Spectrum re-allocation

16. To promote competition, the goal of any spectrum re-allocation should be a level playing field for market participants. No party, and particularly no incumbent, should be given an advantage with respect to spectrum allocation.¹¹
17. While we agree with the draft report’s conclusion that spectrum resources are not scarce ignoring current allocations, the monopolization of the high-quality spectrum (the “beachfront” property in the 850 and 900 MHz bands) by two operators creates artificial scarcity. In this context it is important to emphasize the distinction between allocation of spectrum in the 850 MHZ (CDMA) bands and allocation of spectrum in the 900 MHz (GSM) bands. While both bands demonstrate the superior propagation characteristics that are particularly helpful in reducing network deployment costs, the spectrum blocks allocated in New Zealand accommodate two different network technology platforms.¹² Therefore, they are not equivalent to any potential acquirer of 850 and 900 MHz spectrum.
18. A new entrant will require spectrum in frequencies that match their chosen technology platform, and, where the new entrant has existing spectrum, a spectrum band that matches the technology designated for its existing frequencies. Frequency bands are essentially technology specific; GSM equipment (handsets and base stations) are generally built to operate only in the 900 MHz and 1800 MHz ranges; CDMA equipment in the 850 MHz and 1900 MHz range.¹³ EWNZ has management rights for spectrum in the GSM 1800 MHz band; therefore, EWNZ will need spectrum in the 900 MHz band in order to accommodate its network roll-out.
19. In theory, the 850/900 MHz bands can easily accommodate as many as five mobile competitors, an excessive number for a market this size. In comparison with similar-sized OECD countries, New Zealand should be expected to support three competitors and one substantially smaller competitor. Therefore, in re-allocating these spectrum bands, the MED should keep in mind allocations that accommodate a healthy number of competitors.
20. a. Telecom is reliant entirely on its 850 MHz spectrum, which is heavily (and efficiently) used, while Vodafone has additional spectrum in other bands, and does not appear to make efficient use of its 900 MHz spectrum holdings. Therefore, we do not consider it unreasonable at present to allow Telecom to retain a greater slice of 850 MHz spectrum pending deployment of new technology. In the alternative, we would recommend that 1 block should be reserved for future need or technology developments.
    b. Econet would like the MED to understand that there is a completely distorted mobile market in New Zealand. Econet agrees with the Commerce Commission Mobile Market Review. This imbalanced market at a high level of penetration can create a deadly battle field for new entrants because of the significant market power the incumbents have. Any new entrant has to build a nationwide network which has economy of scale and scalability to battle with the incumbents. However, the incumbents “want a series of pocket networks which have no scalability and can never challenge the incumbents”. Econet needs the MED to understand the safe harbour concept of spectrum allocation to facilitate a break period whereby a new entrant who has not paid off its capital base or even is recovering its opex costs is simultaneously trying to fight off several pocket networks and has an increased SAC cost as a consequence of brand confusion created by retail MVNOs. We request the MED discuss the safe harbour concept with the Commerce Commission and further discuss this principle with Econet. Econet is willing to consider a “building requirement” whereby Econet is obliged to build say 800 cell sites or cover 80% of the country in return for the safe harbour concept. The safe harbour concept is a normal OECD technique to bring competition to a late penetration market.
21. Re-allocated spectrum should be offered first to current holders of 2G and 3G spectrum in the 1800 MHz and the 2.1 GHz bands. Such a right-of-first refusal would ensure that those parties that have already committed to building networks receive the benefit of a level playing field. In addition, transfer of the spectrum to third parties would raise concerns about the number of operators that the market could support – giving too little spectrum to too many players will cause market uncertainty and confusion that may cause the failure of many of the new start-ups.
22. As noted earlier, while TelstraClear also noted that 7 MHz allocations are typical, they propose 10 MHz allocations, one in the CDMA band and one in the GSM band.¹⁴ TelstraClear holds cellular spectrum concentrated in the 1800 MHz (GSM) range (and additional cellular spectrum in the 2.1 GHz band). Under TelstraClear’s proposal, only one of the two potential entrant’s to the mobile market holding other GSM licenses – TelstraClear and Econet – would be able to obtain compatible 900 MHz spectrum. We would prefer that new entrant’s be preserved a choice, and therefore suggest that the 7-8 MHz allocation is more appropriate as it both meets world standards and would accommodate the interests of the two known GSM spectrum managers.
23. Based on the foregoing, it is EWNZ’s opinion that the best means of optimizing the spectrum would be to:
    1. Reduce Telecom’s 850 MHz allocation to under 886 MHz. This would allow full use of the 25 MHz standard GSM spectrum (the spectrum at 890 MHz being nearly unusable due to interference concerns with Telecoms CDMA operations in this band), and would leave Telecom with 16 MHz (paired) in the 850 MHz band.
    2. Divide the GSM spectrum into three equal portions.¹⁵ The approximately 8 MHz blocks would be consistent with the standard 7-8 MHz blocks distributed in most OECD countries. These spectrum blocks should be reserved for GSM and UMTS use.
    3. Distribute the blocks to existing 1800 MHz operators. This would allow all current spectrum owners, Vodafone, TelstraClear, and EWNZ to have 900 MHz GSM spectrum to accommodate their 1800 MHz spectrum, ensuring all entrants can achieve cost-efficient coverage in urban fringe environments and rural areas.

See Diagram 2.

Auctions

24. In re-allocating the spectrum, EWNZ does not believe a re-auction of spectrum would be appropriate. As noted above, in order to ensure effective levels of competition, and in order to create parity between all holders of cellular spectrum, preference should be given to incumbent spectrum managers in the 1800 MHz range, a scenario which does not accommodate auctioning.¹⁶
25. In the event the MED determines that auction should be the preferred means of reallocating the 850/900 MHz spectrum, the MED must introduce spectrum caps or other measures to bar participation from either of the incumbents for the re-auctioned spectrum. Participation by the current spectrum holders would artificially increase bids for the spectrum and would potentially leave the allocation with the same disparity as the auction was meant to remedy.

¹¹ That is not to suggest, however, that some forms of temporary preferential treatment to certain market participants might not be necessary in order to balance out the threats posed by the Significant Market Power of the incumbent operators.

¹² CDMA in the 850 MHz spectrum held by Telecom, and GSM in the 900 MHz spectrum held by Vodafone.

¹³ While there are variances to these standards – for example, the U.S. operates GSM in the 850 band – these standards apply as a general rule, and the majority of equipment is designed to operate only within these standardized classifications. Therefore, the operators of either GSM or CDMA are essentially locked into, respectively, the 900/1800 and 850/1900 frequency ranges.

¹⁴ See TelstraClear submission, paragraphs 13 and 52(c).

¹⁵ In our original submission we errantly listed four portions to accommodate all carriers, thereby including Telecom, who would instead continue using the 800 MHz CDMA spectrum, not the incompatible GSM spectrum.

¹⁶ We would also note that the auction processes used to date have not been particularly successful in allocating critical spectrum in New Zealand in a timely manner. In particular, the Auction 3 (2 GHz) rules used in 2000/2001 allowed the auction process to be dragged on for over half a year, rather than the hours or, at most, days found in auctioning processes in other OECD countries. Further delays in the re-allocation of spectrum would do a great deal of harm to end-users and the introduction of competitive new entrants in the mobile market.