Home > Consultations > Consultation Documents > Award of available spectrum > Executive Summary

1.1 Our Consultation Document published on 11 December 2006 set out a number of proposals relating to the award of available spectrum in the bands covering 2500-2690MHz, 2010-2025MHz and 2290-2300MHz. Key features of the proposals were to:

• hold awards for each of these bands as soon as practical via auction;
• to package spectrum in the 2.6GHz band in such a way as to allow an outcome consistent with the CEPT band plan for 2.6GHz (2x70MHz of paired spectrum and 50MHz of unpaired spectrum before allowing for guard blocks), but to allow flexibility for the split between paired and unpaired spectrum to vary from the fixed arrangement in the CEPT band plan based on relative market demands;
• for the award of the 2.6GHz band, to adopt a two stage auction approach using a Principal Stage in which bidders would compete for generic lots, in a clock phase followed by supplementary bids round, and an Assignment Stage during which these generic lots would be converted into assignments at specific frequencies;
• to package the 2010MHz band as a single lot to be offered as part of the same auction as the 2.6GHz band; and
• to package the 2290MHz band as a single lot to be awarded via a separate auction.

1.2 The responses to this consultation have raised a wide range of issues, including those relating to the technical and non-technical conditions of use and to the design and timing of the auctions. The breadth of the responses reflects the economic significance of the spectrum to the UK and its commercial significance to many individual companies. In light of the responses, we have carried out substantial further analysis, the results of which are presented in this Discussion Document, together with updated proposals where relevant.

1.3 The main focus of this Discussion Document is on technical conditions and auction design for the 2.6GHz band (and the 2010MHz band).

• Technical conditions for use of the bands. In the 2.6GHz band, we consider the significance of compatibility between paired and unpaired use for flexibility in spectrum packaging and auction design. In respect of uses in adjacent bands, we consider the significance of their likely impact for potential variations in technical constraints across different frequencies within the bands for award (and, conversely, the effect of their use on users in adjacent bands).
• Auction design. We review the options for addressing certain technical issues through auction design and propose changes to the December proposals to address issues raised by stakeholders.

1.4 We also provide an updated assessment of Ofcom’s view on timing of awards and on non-technical licence conditions. In the case of award timing, we do so because we recognise that parties who are contemplating participation in these awards need to plan in advance and will wish to understand the possible timetable. In the case of non-technical licence conditions, we do so because of the detailed nature of responses to these aspects of the December proposals, and our updated view is based on the assessment of more detailed arguments from some respondents rather than on new developments relevant to the issues. We emphasise, however, that Ofcom has not made final decisions on these matters and is not using this Discussion Document as a vehicle for announcing decisions. We will make and announce decisions in a Statement which we expect to publish before the end of this calendar year.

1.5 This additional step in the consultation process is intended to facilitate further input into the design of the award and thereby contribute to the development of an award process that is as robust and efficient as possible. However, this Discussion Document is not intended to give a comprehensive review of all comments raised in responses.

Technical analysis of interference conditions

1.6 The most significant concern over interference relates to the potential implications of allowing the market to determine the organisation of the 2.6GHz band, which may result in an outcome different from the CEPT band plan in terms of split between unpaired use (most likely to involve TDD technology) and paired use (most likely to involve FDD technology). Most particularly, this concern relates to the scope for mobile-to-mobile interference from TDD terminals into FDD terminals (the “blocking” effect), under a scenario in which TDD terminals are permitted to operate in the upper part of the 2.6GHz band (2620-2690MHz). We have carried out substantial further work on this issue including the development of Monte Carlo simulations to assess the scale of this blocking effect. The conclusion from this analysis is that the blocking effect can be expected to be very likely to materialise in certain types of situations (where a TDD user and an FDD users are in close proximity and both accessing services at the same time using the 2.6GHz band). However, our analysis suggests that the percentage of overall FDD connections that would suffer blocking would rise by no more than 0.5% under plausible assumptions, similar to the level of lost or failed connections that FDD terminals would typically suffer in their own network regardless of other users. If network operators acquiring 2.6GHz spectrum were concerned about the blocking effect in certain situations, then it is possible that they would be able to take a number of mitigation actions. Indeed, we note that an equivalent risk of a blocking effect may exist for users of TDD technologies (WiMAX in particular), caused by the use of FDD terminals in the 2.6GHz band, even under the CEPT band plan. However, prospective WiMAX operators seem to be confident that, overall, the risk of blocking is low (and/or that they can mitigate the likelihood of blocking).

1.7 In relation to technical conditions within the 2.6GHz band, concerns were also raised about interference between adjacent TDD and FDD users and about whether the use of a 5MHz restricted block was sufficient. This Discussion Document reports our further technical analysis of this issue and sets out the view that the conditions proposed in the December Consultation are likely to promote optimal use of the spectrum, by maximising the amount of spectrum available for use while allowing appropriate scope for interference to be managed efficiently by all users. However, some improvement in the risk of adjacent channel interference could be achieved (particularly in urban areas) if additional constraints were placed on outdoor TDD use of the restricted block adjacent to the FDD uplink and on outdoor TDD use of the restricted blocks between TDD users.

1.8 Turning to uses in adjacent bands, this document includes additional information on the use of radars above the 2.6GHz band.

• We provide new information on MoD radar use above 2700MHz that was not available at the time of the December consultation.
• We are currently carrying out some empirical work to measure the impact of radar use on UMTS and WiMAX equipment which we intend to publish when completed. In the absence of this, we have reviewed the results of some substantive analysis undertaken by one stakeholder which indicates that the presence of radar may require a local increase in cell density, in areas where signals from radars are strongest, but would not prevent effective use of the spectrum.

1.9 Two stakeholders expressed concerns about the impact of new uses of the 2.6GHz band on the ability of PMSE users to use frequencies immediately below 2500MHz and above 2025MHz. We are publishing the results of some empirical work undertaken by ERA on the performance of PMSE equipment. This work indicates that the assessment of interference risks should be based on parameters different from those used to prepare some of the analysis in support of the December Consultation. However, Ofcom’s updated assessment is that no specific restrictions should be placed in respect of the risk of interference between existing PMSE use and new uses in the bands for award. Ofcom plans to consider this assessment with JFMG (the company currently responsible for managing access to spectrum for PMSE use on behalf of Ofcom) and interested PMSE users during the consultation period.

1.10 We have also obtained some further information on the levels of potential interference from neighbouring countries.

• We carried out a series of measurements of received signal strengths in the 2.6GHz band at a number of locations in Northern Ireland, West Wales and North West England (to test signal strengths from Irish MMDS broadcast services) and in the South East of England (to test signal strengths from current use of the 2.6GHz band in France). These showed no, or very weak, signals in all of these locations with the exception of those locations in Northern Ireland close to the border with the Republic.
• We have also received some new information from ComReg on the transmissions from MMDS services in Ireland. It suggests that the level of interference is likely to be slightly less than indicated in the December Consultation.

Technical licence conditions

1.11 In the December Consultation, we set out two alternative approaches to specifying technical licence conditions, one based on the use of spectrum masks and the other based on the use of spectrum usage rights (SURs). Although some stakeholders expressed support for the general SUR approach, there was no support amongst stakeholders for taking this approach forward in the context of the 2.6GHz award itself. In part, this reflects a perception of the time it would take to address outstanding issues. As a result, Ofcom is not proposing to pursue the SUR approach for this award. We therefore intend to specify the technical licence conditions in terms of spectrum masks.

1.12 Based on the technical analysis described above and having taken account of comments raised by stakeholders, this Discussion Document sets out our updated view that the detailed specifications for the spectrum masks included in the December Consultation remain appropriate.

1.13 However, there are three specific areas of possible change in which stakeholders’ evidence in response to this Discussion Document will be particularly helpful:

• extending the out-of-block masks out to an offset of ±20MHz from assigned blocks;
• placing additional restrictions on the use of restricted channels between the FDD uplink and TDD to limit potential interference from their use outdoors (possibly by reducing the in-band power to 18dBm/MHz EIRP); and
• a reduction in mobile station maximum in-band power to 18dBm/MHz EIRP to help a modest reduction in the probability of interference in certain scenarios.

Auction design and related issues

1.14 A central feature of the proposed auction design on which we consulted was to package spectrum in the 2.6GHz band in such a way as to allow flexibility for the split between paired and unpaired spectrum to vary from the CEPT band plan. In response to the December Consultation, the flexibility to depart from the CEPT band plan was supported by stakeholders interested in WiMAX (particularly potential operators and manufacturers) and by one MNO and opposed by other MNOs and several 3GPP equipment manufacturers. Having considered the substance of these responses, in particular the underlying technical issues, our updated assessment is as follows.

• There is significant potential interest in gaining access to 2.6GHz spectrum for unpaired use and the level of interest is likely to exceed the limit of 50MHz of unpaired spectrum that would be available under the CEPT plan. The main interest for unpaired spectrum comes from parties who would generally be new holders of rights to spectrum suitable for the provision of mobile data services, looking to deploy WiMAX based services. This could deliver significant competition and innovation benefits for consumers. This provides a strong argument for retaining flexibility (in terms of relative amounts of paired and unpaired spectrum) within the packaging and auction design proposals.
• The retention of this flexibility creates the possibility that TDD terminals could operate in the upper part of the band and that this could lead to blocking of FDD handsets as noted above. This could reduce the value of the spectrum to FDD users. We have considered allowing FDD bidders to reflect their perception of the impact on value of the blocking effect by modifying the auction rules to allow for contingent bidding (whereby FDD bidders submit separate bids for the case where TDD terminal operation in the upper part of the band is prohibited and the case where TDD terminal operation in the upper part of the band is permitted). However, this modification would add significant complexity to the auction and raise potential concerns in relation to our duty to promote competition and innovation.
• Given that the technical scale and commercial impact of the blocking effect are expected to be modest, as noted above, we consider that any benefits that could result from modifying the auction design to accommodate contingent bidding of this type would not outweigh its risks and disadvantages.

1.15 We note that the proposals do mandate key features of the CEPT band plan, notably the 5MHz channel widths and the 120MHz duplex spacing for paired spectrum. We also note that, depending on market demand as revealed through the bids in the auction, the CEPT band plan could be the outcome of the proposed award. We consider that these proposals, which have been widely supported, provide the opportunity for market-led harmonisation which will facilitate access to the economies of scale of a European handset market and will facilitate roaming.

1.16 A second key feature of the proposed auction design is the use of generic lots (with two types of lot in the 2.6GHz band, one for paired spectrum and one for unpaired). The use of generic lots underpins the division of the auction into the Principal and Assignment stages. Although several stakeholders expressed concerns regarding aspects of the proposal to use generic lots, our view remains that the use of two categories of generic lots for the 2.6GHz band provides the best balance between the technical and auction design issues for three main reasons.

• Our analysis (taking account of the effects of radar use, TDD / FDD adjacencies and interference from neighbouring countries) indicates that the potential for differences in value between lots within one category of lot would not be such as to risk creating substantial inefficiencies under the proposed two-stage auction process with use of generic lots. On the contrary, the proposed auction design, including the use of generic lots, is likely to be the most suitable for furthering the interests of citizens and consumers.
• No stakeholders came forward with suggestions for an alternative approach based on the use of specific lots rather than generic lots that would be likely to result in greater overall efficiency or benefits. Indeed, consideration of an SMRA auction using specific lots (the auction format generally associated with spectrum auctions with specific lots) suggests that this would be substantially more problematic, particularly because of the opportunities for strategic behaviour it creates.
• It would be possible to disaggregate the lots into more than two categories within the 2.6GHz band but this would add significant additional complexity which itself could introduce risks (see further below).

1.17 Although we are not proposing a change in approach on these two substantive features of the auction design (flexibility of the split between paired and unpaired, and the use of generic lots), we are proposing a number of more detailed changes in the auction rules in light of stakeholder feedback and further consideration. The main changes are summarised in the table below.

Table 1: Main changes to the auction design proposals compared to the December Consultation

Stage	Action	Description of change
Principal Stage: primary bid rounds	Primary bids contingent on no split award of unpaired lots	All primary bids for unpaired 2.6GHz spectrum are contingent on receiving contiguous spectrum. Primary bid rounds continue until such time as it is possible to accommodate all remaining bids without a split award of unpaired lots.
	Fungible eligibility	There is a single eligibility rule that applies across all three categories. Switching is allowed between paired and unpaired lots.
	Separate clock prices for the 2.6GHz paired and unpaired	Separate clock prices but the 2:1 price ratio remains for the 2.6GHz paired and unpaired lots, unless demand for unpaired lots falls below nine. In the event that the price for unpaired 2.6GHz lots falls below the 2:1 ratio, but demand subsequently rises above nine lots, the price will be returned to a 2:1 price ratio in the next round.
	Bidding cap	A cap of 18 eligibility points per bidder is proposed, corresponding to a safeguard cap of 90MHz.
	Extension rights replace waivers	Extension rights allow bidders extra time to submit bids in a round in the event that they experience technical problems. Bidders that still fail to submit a bid after using up an extension period will be reduced to zero eligibility.
Principal Stage: supplementary bids round	Supplementary bid submission	Supplementary bids can only be submitted in the supplementary bids round (not during primary rounds).
	Supplementary bids round always run	Supplementary bids rounds will always be run after completion of the primary bid rounds, even if supply and demand are exactly in balance in the last primary round.
	Two types of supplementary bids	In addition to ‘standard supplementary’ bids, bidders for unpaired 2.6GHz spectrum may submit ‘split supplementary bids’ for packages with specified splits between the lower and upper unpaired areas.
	Pricing rule	Second price rule for the Principal Stage has been refined.
Assignment Stage	Bid options	Rules for determining bid options have been refined and now cover the case where some lots from the Principal Stage are unsold.
	Split assignments	Identity of the bidder receiving a split assignment of unpaired 2.6GHz lots (if there is one) will be determined by the Principal Stage, and this will be taken into account in the available bid options for each bidder in the Assignment Stage. A bidder may only get a split assignment if it has placed a bid for the corresponding option.
	Pricing rule	A second-price rule analogous to that being use for the Principal Stage is proposed (instead of a pay-as-bid rule).

1.18 The auction design proposals set out in this Discussion Document also include more information on certain steps not previously covered in detail (for example on application and qualification to bid) and other specific aspects (such as deposit requirements or provisions for unsold lots).

1.19 Consultation respondents made a number of suggestions for other changes to the auction rules to:

• make it easier for MNOs holding TDD spectrum at 1.9GHz to bid for spectrum at 2600-2620MHz in order to use these lots for external pairing with their blocks at 1.9GHz;
• add extra categories of lot (such as distinguishing FDD channels which are immediately adjacent to the restricted block between TDD and FDD); and
• allow for a distinction between TDD lots, dependent on whether or not the lots are in the upper part of the band or are in, or below, the centre of the 2.6GHz band.

1.20 In preparing auction design proposals, there is a balance to be struck between the competing requirements expressed by interested parties: it is not possible to construct an auction that can facilitate all possible outcomes in a similar way without introducing efficiency risks and/or opportunities for strategic behaviour. In the case of spectrum at 2600-2620MHz which could be paired externally, our analysis suggests that this would have a number of disadvantages, including the introduction of significant additional complexity into the auction. We consider that any benefits that might flow from this are not sufficient to justify incurring these adverse effects. Similarly, whilst it would be possible in the other cases to amend the auction design to accommodate additional categories of lots, each additional category would add considerably to the complexity of auction design and execution. With additional complexity comes the risk that auction participants may not fully express their preferences for all the packages of lots in which they might be interested, thereby creating the possibility that the auction result becomes suboptimal. We note that some stakeholders who suggested such modifications also commented that they felt the auction design was already complicated. As such, we are not proposing to implement any of the suggestions listed above in paragraph 1.19.

Timing and linkages between awards and non technical licence conditions

1.21 As noted above, this Discussion Document provides an updated assessment on timing and on linkages between the awards of the different bands. The main focus is on the timing of the 2.6GHz award.

1.22 The substance of stakeholder feedback, both in responses to the December Consultation and in various meetings, together with further analysis which we have undertaken, have reinforced our view that it is desirable to hold the award of the 2.6GHz band as soon as is practical. Key aspects in this updated assessment are as follows.

• As a result of meetings with stakeholders, we consider it very likely that if the 2.6GHz award was held in 2008 then there would be parties who would participate in the award and who, if successful, would intend to bring the spectrum into use without delay.
• The main interest in imminent use of the spectrum comes from parties who would be new providers of mobile data services and who intend to deploy WiMAX systems. This raises the prospect that the combination of innovation and competition benefits for consumers could be substantial, in part because the advent of the WiMAX standard seems to have prompted a competitive response from the 3GPP community to advance their own next generation standard (LTE).
• The majority of the existing MNOs would prefer a delay in the award until they have a more immediate need for 2.6GHz spectrum themselves and/or various sources of uncertainty were reduced. We accept that if the award were held a long time in advance of need for some potential users then this could present those users with some uncertainty over how much spectrum to bid for and this could be a source of inefficiency under certain assumptions (i.e. if it led to an outcome which was different to one that would have been arrived at had they had better foresight of their future spectrum requirements and one they could not replicate through subsequent trading of spectrum). However, we are not persuaded that this would be a source of substantial inefficiency. In addition, Ofcom notes that new information is available and expects further information to become available on such issues as 2G spectrum and European discussions on the WAPECS concept and its potential application to the 2.6GHz and 2010MHz bands. Any potential source of inefficiency also needs to be balanced against the potential loss of benefit from denying other users (and their consumers) access to the spectrum without delay.
• On this last point, the potential loss of benefit from denying earlier access relates not only to the direct loss of benefit from use of the spectrum over the period of delay. The loss of innovation and competition benefits could be magnified if, as seems plausible, there is a short term window of opportunity for new entry by providers of high bandwidth dependent mobile data services using new technology. This likely window of opportunity highlights the potentially large risks that would be associated with relying on demand from some MNOs that may exist in future rather than on clear demand that has been expressed for an award without delay and which may reduce or disappear in case of delay.

1.23 On the question of linkages between the awards of the bands under consideration:

• Stakeholders did perceive a linkage between the 2.6GHz and 2010MHz bands, primarily as substitutes for unpaired spectrum. Although we believe that the linkage may be relatively weak, we are minded to continue with the award of the 2010MHz band as part of the same auction, and therefore on the same timescale, as the 2.6GHz award.
• There was very limited interest in the 2290MHz band and varying views on whether any award should be held before the 2.6GHz auction as proposed in the December Consultation. We therefore intend to separate the 2290MHz band from the 2.6GHz award process and consider it again at a later date.

1.24 As noted above, a number of the existing MNOs made submissions relating to the proposals for non-technical licence conditions relating to roll-out obligations, technology neutrality, tradability and duration / tenure. We have carefully considered those submissions and comment in Section 8 on our preliminary conclusion that they do not provide grounds to alter the approach which was set out in the December Consultation. That approach is consistent with the policies in a number of spectrum policy decisions made by Ofcom over recent years, including those set out in the Spectrum Trading and Spectrum Liberalisation and the Spectrum Framework Review statements.

Next steps

1.25 We are inviting comments from stakeholders by 28 September on the issues raised in this Discussion Document. We intend to hold a seminar on the updated proposals during September, before the close of this period for comment. As noted above, the main focus of this Discussion Document is on technical interference issues and any implications for auction design, and it is in these areas that we particularly seek comment. However, stakeholders are welcome to comment on any issues raised in this Discussion Document. We would encourage stakeholders to substantiate comments on technical issues with supporting technical analysis where possible and to link comments on auction design with concrete proposals for modifying auction rules where relevant.

1.26 We intend to hold some more detailed workshops on auction rules during the autumn targeted at those parties that have a prospective interest in participating in the auction. If feasible, we may also hold some demonstrations with these parties during this period.

1.27 We aim to reach policy decisions on this award, to make a Statement and to publish an Information Memorandum and draft regulations (in particular the regulations that will define the award process and allow it to take place) before the end of this calendar year. If the decision is to proceed to award as soon as practicable then, on this timetable, we would expect to:

• make decisions for the regulations, publish statements on those decisions and make the regulations in March or April 2008;
• invite applications to participate in the award by the end of the first quarter of 2008/09 (after the auction regulations have come into force); and
• commence the bidding process of the auction as soon as possible thereafter.

The full document is available below

• Award of available spectrum: 2500-2690MHz, 2010-2025MHz [pdf]
  Full Print Version - discussion document published 1|08|2007

Back to top