (hence sector-specific regulation would be more efficient than competition law). The first situation applies when the cost structure and the level of the demand are such that they create asymmetric conditions between market players, the topical situation being a natural monopoly case. That may be the case for local fixed infrastructure, in particular in rural areas where the level of fixed sunk cost may be such that only a single network provider could be profitable63. This situation covers the essential facility cases in antitrust64, such that 61

Note that the issue of protecting emerging markets should not be confuse with the issue of protection of new entrants on non-emerging markets and addressed in Annex of the Common Position on remedies. Regulators may protect new entrants by imposing on them differentiated (and lighter) remedies than those imposed on incumbents, like in Sweden for fixed and mobile call termination markets (Commission decision of 9 July 2004, SE/2004/50 and Commission decision of 9 July 2004, SE/2004/52), or in Finland for the same both markets (Commission decision of 17 December 2003, FI/2003/29 and Commission decision of 17 December 2003, FI/2003/31). 62 In this perspective, regulation focuses on a sub-set of dominance cases where market power is very large. In such cases, the risk of false condemnation is relatively low and the risk of false acquittal is high. Therefore, a public intervention instrument with a low burden of proof (and relatively lower than antitrust) is justified. In a different and somewhat opposed perspective, it may be argued that regulation should focus on the sub-set of dominance cases where market power is low or at least difficult to prove (like collective dominance cases). In such case, antitrust remedies are difficult to apply and regulation would be more efficient. The first perspective places emphasis of the two first selection criteria (and nearly ignores the third one). The second perspective places emphasis of the third criterion and may help to understand Garzaniti’s critique at note XXX. 63 The question whether local loop is still a natural monopoly is heavily discussed in the economic literature and the answer may depend on the models used. Compare Fuss and Waverman (2002) with Sharkey (2002). It remains that in practice, there is very little competition in the local loop: See Communication from the Commission of 19 November 2003, Ninth Report on the Implementation of the Telecommunications Regulatory Package, COM(2003) 715. 64 It is debatable if the essential facility doctrine is limited to the control of natural monopoly (Lipsky and Sidak, 1999:1220; Werden 1987:476) or goes beyond (Doherty, 2001:424), but for the simplicity of the argument, let’s say that it is the competition law tool to regulate natural monopoly.

16

at the very beginning of the SMP regime, a strong filter limits regulation (and in particular the imposition of compulsory access) to the special cases of essential facilities. Contrary to some authors65, I think that the imposition of compulsory access is not easier for under sector regulation than under competition when considering step 1a and step 3 together. The second situation is the presence of externalities where one player imposes a cost or a benefit to others without having to pay/being rewarded accordingly. These externalities may due to network effects that create a sort of natural monopoly with demand side economies of scale66. That is the case for network interconnection now, and may be the case for some bottlenecks in Next Generation Networks in the future. These externalities may also be due to tariff principles like the calling-party pays principle that justified the regulation of fixed and mobile call termination in Europe67. Is there the need to prove in addition to a natural monopoly or important externalities that the infrastructure has benefited from past legal protection? In other words, is the ‘original sin’ part of the 2003 Framework68? On an economic perspective, the original sin requirement may make sense. Some economists69 argue that there is no need to intervene if there was competition for the market (i.e. no legal entry barrier), even if there is afterward no more competition in the market. To be more precise, there is no need to condemn exploitative abuse (like excessive pricing) so that the monopolist should be free to reap its monopoly rent that is the premium for the investment risk and the motive for future investment. There is still a need to condemn exclusionary abuse (in particular foreclosure devices whereby the monopolist leverages its dominant position to another related markets), but that may be done by competition law alone without recourse to sector regulation70. From a legal perspective, there are two –weak- arguments in favour of the original sin requirement. Firstly, several provisions of the 2003 framework71 instruct the NRAs to protect investments and some economists argue that the best way to protect investment incentives is to limit regulation to networks built under legal protection. Secondly, the second specific selection criterion may be read as instructing the authorities not to select a market when there was a competition for the market. However, there are several –more convincing- arguments against the original sin requirement. Firstly and fundamentally, as markets should be defined according to substitutability among services and not according to technologies, if a regulated service may be provided over a legacy network and a non-legacy network, the latter will be regulated in the same way as the former. In other words, an original sin requirement directly conflicts the technological neutrality of the 2003 framework. Secondly, Annex I of the 65

These authors submit that it is easier to impose compulsory access under sector regulation than under antitrust because they focus on the third step of the SMP regime without taking into the first step: Cave (2004a:36), Geradin and Sidak (2003:24) focusing only on Article 12 of the Access Directive and stating that “under sectorspecific regulation, SMP operators could thus be forced to grant access to non-essential facilities on the ground that such access would be desirable to stimulate competition on the retail markets”. 66 On network effects, see Liebowitz and Margolis (2002). 67 On regulation of call termination, see Armstrong (2002), Laffont and Tirole (2000). 68 We refer here to the harmonisation directives of the 2003 Framework, as the original sin is present in the liberalisation directive: Commission Directive 2002/77/EC of 16 September 2002 on competition in the markets for electronic communications networks and services, O.J. [2002] L 249/21. 69 See Motta (2004:section 2.5). One of the first economists to argue that were Demsetz (1968) and Posner (XXX). 70 Even though leverage strategies may be easier to implement in ICT’s market because many markets are related and monopoly is quickly acquired due to network effects (bandwagon effect), I don’t think that alone justify to police leverage with sector regulation instead of competition law. 71 For instance the Framework Directive, Article 8(2c); and the Access Directive, Article 12(2c).

17

Framework Directive selects several mobile markets (call origination, call termination, and roaming) whereas most mobile networks have been deployed under special or exclusive rights. Thirdly, Recital 13 of the Access Directive cited above specifically refers to new bottlenecks to be regulated. Fourthly, the very initial draft of the Framework Directive proposed by the Commission contained two conditions for an SMP designation: (1) a dominant position and (2) original sin requirement or vertical integration72. However, during the negotiation the second condition has been dropped out. Indeed in practice, some policy positions of the NRAs go against the original sin requirement as they are ready to regulate emerging market after sufficient time elapsed for the market not to be emerging any more. Thus, the new substantive approach is based on an efficiency criterion, hence may lead to better and more focus regulation. However, it carries also several risks. First, there is a danger of legal uncertainty because efficiency of antitrust is an untested criterion and it is not easy to apply in practice. That is an application to the more general problem for legal certainty of the transposition in law of economic concepts. Second, there is a risk that the criteria are not applied carefully and individually in each Member State. Instead, some NRA takes the lead, other follows and the diverging minority is forced to align under a Commission veto. Moreover, the Commission periodically updates its Recommendation on relevant markets on the basis of the markets that have been (or not been) regulated by the majority of Member States during the previous iteration of market analysis73. That leads to the ‘law of the majority’ which does not necessarily coincides with less regulation, or even better focus regulation. Third, there is a danger of regulatory creep. Indeed, efficiency of antitrust may well be interpreted in different manners because it is so uncertain. If this interpretation is left to institutions having incentives to regulate, they may use it broadly and extend regulation beyond necessity. That leads me to the reform regarding the institutional structure. The new approach places the selection of markets in the hands of the Commission and the NRAs. That enables more flexibility, but that may lead to over regulation. Thus a framework announced to be de-regulatory might well lead to more (and sometimes unjustified) regulation74. Indeed, NRAs are there to regulate, hence have incentives to maintain regulation. More critically, they may tend to maintain and extend the same sort of regulation they are used to and that was designed for old legacy networks to new Schumpetarian infrastructures deployed under competitive conditions75. Indeed, several authors have already warned the NRAs against the tendency to apply regulation developed for narrowband network into the broadband world76.

72

See Article 13(2) of DG Information Society Working Document of 27 April 2000 on a Common regulatory framework for electronic communications networks and services. This working document added to the present SMP definition the following text: “and, where (a) undertaking has financed infrastructure partly or wholly on the basis of special or exclusive rights which have been abolished, and there are legal, technical or economic barriers to market entry, in particular for construction of network infrastructure; or (b) the undertaking concerned is a vertically entity owning or operating network infrastructure for delivery of services to customers and also providing services over that infrastructure, and its competitors necessarily require access to some of its facilities to compete with it in downstream market.” 73 As suggested by Recommendation on relevant markets, recital 16 in fine. 74 Tellingly, if the first speeches of the Commissioners in charge of Information Society or Competition were referring to ‘less regulation’, they are now referring to the more cautious notion of ‘better focus regulation’, Monti (2004). 75 This danger is all the more important than the 2003 Framework has considerably extend the scope of regulation from telecommunications to all electronic communications networks. Thus, the telecom model has been extended to media infrastructure (Nihoul and Rodford, 2004:40). 76 Crandall and Alleman (2003); Gual (2002); Criterion Economics and Dotecon (2003).

18

This regulatory incentive is not pure speculative theory but has been observed in practice77. Under the 1998 framework, there was an extension of regulation in the fixed segment (from local loop unbundling to bitstream access to part circuits) as well as in the mobile segment (cost orientation of termination charges)78. This extension has continued during the initial application of the 2003 framework. It is true that some deregulation has taken place. For instance, the minimum set of leased lines has been reduced from seven to five types of lines, or the mobile access and origination market has been de-regulated in the UK or in Austria. Even though these steps have been welcomed and (too) much underlined by the Commission79, that should not hide the fact in many other market segments, regulation is actually increasing. For instance, the ERG suggests that in the fixed world that bitstream will be strictly regulated and that non-reciprocity interconnection charges might be allowed, and in the mobile world that termination and roaming charges will have to be cost-oriented80. The thrust of the Common Position on remedies is also fairly regulatory81. More fundamentally, the whole remedy discussion is focused on the five remedies of the Access Directive, and is not very creative in reflecting upon remedies that may really lead to de-regulation (like spectrum trading, bill-and-keep or receiver party pays principle) and to infrastructure competition (like removing any impediment to the development of VoIP such as availability of numbering). To balance the regulatory incentives, every regulatory brake at the market selection step but also at the other steps of the SMP regime should be used to the fullest extent possible. Thus, the other institutions that are involved in market selection and that may have less incentive to regulate should play their check and balance role: the Commission should be wary to remove markets from its next Recommendation as explained below and strictly control any addition by individual NRA under Article 7 review, the NCA should actively co-operate with their regulatory counterparts and the Courts should review strictly NRAs decisions in case of legal dispute. In addition, during the SMP assessment (step 2), the NRA should undertake a comprehensive analysis of market conditions and not simply calculate market shares. They should in particular focus on entry barriers. During the choice remedies (step 3), the NRA should take care to impose only proportionate remedies82. Only then, we may expect some deregulation.

4.2. The next Recommendation on relevant markets The ERG argues that the three selection criteria have already been applied by the Commission for the first Recommendation of February 200383. However, I submit that is not the case because the Commission was bound to include all the markets listed in the Annex I of the 77

See Cave (2004b), Stern (2004), Waverman (2003). In some way, the exchange of best practices under the 1998 package seem to be one-sided. It refers to regulatory practices, and not to de-regulatory ones. 79 IP/03/1114 of 25 July 2003, and IP/03/1203 of 5 September 2003. 80 ERG Consultation Document of July 2003 on a Proposed ERG common position on FL-LRIC modelling ; PIBs of 20 November 2003 on the application of remedies in the mobile voice call termination market; ERG Common Position of 2 April 2004 on Bitstream Access, ERG(03) 33rev1. 81 Possibility of regulating emerging markets, extensive use of cost-orientation obligation, extensive reading of the possibilities to regulated offered by Article 9 of the Access Directive. 82 Article 8 of the Access Directive; Guidelines on market analysis, para 117-118; Common Position on remedies, section 4.2.1. 83 Common Position on remedies, p. 20. 78

19

Framework Directive and there is no guarantee whatsoever that these markets would meet the selection criteria. Indeed the Annex I was not drawn on the basis of the three criteria (which even did not exist at the time of the drafting of the Framework Directive), but included all markets regulated under the 1998 framework plus one additional market (international roaming) inserted under the political pressure of the European Parliament. Thus, the Recommendation might have counted fewer markets where the criteria had been fully applied84. In any case, for the second Recommendation foreseen at the end of 200585, the Commission will not any more be bound by Annex I of the Framework Directive, but instead by the three selection criteria only. As stated earlier, the Commission may be tempted not to apply such difficult criteria and instead include in its next Recommendation the markets that are regulated by a certain proportion (say the majority) of Member States. That would simple (thus probable) but unfortunate. Indeed, it may lead to illegality if a regulated market does not meet the three selection criteria or vice versa, and/or lead to over-regulation if under the leadership of some NRAs, the majority has followed by inappropriately regulate a market. Thus, the Commission should strictly apply the selection criteria and it is thus useful at this stage to sketch their practical implication to determine which markets may possibly be removed or added. As market definition is based on demand and supply substitutability of services provided and not on the underlying technologies, the starting point to determine the scope of SMP regulation is thus the service and not the technology. Note that this approach differs from the equally valid approach adopted by the Commission in recent policy paper or studies focusing on emerging technologies and aims to provide clarity for investors in such technologies (like 3G86, Voice over IP87, Power-line Communications88, or the more generic concept of Next Generation Networks89) Fundamentally, the Commission will need to assess if there is facilities-based competition, i.e. if different networks compete against each other (being fixed telecom, fixed cable TV, mobile 2G or 3G, …), or in other words if the much discussed convergence (fixed and mobile as well as telecom and broadcasting) is really taking place in the market. Thus, if the Commission were to conclude that there is substitution between fixed and mobile for narrowband and broadband services90, then many markets should probably be removed from the next Recommendation. Similarly, if the Commission was to conclude there is substitution between telecom, cable and say powerline networks, much of the fixed markets may be removed. At the current pace of technological and market developments, the most probable candidates for de-regulation are the following. First, the retail market for fixed access and telephone services (market 1 to 691). If the wholesale markets are efficiently regulated, there remain no 84

For instance, all of some of the markets that are considered below for removal could possibly not have been include in the initial Recommendation on relevant markets, were the three criteria had been fully applied. 85 As stated by the Commission on the 30 June 2004. 86 Communication of 11 June 2002, Towards the Full Roll-Out of Third Generation Mobile Communications On 3G, COM (2002)301. 87 Communication from the Commission on the Status of voice on the Internet under Community law, OJ [2000] C 369/3; Commission Staff Working Document of 14 June 2004 on The Treatment of Voice over Internet Protocol under the EU Regulatory Framework; Analysis (2004). 88 Draft Commission Recommendation on broadband communications through powerlines, COCOM-04-50 of 30 June 2004. 89 Devoteam and Siticom (2003); Elixmann and Schimmel (2003). 90 For the UK, see Dotecon (2001). 91 The numbering refers to the Recommendation on relevant markets of February 2003.

20

important entry barriers on the retail market, hence the first specific criterion is not met. This rationale coincides with the Universal Service Directive92 stating that retail regulation should only be imposed if wholesale regulation is inefficient93. In addition, the development of VoIP and its radically different business model (and retail tariffs) will probably undermine even more the need for regulation in local, national and international call services (market 3 to 6). Second, the wholesale market for fixed transit (market 10). Except maybe on thin routes, the common cost of the core network may be shared between important traffic and the economic entry barriers are limited, hence the first specific criterion will not be met94. Indeed, the Commission noted that depending on the experience gained over the implementation, the need for the transit market’s continued inclusion will be assessed95. Third, the wholesale market for trunk segment of leased lines (market 14). The same rationale than in the case of fixed transit applies here96. Fourth, the wholesale market for mobile access and call origination (market 15). Member States have on average three to five mobile operators and the number of operators and the competition will increase with 3G rollout, hence the economic entry barriers are not so high. There are legal barriers due to spectrum scarcity, but there is enough competition behind the barriers, hence the second specific criterion will not be met97. In any case, the best way to improve competition in the mobile sector would be to introduce spectrum trading outside the SMP regime98 and not to regulate the existing mobile operators. Indeed, the Commission anticipated that this market could be removed in the future recommendations99. Fifth, the wholesale market for international roaming (market 17). The rationale to remove such market is similar to the mobile access and call origination market. There are legal entry barriers due to spectrum scarcity, but there is enough potential to get competition behind these barriers, hence the second specific criterion will not be met. Note that this argument supposes that the international roaming market comprises all the national mobile networks, which is the definition adopted by the Commission in the Recommendation on relevant markets100. If some markets may be removed from the next Recommendation, others may also be added With regard to non-emerging markets, two possible candidates are: wholesale market for premium rate services and wholesale market for SMS termination which exhibits the same economic characteristics than mobile voice call termination. Moreover, the wholesale market 92

Universal Service Directive, Article 17(1). Cave (2004a:40) shares this opinion. However, Laffont and Tirole (2000:163) argue that if wholesale markets are regulated, it is better to regulate the retail markets as well through a global price cap instead of leaving them unregulated. Indeed, if regulation apply to wholesale markets alone, it create strong incentive for the dominant operators to reap its monopoly rent on the retails market using foreclosure devices (on price and non-price) that are very difficult for a regulator to detect. 94 See Koboldt (2003). 95 Explanatory Memorandum of the Recommendation on relevant markets, p. 19. 96 This market is not regulated in Austria (Commission Decision of 8 July 2004, AT/2004/74). 97 This market is not regulated in Austria (Commission Decision of 25 June 2004, AT/2004/63), and in United Kingdom (Commission Decision of 29 August 2003, UK/2003/1). 98 Framework Directive, Article 9(3). On the usefulness of spectrum trading, see Analysis, Dotecon, Hogan & Hartson (2004); Cave (2002); Valletti (2003). 99 Explanatory Memorandum of the Recommendation on relevant markets, p. 30. The possible Commission veto against the Finnish NRA that wanted to regulate such market goes in the same direction. 100 Moreover, in its Statement of Objections against O2 and Vodafone cited at note XXX, the Commission implies that for the future, roaming market tend towards effective competition. If however, the market delineation of international roaming would be limited to each individual network, that means that there is no competition between different networks, hence it automatically implies that such markets should continue to be selected for regulation, see note XXX. 93

21

for broadcasting transmission services to deliver broadcast content to end users (market 18) should be refined and possibly be broken down according to technological platforms101. With regard to emerging market, as explained above, if an emerging technology is used (partially or totally) to provide a existing regulated services (case 3), then either the existing service is deregulated or the new technology is regulated (as the existing service continues to be regulated). If the emerging technology is used to provide a new service, there should not be regulation for a certain period of time. Thus if operators decide to place fibre until home of end-users to provide new broadishband services like Video-on-Demand, their markets should not be included in the Recommendation. Similarly, if Next Generation Networks were set up to provide emerging services, they should not be included in the Recommendation. To sum up, precise indications on possible future regulation may only be given with regard to service and not technology. As it is not probable that a full convergence will take place in the near future, many markets will continue to be included in the next Recommendation. However, some of them in the fixed and mobile sector may be removed. Finally, new infrastructures that provide new services will not be included in the next Recommendation.

A Possible Next Recommendation on relevant markets Candidates for removal: 1. Access to the public telephone network at a fixed location for residential customers 2. Access to the public telephone network at a fixed location for non-residential customers 3. Publicly available local and/or national telephone services provided at a fixed location for residential customers 4. Publicly available international telephone services provided at a fixed location for residential customers 5. Publicly available local and/or national telephone services provided at a fixed location for nonresidential customers 6. Publicly available international telephone services provided at a fixed location for non-residential customers 10. Transit services in the fixed public telephone network 14. Wholesale trunk segments of leased lines 15. Access and call origination on public mobile telephone networks 17. International roaming on public mobile networks Candidate for re-definition: 18. Broadcasting transmission services, to deliver broadcast content to end-users Candidates for addition: - Wholesale premium rate telephony - Wholesale SMS termination on individual mobile networks

5. Conclusion

101

Note than an additional candidate would be the wholesale market for satellite transmission, but given its trans-national nature, it should follow another procedure. It should be identified in a Commission decision that NRAs are bound to follow (Framework Directive, Article 15.4).

22

Sunset clauses and flexibility were two buzzwords of the 99 Review and explain the radical reform of the SMP regime. If much of the literature focused on the two last steps of the regime, the most radical reform took place at the hidden first step, i.e. the selection of markets for possible regulation that embodies the rationale for sector regulation. The 2003 framework changes the substantial as well as the institutional (or procedural) approach. On substance, an opaque and political test is replaced by a more transparent generic test based on the relative efficiency of antitrust compared to sector regulation. On institution and procedure, a selection by the European legislator (European Parliament and Council) with the heavy co-decision procedure is replaced by a selection by the Commission and the NRAs (in co-operation with the NCAs and under the control of national and European Courts) with a light procedure for revision. The reform, which introduces market-by-market sunset clauses102 based on an efficiency criterion, may well entail a progressive removal of sector regulation to the benefit of a mere application of competition law, especially if the much trumpeted convergence takes place and leads to facilities-based competition. However, the reform may also (and more probably) not radically change the trends observed under the 1998 Framework entailing more regulation, in particular if the Next Generation Networks multiply the number of bottlenecks that would be inefficiently regulated by antitrust. That should not be problematic if the selection criteria are applied strictly, and the emerging market safeguard clause is respected. However, that may become a problem if criteria are applied so loosely that markets are not removed quickly enough from regulation, and the emerging market clause not respected such that new markets are included too rapidly in the realm of regulation. These risks are important because NRAs are used to a specific regulation adapted for the old legacy networks and may want to transpose such regulation to new Schumpetarian infrastructures developed under competitive conditions. To counterbalance this over-regulation incentive, all regulatory brakes in place in the directives should be used. Thus the NRAs should justify the application of the selection criteria. More importantly, authorities that have less incentive to regulation, like the Commission, the NCAs and the Courts should control strictly the work of the NRAs. In particular, I suggested some markets that are candidates for removal of the next Recommendation on relevant markets. La route de l’enfer est pavée de bonnes intentions. Surely, the change of the European regulatory paradigm is based on sound premises. It should deliver a more flexible and efficient economic regulation. But it carries the dangers of legal uncertainty and regulatory creep, which may lead the sector to the Hell. The application of the new directives might end up like the much criticised application of the 1996 US Telecom Act with pervasive (and perverse) regulation, multiple legal challenges and a surge in regulatory costs. Conscious of these risks, NRAs, the Commission and the national and European judges should apply their new powers with intelligence and restraint. Then, progress will be made on the way of the so promising e-Society!

References

102

Cave (2004a:40).

23

Abernathy K.Q. (2002), “The Nascent Services Doctrine”, Speech 11 July. Analysis (2004), Internet Protocol voice and associated convergent services, Study for the European Commission. Analysis, Dotecon, Hogan & Hartson (2004), Conditions and options in introducing secondary trading of radio spectrum in the European Community, Study for the European Commission. Armstrong M. (2002), “The Theory of Access Pricing and Interconnection”, in Cave M., Majumdar S., Vogelsang I (eds), Handbook of Telecommunications Economics V.I, North-Holland, 295-384. Australian Productivity Commission (2001), Telecommunications Competition Regulation, Final Report, AusInfo. Bavasso A. (2003), Communications in EU Antitrust Law: Market Power and Public Interest, Kluwer. Bavasso A. (2004), “Electronic communications: a new paradigm for European regulation”, Common Market Law Review 41, 87-118. Bishop S. and Walker M. (2002), The Economics of EC Competition Law, 2nd ed., Sweet & Maxwell. Bourreau M. (2003), “Local loop unbundling: the French case”, Communications & Strategies 49, 119-135. Cave M. (2002), Review of Radio Spectrum Management, Report for the DTI and HM Treasury. Cave M. (2003), Remedies for Broadband Services, Paper for the European Commission. Cave M. (2004a), “Economic Aspects of the New Regulatory Regime for Electronic Communication Services”, in P. Buiges and P. Rey (eds), The Economics of Antitrust and Regulation in Telecommunications, E. Elgar, 27-41. Cave M. (2004b), “Ofcom and light touch regulation”, in C. Robinson (ed), Successes and Failures in Regulating and Deregulating Utilities, E. Elgar, XXX. Cawley R.A. (2004), “The new approach to economic regulation in the electronic communications sector in Europe: the application of regulatory remedies”, Jour. of Network Industries 5, 3-21. Crandall R.W., Sidak J.G., Singer H.J. (2002), “The Empirical Case Against Asymmetric Regulation of Broadband Internet Access”, Berkeley Technology Law Journal 17, 953987. Crandall R.W. and Alleman J.H. (2002), Broadband: Should We Regulate High-Speed Internet Access?, AEI-Brookings Joint Centre. Criterion Economics and Dotecon (2003), Competition in broadband provision and its implications for regulatory policy, Report for the Brussels Round Table. Demsetz H. (1968), “Why regulate utilities?”, Jour. Law and Economics, XXX. Devoteam and Siticom (2003), Regulatory implications of Next Generation Networks and other developments in electronic communications, Study for the European Commission. Doherty B. (2001), “Just what are essential facilities?”, Common Market Law Rev. 38, 397436. Dotecon (2001), Fixed-Mobile Substitution, Report prepared for BT. Elhermann C.D. (2000), “Introduction”, C.D. Ehlermann and L. Gosling (eds), European Competition Law Annual 1998: Regulating Communications Markets, Hart. Elixmann D. and Schimmel U. (2003), “Next Generation Networks and Challenges for Future Competition Policy and Regulation”, Communications & Strategies 50, 239-268. Fuss M.A. and Waverman L. (2002), “Econometric Cost Functions”, in Cave M., Majumdar S., Vogelsang I. (eds), Handbook of Telecommunications Economics, V.I, North-Holland, 144-177.

24

Garzaniti L. (2000), Telecommunications, Broadcasting and the Internet: EU Competition Law and Regulation, 1st ed., Sweet & Maxwell. Garzaniti L. (2003), Telecommunications, Broadcasting and the Internet: EU Competition Law and Regulation, 2nd ed., Sweet & Maxwell. Geradin D. and Sidak J.G. (2003), “European and American Approaches to Antitrust Remedies and the Institutional Design of Regulation in Telecommunications”, in M. Cave, S. Majumdar, I. Vogelsang (eds), Handbook of Telecommunications Economics v.II, North-Holland, forthcoming. Gual J. (2002), “Regulation and the development of electronic communications in Europe”, Info 4, 42-49. Harbord D. and Hoehn T. (1994), “Barriers to Entry and Exit in European Competition Policy”, Int. Rev. of Law and Economics 14, 411-435. Hocepied C. (2002), “The new regulatory framework for electronic communications: From sector specific regulation to Competition Law”, Speech 20 May. Intven H. (2001), Telecommunications Regulation Handbook, World Bank. Kerf M. and Geradin D. (2003), Controlling Market Power in Telecommunications: Antitrust vs Sector-Specific Regulation, Oxford University Press. Koboldt C. (2003), Regulatory obligations to be imposed on operators with significant market power: narrowband services, Paper for European Commission. Krüger R. and Di Mauro L. (2003), “The Article 7 consultation mechanism: managing the consolidation of the internal market for electronic communications”, Competition Policy Newsletter 3, 33-36. Laffont J.J. and Tirole J. (2000), Competition in Telecommunications, MIT Press. Larouche P. (2000), Competition Law and Regulation in European Telecommunications, Hart. Larouche P. (2002), "A closer look at some assumptions underlying EC regulation of electronic communications", Jour. of Network Industries 3, 129-149. Larouche P. (2003), “What went wrong: the European perspective”, in Report of the Second Conference of the Round Table Expert Group on Telecommunications Law, IviR University of Amsterdam. Liebowitz S.J. and Margolis S.E. (2002), “Network Effects”, in Cave M., Majumdar S., Vogelsang I (eds), Handbook of Telecommunications Economics V.I, North-Holland, 7596. Lipsky A.B. and Sidak J.G. (1999), “Essential Facilities”, Stanford Law Rev. 51, 1187-1248. Maxwell W. (2002), Electronic Communications: The New EU Framework, Oceana Publications. Monti M. (2000), “Competition and Information Technologies”, Speech 18 September. Monti M. (2004), “Remarks at the European Regulators Group Hearing on Remedies”, Speech 26 January. Motta M. (2004), Competition Policy: Theory and Practice, Cambridge University Press. Motta M. and de Streel A. (2003), “Exploitative and Exclusionary Excessive Prices in EU Law”, in C.D. Ehlermann and I. Atanasiu (eds), European Competition Law Annual 2003: What is an Abuse of Dominant Position, Hart, forthcoming. Nihoul P. and Rodford P. (2004), EU Electronic Communications Law: Competition and Regulation in the European Telecommunications Markets, Oxford University Press. Posner R. (XXX), “Natural monopoly and its regulation”, XXX. Scherer J. (1998), Telecommunications in Europe, 4th ed., Sweet & Maxwell. Sharkey W.W. (2002), “Representation of Technology and production”, in Cave M., Majumdar S., Vogelsang I (eds), Handbook of Telecommunications Economics V.I, North-Holland, 180-222.

25

Sidak J.G. (2003), “The Failure of Good Intentions: The WorldCom Fraud and the Collapse of American Telecommunications After Deregulation”, Yale Jour. of Regulation 20, XXX. Squire-Sanders-Dempsey and Analysis (1999), Consumer demand for telecommunications services and the implications of the convergence of fixed and mobile networks for the regulatory framework for a liberalised EU market, Study for the European Commission. Squire-Sanders-Dempsey and WIK (2002), Market Definitions for Regulatory Obligations in Communications Markets, Study for the European Commission. Stern J. (2004), “Regulatory forbearance: why did Oftel find it so hard?”, Telecommunications Policy 28, 273-294. de Streel A. (2003a), “Market Definitions in the New European Regulatory Framework for Electronic Communications”, Info 5, 27-47. de Streel A. (2003b), “The Integration of Competition Law Principles in the New European Regulatory Framework for Electronic Communications”, World Competition 26, 489514. de Streel A. (2004), “Remedies in the Electronic Communications Sector”, in D. Geradin (ed), Remedies in Network Industries: EC Competition Law vs. Sector-specific Regulation, Intersentia, 67-124. de Streel A. and Queck R. (2002), “La puissance significative nouvelle est arrivée! L’évolution de la réglementation asymétrique dans le secteur des communications électroniques”, Revue Ubiquité – Droit des technologies de l’information 13, 43-71. Veljanovski C. (2001), “Assessing EC Communications Regulation”, Mimeo. Valletti T. (2003), Obligations that can be imposed on operators with significant market power: Access to public mobile networks, Paper for the European Commission. Vernon Y. (2002), L’accès aux réseaux et le partage des ressources essentielles dans l’Union européenne: L’exemple des télécommunications, Ph.D. University Rennes I. Walden I. and Angel J. (2001), Telecommunications Law, Blackstone Press. Waverman L. (2003), “Regulatory incentives and deregulation in telecommunications”, in C. Robinson (ed), Competition and Regulation in Utility Markets, E. Elgar, 138-159. Werden G.J. (1987), “The law and economics of the essential facility doctrine”, St Louis University Law Jour. 32, 433.

26