Review

Radio and television broadcasting

Approximate the true market value of the spectrum rights

Fair return to the crown

Cellular telecommunications

Multipoint distribution services (MDS)

Other issues: Averaging of base auction prices

Conclusion

The price setting formulae are to be applied to four classes of use: broadcast radio and TV, cellular telecommunications and Multipoint Distribution Services (generally used for television transmission). These are considered below in turn.

Radio and television broadcasting

At present, the radio and television broadcasting sector is relatively mature and stable (compared, say, to cellular telecommunications).³ Revenue growth is derived from growth in advertising revenue in each of radio and TV since licensing.

Historical industry revenues were obtained from figures published by CAANZ. Covec demonstrate a high degree of correlation⁴ between growth in advertising spend and population growth and the historical revenue figures are therefore projected forward on the basis of Statistics New Zealand's medium population growth forecasts. Non-national licence growth factors are adjusted to take account of population growth of the relevant locality compared to the national growth rate. Note that simplifying assumptions have been used to adjust for regional population growth rates:

• rates are based solely on Statistics New Zealand's projections of population growth and do not incorporate historical population growth; and
• the national z factor is adjusted for the difference between regional and national population growth, rather than being recalculated on a regional basis. Again, there is a trade-off here between simplicity and accuracy: in this case a single national z factor is calculated, then adjusted for regional variances instead of needing to calculate a new z factor for each.⁵

As the industry is mature, relatively stable and there is publicly available data, it is reasonably suited to the formula-setting approach.

The formulae are mathematically consistent with the approach described in Covec's paper.

The approach is considered below against MED's objectives.

Approximate the true market value of the spectrum rights

There will always be limitations on the extent to which a formula-based approach can approximate a market sale. Within these constraints, changes in advertising revenue (and basing projections of advertising revenue on population growth) are a simplistic but not unreasonable approach to take. Note that, in addition to the general assumptions outlined above, the following assumptions are implicit in the approach:

• Advertising revenues are the dominant source of revenue (and profit) for broadcasters;
• The econometric forecasting model provides an accurate forecast for revenues and corresponds closely to rights holders' own expectations.

The price setting formulae should be simple to calculate and administer

The price setting formulae are simple, using few variables. The renewal prices can be easily calculated in a simple spreadsheet and the formulae are simple to revise for updated information. All variables for the price setting formulae are derived from publicly available information.

Fair return to the crown

A high degree of caution needs to be used when considering whether a "fair" price would be achieved. As noted above, there are a large number of assumptions implicit in the formula-based approach, which affects the accuracy with which the formula may reflect market reality.

Assuming that the formula proxies changes in the value of the spectrum (see above), the formula should produce a reasonable return to the Crown. It is not possible to say whether an auction would produce a higher or lower return, but the formula approach does provide a higher certainty for the Crown as to its return from the spectrum.

Please also note that the discount rate used to bring the licence payment forward five years (Covec use 11% to illustrate calculations in their paper) will also affect the actual cash sums raised (the higher the discount rate, the lower the cash return to the Crown).

Cellular telecommunications

When licensed, the cellular market was in the earliest stages of development (it was an immature market). Technological changes and new business models might well have completely invalidated the assumptions on which licensees originally made their bids. The value of the 2G spectrum that is subject to renewal is also likely to be closely linked to the value of 3G spectrum (which is not subject to renewal at the same time as the 2G spectrum).

The industry does not meet the basic tests outlined above for the applicability of a formula-based approach. Any formula-based approach is therefore likely to have a much lower level of certainty as to whether it approximates changes to the spectrum's market value.

Acknowledging these difficulties, Covec propose that MED either:

Approach 1 retains the benefits of simplicity and transparency. However, there is likely to be a relatively low level of certainty as to whether the formula's result approximates market value (and hence produces a fair return to the Crown). This uncertainty arises from:

• Whether the base price remains a valid starting point, given subsequent changes in the market (auction bids would have been made on assumptions that have likely proven highly inaccurate);
• Whether it is possible to calculate a valid growth factor for the relevant period (40 years), given that revenues have not shown a reasonably stable growth path.

Covec propose a growth factor based on a logarithmic projection of current revenues over the next 30 years. Whilst this produces a simple and transparent result, given the issues outlined above, we cannot say whether this would produce a result that approximates the change in market value of the spectrum since the original licensing.

Approach 2, essentially undertaking a full valuation of the business in order to value the spectrum, would be likely to generate a better approximation of the true value of the spectrum. It would also be a complex and time-consuming process (even with only two licensees), and would not have the advantages of simplicity and transparency. The approach would also mean that cellular was treated differently to the rest of the spectrum users (although given the very different nature of the businesses, this would not be unreasonable).

This approach has been successfully used in the UK for re-licensing spectrum for radio broadcasters.

Multipoint distribution services (MDS)

MDS also suffers from the problem of not being entirely suited to the formula-based approach, as Covec were unable to obtain data on revenues derived from use of spectrum for MDS.

Instead, a producer price index output series for the communications sector (provided by Statistics New Zealand) was used as a proxy. This approach retains the objectives of simplicity and transparency. It does however have a lower level of certainty as to whether the results of the formula are representative of the actual market value of the spectrum as a further major assumption has been introduced: whether the PPI output index is a valid proxy for revenues. Nevertheless, in the absence of revenue data, the approach is sensible.

Other issues: Averaging of base auction prices

Table 3 of the Covec report summarises the advantages and disadvantages of averaging base auction prices. We agree with the Covec position that average base auction prices can have perceived equitable and inequitable effects. While applying a similar price to similar assets may potentially appear more even handed, it is clear that from the initial auction results that different prices did result for similar assets, caused by the heterogeneous demands of bidders and the auction design.

Averaging and other adjustments to base prices has potential to cause greater distortion to the market for spectrum, ranging from welfare losses, potential market exit, and a reduction in income from spectrum licences.

The decision on whether to average base prices is outside the scope of the Covec report and of this Peer Review. However, we encourage the MED to consider adjustments to base prices only when correcting for perceived market abnormalities stemming from the original auction designs.

Conclusion

Our conclusions may be summarised as:

3 The introduction of digital television may have some impact (although will not affect the number of people that can be targeted by advertisers).

4 Note that we have not evaluated or audited Covec's econometric or statistical calculations.

5 Also, if the national z factor was zero, the formula would not be able to adjust for regional differences (in the event that e.g. positive and negative regional growth rates netting out to a zero national growth rate).

6 This creates an assumption isolating the value of 2G spectrum from that of 3G spectrum. Also, data on industry revenues is only available since 1999.