Does network sharing conflict with competition?

21 November 2012

Colin Long

This article was first published in Policy Tracker in November 2012.

So far regulators think the positives outweigh the negatives but will that change in the era of cognitive radio?

There was a time in the very recent past of the development of modern telecommunications when the great beacon for regulators was network and inter-platform competition because it was thought that only through differentiation at the infrastructure level would real and independent competition develop. This was a model which gained much popularity during the 1990s and 2000s, particularly in relation to fixed networks. As regards to mobile networks, the model also had the added attraction for governments looking to raise money on the back of spectrum auctions that, the more participants, the greater the number of auction bidders and the higher the prices paid. 

In the present era of severe budget constraints, mobile network operators are understandably looking to make cost savings (in some cases of up to 30% of CAPEX and/or OPEX), not only to maximise the return on their assets, but also to: conserve funds to bid for yet further spectrum in future auctions, speed up coverage, and limit the impact of network installations on the environment.

Infrastructure sharing can also bring societal benefits, in terms of reaching underserved areas, expanding coverage and minimising unnecessary duplication.

Regulators have meanwhile become more sympathetic to network sharing in its current manifestations. For example in the EU the revised EU regulatory framework not only encourages sharing of passive infrastructure as being beneficial for town planning, public health and environmental reasons, but also requires that NRAs should mandate sharing in certain circumstances.

Thus, sharing ‘passive’ [1] , and, increasingly, ‘active’ [2] infrastructure between multiple mobile network operators has become more and more fashionable. 

This article looks at some of the main network-sharing deals to date, the lessons learned, the reactions of regulators and the legal/regulatory pointers emerging for future sharing arrangements as, in all probability, they become more sophisticated and penetrating. 

In the beginning, and until recently, sharing was mostly limited to passive infrastructure. Now, in a growing number of instances, mobile operators are extending the concept to active infrastructure in the form of the radio access network that connects base-stations to customers’ handsets.  In a report by BEREC [3] and the RSPG [4]  to the European Commission in 2011 [5], BEREC noted that operator responses to its questions of them indicated the number of sharing agreements would increase in the future. The following table illustrates this trend:

TerritoryPartiesAssets sharedDate
SwedenTele 2 and TeliaActive elements2001
UKO2 UK Ltd and T-Mobile UK LtdPassive elements2003
GermanyT-Mobile Deutschland and O2 GermanyPassive elements; possibility of future RAN-sharing2003
SpainOrange and VodafoneRAN sharing (3G)2006
UKT-Mobile UK and 3UK(Multi-operator) MORAN 

UK, Spain,
Ireland, Germany

Telefonica O2 and VodafonePassive Elements 2009
FranceOrange, Bouygues, SFR and Free3G RAN-sharing (to cover smaller towns)2010
DenmarkTelia and TelenorRANs (2G, 3G, 4G+)2012
PolandT-Mobile and OrangeRAN sharing on GSM/HSPA networks2012
IrelandH3GI and VodafonePassive elements2012
UKVodafone and Telefonica UKSite management businesses merged; passive and active elements (MO-RANS)2012

The whole concept of sharing immediately suggests compromise and coordination, so, unsurprisingly, regulators, competition authorities and the European Commission have taken a cautious interest in these arrangements from the legal as well as competition policy perspective. Some regulators have issued their own general guidance [6], e.g. as to sharing network intelligence and sensitive customer data, and some have issued actual case decisions, whilst the Commission’s approach can be gleaned from its own cases.

As regards agreements with an EU dimension, the early O2/T-Mobile 3G network sharing agreement was reviewed by the European Commission [7] in 2003. It noted, having analysed the effects of the agreement in the economic context, that the parties were direct, well-established competitors, in a strong position to roll-out their networks individually. Therefore the agreed site sharing arrangements needed to be examined for compatibility with EU competition rules on restrictive agreements [8].

Nevertheless, the Commission found that in light of the parties:

• sharing a limited number of passive elements of the access network, but retaining control of their networks, including the core network;

• being able to differentiate their services downstream as they also retained independent control of their service platforms, there were no grounds for challenging the agreement.

At a national level, much more recently, the second phase of the network sharing between Vodafone and Telefonica UK, now enshrined in a company jointly owned by the parties, came up for review in September/October 2012 by the UK competition law enforcement authority, the Office of Fair Trading. The case is not so typical because its validity was considered only under UK merger control legislation [9] which is very technical and applies a different criterion (essentially whether there will be substantial lessening of competition) compared to a pure competition analysis under EU competition rules. The formation of the Joint Venture (JV) company constituted a merger qualifying for review because of the parties’ share of the market in supplying access to base station sites to other mobile network operators (MNOs).

The essential objectives of the joint venture were the parties’ management, through the joint company, of their passive site infrastructure, and their division of RAN design management and maintenance, as well as installation of new multi-operator RAN equipment by each of them, in their contractually agreed separate geographic regions of the UK.

The OFT cleared the Vodafone/Telefonica JV, concluding that the passive arrangements did not (a) give rise to a realistic prospect of inter-party coordination; nor (b) decrease the parties’ incentives to expand their downstream activities in mobile call access and origination. (For legal technical reasons [10] the RAN-sharing arrangements did not qualify as part of the merger and so escaped scrutiny). Thus, in the OFT’s view, the ‘merger’ did not give rise to a substantial lessening of competition.

The OFT was clearly influenced and encouraged in this decision by the fact that third parties like the other MNOs did not register concerns with the OFT regarding either the unilateral effects of the passive infrastructure arrangements or incentives on the parties to expand downstream (though third parties apparently were concerned about potential coordination; the OFT, however, took the view coordination would not be increased beyond the parties’ existing sharing arrangements). 

Whether the OFT should possibly have been more concerned about the competition aspects will be interesting to review once the UK 4G auction has been completed in early 2013. At the moment there are in practice only two infrastructure owner/operator groupings in the UK, EE/H3G and Vodafone/ Telefonica. If a new entrant was to bid successfully for the auction spectrum reserved to a ‘fourth’ operator, this might raise some competitive tensions, such as in the context of gaining access to the other operators’ infrastructure.

Network sharing in Denmark

A case earlier in 2012, a RAN-sharing agreement between Telia Denmark and Telenor, is also instructive as to competition authority thinking in this area. Again via a joint venture company, the parties agreed to own, control and develop the active (RAN) infrastructure required for their respective mobile businesses (up to and including LTE-advanced) in Denmark, in order to achieve cost reductions and better coverage and technology. In principle the agreement could have infringed EU and Danish competition rules.

The Danish Competition Council (DCC) flagged several prima facie competition concerns, such as the risk of collusion on the wholesale market, reduced incentives to compete, the parties’ combined frequency resources exceeding their competitors, a reduction in masts and antennas available to competitors, risk of exchange of commercially strategic information, and reduction of competition in coverage and technology. However, after the parties gave binding commitments/conditions to the DCC designed to neutralise these concerns, the agreement was approved subject to these conditions.

This far and no further?

So we can see that to date in most, if not all, instances the competition authorities have been willing to allow active as well as passive mobile network infrastructure sharing, where (as the competition rules in essence require) there are at least clear benefits to competition and customers. The cases nevertheless also indicate several ‘bright lines’: for although competition and regulatory authorities’ thinking appears gradually to be evolving with the development of network sharing, if parties were to try to go beyond them, this might mean the agreement would be struck down unless the parties could come up with suitable modifications or undertakings. These bright lines for the time being appear to include e.g:

• Information exchange: sharing of data not to go beyond what is strictly necessary for the passive/active sharing, nor extend to network intelligence and sensitive customer data

• Competition: competition in the market for sites and site infrastructure, as well as in downstream markets, to be maintained or not adversely affected

• Coordination: coordination between the parties strictly limited to that necessary for the efficient planning, roll-out, installation, management and operation of their joint network sharing arrangements

• Control: the parties to retain independent control over their core networks/network intelligence, as well as service platforms

• Costs: tariffs for access to jointly-owned or managed radio-access capacity to be economic and not disincentivise the parties from competing.

Technology: a cognitive approach to network sharing?

In the future, undoubtedly new developments in technology will push the boundaries of network sharing and test the adaptability and efficacy of competition laws and regulations. For example, dynamic or cognitive sharing of spectrum, or virtual base stations - where dedicated software allows operators to operate one physical base station as several base stations at the same time. Nevertheless the reviewing authorities will continue to apply the same methodology, analysis and principles to assess whether there is likely to be a distortion or restriction of competition, involving not just those bright lines mentioned above but also the:

• degree of control retained by the parties over radio planning and freedom to add sites, and

• parties’ continuing ability to differentiate in terms of price, quality and variety of services.

As network sharing develops in sophistication so the tension will increase between the commercial objectives of the parties and the competition implications. This in turn will put pressure on the parties to anticipate the authority’s likely concerns and (as is commercially advisable anyway) build into their agreements sufficient protections for their own independent freedom of action in such a way as not to compromise the cost savings and efficiency gains from the venture. That said, the subject will remain topical as it does seem likely from the industry view that the number of operators seeking to enter into mobile network sharing arrangements of varying complexity will very likely increase in the future.

[1] Passive infrastructure, typically, is limited to sites, towers, masts, antennas, aerial support structures, base-station cabinets, cooling equipment and power supplies.
[2] Active infrastructure usually encompasses the radio access network (RAN), including base stations, radio network controllers and backhauls to the BSC/RNC.
[3] Body of European Regulators for Electronic Communications
[4] Radio Spectrum Policy Group 
[5] RSPG11-374 final
[6] e.g. France, Germany, The Netherlands, UK
[7] as the agreement had inter-Member State trade effects
[8] Article 101 TFEU
[9] Enterprise Act 2002
[10] The active arrangements did not form part of a ‘business’ as such, being confined to the parties themselves

This article was written by Of Counsel Colin Long.