2. The recommended pricing approach

Issue 2.1

Is the proposed incremental ODV approach, benchmarked against New Zealand and overseas spectrum values, a reasonable and appropriate approach in approximating a market value for management rights in the 800 MHz and 900 MHz bands? Why or why not?

41. Telecom found it difficult to assess IODV model, without further guidance from the Ministry. This is because:
    1. We do not know what generic pricing inputs the Ministry will determine appropriate. Although we were asked to choose inputs based on a ‘generic’ model, there is a range of possible inputs, and the choice of input could result in very different output values. For example, if we chose 1.25MHz that would produce quite a different result from if we chose 5 MHz.
    2. As Telecom does not have an actual model, we do not know how the Ministry will combine the various inputs. Again, the way in which the Ministry applies the inputs may result in quite different outcomes. For example, a length of renewal period could range from 1 to 20 years. E.g. In year 1 ‘X’ we require 1.25MHz of spectrum, which is purchased for 20 years, in year 2 ‘Y’ we require another 1.25MHz of spectrum which is also purchased for 20years. We now have two blocks each requiring a separate PV calculation (‘X’ for 19 years, and ‘Y’ for 20 years).

This is assuming, that is how the model works, or do all renewal periods end after 20 years from year 1?

3. IODV is technology dependent. It may therefore unfairly disadvantage technologies such as CDMA, which has different spectral requirements from GSM.
42. We are not saying this cannot be worked out. Once the values and how they are to be applied are known, it can be done. Our key concerns with the model are that it:
    1. is unlikely to produce a fair market value. This is because to reflect the market there must be a scarcity of spectrum and, as noted in the Ministry’s advisors’ reports, this does not reflect the New Zealand situation;
    2. will be difficult to apply as it relies on future forecasts and shift the value of spectrum, according to the model, significantly. Therefore, there will be significant debate over the correct values; and 
    3. without worked examples, incumbent right holders and other stakeholders will arrive at differing outputs. This would bring into question the validity of their conclusions, comments and suggestions.
43. It is Telecom’s view that IODV, as currently presented in the paper, is confusing and open to misinterpretation and therefore not simple or transparent.
44. Contrary to what is suggested in paragraph 40 of the discussion paper, until the inputs and their application are determined, any outcomes would be speculative.

Top

Does this approach meet all policy objectives (fair market value, fair return to the Crown, transparent and simple to administer)? Why or why not?

45. As outlined above, we do not consider IODV meets the policy objective of being transparent and simple.
46. Nor do we think IODV will result in a fair market price. We note that this view is reinforced by comments made by CRA International:
    1. The CRA paper highlights that the ‘ODV has some significant limitations in that it is impossible to fully replicate the reality of the market place’.³ We therefore argue that IODV will not produce a fair market value.
    2. The CRA paper goes on to highlight that ‘[IODV’s] output will necessarily be dependent upon a number of judgements and assumptions made without full information’⁴ Although CRA concluded that the methodology was reasonable, the acknowledgment confirms our view that IODV is unlikely to produce a fair market value.
    3. CRA also states that ‘to the extent that the actual licence up for renewal is significantly different from the generic licence used in the analysis, the calculated spectrum value could differ significantly from the value that might be obtained at auction’.⁵ Again, this indicates that IODV will not produce a fair ‘market’ value, as would be achieved by an auction. 

Issue 2.2

Are there factors or inputs that should or should not be considered in the proposed methodology? Why or why not?

47. Telecom recommends that further discussion or a workshop be held to agree a ‘generic model’, the associated input values and how they are applied. Security of tenure issues would have to be addressed.
    

Top

Issue 2.3

Do you have any comments on the benchmarking to be applied to the ODV-calculated value?

48. The discussion paper draws a link between spectrum in New Zealand being ‘non-scarce’ and so the value of a marginal piece of spectrum to the operator is likely to be low. We recommend that the steps outlined for benchmarking should include a selection of countries with similar ‘nonscarce’ spectrum supplies to ensure that prices obtained in those countries are consistent with those expected in New Zealand.
49. Arguably the most relevant data source for any spectrum will be those auctions that have been carried out in New Zealand in the past. These provide the most up to date and comparable data set for valuing the current spectrum band, and the Covec methodology is a benchmark that is consistent with that approach.
     

Issue 2.4

Do you have any other comments on the recommended approach?

50. No further comments.

Do you have any other comments on the consultants’ report and peer reviewers’ feedback released with this discussion paper?

51. In the paper Renewal of Spectrum Rights for Cellular Services Pricing Methodology (“the PwC/NZIER paper”), PricewaterhouseCoopers and NZIER conclude that IODV is appropriate for determining a price for management rights of the 850 MHz and 900 MHz to be offered to the existing rightholders. The peer review by CRA supports this conclusion. We disagree with the reasoning in the PwC/NZIER paper because:
    1. It does not consider the incentives of the bidders in an auction for the renewal of spectrum rights;
    2. It does not consider the effect of renewal on the welfare of consumers; and
    3. IODV is inconsistent with the allocation of spectrum in New Zealand.
52. The following discussion starts by setting out Telecom’s view regarding the policy objective of renewal process. It then moves onto a critique of the PwC/NZIER paper, concluding with some observations regarding the determination the appropriate setting of the renewal price.
     

The policy objective of rights renewal

53. The various documents containing the policy behind the renewal process are set out above at paragraphs 30 to 32.
54. In summary, the general objective of the renewal process is to maximise the value of the spectrum reallocation to society as a whole. Further, the Government’s fair financial return for the use of the spectrum is subject to this general objective (see paragraph 30(a)). Therefore, the price-setting formula for the renewal of the spectrum, which is to ensure the government earns a fair financial return, must maximise the value of the spectrum reallocation to society as a whole.
55. This is supported by the decision to auction the spectrum band if the existing rightholders do not accept the Government’s calculation of the renewal price. Reallocation by auction, taking account of potential switching costs, maximises the value of the spectrum to society, as the auction is a competitive process that assigns the spectrum to the bidder with the highest valuation for the spectrum. Furthermore, the bidder pays society’s opportunity cost of the spectrum for the spectrum, where the opportunity cost is value of the alternative use of the spectrum which is next highest valuation.
56. This will also maximise the revenue that the Government can expect to earn from the assignment of the spectrum. Should the Government make an offer that is greater than the expect auction price of the spectrum, then it would not be rational for bidders to accept this offer if the downstream markets are competitive. If a bidder accepts an offer that is greater than the opportunity cost of the spectrum then they will incur a loss. The assumption that the downstream markets are competitive is consistent with PwC/NZIER’s assumption that the purpose of the assignment is to limit profits or to extract economic rents. The Government has more targeted and efficient instruments for examining competition related issues.
57. This leads to two possible approaches for thinking about the price-setting formula. Firstly, the price-setting formula may calculate the price that maximises the value of the spectrum to society as a whole. This is essentially a cost-benefit calculation, where the price-setting formula calculates the price that maximises society’s total welfare. This would require evaluating potentially a large number of alternative options to determine the option with the greatest value to society and the opportunity cost. Telecom is not advocating this approach as it would be computationally complex and inaccurate, whereas the auction mechanism is much lower cost is far more accurate at valuing the spectrum. However, thinking about the problem in this way helps highlight some of the issues to be considered in a price-setting formula. The second approach is that the price-setting formula is the Government’s expectation of the market price for the spectrum given a competitive auction. This seems to be the approach taken by PwC/NZIER for the reasons set out below.
     Top

Maximising total welfare

58. The following discussion of a cost-benefit analysis is to highlight some key issues. If a hypothetical cost-benefit analysis were carried out, the renewal price would maximise the net benefits – that is the total benefits less total costs, where total benefits and total costs are the costs and benefits of the service providers and consumers of mobile services over time.
59. Some of these costs and benefits were considered by the Ministry in the decision to reallocate existing rights at least five years in advance of their expiry. The Ministry notes in a background information paper on its website⁶ that a lack of certainty for rightholders and other industry stakeholders for dealing with rights on expiry can have a negative impact on economic growth and innovation. It emphasises the difficulty in making good investment decisions and being able to plan effectively. The Ministry notes that “from an economic perspective, the reallocation of commercial spectrum rights should:
    1. Optimise incentives to invest – ie long-term tenure provides certainty for investment;
    2. Minimise stranded investment – this could occur if existing rightholders lose their spectrum rights and are required to scrap any investment; and
    3. Minimise supply discontinuity – this could occur if existing rightholders lose their spectrum rights and as a result must stop providing services to their customers.”
60. Similarly, we proposed that if the pricing formula estimates the market value of rights it is arguably more efficient to offer the spectrum to existing rightholders first, rather than auction the rights in the market, because it avoids the following costs:
    1. The costs an existing rightholder would incur due to switching to another spectrum band or even ceasing to provide the service, should they lose the spectrum at the auction;
    2. The additional cost that customers of an existing rightholder would incur due to the rightholder having to switch to another spectrum band or even ceasing to provide the service, should they lose the spectrum at the auction; and,
    3. The costs of administering an auction, and the resulting delays and uncertainty that this creates.
61. One important implication of these costs is that the economic problem for “renewing” spectrum differs from the economic problem of assigning "new" spectrum. To illustrate, if new spectrum is being auctioned then all bidders are essentially starting from the same position, whereas if the spectrum is currently assigned there will be downstream markets, and the existing rightsholder will be starting from a different position than an entrant. We think the Ministry should take this into account.
62. In particular, in the case of renewal, if the existing rightholder loses its management rights it may have to write-off sunk assets whereas an entrant does not have any sunk costs to write-off. This means that the existing rightholder's valuation of keeping the spectrum is likely to be greater than an entrant’s valuation of winning spectrum by the value of the write-off. However, the efficient price is not the incumbent’s valuation but the opportunity cost to society, which in this case would be the entrant's valuation less the cost consumers would incur if the existing rights holder lost the rights.
63. From a policy perspective, this asymmetry of the economic effect of the existing rightholder keeping or losing rights extends to the consumers of the service in the downstream market for mobile services. Given the Cabinet decision to consider the effect of renewal on the whole of society, the Ministry should also include the costs to consumers of renewal or the failure of renewal. In particular, if the existing rightholder lost its rights, then consumers are likely to incur costs when, say, switching to another service provider or perhaps a new service developed by existing provider operating on different spectrum bands. Such costs may include consumers having to write-off current mobile phones as well as the time and effort to make the switch.
64. The entrant or incumbent is not likely to bear these costs, therefore these costs are not likely to be factored into the auction bids. As a result, there is a risk that an auction will not result in an efficient allocation of spectrum. That is, there is an increased possibility that the existing rightsholder does lose the auction when it may more efficient for the existing rightsholder to keep the spectrum. The Ministry can reduce this risk by taking into  account the consumers switching costs when calculating the renewal price for the existing rightholders.

 Top

PwC and NZIER’s pricing methodology

1. PwC and NZIER’s rationale for the recommending IODV is that it calculates a price that will earn the government a “fair financial return” on the reallocation of the spectrum. PwC and NZIER also claim that the government does not provide guidance regarding the definition of fair financial return other than it is a return that is consistent with the outcome of a price that is consistent with an auction.
2. PwC/NZIER also make some “other assumptions”:
   1. The renewals process is not intended to be a regulatory exercise (i.e. to limit profits or to extract any economic rents that the current spectrum users may be enjoying); and
   2. The Government’s overall approach to management of radio spectrum is to seek to maximise the economic and social welfare of the country as a whole. It would therefore not seek to place artificial constraints on the supply of spectrum.
3. Having made these assumptions, PwC/NZIER do not seem to refer to them in their subsequent analysis. Although it is not clear, they seem to pursue a line of reasoning which supports the objective of with extracting the existing rightholders’ willingness to pay for these blocks of spectrum as they conclude:

“Unfortunately, in the final analysis, all that will be certain is that if the operators accept the offer price [as determined by the pricing methodology], then it is lower than they were actually prepared to pay.”

68. It is unclear why PwC/NZIER should consider that it is unfortunate that an operator will accept a price when it is prepared to pay more. In an auction the winning bidder will always pay less (or in some rare situations the same) than they would be willing to pay. In the case of auctions for new spectrum allocations, if the winner is rational they will pay the opportunity cost of spectrum, which is the highest losing bid.
69. Telecom submits that the output of any price setting formula should represent the Government’s view of a fair price for the spectrum that would be determined by an auction, taking into account the potential switching cost to consumers. There is nothing unfortunate if the existing rightholder accepts this price, as this is an indication that this price is consistent with the existing rightholder’s expectation of the auction price.

In fact, it would be unfortunate if the existing rightholders did not accept the Government’s price as it would indicate that the calculated price is not considered fair, and that rightholders are willing to accept the significant risk of an auction process.

What is the basis for the price-setting formula?

70. The pricing-setting formula should produce the Government’s estimate of the market value of the spectrum rights for renewal, with an adjustment to take account of the potential switching costs for consumers. If this price is greater than the existing rightholder’s expected auction price to renew then the existing rightsholder may progress to an auction. This will depend on the extent to which the formulae price deviates from the existing rightholder’s expectation of market value.
71. Therefore, Telecom submits that the best estimate of an expected auction price is the auction price received for the rights up for renewal modified using the Covec methodology, as well as the auction prices received for other comparable rights. Telecom has presented a substantial body of evidence in this submission on both the Covec methodology and the auction of other comparable rights.
72. PwC/NZIER recommends that the appropriate methodology is IODV. Telecom does not agree with this recommendation. The basis for PwC/NZIER’s recommendation seems to be that it is a simplification of a more general ODV approach, and that IODV reflects the commercial decisions that a mobile service provider will make between building an additional cellsite or purchasing another unit of spectrum. However, the basic flaw in this approach is that IODV assumes firms actually trade-off cellsites and spectrum at the margin of the network, which is not the case.
73. One reason that firms do not make commercial decisions in this way is that spectrum is not allocated in the way that PwC/NZIER assume. The
    definition of the management right has a significant bearing on the value of the spectrum. As Peter Cramton notes:⁷

“There are two steps in making spectrum available to companies. The first step is the allocation of the spectrum for licensing. The allocation defines the licence (the frequency band, the geographic area, the time period, and the restrictions on use). The second step is assigning the licenses to particular companies. … Arguably the greatest economic gains will come from better allocation of spectrum, rather than from improved methods of assigning spectrum. ...” 

74. The proposed IODV methodology essentially assumes that spectrum rights are traded in some small increment and that operators make a trade-off between whether to build a cellsite or to buy an additional increment. This is clearly inconsistent with the spectrum allocation under the current management rights and how these rights are traded. Therefore, no conclusion can be drawn regarding the price that the IODV calculates per se and the opportunity cost of the management rights, other than whether or not a calculated price is consistent with individuals’ expectation of a future auction price for the management rights.
    

Issue 2.5

If you do not agree with the proposed approach, what alternative approach would best meet all policy objectives?

75. As outlined above, we believe the Covec model is the only methodology that meets the policy objectives at present.

³See the CRA paper, page 2, paragraph 4.

⁴See the CRA paper, page 2, paragraph 4.

⁵See the CRA paper, page 4.

⁶The reallocation of commercial spectrum rights at expiry - Background information

⁷Handbook of Telecommunications Economics, Chapter 14 Spectrum Auctions, page 631