5. There are dangers in the MED directly influencing spectrum allocation

46. TelstraClear and Econet are proposing that the Ministry effectively allocate 900 MHz spectrum to them at less than Vodafone is willing to pay for it. ¹
47. The main argument for not allocating the spectrum to the highest value user is that there will be broader competition benefits to New Zealand by allocating spectrum to others, and that these benefits are large enough to outweigh the costs and risks associated with more direct Ministry control of spectrum allocation.
48. In this section, we offer some comments on this argument. We make two main points:

• The argument is inconsistent with the statutory provision for renewal rights and their underlying policy. Vodafone has been entitled to proceed on the basis that a rights holder who invested heavily and used the spectrum previously allocated as expected could expect it to be renewed.
• Even if permissible, this approach could only generate significant competitive benefits if a new entrant required 900 MHz spectrum to enter, and would actually use the spectrum in rolling out a network outside the cities.

49. We also consider what we might expect TelstraClear or Econet to do if they were to acquire this spectrum.

A market management approach is legally questionable

50. A more interventionist approach by the Ministry in the allocation of spectrum would be a significant departure from published policy. Certainly any decision not to renew would need to be justified by clear reasons that explain precisely why Vodafone is thought not to be the highest value user of this spectrum.
51. It is very late in the renewal process to be suggesting non-renewal of rights for its own sake and irrespective of the likelihood of Vodafone’s valuable use of the spectrum in the future based on its actions in the past.
52. Cabinet policy and the articulation of this policy by the Ministry favour offering the rights to the existing right holder at an estimate of the market price unless there is a clear reason for not doing so.² These reasons could include a lack of use of the right, or a lack of investment in the spectrum.³
53. In our view Vodafone has met the conditions for renewal. We are currently using all the spectrum that we hold, and society is getting significant value from our use of the spectrum, which has been the subject of very significant investments by Vodafone to provide an important part of New Zealand communications infrastructure.
54. On this basis we have always expected that our current rights to 15 MHz of 900 MHz spectrum will be offered to us for renewal for 20 years.

900 MHz spectrum is not required for a metro mobile build

55. Leaving aside the methodology of the renewal process, Econet and TelstraClear’s argument is that there would be a significant boost to competition in the mobile services market if they had some 900 MHz spectrum.

• TelstraClear goes so far as to say that their lack of access to 900 MHz spectrum is a “likely reason” why there are only two operators in the New Zealand market and why retail mobile prices rank so poorly on OECD comparisons (para 8). ⁴
• The Commission also looked at this issue in its recent report on mobile market entry issues. It concluded that the lack of availability of 900 MHz spectrum is a barrier to entry for new entrants because it raises the costs of entry (see para 13 of its report).

56. We agree that, if an entrant can only enter the market with access to 900 MHz spectrum, then our ownership of the rights to all 21 MHz of 900 MHz spectrum for the next 25 years would be a very significant barrier to entry.

• In that event, no entrant would value the spectrum as highly as we would. The value for us would not just be the value of reduced network operating costs from using 900 MHz spectrum rather than 2.1 GHz in our 3G network, but also the value of avoiding retail market competition from another UMTS entrant for the next 25 years. But this is not the case.

57. An entrant clearly does not require access to 900 MHz spectrum to build a 3G network in metro areas.

• We have built a 3G network out to almost 50% of the population using 2.1 GHz spectrum. We started commercial services on this network in August 2005 and we continue to build it out.
• In addition, TelstraClear has announced they will offer commercial services on a new 3G network in Tauranga from the middle of 2007. This network will use their 2.1 GHz spectrum (see para 7 of their submission).

58. This point is supported by the results of our model of the costs of entry for a five-city network by a new 3G operator. Our numbers suggest that the capital expenditure savings from using 900 MHz rather than 2.1 GHz spectrum are not large relative to the total costs of entry, i.e., 900 MHz spectrum does not make a major difference to the business case for metro entry.⁵

• In addition, 900 MHz spectrum does not seem to generate significantly reduced site counts in the cities. As an example, we have [ ] 900 MHz 2G coverage sites in Auckland at present. And we have [ ] 2.1 GHz sites to provide 3G coverage to Auckland.
• Others also recognise that 900 MHz spectrum is most us eful for rural areas (see paras 17 and 18 of Telecom’s submission, and Econet on page 5). As far as we know, neither Econet nor TelstraClear have any public plans to build outside the cities. Indeed, it seems that Econet may be concerned at building even to 30% of the population (page 7).

59. If access to 900 MHz spectrum is not absolutely required for entry, then the Ministry’s standard approach to renewal will work as planned.

• The Ministry will propose a price for renewal.
• If this is less than the additional costs for us of operating without the renewing spectrum, then we will renew.
• If an entrant values the 900 MHz spectrum more highly than we do on the basis of the cost savings to them from using that spectrum as opposed to 2.1 GHz, then they will buy it from us after renewal.
• If the Ministry’s proposed price is greater than the value we place on the spectrum, then an entrant can buy it at the subsequent auction.

We question whether Econet and TelstraClear would actually use this spectrum if they had it

60. The likely alternative holders of 900 MHz spectrum are Econet and TelstraClear.

• Econet has promised to invest $200 million. From the information published in the media Vodafone understands that Econet is in the process of securing the funding but has not done so yet.
• Econet says it has plans to build a network to par with Telecom and Vodafone. Econet has been planning this network for some years. Econet has access to 2.1 GHz spectrum that it could use to start roll out a 3G network.
• Econet has said it will not build until stronger regulation forces Vodafone to provide it with roaming at a very low price, and with cheaper access to our sites for it to put its radio equipment on, as well as other controls on our retail pricing behaviour.
• Econet is prepared to pay only a “nominal fee” for spectrum (page 6 of its submission). Econet’s failure to make use of considerable spectrum it already has (including 1800 MHz rights) may raise questions about whether it could put more spectrum to good use.

61. TelstraClear is a much more credible entrant given its significantly greater scale at present, and an experienced parent company in Australia that is rolling out a 3G broadband network in 850 MHz spectrum.

• TelstraClear is in the process of trialling its “Unplugged” service. It has not specified its investment plans beyond Tauranga and has not indicated that is has any commitment to build outside the cities.

62. In addition, both TelstraClear and Econet had opportunities to secure access to 900 MHz spectrum before now that they did not take up.

• Telstra held rights to 900 MHz spectrum when the mobile market was only 10% penetrated. At that point it decided to sell the spectrum to BellSouth.
• In 2001 both Econet and TelstraClear had rights to 1800 MHz spectrum suitable for GSM when the market was around 60% penetrated. These companies then decided not to roll out a network. There may be concerns about the feasibility of rolling out a network now, when higher mobile penetration is widely thought to make entry more challenging.
• At the time of the 2002 900 MHz auction both Econet and TelstraClear would have been aware that the earliest 900 MHz spectrum would be available was 28 November 2011 if they did not get 900 MHz spectrum in that auction.

63. Allocating 900 MHz spectrum to an alternate provider who has no plans to build in rural areas could lead to the spectrum being sold to Vodafone at a later date. In this scenario the New Zealand taxpayers may be worse off.

• Although the highest value user of the spectrum (presumed to be Vodafone) will eventually be the owner, Econet or TelstraClear may get a windfall gain from the sale that would otherwise go to the Crown.

¹Econet suggests that the 900 MHz GSM spectrum be cut into four equal lots and divided amongst the current holders of 1800 MHz spectrum (page 5 of its submission). It seems that Econet is willing to pay no more than a “nominal fee” (page 6) for spectrum. TelstraClear is happy to have an auction to reallocate rights. But it is prepared to bid only against Econet, unless Vodafone or Telecom both give up all claim to renewal on all spectrum rights they already hold (para 14 of its submission).

²The Cabinet policy states as two of the key elements of the policy (POL Min (03) 9/1): “that commercial spectrum rights be reallocated five years before expiry for a further 20 years, subject to review on a case-by-case basis to ensure consistency with New Zealand’s international radio obligations and with the general objective of maximising the value of the spectrum to society as a whole; … that spectrum rights be reallocated to existing rightholders based on price-setting formulae that estimate the market value of the rights, and that, if existing right-holders do not wish to pay this price, the respective rights be reallocated by auction.”

³A section of a discussion paper released by the Ministry in June 2004 looks at reasons for not offering new rights (“Implementation of a Price-Setting Formula for Reallocation of Commercial Spectrum Rights”, para 29): “The Ministry will consider the current and likely future use of the particular spectrum, based on available market and technical information, and whether reallocation of the rights in the current form will maximise the value of the spectrum. If a spectrum right has not been used over its lifetime, or there has been little or no spectrum-related investment in the band as a whole, offering the right to the present user or in the present form is unlikely to maximise the value of the spectrum.”

⁴It is also worth noting that TelstraClear’s information is based on February 2005 data. The most recent data (from August 2006) shows dramatic reductions in prices in the last 18 months, with New Zealand now ranking 21st for low users, 24th for medium users, and 21 st for high users compared with 23rd, 30th and 29th on the February 2005 data. Vodafone has publicly stated that it will lead New Zealand to the OECD average (i.e., in 15th place or better) by 2010, in line with the government’s goals for telecommunications generally in the Digital Strategy.

⁵In the model, cell sites make up 40% of an entrant’s capital costs, a total of $93 million. Assuming 900 MHz spectrum uses 60% of the number of cell sites of 2.1 GHz spectrum, then the savings amount to $37 million or 16% of the $238 million total capital expenditure. This is not insignificant, but pales in comparison to the $1.1 billion or so we estimate a new entrant would need to commit to build a five city network and generate 15% market share over 5 years.