think digital
European Telecommunications Network Operators' Association



ETNO response on BEREC consultation BoR (13) 186

PDF here

ETNO welcomes the review of BEREC’s common position on geographical aspects of market analysis.

ETNO believes that BEREC and NRAs should adopt a more proactive approach to reflecting the geographic differences within a national territory in their definition of relevant markets in order to effectively take into account the competitive dynamics of today’s telecommunications markets.

Executive Summary

ETNO welcomes the review of BEREC’s common position on geographical aspects of market analysis.

ETNO believes that BEREC and NRAs should adopt a more proactive approach to reflecting the geographic differences within a national territory in their definition of relevant markets in order to effectively take into account the competitive dynamics of today’s telecommunications markets.

Differences in the competitive conditions between areas within the national territory are an essential feature of today’s broadband markets in many EU member States. Delineating markets at the sub-national level allows identifying the right addressee for regulation, tailoring regulatory remedies and reducing regulation to the minimum necessary. It thereby contributes to a more investment-friendly framework which is conducive to improving the NGA broadband coverage in Europe.

NRAs have handled specific aspects of geographical segmentation differently in different EU Member States in the past.  The clarifications in the draft common position are therefore appreciated.  We support a number of BEREC’s findings in the consultation document. Our main comments are as follows:

  • As highlighted by BEREC, the absence of geographic pricing by the nationwide incumbent operator does not per se indicate that the market should be defined as national. If there is one nationwide incumbent with a national pricing but there are local or regional alternative operators in some parts of the territory, there clearly are geographic differences in prices from an end-user’s point of view who either has no, some or many choices in the respective areas.
  • The Ofcom approach of a geographic segmentation of market 5 is an example of a sustainable and transferable model for other EU countries. The application of geographic segmentation in the UK and also in Portugal gives evidence of the positive effects of such an approach on both investment and competition. The general principles used by those NRAs are in line with BEREC’s findings and should be used more consequently by all EU NRAs.
  • Broadband markets have developed in a very competitive manner, often based on several NGA networks being in place, such as cable and alternative fibre networks. NGA transition of one specific platform is not necessarily a barrier to implementing a geographic segmentation based on today’s infrastructure. As the UK case of market 5 shows, the impact of NGA deployment has to be assessed based on the concrete facts, not on speculations such as on potential future MDF-close downs.  The fact that adjustments to the geographic unit might become necessary at one point in the future does not impede the necessary delineation of sub-national markets. 

BEREC has launched a public consultation in December 2013 with the objective to provide new guidance and an updated Common Position on geographical aspects of market analysis. ETNO welcomes this initiative which is both timely and important.

ETNO believes that BEREC and NRAs should adopt a more proactive approach to reflecting the geographic differences within a national territory in their definition of relevant markets in order to effectively take into account the competitive dynamics of today’s telecommunications markets. Delineating markets at the sub-national level is in many cases a precondition for addressing regulation to the right entity, tailoring regulatory remedies to the actual competition problem in the market and reducing regulation to the minimum necessary. Geographical segmentation thus promotes a more investment-friendly framework which in turn is conducive to improving the NGA broadband coverage in Europe. The concept is applied successfully by several NRAs in Europe and has been discussed favourably by competition bodies. [1]

1. The rational for a focus on geographic market analysis

We fully agree with BEREC’s finding that the importance of geographical aspects has increased over the past years. More and more competing network infrastructures have developed, such as cable and local fiber networks. Alternative network operators, both those relying on own access infrastructure and relying on LLU, increased their coverage, performance range and market shares. This trend is expected to continue in the coming years. This in many markets leads to a ‘patchwork’ of areas with different numbers of infrastructures and operators posing significant competitive constraints at retail and wholesale level. The competitive conditions in these markets are strongly affected by these geographic differences and render the definition of a nationwide market unjustified. In this case, it is appropriate to define markets on sub-national level, to identify the right addressee for regulation, where applicable, and, within competitive areas, to abolish regulation. Infrastructure-based competition can thus be sufficient on its own to lead to effective competition on downstream markets.

As BEREC points out the definition of one national market with differentiated remedies does not constitute an equivalent to geographic market definition in the presence of uneven development of competition across the territory. Where no regulation would be warranted due to a significantly differing degree of competition, maintaining regulation in competitive areas, even using less intrusive remedies, would nonetheless distort competition in the market. The differentiation of remedies within a national market remains an alternative option, if sound and objective reasons exist which render the geographic segmentation into different local markets inappropriate. Sub-national markets should be defined whereever competitive conditions are not sufficiently homogenous across the national territory.

Many broadband markets in Europe have developed in a very competitive manner. A large number of broadband providers are offering services, often based on proprietary infrastructures. In many areas with alternative operators ETNO Members are competing with infrastructures such as cable and alternative fibre and don’t have significant market power anymore in such areas if looked at in a differentiated manner. In particular, as BEREC recognizes in the consultation document, cable operators have achieved strong regional or even national positions in a number of European markets where the investments on the upgrade of their networks have been substantial in recent years. [2]

2. Indicators for geographic segmentation and the role of national pricing

As stated by BEREC, the findings on relevant infrastructures are dependent on the prior definition of the relevant product market. ETNO shares BEREC’s view that on that basis (i) the coverage of competing regional or local infrastructures and (ii) the number and size of alternative operators based on these infrastructures are the leading indicators of the need to undertake a geographical analysis. When alternative infrastructures and operators are competing in the market this should be treated as clear evidence for the need of conducting such type of analysis.

Alongside the two above mentioned criteria indicating the need for geographical analysis, BEREC is also giving guidance as to whether pricing and price differences are indicative for sub-national markets. We agree with BEREC that the analysis of prices needs to be assessed and interpreted carefully. A nationwide uniform pricing does not in itself indicate that sub-national markets would be inappropriate.

When investigating geographic aspects and analysing pricing on respective markets, NRAs should  consider the closely linked aspect of geographic market share developments over time. For a geographic analysis it is more important to analyse whether there are different market shares and products (i.e. UBB offers only in some geographic areas)  in different regions and thus to reflect the customers’ perspective. Significant deviations in market shares between compared regions are a clear indication for separate markets. The question whether pricing is national or differs by geography is therefore not a conclusive indication of the nature of competition and price differences no precondition for a delineation of geographic markets.

BEREC’s clarification on this point is very important, as some national regulators have given too much weight on the presence of national pricing in the past and abstain from implementing a geographic segmentation just because of the absence of regional pricing. Moreover, BEREC states that geographic differences in pricing is already present from the customer’s point of view, when there is a national pricing of an incumbent operator, but different geographic presence of alternative operators. In the past, NRAs sometimes have narrowed that view on the geographic difference in the incumbent’s pricing.

3. Markets susceptibe to geographic segmentation and experience in Europe

ETNO agrees that the retail market analysis should be the starting point of every market analysis. The retail markets should be examined in detail, in particular as to whether the market exhibits strong and sustainable inter-platform competition. If present, it creates either direct or indirect competitive constraints on the upstream wholesale markets and must not be neglected. Most national regulators already follow this approach. We strongly agree on taking account of such competition, which becomes especially relevant in markets 5 and 6. [3]

In the wholesale broadband access market, the implementation of geographic segmentation has already been proven to be adequate and necessary. The approach taken by Ofcom in the UK gives a good example of an application of a geographic assessment within the market analysis procedure.

  • Ofcom is currently conducting the next market analysis for the wholesale bitstream access market in UK. Ofcom proposes to continue defining sub-national markets and to abolish regulation in areas in which the competition conditions are deemed sufficient. By doing so, Ofcom makes consequent use of the corresponding scope of the EU legislative framework and tailors regulation to the actual competitive developments in the market. The overall objective is to focus on the prevailing local competitive conditions and to effectively assess whether these differ between different geographies.
  • According to the draft consultation document, the WBA market is split into two groups.[4] The group of competitive areas consists of exchange areas in which BT and at least two more competing operators are present, whereby the latter are based on BT’s wholesale products or on own infrastructure. In this group Ofcom did not find SMP and would abstain from imposing remedies. The deregulated area would cover in total approx. 90% of the market.
  • Ofcom also applied geographic segmentation in Market 6 with the individuation of specific areas within London (Wecla areas: western, eastern and central London areas) where the conditions for the wholesale market were found competitive. Within the Wecla area Ofcom did not find SMP for Traditional Interface leased lines (8 to 622 Mbit/s), and for Multiple interface lines.
  • More recent academic work reveals important positive effects from the geographic segmentation in UK: more competition and more investment in deregulated areas.[5] This is further underlined by market developments. In 2013, FTTx-based broadband lines have been available to more than 55% of the market; 65% planned until April 2014. Average broadband speed has grown by 22% up to 14.7 Mbps. [6]

Portugal is another successful example of geographic segmentation of market 5, providing strong empirical evidence of the benefits of such an approach. Following the decision in 2009 to deregulate the wholesale broadband market in Competitive (Local Exchange) Areas, ANACOM also signalled that it would take a segmented and phased approach to fibre regulation: in competitive areas the focus would be put in the access to civil infrastructures (remedies already imposed) and fibre access regulation would only be considered if, for technical reasons such as capacity constraints, ducts and poles can’t be shared and access to dark fibre is also not possible. This deregulatory context not only intensified competition in competitive areas (as Portugal Telecom became able to match the prices of its main competitors) but also stimulated significant investments in next generation infrastructures: while the cable operators upgraded their networks to DOCSIS 3.0 (the biggest operator with 3.1 million homes passed in 2012), PT began the roll-out of its GPON fibre network, which reached 1.6 million homes in 2012. Two other operators co-invested since 2010 in a fibre network covering 0.4 million homes and, very significantly, in November 2013, Vodafone detailed its objective (framed within the Vodafone Project Spring) to "accelerate and extend FTTH build to over 1,5 M households" in Portugal, which will put its fixed NGA network at the same level as PT’s. In 2013, Altice has also announced its willingness to invest around €150 M in the next three and a half years in a NGA network.

In a number of other countries, such as Germany and Bulgaria, regulators have carried out or announced a collection of data by geography in order to provide a basis for their forthcoming round of market analysis in broadband markets and to allow a possible geographic segmentation.[7] 

In addition to the wholesale broadband access markets, we welcome BEREC’s efforts in bringing other markets of the Recommendation to attention, which are suitable for geographic segmentation, such as Market 1 – the market for retail fixed voice telephony access, which is due for deregulation in the ongoing review of the recommendation on relevant markets - and market 6 (terminating segments of wholesale leased lines) or potentially other markets that the regulator would deem necessary to analyse under specific national circumstances[8].

4. Proposed analysis to assess differences in the competitive situation

We broadly agree with the approach to analysing differences in the competitive situation. We would like to comment on two aspects, i) the retail market as starting point and ii) the choice of the geographical unit.

Retail market as the starting point: above, we already referred to the importance of competition conditions at retail level. We basically agree with the typical situations described by BEREC which need to be taken into account in the geographical analysis, that is i) retail competition driven by wholesale access and/or ii) driven by inter-platform competition. In particular the existence of operators with own infrastructures such as cable or fiber needs to be acknowledged. If the corresponding products are substitutes at retail level these operators pose direct and indirect competitive constraints which are important for the decision on sub-national markets. If inter-platform competition poses sufficient competitive constraints at the retail level, there is no need for maintaining or introducing regulatory interventions on the wholesale level.

Choice of geographical unit: we basically agree with BEREC in that the choice of the relevant geographical unit may to some extend depend on the particular situation. In case competition is based on Local Loop Unbundling, areas served by an exchange or main distribution frame are likely to be the appropriate unit. In case of inter-platform competition, alternative network’s structure is likely to be appropriate. However, often there is a mixed situation, i.e. competition on retail level stems from both wholesale access and the presence of alternative infrastructures (inter-platform competition). Experiences from advanced broadband markets such as Germany and the United Kingdom show that exchange or main distribution frame areas can be an appropriate and sustainable geographical unit.

What is the most appropriate unit and granularity is best analysed by the national regulator on a case-by-case basis under consideration of the specific market situation within the Member State. It is important to stress that – as BEREC also acknowledges - the indication contained in EC’s guidelines on market analysis (market comprises the area of the SMP operator; legal or other regulatory instruments) may not be flexible enough to adequately deal with today’s and future market development.

It is sometimes argued with respect to practicability that the costs for gathering relevant data for geographic differentiation would justify a wider market definition. However, we do not see that obtaining the necessary network-related information constitutes an insurmountable hurdle. Market analysis - not only in the course of geographical segmentation) - requires a procedure that by its nature generates a certain degree of complexity and involves extensive data gathering. The administrative burden of gathering data alone cannot justify ‘jumping’ directly to a national market definition. In most cases, technical data is linked to geographic information (such as the connected MDF and its address or coordinates) for maintenance and billing purposes. Such data operators have to keep  available. 

This perceived complexity has arguably resulted in many NRAs being reluctant to proceed with geographic segmentation even though the definition of the relevant geographic markets is a natural task of NRAs. Distortions of competition and over-regulation of incumbent operators have been the consequence, resulting in inefficiency. It is therefore essential to work towards application of the guidelines outlined in the draft Common Position by all NRAs in all cases.

5. Criteria for assessing differing competitive conditions

BEREC suggests a couple of criteria to be used to assess differences in the competitive conditions between different areas, such as barriers to entry into the market, number of suppliers, market shares, price differences and other aspects, e.g. marketing strategy, commercial offers, nature of demand etc.[9]

We agree that these criteria are useful for the task of assessing geographic differences in competition.  In particular  barriers to entry, number of suppliers and market shares should be used in the assessment as they are straightforward. Pricing issues on the other hand are less obvious when it comes to the question whether a market should be delineated as nation-wide or subnational.

As stated above, we welcome BEREC’S clarification on the question that if there is one nationwide incumbent with a national pricing and local or regional alternative operators, there clearly are geographic differences in prices form the customer’s point of view having either none, some or many choices in the respective areas, leading to a heterogenous pricing from a national perspective. Other criteria such a quality of service / service availability should also be taken into account if the national circumstances require it.

We agree upon BEREC’s finding, that “a segmentation based on a single criterion (e.g. the number of operators) will usually not be appropriate”. OfCom’s approach on market 5 is based on a well-suited (but not necessarily exhaustive) set of criteria, which can be transferred to other national markets:

  • Number of operators
  • Market share of the incumbent
  • Structure of the geographic units (such as the number of access lines per MDF)

6. Implications on NGA networks

In the consultation document, the transition from the copper-based network to NGA is discussed. It is implied that such a transition involves an active decomposition of the traditional network with MDF-site closures. First of all it has to be clarified that such a quick transition and the removal of the traditional network elements is very unlikely to happen, especially in markets with ULL-driven competition and high numbers of ULL-lines. The transition will take a long time and for the coming years an active removal of traditional networks will not be considered. The closing of MDF-sites on a significant scale is not foreseeable in most member states today. Furthermore, where the phase-out of MDF and other structural changes in the network are already foreseen, they are subject to a longer pre-notice period to allow market players to adapt.

NGA plays an important role in the upcoming (and past) market analysis, as the rollout of NGA networks might shift the market structure in various ways. BEREC successfully brings together possible NGA scenarios and geographic segmentation. We agree on the general point, that NGA and the upcoming developments do not produce such an uncertainty that the consequence would be to refrain from geographic segmentation.

The best proof can be found in the current review of Market 5 in the UK. Ofcom does not see any necessity in changing the basic principle of defining MDF-wise areas, even if the network structure might change in the future. Moreover, NGA even is not seen as a separate market apart form the traditional networks. Ofcom just integrates NGA into the present market and basically finds, that any changes will be assessed, when they are about to happen. Additionally, Ofcom’s approach is flexible enough to deal with “moving” areas, i.e. areas that have become effectively competitive as well as formerly competitive areas that are now grouped within the regulated sub-national market. Even for a forward looking approach on the market analysis, BEREC should underline that any changes involving NGA should be clearly foreseeable and certain when taken into account. As there is no long-standing experience in NGA transition, it is difficult to foresee if any MDF-reduction will take place in the next years and how fast the “traditional” ULL will vanish. In countries with strong ULL-based competition, this might take a very long time to migrate millions of ULL-lines also considering the issues related to the migration of all services currently provided on copper (i.e. ISDN, analogic leased lines, uninterrupted access to emergency services, etc). In the meantime, this must not hamper an appropriate geografic delineation. NGA-driven changes in the market are moreover more relevant in scenarios where competition relies mostly on LLU competition. In case of alternative competing infrastructures, an NGA migration of one platform will not be a hurdle for defining sub-national markets in any case.

We agree with BEREC that absence of a commercial wholesale offers (fiber unbundling) does not necessarily signal a non-competitive environment at the retail level. This issue should be analysed by NRAs on a case-by-case basis.

ETNO welcomes that BEREC recognizes the importance of the availability of wholesale offers for access to civil infrastructure and of the feasibility of access by operators to in-house wiring infrastructure which, as highlighted in the consultation document (footnote 71), are a key regulatory condition to ensure the prospects of effective retail competition in certain Member States.

7. Risks and benefits of a geographic analysis

BEREC discusses a number of perceived risks associated with the definition and deregulation of sub-national markets, e.g. with regards to business users.

We agree with BEREC, that business customers’ may have a nation-wide demand but that this does not per-se require a nation-wide market. Business customers can rely on multiple commercial offers within competitive areas and regulated products elsewhere.

In some EU-countries, NRAs have defined separate markets for business customer products and mass market products mainly for retail customers. In those cases, the nation-wide demand for business products has to be assessed within the respective productl market. If there is one market for both mass market and business customers, the relevance of business products might be minor compared to the great quantity of mass market demand, rendering the effect of nation-wide demand on such products less relevant confirming the BEREC’s opinion that the presence of a nation-wide wholesale demand is not per se an indicator of a national wholesale market.

It is appreciated that BEREC classifies geographical segmentation as a means to limit ex-ante regulation to the minimum extend needed and finds that geographical segmentation can be beneficial because the former SMP operator is now more flexible to innovate, invest, and offer better prices and quality. Geographical segmentation thus constitutes an important lever to promote a more investment-friendly framework and ultimately to improve NGA broadband coverage in Europe.

We welcome BEREC’s discussion on type 1 and type 2 errors. In areas with intense competition, to not revoke regulation completely can hinder innovative behavior of the regulated operator in designing its products and setting prices. As a result, in these areas the market and end-users miss out on competitive impulses than would be possible in a deregulated environment.

[1] For example, geographic segmentation is supported by the German Monopolies Commission. which in its recent special report on competition in the telecoms market explicitly invites the NRA to consider a geographically differentiated regulation in the forthcoming analysis of market 5 and stresses that geographical segmentation opens up the potential to reduce regulatory intervention in a faster manner than it would be possible within a national market. The Monopolies Commission finds that, the risks for competition and infrastructure rollout sometimes associated with geographic segmentation as well as practical difficulties of such an approach are being exaggerated (Monopolkommission, Sondergutachten 66, Dezember 2013).

[2] In Portugal, the cable footprint reached 77% of households in 2012, 98% of which fully upgraded to DOCSIS 3.0, making one of the cable operators (now merged with a mobile operator owning a LTE network with 80% coverage of the population) the market leader in many areas and the sole provider of fixed NGA in many others. In Germany, cable operators have doubled their number of broadband lines since 2009, now approaching 4.8 million. In Belgium, cable operators have also a very high penetration (circa 95%) and have almost fully upgraded their network to DOCSIS 3.0. In Flanders, which represents 2/3 of the Belgian territory the cable operator (Telenet) has become the market leader.

[3] Though it may also impact market 4.   

[4]    Another group covers the HULL area.

[5]    Falck, O. und Fabritz, N.: „Investment in Broadband Infrastructure Under Local Deregulation: Evidence from the U.K. Broadband Market”, CESifo Working Paper No. 4277, 06/2013.

[6]    Ofcom: “UK fixed-line broadband performance”, 05/2013; Point Topic: “Mapping broadband in the UK – Q2 2013”, 08/2013.

[7]   The Bulgarian NRA CRC plans to survey the retail broadband market both at national level and regional levels - in 6  large regions and 28 administrative districts - and has already distributed invitations for indicative offers to market research agencies at the end of 2013. In Germany, BNetzA has requested data on a geographic basis for its M5 analysis.

[8] In Belgium, the regulator has defined a broadcast market and identified five differentiated broadcast markets corresponding to each cable operator’s geographic foorprint.

[9] In Greece, OTE has provided in the past data and has shown concrete evidence on the difference in competitive conditions in different geographic areas. Data shows that 69% of broadband subscribers are served by at least four (4) operators and OTE’s broadband market share in these areas is 38% and retail access market share is 49% (data: end 2013, analysis carried out by OTE). But even though evidence is tangible, the EETT (NRA) has never taken into consideration such data and has defined OTE as having SMP nationwide while remedies have not been geographically differentiated, prohibiting OTE from competing in these areas on an equal footing.

PDF here

Members & Observers - View companies map

  • Albtelecom
  • Proximus
  • Eir
  • Elisa Communications Corporation
  • GO Plc (Malta)
  • Hrvatski Telekom
  • Koninklijke KPN
  • Lattelecom
  • Magyar Telekom
  • Makedonski Telekom
  • Orange
  • OTE
  • Portugal Telecom
  • Síminn (Iceland Telecom Ltd.)
  • Slovak Telekom
  • Swisscom
  • TDC
  • TDF
  • Telecom Italia
  • Telefónica
  • Telekom Austria
  • Telekom Slovenije
  • Orange Polska
  • Telenor
  • Telia Company
  • Turk Telekom
  • Vivacom
  • Deutsche Telekom AG
  • CYTA (Cyprus Telecommunications Authority)
  • BT
  • BH Telecom
  • POST Group

ETNO aisbl - Boulevard du Régent 43-44 - 1000 Brussels - Belgium

Phone: +32(0) 2 219 32 42 -