Advice to Government on the consumer and competition issues relating to liberalisation of 900MHz and 1800MHz spectrum for UMTS

Advice to the Secretary of State for Business, Innovation and Skills

Executive summary

1.1 This paper is a summary of Ofcom's analysis of the consumer and competition issues that might arise from liberalisation of 900MHz and 1800MHz spectrum for 3G mobile services (UMTS), and the actions that might be taken to address any material distortions of competition arising. This builds upon our February 2009 consultation , including responses thereto, and developments since then.

1.2 This advice was prepared in response to a request by the Secretary of State for Ofcom's analysis of the impacts on consumers and competition of liberalising the 900MHz and 1800MHz spectrum for UMTS use.

1.3 In summary, Ofcom considers that liberalising 900MHz and 1800MHz spectrum in the hands of the existing licensees is likely to benefit consumers and is unlikely to result in a material distortion of competition that requires further action to be taken.

2G liberalisation

1.4 The 900MHz and 1800MHz spectrum is currently licensed for use in the UK for the provision of GSM services only. This advice considers the likely effects of varying the current restrictions on the use of that spectrum, to allow UMTS services in addition to GSM services.

1.5 In considering this question, in light of our principal duties under the Communications Act 2003 we have looked at both the likely effects on consumers and on competition of liberalisation of these spectrum bands.

Consumer benefits of liberalisation

1.6 Allowing 900MHz and 1800MHz spectrum to be used to deliver 3G (UMTS) mobile services is likely to bring significant benefits to consumers in some or all of the following ways:

• Greater network capacity allowing more customers to be served and to enjoy higher mobile broadband speeds (both 900MHz and 1800MHz spectrum);
  
• Improved quality of coverage allowing customers to use mobile broadband in more locations with greater consistency (900MHz spectrum);
  
• Improved in-building coverage (900MHz spectrum);
  
• Wider coverage of rural areas (900MHz spectrum).

1.7 UMTS networks using 900MHz spectrum are already deployed and in operation in a number of European markets. An increasing number of handsets and other user devices being sold in the UK today are UMTS900-ready. The only thing that is now stopping consumers with such devices from enjoying the benefits of UMTS900 in the UK is the regulatory restriction that limits the technology that can be deployed in the 900MHz band to GSM.

1.8 The situation as regards use of 1800MHz spectrum for UMTS is less well developed in terms of both likely speed of availability of equipment and services, and hence the likely effect on consumers, but equally in our opinion the risks to competition of liberalisation of this spectrum for UMTS are, as we set out below, also considerably less than for 900 MHz spectrum.

1.9 In both cases we therefore consider that from a consumer benefit perspective, it is imperative that decisions are taken quickly allowing this spectrum to be liberalised for UMTS use. The UK is of course also required to liberalise this spectrum for UMTS use by two pieces of European legislation - the amended GSM Directive and the associated Radio Spectrum Decision .

900 MHz liberalisation

1.10 In our February 2009 consultation we recognised the likely benefits of 900 MHz spectrum in providing higher quality mobile broadband services compared to the 2100 MHz spectrum currently used. However, we were concerned that, as a result, liberalising 900 MHz spectrum in the hands of the incumbent holders, without constraint, could lead to competitive distortion and/or inefficiencies for around 2-4 years, until services using 800 MHz spectrum could offer a competitive constraint. This could result in consumers facing higher prices and/or lower quality mobile broadband services than they might if the benefits of 900MHz spectrum were available more widely (but the services should still be better and potentially cheaper than those they would receive if 900MHz spectrum were not liberalised at all). Whilst noting that the evidence in favour of any one policy option was not overwhelming, our judgement at that time was that requiring the 900 MHz incumbents to release one block of spectrum between them (2x2.5 MHz each) appeared to be the best option overall to address our competition concerns.

1.11 Since our consultation, demand for mobile broadband services in general, and for use of the 900MHz band to deliver 3G services more specifically, has become clearer. However, our view of the likelihood and size of a competitive distortion arising has, significantly, reduced. This is largely because of the merger between Orange and T-Mobile creating Everything Everywhere (EE). Of the operators today, EE, and to a lesser extent H3G through its network sharing arrangement with EE, are in the strongest position in terms of network capability for providing UMTS services. They have the largest amount of 2100 MHz spectrum and access to the largest number of base station sites. Accordingly, we believe these operators would be able to improve their coverage (if required) and consequently reduce any competitive advantage that O2 or Vodafone might realise from using 900MHz spectrum for the provision of 3G services, albeit at potentially greater cost.

1.12 Updated technical analysis taking account of the availability of these extra sites, and also including refinements made following stakeholder feedback in response to our February 2009 consultation, suggests that a UMTS 900MHz network deployed by O2 or Vodafone could still provide improved quality of coverage to some indoor locations when compared to what Everything Everywhere or H3G could provide with 2100MHz. However, the extent of the improved quality of coverage is relatively small. The extent of this advantage will be dependent on the construction of buildings and the location of the user within the building. Little or no advantage would exist in many easier to serve indoor locations. In addition, other ways of dealing with poor indoor coverage, such as in-building repeaters and femtocells have become a more plausible strategy for EE/H3G to address residual areas of coverage disadvantage since our February 2009 consultation.

1.13 Overall we now consider the risk and extent of any competitive advantage for O2 or Vodafone arising from liberalisation of the 900MHz spectrum for UMTS to be low and significantly less than our analysis suggested in February 2009.

1.14 In light of this change in our view of the risk and likely magnitude of any competitive distortion, we have re-assessed what measures, if any, it might be appropriate to take to address any residual competition concerns. Firstly, as a result of the lower risk and likely reduced magnitude of any competitive distortion, we now consider that an option of requiring release of one block of spectrum, with costs of around 60m-210m (revised slightly upwards from our previous estimates) is likely to be disproportionate - the benefits are unlikely to outweigh the costs. In addition, following further technical research, we continue to believe that there are considerable implementation challenges that would need to be resolved to ensure that regulated access was effective and did not have negative unintended consequences. Consequently, we are not certain that regulated access would be a timely, effective and proportionate option. Therefore, given our judgement that there is a reduced risk of a material competitive distortion and concerns over the leading alternative options, we consider that liberalising 900 MHz spectrum for UMTS in the hands of the current licensees, without imposing conditions (beyond essential technical requirements), is now likely to be the best option.

1.15 As regards any concern that the likely lower costs of providing 3G UMTS services using 900MHz spectrum, as compared with higher frequencies, might indirectly give rise to a competitive distortion (for example by virtue of a windfall gain to those operators with 900MHz licences), we firstly note that charging annual licence fees for 900MHz spectrum that reflect the full market value of that spectrum ought to mitigate if not entirely eliminate any such distortion. Furthermore, even if annual licence fees were not accurately to reflect full market value (and in particular the differential in value between 900MHz and higher frequency spectrum) we consider the impact on consumers to likely be limited. This is because we consider it unlikely that any difference in cost would have a material effect on competition, for the reasons set out in our February 2009 Consultation, and almost certainly would not lead to any existing player having to exit the market before alternative spectrum becomes available.

1800 MHz liberalisation

1.16 Our analysis in February 2009 showed that use of liberalised 1800MHz spectrum for 3G provided no material advantage relative to 2100MHz spectrum for providing improved mobile broadband services, in terms of speed or coverage. Although liberalising the 1800MHz band for 3G could in principle offer significant extra capacity to T-Mobile and Orange (as they were then), in practice there was a lack of momentum in relation to compatible equipment, and operators had other options for increasing capacity such as acquiring the right to use additional spectrum in other bands and deploying more base stations.

1.17 Our views on these issues have not materially changed. In addition, as a result of the merger review process T-Mobile and Orange have agreed to divest 2x15 MHz of Everything Everywhere's 1800MHz spectrum (to address European Commission concerns in relation to the provision of LTE services using 1800MHz spectrum in future, rather than UMTS services today) which will therefore be available to another operator in the future. Our view therefore remains that there is little risk of a material competitive distortion arising as a result of liberalising the 1800MHz spectrum for UMTS in the hands of the current holders, without additional conditions (beyond essential technical requirements), and this is still likely to be the most appropriate option.