Telecom New Zealand

TELECOM NEW ZEALAND SUBMISSION ON MINISTRY OF ECONOMIC DEVELOPMENT DISCUSSION PAPER “RADIO FREQUENCY AUCTION: 2.3 AND 2.5 GHz BANDS” 14 SEPTEMBER 2007 (Public Version)

TABLE OF CONTENTS

A Executive Summary

B Submission Structure

C Background

D Issues

E Competition

F Spectrum Auctions

G Terms of Sale and Management Rights

H Managed Spectrum Parks

QUESTIONS

A Question 1 Considerations Relating to Lot Design

C Question 2 Proposed Lot Design

E Question 4 Eligibility to Bid

F Question 5 Duration of Acquisition Limits

G Question 6 Use or Lose Provisions

H Question 7 Settlement

I Question 8 Managed Spectrum Parks: Proposed Allocation Procedures and Usage Rules

APPENDIX A

A Executive Summary

1. Telecom supports the Government’s policy objectives for allocating management rights to radio spectrum. This paper has been developed with those objectives in mind.
   
2. The Government’s proposals on the 2.3/2.5 GHz auction followed the consideration of submissions on the lot and auction design for a proposed 2.3 GHz auction. Auctioning the 2.3 and 2.5 GHz frequencies at the same time raises a different set of issues to those that existed when the 2.3 GHz spectrum was being auctioned on its own. This is because both WiMAX and UMTS-LTE (cellular) technology are likely to be deployed in the 2.5 GHz spectrum. The 2.3 GHz spectrum was only to be used for WiMAX.
   
3. The main focus of the Cabinet decision and the Ministry of Economic Development (“MED”) discussion paper (the “discussion paper”) is competition between WiMAX operators. Focusing too heavily on this objective has the potential to undermine the Government’s economic transformation agenda and its digital strategy. This is because of the existence of WiMAX and UMTS-LTE which can and should be made available to New Zealand for wireless services.
   
4. We believe both technologies should co-exist in the 2.5 GHz band. WiMAX and UMTS-LTE technologies will serve different market needs (with the potential for WiMAX to have advantages in fixed broadband applications and for UMTS-LTE to have advantages in mobility-oriented applications), so both need to be made available to New Zealand consumers. This approach has been adopted by Sprint in the US. Sprint has deployed both CDMA and WiMAX based services.
   
5. The lot and auction design should be carefully structured to avoid investment in UMTS-LTE technology being limited. Telecom believes it is essential to take account of the different requirements of the WiMAX and UMTS-LTE technologies (in particular the difference in spectrum blocks and likely deployment date). It is important that MED manages this risk to ensure that the long term benefits to end users are maximized.
   
6. Given that there are now two technologies that could be deployed, the regulatory requirement which causes the greatest concern to potential bidders is the proposed spectrum acquisition cap. This aspect of the auction design forces parties to choose between WiMAX and UMTS-LTE at a time when there is a great deal of uncertainty about the future of these technologies. This has the potential to exclude investment in one of these wireless technologies by proven New Zealand infrastructure investors, who are likely to have greatest capacity to build national networks and bring benefits to consumers.
   
7. Telecom suggests two alternatives, either of which would help to mitigate the risks associated with auctioning this spectrum at a time when investors will only have limited information about the technologies.
8. The key feature of the first alternative is what we describe as a technology cap, as opposed to a spectrum cap. The technology cap is still consistent with the Government’s objective of preventing anti-competitive behaviour while still ensuring that a company is able to invest in spectrum for both WiMAX and UMTS-LTE technology. A technology cap would preserve the Government’s objective of widely-held WiMAX spectrum and the associated competitive benefits this brings. It would ensure that established New Zealand wireless infrastructure investors can compete using WiMAX technology while also protecting their ability to upgrade the capability and capacity of their existing cellular networks.
   
9. Under the second alternative, the sale of the spectrum that has been allocated for UMTS-LTE would be deferred until there is more certainty in major markets about a dominant frequency band for the deployment of this technology. This second option would allow a company to invest in WiMAX now and invest in the frequencies allocated to the UMTS-LTE technology if the MED chose to sell this at a later date.
   
10. We note that when Ofcom in the United Kingdom was faced with the same dilemma to the one MED currently faces, it chose to defer the auction of the 2.5 GHz spectrum. Ofcom chose to undertake another round of consultation on the technical conditions for use of the 2.5 GHz band and on issues relating to spectrum. Further consultation is being undertaken to ensure that “a robust and efficient award process” is developed. It would seem pragmatic for MED, at the very least, to follow this approach.
    
11. Telecom would be happy to discuss its current thinking and tentative proposals and the information it has on WiMAX and UMTS-LTE technologies with the MED.

B Submission Structure

12. This submission is divided into two parts, in the first section of this submission Telecom outlines the main issues, options and risks that arise from the discussion document.
13. The second section answers questions from the discussion paper. The questions Telecom have answered are the ones that are relevant to its situation, but have not been dealt with in the first section of the paper.

C Background

14. The Government has committed itself to an economic transformation agenda for the next decade. The agenda seeks to progress New Zealand to a high income, knowledge-based market economy, which is both innovative and creative, and provide a unique quality of life to all New Zealanders.
    
15. A key component to the transformation agenda is the world class infrastructure theme, which is guided by the Digital Strategy. The Government believes that information and communications technology plays a fundamental role in providing capacity for New Zealand to transform its economy through sustainable economic growth.
16. The MED discussion document on the auctioning of the 2.3 GHz and 2.5 GHz bands raises issues with this policy setting. In particular, if there is a failure to fully consider the significant risks which are attached to this spectrum auction, there is scope for stunting the technological evolution of mobile telecommunication services in New Zealand. This submission provides some policy alternatives which would enable the Government to manage these risks, while still achieving welfare maximising outcomes.
    
17. 17. In identifying the issues which we have presented we have adopted the same analytical framework adopted by the MED. The overall policy settings the Government is attempting to facilitate are expressed to be:¹
    
    
    17.1 Efficient use of spectrum and maximum value to society;
    17.2 Investment in infrastructure, particularly broadband services; and
    17.3 Competition between service providers.
    
    
    
    ¹ See Ministry of Economic Development, Radio Frequency Auction: 2.3 GHz and 2.5 GHz Bands, August 2007, see page 7

D Issues

18. Many of the issues that arise from the proposed auction of the 2.3 GHz and 2.5 GHz bands of spectrum are caused by the current level of uncertainty associated with the two main technologies that are expected to use these spectrum bands. For example, it is uncertain which spectrum bands global operators will deploy the new UMTS–LTE technologies in. Also, as the MED outlines in its discussion document, it is uncertain whether mobile WiMAX will be able to offer features comparable to those offered by the UMTS-LTE technologies. The decisions Government is requiring firms to make today are subject to a number of factors which are outside of the control of the Government and of the companies who are intending to deploy technology in the 2.3 GHz and 2.5 GHz bands. In addition to this, New Zealand is a small market and must look to deploy systems that achieve scale in major markets.
    
19. Because of these technological uncertainties, wireless infrastructure investors have a need to keep their options open. However, some of the proposals contained in the discussion document do not allow them to do this. Specifically, spectrum acquisition caps will force companies to prematurely choose between the two technologies based on incomplete information. The current MED proposals are meant to manage the risk of monopolisation and anticompetitive behaviour. Unfortunately they may unintentionally result in monopolisation and generate additional risks for the economy. We believe our recommendations can assist the government to manage these risks.
20. Based on the current MED proposals we believe that New Zealand’s economic outcomes will be dependent upon a gamble that specific, yet uncertain, assumptions eventuate. For example, operators may all decide to invest in spectrum blocks designed for WiMAX as a low risk option but discover in a few years time that this was not an optimal technology path. Conversely there are risks to the market if prospective bidders are all interested UMTS–LTE capable spectrum. Either way, the policy dilemma is to manage the risk that firms will opt for only one technology.
    
21. In light of the technological uncertainty we believe that there are a number risks that are created by MED’s proposed auction process. Telecom believes that these risks have the potential to undermine the Government’s broader policy objectives. We summarise these risks under the following headings:
    
    
    a) Lot and auction design; and
    b) Spectrum acquisition limits and the impact on competition.
    
    
    
    We also analyse various other aspects of the auction design. We have assessed:
    
    
    
    21.1 Infrastructure investment;
    21.2 Spectrum auctions;
    21.3 Terms of Sale and Management rights; and
    21.4 Managed spectrum parks.
    
    
    
22. These key areas are analysed after the discussion of Telecom’s proposals for mitigating the risks we identify in this submission.

Telecom’s proposed options for auctioning the 2.3 and 2.5 GHz spectrum

23. Telecom considers it would be prudent for the Government to use one of our proposed approaches discussed below. The first option appears to be the most consistent with the Government’s current policy for radio spectrum, but Telecom considers that either option would help to mitigate the risks that are discussed in this submission. These options will ensure that the benefits received by long term end users of these technologies are maximised and the risks faced by long term end users are minimised.
    

Option 1: Purchase of TDD and FDD spectrum or the “technology acquisition limit option”

24. This option has two key components:
    
    a) It is based on the assumption that the Government will be dividing the spectrum to ensure that spectrum is available to deploy both TDD (WiMAX) and UMTS-LTE technology; and
    b) The acquisition limits that are in place allow firms to buy spectrum for both WiMAX and UMTS-LTE technology (i.e. one block of TDD spectrum and one block of paired spectrum).
    
    A lot design that allows for both WiMAX and UMTS LTE spectrum blocks
    
25. For reasons that will be discussed later in the submission, Telecom considers that it is essential that the lot design accommodates both WiMAX and UMTS-LTE technology. Telecom’s preferred lot design for the 2.5 GHz spectrum is attached at Appendix A. This design is similar to MED’s option C (i), with a major difference being that it makes provision for one managed spectrum park.
    
26. If the Government is restricting its considerations to the lot designs which are set out in the discussion paper, Telecom would opt for option C (i). However, as we will detail below, Telecom considers that this lot design has inherent limitations because nearly a third of the available spectrum is set aside for managed spectrum parks, and represents a sub-optimal allocation. Acquisition limits that allow firms to purchase spectrum for both WiMAX and UMTS LTE technology
    
27. Telecom considers that companies should be able to buy a 35 MHz spectrum block, in either the 2.3 GHz or 2.5 GHz band, to enable them to deploy WiMAX, as well as a paired 2 X 20 MHz block of spectrum in the 2.5 GHz band to enable them to deploy UMTS-LTE technology. The paired spectrum should be fully useable and not be compromised to create guard bands from within the acquired spectrum.
    
28. The current acquisition caps proposed by MED will force firms to make long term investment decisions based on incomplete information. Companies will be forced to choose one technology over another. Given the high level of uncertainty around a dominant frequency band for UMTS-LTE technology, the Government should be adopting options that enable investment in the paired spectrum. The current process creates a real risk that no firms, or potentially only one firm, may opt for the paired spectrum necessary for this technology.
    
29. If the Government encourages one technology over another it may mean that New Zealand consumers will not have access to the technology that is available to their contemporaries in other OECD countries.
    

Option 2: Purchase of WiMAX spectrum and delaying the sale of cellular spectrum or the “delay option”

30. The Government could consider delaying the sale of the spectrum that is to be used for the UMTS-LTE technology. Again, this option is based on Telecom’s preferred band plan that is set out in Appendix A, or if this is not an option, the MED’s band plan C (i).
    
31. Under this option, the sale of the spectrum that is to be used for WiMAX would take place in December and the sale of the spectrum that is to be used for the UMTS-LTE technology would be delayed until the current uncertainties associated with this technology are resolved. Telecom considers that it would be prudent to delay the auction until the auctions of this spectrum are held in Europe. Operators will initially be able to purchase spectrum for WiMAX and would also be able to buy spectrum at a later auction if paired bands for UMTS-LTE technology were offered at that stage. Appropriate decisions about acquisition limits could be made based on an assessment of the market and better informed indications as to predominant spectrum choices elsewhere at that time.
    
32. As stated earlier, it is still not clear whether the UMTS-LTE technology will use the 2.5 GHz band internationally. Delaying the auction of the paired bands ensures that the Government retains ownership until this is clear. If it becomes clear that the 2.5 GHz band is not going to be required for UMTS-LTE technology it could be made available for use by WiMAX operators either as a managed spectrum park or as a management right.
    
33. Our second option will enable companies who want to invest in spectrum for WiMAX technology to make that investment now, while at the same time ensuring that New Zealand does not miss out on the potential benefits of UMTS-LTE technology (which are uncertain at this time).
    
34. This option allows the MED and the market to manage the risks, which are outside of the control of the industry and Government, and to make an informed decision on the best allocation of spectrum. Telecom considers that this option would be a pragmatic one for the Government to take in light of the uncertainties that are highlighted in this submission.
    

The Approach taken by Ofcom in the United Kingdom

35. We also note that delaying the spectrum sale is consistent with the approach taken by Ofcom in the United Kingdom.
    
36. When Ofcom was faced with the dilemma that the Government currently faces it reconsidered its initial spectrum allocation and decided to delay the entire spectrum award.
    
37. Ofcom chose to undertake another round of consultation on the technical conditions for use in the 2.5 GHz band and on issues relating to spectrum and packaging. This further consultation is being undertaken to ensure that “a robust and efficient award process” is developed.
    

Issues associated with the lot and auction designs that are set out in the discussion document

The availability of WiMAX and UMTS-LTE technology

38. The Cabinet minute which provides the basis for MED’s consultation is entirely focused on providing competition between WiMAX operators. It does not discuss UMTS-LTE technology and does not suggest that the paired bands that are necessary for the UMTS-LTE technology should be made available. This is consistent with the fact that this decision came out of the submissions on the proposed 2.3 GHz spectrum auction and also goes some way to explain how some of the concerns that Telecom have identified have arisen.
    
39. The availability of WiMAX and UMTS-LTE technologies is important to the Government’s economic transformation agenda and digital strategy. If competition between a large number of WiMAX operators is at the expense of UMTS-LTE technology, in Telecom’s view the transformation and digital strategy outcomes will not be realised. It is important that both technologies are able to co-exist. Such an outcome would be beneficial for consumers. An example of this approach is seen with Sprint’s deployment in the US, where both CDMA and WiMAX based services will be offered to consumers.
    
40. In Telecom’s view, the relative economic benefit to New Zealand associated with a 6th or 7th independent holder of TDD spectrum is much smaller than the benefits that flow from ensuring that existing cellular network operators to upgrade their capability and the benefits that will come from these operators competing with WiMAX using UMTS-LTE.

Competition from managed spectrum parks

41. Telecom also notes that some of the concerns about the level of competition between WiMAX operators can also be allayed by the potential for competition from parties operating out of managed spectrum parks. Telecom considers that MED has underestimated the levels of competition that may come from these parks.
    
42. Regional providers will be able to compete with national providers of WiMAX services on a regional basis.
    
43. Telecom also notes that the discussion paper does not specifically address whether these regional players will be able to enter into roaming agreements. A possible scenario would be that the local council in Invercargill may access a managed spectrum park to provide WiMAX services to rate payers. Local councils in Wellington and Auckland may do the same. A roaming arrangement would mean that customers would be able to roam across the spectrum held by the different local councils. Precedent for this type of a commercial arrangement is already gaining popularity in the WiFi hotspot market space.
    

Proposed lot designs

44. The spectrum band plan, along with the acquisition caps, will indirectly shape the wireless market. Despite the possibility of six firms competing for spectrum, a likely long–term outcome could be narrow market segments. For example, this will occur if only one firm purchases spectrum for UMTS–LTE and the remainder of the spectrum is purchased for WiMAX use. If the UMTS-LTE technology ends up using the 2.5 GHz band and offering significantly different functionality to WiMAX, the UMTS-LTE market will be limited while the WiMAX market will experience a high degree of competition.
45. The marginal benefits of additional competition in WiMAX technologies are likely to be substantially less than the opportunity cost of another UMTS–LTE option. We believe that by allowing for technology caps and the sound spectrum band plans which we have recommended, these problems can be overcome.
    
46. 46. Annexed to this submission is Telecom’s preferred band plan. Telecom considers that this spectrum allocation better recognises the potential for both WiMAX, UMTS–LTE and other technological developments. Telecom considers that the band plan should also recognise that the 2.3 and 2.5 GHz bands are not companion bands. In companion bands, equipment works in both bands in a seamless manner for the user. For example GSM 900 & UMTS 2100 are companion bands. One could deploy UMTS in either bands, or both. This is not the case for the 2.3 and 2.5 GHz bands. Telecom discusses the band plans that MED has proposed below.
    

MED’s Band Plan - Option A

47. The current spectrum allocation outlined in the MED’s “Option A” will not allow for both WiMAX and UMTS-LTE technology. This option seems to be a throwback to the 2.3 GHz auction. For the reasons we have outlined, focusing only on WiMAX technology would undermine the Government’s broader policy objectives.
    
48. If UMTS-LTE does use the 2.5 GHz spectrum, Option A has the potential to inhibit the evolution of the mobile market in New Zealand and to undermine the Government’s digital strategy.
    

MED’s Band Plan - Option B and C (ii)

49. Telecom is concerned that proposed lot designs B and C (ii) present a number of risks. Neither of these options guarantees that spectrum will be available for UMTS-LTE technology. The rationale behind these options is that the “market” will decide how the spectrum should be allocated. This could work well if the auction was for substitutable spectrum lots and the market was fully informed about both technology options. However, at present the market has limited information about UMTS-LTE technology and may decide that it is currently too risky to warrant investment. Telecom considers the risks posed by parties choosing not to invest in UMTS-LTE technology are serious.
    
50. Another risk of these band plans is the potential for an operator to use TDD technology in an FDD part of the band. This will result in interference and loss of performance.
    
51. These designs also present a particular economic challenge. In this example we assume that there are only two bidding firms for two pieces of spectrum. The UMTS–LTE provider would see the two pieces as complementary (cannot have one without the other) and the WiMAX provider supplementary (can have either one). If the complementary pair was valued by the UMTS–LTE provider at $2 million but that WiMAX provider valued what it saw as substitutable portions at $1.5 million it would be unwise for the UMTS–LTE provider to bid. This is because the UMTS–LTE technology provider would need to bid at least $3 million to secure the spectrum
    pair to mitigate the risk the WiMAX bidder would win one piece of spectrum for its valued $1.5 million. The UMTS–LTE provider will not bid for the spectrum if it is not prepared to pay $1 million or more in excess of its original valuation ($2 million) of the spectrum. Also, a UMTS–LTE provider will not bid if it believes that there is a huge risk that it will only be able to secure one piece of spectrum. There is a risk of one spectrum lot being unused representing an inefficient allocation of spectrum. If the MED adopted an auction structure requiring package bids then this risk is overcome.
    
52. Similar issues are raised in an MED report² where it outlines that package auctions are important when spectrum allocations are complementary in nature.
    

MED’s Band Plan - Option C (i)

53. In Telecom’s view this is the only appropriate MED option. The paired lots are essential to ensuring the availability of both WiMAX and UMTS-LTE technology. Although, we note there is still a risk of someone operating a TDD technology in an FDD part of the band.
    
54. Telecom’s concern with this proposal is that it creates an inefficient use of spectrum because of the proportion of the spectrum that is set aside for managed parks. This is discussed in greater detail below.
    
55. The MED proposal provides for paired spectrum in the managed spectrum parks. Telecom thinks this is unnecessary for two reasons. Firstly, Telecom considers it unlikely that it will need these paired bands within the managed spectrum parks to operate CMAR. Telecom should be able to use a band it purchases along with a managed spectrum park. Secondly, it thinks it is unlikely there will ever be a localised UMTS-LTE operator.
    
56. These issues are answered in greater detail below.

E Competition

Spectrum acquisition limits and the impact on competition

57. In light of the proposal that firms be subject to spectrum acquisition limits, a company will be forced to choose which technology path it wishes to take. When making this choice a company will be faced with a number of uncertainties.
    
58. 58. As identified in the discussion document, there is currently uncertainty around the direction that WiMAX technology will take. In particular, because of rapid technological change, it is unclear whether products which result from mobile WiMAX will effectively compete with products using UMTS-LTE. Unfortunately it is unlikely that this uncertainty will be resolved within the next two years.
59. A company will ultimately be faced with the uncertain choice between the WiMAX and UMTS–LTE technologies. Because of the likelihood that products using WiMAX will not be substitutes with UMTS–LTE based products (as recognised in the MED paper), a firm’s capacity to compete will be restricted to those firms who selected the same technology option as them. Accordingly, the number of competitive options for the wireless market will be restricted. The outcome of this will be segmentation of the wireless market on the basis of technology.
    
60. The competitive restrictions which would arise because of spectrum acquisition limits can be overcome through the use of the “technology acquisition limit proposal”. With technology allocation limits a firm would be able to engage both WiMAX technology and UMTS–LTE technology to compete across the entire wireless market. This would enhance the capacity for product growth and likely benefit end users.
    
61. However, when the actual nature and capabilities of new technologies becomes clear, it may eventuate that new technologies such as UMTS-LTE are not widely deployed in the 2.5 GHz band, thereby reducing the number of firms who want to use the 2.5 GHz spectrum. The delay option would enable the Government and the companies looking to invest in UMTS-LTE technology to make a more informed assessment of the viability of this technology.
    

F Spectrum Auctions

Simultaneous ascending auction

62. Auctioning of spectrum will tend to allocate it to a firm which is best able to use it. However, spectrum auctions do present some economic challenges because they typically involve the sale of multiple interrelated licences. If these issues are not well managed, the auction can result in sub optimal outcomes for both government and the industry.
    
63. The Federal Communications Commission in the United States has used a number of sale methods to enable efficient spectrum allocation and revenue maximisation. The most successful auction structure has been a simultaneous ascending design. Simultaneous ascending auctions are also used in Australia, Canada, Mexico, the Netherlands and the United Kingdom.
    
64. In its report on spectrum auctions in New Zealand the MED states that simultaneous ascending auctions have been successful and remain the most appropriate method for allocating spectrum lots with complementary or substitutable characteristics. ³However, the MED does note some issues such as exposure, efficient allocation, parking and collusion, although notes that these are not always experienced in New Zealand.
    
    ³ See Ministry of Economic Development Spectrum auction design in New Zealand
    
65. In order to manage these risks, the FCC in the United States opted for full information disclosure so long as collusion or predatory bidding did not take place. The knowledge of all information would benefit bidders through improved efficiency gains and transparency while also providing necessary incentives to ensure parties do not collude. The capacity for post-auction transactions would allow any inefficiency which may have arisen through initial assignment to be overcome.<sup>4
    
    4</sup> See Cranton, P., et al (2002), Handbook of Telecommunications Economics: Structure, Regulation and Competition (Volume 1), Elsevier Ltd, Oxford, UK at p. 608
    
    
66. We believe it prudent to retain a simultaneous ascending auction as it has previously been used successfully in New Zealand and has also been successful in a number of jurisdictions.
    

G Terms of Sale and Management Rights Secondary markets

67. We support the principle behind MED’s desire to ensure that there is access to spectrum by all market participants who are interested in the 2.3GHz and 2.5GHz bands. An appropriate framework for ensuring competitive allocation will mitigate the risk of spectrum monopolisation. However, we believe that following the auction firms should be afforded the capacity to sell their spectrum to their competitors. This would enable the market to continually evaluate efficient allocations of spectrum and allow for dynamic market growth.
    
68. We believe that Telecom’s proposed auction structure will protect the competitive nature of the market through regulating spectrum access in a manner that will inhibit monopolisation. Following this, competitive pressure should have taken hold and firms should be able to determine which markets will provide commercial benefit and to purchase or sell corresponding spectrum as they see appropriate. Applying the acquisition caps and associated party restrictions on an ongoing basis creates a risk that dynamic forces will not play themselves out.
    
69. If the acquisition limits were to apply well after the initial auction, Telecom considers that parties should be able to enter joint ventures without breaching acquisition limits and/or any associated party rules.
    
70. Finally, the discussion document does not specifically state that the reselling or wholesaling of services will be permitted between parties that acquire 2.3 GHz or 2.5 GHz spectrum rights. Telecom considers that the Terms of Sale management rights or the associated party rules should not preclude it from buying wholesale WiMAX services. Preventing this type of commercial behaviour is unnecessary and has the potential to stifle the market.
    

Use it or lose it

71. As we noted in our previous submission on MED’s discussion document Radio Frequency Auction Number 9 Overview, we do not support use it or lose it requirements (otherwise known as implementation requirements). We recognise the policy reasons for having implementation requirements to avoid spectrum not being allocated to its highest value use and to avoid spectrum hoarding and speculation. However, we believe that firms will have the incentives to use the spectrum and that there are a number of risks which are attached to use it or lose it requirements.
    
72. With an appropriate auction structure firms will face a competitive bidding process that will ensure that the spectrum is won by the firm which places the highest value on the spectrum use. Given the substantial outlay by a firm for spectrum and the fact that other firms will also own spectrum, a winning bidder will have the necessary incentives to use the spectrum so as to begin recovering the cost of the spectrum purchase and then earn returns. To not use the spectrum would result in a loss because there is no capacity for a single company to monopolise the entire spectrum band.
    
73. The key risk of a use it or lose it requirement is premature investment. To ensure retention of one’s right to spectrum, a firm will be incentivised to rush investments because of the enforced time period. However it may be commercially sensible and prudent not to invest on the basis of a regulatory timetable because of a number of variables which may be outside the control of the company concerned. An outcome of a regulatory induced incentive to prematurely invest may be that investments could become stranded or precluded.
    
74. Premature investment will not only harm the company being forced to invest but also results in inefficient allocations of resources across the market. This would negatively effect dynamic development and inhibit welfare maximisation.
    
75. For example there is uncertainty about the technological path of WiMAX technology but an even greater degree about UMTS–LTE technology. In regard to UMTS–LTE technology much of this uncertainty is unlikely to be removed within the next two years. Therefore to require use within 5 years could be commercially impractical when one factors in the time and investment in product development and then deployment of technology. The use it or lose it provisions may actually reduce the incentives for firms to bid for UMTS-LTE capable spectrum due to the risks involved.
    
76. If the terms of sale of management rights were to include use it or lose it provisions, it would seem essential that there were different timeframes for WiMAX and UMTS-LTE technology. This is because the development of UMTS-LTE technology is currently behind WiMAX technology.
    
77. As noted in previous submissions, we recognise defining what constitutes ‘use’ as problematic. This is further exacerbated by the fact there is considerable uncertainty regarding technological evolution. Setting fixed ‘use’ thresholds runs the risk of implementing black and white rules that are incompatible with actual technological development on which companies should be making their investment decisions.

CMAR and use it or lose it

78. Telecom would like MED to create a specific exception from the use it or lose requirements for circumstances where spectrum is being used for the provision of CMAR services. This exception would be in recognition of the fact that, while the provision of CMAR services may not fulfil the Ministry’s evidentiary criteria, it is a necessary service that is of great value to rural New Zealand.
    

Regulatory certainty

79. The spectrum auction and the requirements attached to each block of spectrum should be clear and not subject to change following the auction. Firms will be making significant outlays and investment to purchase spectrum access. Altering rights attached to lots of spectrum following the auction could significantly devalue the firm’s purchase. This will have implications for shareholders and raise uncertainty around future investments which could be made.
    

H Managed Spectrum Parks

Optimal use of managed spectrum parks

80. It is proposed that a fixed proportion of the total spectrum be allocated as managed parks (MSPs) to facilitate the provision of local and regional services. The Government’s objectives in creating MSPs are to encourage:⁵
    
    
    a) Efficiency;
    b) Flexibility;
    c) Availability;
    d) Innovation; and
    e) Competition policy.
    
    

81. This policy framework provides the context for discussions relating to MSPs.
    
82. Telecom supports the Government’s principle of availability and supports the Government in its consideration of frameworks which will enable niche providers to access spectrum to deliver services to their constituents.
    
83. However, we also recognise that this is a new area of spectrum policy. To ensure the successful implementation of MSPs we believe there are some issues which should be raised. 5 See Ministry of Economic Development, Radio Frequency Auction: 2.3 GHz & 2.5 GHz Bands, August 2007, see paragraph 111
84. Telecom does not believe it is allocatively efficient to have three MSPs. Given that there will already be two MSPs that are available for WiMAX operators (one in the 2.3 GHz band and one in the 2.5 GHz band), Telecom would suggest that the spectrum that is currently allocated to the third MSP should be made available for use by a commercial operator offering UMTS-LTE technology. Telecom does not think it is necessary to have 28% of the frequency available for managed spectrum parks.
    
85. One of the strongest reasons for this conclusion is that the total number of potential access seekers in MSPs can be compounded at a rate greater than that experienced for those purchasing spectrum with the intention of providing national coverage. This is because there is sufficient geographical isolation between niche access seekers to enable spectrum to be used concurrently across New Zealand. For example, a niche player accessing a segment of spectrum in Invercargill can access the same segment of spectrum being used by an access seeker in Wellington or Auckland. There is sufficient geographical isolation between these access seekers. If MSPs are well designed, total capacity is compounded.
    
86. As previously set out, Telecom does not need a “pair” of managed spectrum parks to operate CMAR. This is because Telecom should be able to achieve the separation required (64 MHz) to operate CMAR through using the frequency it purchases and the managed spectrum park. It submits there should only be one managed spectrum park in the 2.5 GHz band.

QUESTIONS

This section answers questions from the discussion paper. The questions Telecom have

answered are the ones that are relevant to its situation, but have not been dealt with in the first section of the paper.

A Question 1 Considerations Relating to Lot Design

Expected technical developments and likely demand

As discussed in the paper, many of these issues are still unclear. We attach a snapshot of current international developments:

• The 2.3 GHz band is used for 802.16 based networks. Korea has three operators in this band, each with 30 Mhz of spectrum;
• The 2.3 GHz band is also a candidate band for IMT advanced;
• The 2.5 GHz band is used in the US for 802.16 based systems. One operator has access to almost all the spectrum;
• A part of the 2.5 GHz band is likely to be used for BWA (802.16) in Japan. No decision on the exact frequencies and the number of operators is yet available;
• UK is considering an auction in early 2008 but has yet to finalise a band plan; and
  
• Other European countries have not yet declared their band plans and auction timings but are likely to consider the CEPT band plan that allows for 2 x 70 MHz of FDD spectrum and a 50 MHz TDD block. The FDD spectrum is likely to be used for UMTS LTE. All the UMTS operators are expected to adopt this standard as part of their on- going developments. Anecdotal evidence suggests that UMTS LTE systems will be in commercial existence in Europe 2012 and onwards

Acquisition limits

Key principles – Spectrum Blocks

In Telecom’s view, a minimum spectrum block allocation of 30MHz is required for WiMAX TDD technology. A viable system using 802.16e technology needs 3 x 10 MHz carriers. This will enable;

1. A repeat pattern of three that is essential to mitigate interference that arises in wide area deployments.
2. Average gross throughputs per 10 MHz is about 15-20 Mbps. This throughput is to be shared amongst multiple users.
   

This compares to the requirements for UMTS technology. The UMTS evolution which may be deployed in the 2.5 GHz band is the Long Term Evolution (LTE) standard. LTE is an FDD technology and requires 2 x 20 MHz carriers in order to achieve its targeted performance goals – supporting rates of up to 100 Mbps eventually. Allocations of less spectrum than this can be supported, but will effectively undermine the performance capability of the technology. Telecom therefore recommends the following minimum spectrum allocations:

a) WiMAX TDD: 30 MHz

b) Cellular FDD: 2 x 20 MHz

Key Principles – Acquisition Caps

It is Telecom’s view that if caps are to be utilised, they should generally be set at a level above the minimum carrier allocation size – and should be used to ensure that no one party can dominate the allocation of spectrum for a particular technology. The only constraint on this principle should be the total amount of spectrum available and therefore the degree to which larger spectrum allocations can be accommodated without compromising the key objectives of the Ministry and the Government.

Technical issues arising from accommodating TDD and FDD technologies and the necessity for Guard bands

Telecom believes that the Government’s objectives can only be met through a combined use of the 2.5 GHz band. We strongly urge, however, that the Ministry plan for FDD/TDD interference issues and consider what steps may be available to resolve these problems if/when they develop.

Telecom would also like to highlight some of its experience with interference issues to assist MED. We have experience of AMPS/GSM interference due to transmitters and receivers being in adjacent bands. In order to mitigate the interference we have used a guard band.

We also note that the ITU R WP 8F has conducted sharing studies between 802.16 and UMTS in the 2.5 GHz band and recommends that even with 5 MHz guard bands there will be a need for isolation contours.

Telecom considers 5MHz Guard bands to be essential. Telecom’s preference is that these guard bands be owned by MED.

If the guardbands are to be purchased as part of the management right, they should only be able to be used by the owner where they are not required to deal with interference issues. The needs of the owner of the adjacent spectrum should be paramount. An adjacent management right owner should not have to justify a request for a guard band, a guard band should be an absolute right.

C Question 2 Proposed Lot Design

Telecom’s preferred lot design

As discussed, previously Telecom submits that the band plan would allow for the optimal use of the 2.3 GHz and 2.5 GHz spectrum.

The 2.3 GHz Band

Telecom retains the view that the following band plan is optimal for this spectrum reflecting the reality of guard bands which effectively prevent more than two commercial
operators from being accommodated in this spectrum range.

We recommend the following option:

• 2 x commercial operators @ 30 MHz with 1 x Managed Park @ 26.5MH

The 2.5 GHz Band

The ITU-R has prepared a number of band plans for this band. In choosing a band plan for New Zealand, we recommend that the Ministry elect a mixed use plan which can
accommodate both WiMAX TDD and UMTS FDD uses.

Telecom believes that LTE is the natural UMTS FDD technology for use in the 2.5 GHz band. The FDD equipment requires a duplex spacing of 120 MHz and a centre band
spacing of at least 50 MHz.

Clearly, the critical element to be considered is the number of separate FDD operators who will be accommodated within the 2.5 GHz band. The other key consideration is the number of separate WiMAX TDD operators to be supported, and the need to establish a MSP for special interest and regional stakeholders.

2 x UMTS LTE Operators / 2 x WiMAX Commercial Operators / 1 x Managed Park

E Question 4 Eligibility to Bid

Do you agree that no restrictions should be placed on the eligibility of parties to bid for lots?

Telecom agrees that there should be no restriction on parties’ eligibility to bid in this auction. We have previously outlined our reasons for this to the Ministry and therefore do not repeat our reasons here. If MED would like further information, we will be happy to provide it.

F Question 5 Duration of Acquisition Limits

Do you consider the expiry of acquisition limits 1 year before the (December) use or lose date to be satisfactory? Do you prefer a different acquisition limit date? Please explain reasons for your view? Do you have any other comments on the time for acquisition limits to expire?

For the reasons set out above, Telecom does not think that spectrum acquisition limits are necessary but rather technology acquisition limits.

If the Government adopts our proposed “technology cap” option or the delay option, Telecom would not have a problem with the acquisition limits expiring one year prior to the use it or lose it date.

However, if the Government decided to retain a common spectrum cap across technologies Telecom would have to reconsider this position. This is because the spectrum cap would confine firms to only one technology option inhibiting competition across the entire wireless market. To achieve long-term competition across the wireless market for the benefit of end users, with the existence of a common spectrum cap, rationalisation of industry participants would likely be necessary. In addition to this, as identified in the MED paper, consolidation may also be necessary to improve the commercial viability of the services that are being delivered. Accordingly, the longer the duration of spectrum acquisition limits the longer it will take for welfare maximisation to occur. Therefore we would recommend that spectrum acquisition limits fall away sooner than is currently proposed.

Telecom is also concerned about the effect that acquisition limits and associated party rules may have on commercial arrangements with other companies. This issue arises because of the uncertainty as to how acquisition limits are defined. We recognise this issue can be resolved further into the policy development process but believe it is important that we raise the issue now. Telecom is concerned that these limits will interfere with arrangements between competing companies to enter into joint ventures, or to wholesale their services. Interfering with these types of arrangements would impact on the dynamic operation of the market and likely impede competitive pressures which enable efficient resource allocations for the benefit of end users.

G Question 6 Use or Lose Provisions
Do you prefer a date of December 2012, 2014, or 2106 [sic], for applying the use or lose test? If not what alternative date would you prefer to implement the Cabinet decisions on use or lose? Please explain.

For the reasons specified above, Telecom opposes use it or lose it provisions. This type of regulatory tool can lead to firms making premature investments so as to ensure they do not run the risk of losing their spectrum right. This negatively impacts upon companies and the economy due to irrational investment decisions which alter dynamic market outcomes. Reasons as to why investments would be premature are likely to be the result of factors which are outside the firm’s control. For example, the technical capability may not yet be available to make the most optimal investments.

However, if the Government feels that use it or lose it requirements are necessary Telecom considers that there should be different use it or lose it dates for WiMAX and UMTS-LTE technology. This is because of the relative degrees of uncertainty around each technology. At present there is greater market knowledge around the capabilities and likely global deployment profile of WiMAX technology in comparison to UMTS–LTE technology. As such, assuming the government proceeds with this tool, the use it or lose it period for UMTS–LTE should be greater than that for WiMAX given the technological uncertainty.

Do you prefer Option A or Option B or some other option regarding the test for ‘use’? If you do not agree with the proposals, what ‘use or lose’ provisions do you propose?

If Telecom were required to choose an option it would prefer Option B. Option B is preferable because it would provide management right holders with more certainty. This is because it provides details about what the company is expected to achieve.

Should financial consequences result from not using rights, either in addition or as an alternative to loss? How should the amount of any financial consequence be calculated?

The discussion paper proposes various financial penalties for non-use. As previously set out in Telecom’s submission on the 2.3 GHz auction paper, if implementation requirements are imposed, Telecom is opposed to the introduction of financial penalties in addition to the loss of spectrum. We consider that the loss of spectrum (along with any infrastructure investment that may have taken place) is sufficient (in fact in our view it is unnecessary) incentive to invest and that nothing is gained by the introduction of financial penalties. The inclusion of penalties can only reduce the revenue that the government can earn from the auction of the rights, as bidders will factor this cost into their expected of the auction.

The Ministry has suggested that a “buy out” option might be attractive. Telecom would consider that this type of option to be unnecessary if realistic implementation requirements have been set.

H Question 7 Settlement

Do you agree with the proposed settlement terms (30 days following completion of the auction)? If not, what other factors do you see as relevant?

7.1 Telecom agrees that the MED’s proposed settlement terms are appropriate. We note that the proposed terms are also consistent with the approach that has been taken in previous auctions.

I Question 8 Managed Spectrum Parks: Proposed Allocation Procedures and Usage Rules

Do you agree with the suggested eligibility criteria for access to MSP?

Telecom would like to ensure that CMAR is explicitly dealt with in the eligibility requirements. Telecom would like to ensure that where it is necessary, it can use managed spectrum parks to deliver CMAR services to rural customers.

Telecom agrees that, except for the provision of CMAR, holders of management rights in 2.3 GHz and 2.5 GHz spectrum and associated parties should not be eligible to participate in the managed spectrum parks.

In addition to this whether regional operators can enter into roaming agreements with other operators should be clarified.

Do you agree with the suggested core technical and usage requirement?

In light of its importance to rural customers, special provision should be made to ensure, that where necessary, CMAR can operate in MSPs.

Of the three options outlined above, which is your preferred method for implementing a MSP? Why?

Telecom considers that Option 1 is not viable because of the “General User Licence.” This type of licence would impact on the provision of CMAR services. This is because, where a general licence was in place, Telecom would not be able to guarantee that it could maintain these services at a commercially acceptable level.

Telecom does not have strong views on whether Option 2 or 3 should be preferred.

What fees or resource charges should be levied (particularly for your preferred option)?

Telecom considers that users of frequency in managed spectrum parks should be charged to reflect the value of using a limited resource. This is important to ensure there is a level playing field between all WiMAX operators. If this were not the case the WiMAX providers who are operating out of managed spectrum parks would have a competitive advantage over providers who were required to purchase their spectrum. While this may not be a problem where providers are servicing smaller centres, this type of competitive advantage could not be justified in a market like greater Auckland.

APPENDIX A

2 x UMTS LTE Operators / 2 x WiMAX Commercial Operators / 1 x Managed Park