Review and Update of the Spectrum Pricing Models

Executive Summary

As part of the preparations for the introduction of spectrum pricing, the Radiocommunications Agency (RA) brought together an industry/government group, the Spectrum Pricing Preparatory Group (SPPG), to review and discuss the detailed proposals. The group met from December 1997 to April 1998. Similar discussions were held with the fixed links community via an ad hoc group of the Microwave Fixed Links and Satellite Committee (MFLSC).

Proposed prices for spectrum were based on the results of modelling work originally carried out by Smith and NERA over two years ago, published by the RA as the Study into the Use of Spectrum Pricing in April 1996. This report revisits that original work. It presents a review and update of the models, data and assumptions underlying the pricing of mobile radio and fixed link spectrum.

This review was commissioned by the RA as one of the inputs into the SPPG and the wider public consultation process. It took its cue largely from the discussions and issues raised within the SPPG meetings and the ad hoc MFLSC group. It covered a number of separate tasks which re-examined various aspects of the original analysis as well as extending it in other areas. Specifically, it covered:

review/update of the data in the original models for least cost alternatives for mobile radio and cellular services;

review of the pricing proposals for fixed link spectrum;

analysis of the sensitivity of the model results to the assumptions of discount rate and equipment lifetime;

review of the current availability of narrowband (linear modulation) systems for mobile radio;

alternative approaches to the pricing models for cellular spectrum.

This report presents the results from these various tasks in the order in which they were carried out over the 6 month period of the review.

Application of spectrum pricing to mobile radio

The applicability of spectrum pricing to all the various mobile radio services was reviewed in the light of market developments over the last couple of years (§2.1). The options available to PMR, CBS and PAMR users for relieving congestion were also reviewed (§2.2). No substantial changes to the original model assumptions were suggested by this review.

The main development has been in narrowband technology, particularly Linear Modulation technology (see §6) and TETRA. Now that TETRA equipment is commercially available, actual prices can be used in the revised models instead of the estimates necessary two years ago.

Review of marginal values

The prices for base stations, mobiles, other equipment and services were updated in line with current market information in the spectrum pricing models (§3). The updated spreadsheets are presented in Appendix A.

The differences in cost for PMR users between the various options (moving to a CBS, PAMR or cellular service) have changed in response to the revised prices. However, the option of moving to a PAMR service remains, as in the original study, the least cost option and therefore the basis for calculating the marginal values used to define spectrum prices (§3.1).

With revised prices for TETRA equipment based on actual market data, the marginal values for PAMR spectrum are much reduced compared to the original study (§3.2.3). However in setting spectrum prices, other issues have to be considered, not least that CBS, PMR, PAMR and cellular services are substitutes for one another, which argues strongly for setting a common average spectrum price (as the RA currently proposes to do).

A preliminary review of the original pricing models for cellular was carried out (§3.3). A more substantial remodeling exercise was also carried out later on in the review (see §7). The preliminary review suggested that the models and resulting marginal values for cellular spectrum should remain unchanged. It was noted that in terms of moving to new technology half-rate codecs were much less probable now than using dual-band equipment, though either was a much more costly option than using smaller cells to relieve congestion (§3.3.3).

Review of the fixed link spectrum tariff unit

The price model for fixed link spectrum was also updated as part of this review (§4.1). Reductions in the costs of cabling and in equipment costs produced marginal valuations which were about 10% lower than the values in the original study. This level of price variation, however, is well within exchange rate fluctuations which affect equipment prices and it was concluded that overall there was no strong justification for changing the spectrum prices currently proposed by the RA.

The assumptions behind the calculation of the spectrum tariff unit (STU) were also reviewed (§4.2). There is no meaningful economic argument for direct comparability between fixed-link spectrum prices and mobile spectrum prices. But the proposed STU values do come out at roughly the same level. Arguably, this has the merit of keeping the pricing regime relatively simple and understandable.

Sensitivity analysis

The pricing models include an assumption about the lifetime of equipment and the discount factor to be applied to future capital expenditure (set at 10 years and 10% respectively in the original models) (§5.1). An analysis of the impact of changes in these assumptions on which alternative is ‘least cost’ and on marginal values was carried out.

For the PMR pricing model, the conclusion that PAMR is the least cost alternative (and should therefore be used as the basis of the marginal valuation for spectrum pricing) is very robust to changes in discount rate and/or lifetime (§5.2). Unrealistic discount rates (more than 23% over 10 years) or lifetimes (less than 5 years at 10% discount rate) would have to be assumed to change which alternative is ‘least cost’.

Changes in these assumptions do directly affect marginal values. However, the original assumptions were conservative. Moreover, the RA’s proposal is to phase in fees over three years to reach a level equivalent to roughly half the marginal valuations. To produce lower prices than those proposed, highly unrealistic assumptions about discount rates (less than 1% over 10 years) or lifetimes (more than 25 years at 10%) would have to be made (§5.3).

Developments in linear modulation technology

During the period of the review there were significant developments in linear modulation (LM) technology (§6.1). Specifically, ETSI standardised a narrowband specification and several manufacturers have started to make LM equipment widely and cheaply available.

Current prices for LM technology were plugged in to the PMR valuation model (§6.2). This suggested that for small system users (less than 25 mobiles), while there was already a price differential in favour of CBS systems, LM now offered a cost competitive alternative for those seeking to renew their systems. For larger systems the lowest cost alternative remains PAMR or cellular though LM falls within much the same cost range.

Overall, PAMR remains the least cost alternative to be used as the basis for marginal valuations. That said, the availability of cheap standard LM equipment is a significant development in the mobile industry. LM costs are now lower than other PMR systems and users of such systems already have a price incentive (which will be that much greater when spectrum pricing is introduced) to consider migrating to LM when renewing their systems (§6.2).

Alternative valuation models for cellular spectrum

In line with comments received by the RA over the last 2 years in response to the White Paper and the original study, as well as discussions with the SPPG, more detailed modelling of the use of smaller cells as the basis for valuing cellular spectrum was carried out (§7.1)

Two new models were developed. The first is representative of the approach of reducing cell sizes by changing transmitter powers or sectoring existing cells, for example. The second aims to capture the costs of overlaying a microcell network on top of the existing cell structure. The marginal value of spectrum was calculated by taking the difference in the cost of expanding network capacity between using more spectrum and using smaller cells, divided by the amount of additional spectrum used (§7.3).

Again, the results are sensitive to the assumptions made. Every attempt was made to use baseline data supplied by, or at least sanctioned by, the operators themselves. The marginal values resulting from the lower cost of the two models was £1.72m for 2*1MHz of national spectrum per year, slightly higher than the £1.625m produced by the original model (§7.6). The results from the other model were hugely higher (at £15.8m).

The RA currently propose to introduce spectrum prices for cellular operators which will increase over a three year period to a level which is roughly half the original NERA/Smith valuation (of £1.625m), at which time prices will be reviewed. It was concluded that these proposed prices are firmly at the low end of the marginal values that could be derived. Under most assumptions and models, estimates of the marginal value of cellular spectrum would be significantly higher.

Conclusions from the review

Estimating the marginal value of spectrum is not an exact science. The review has tested the robustness of the valuations which underpin the proposed spectrum prices in a number of different ways and the models have been updated to reflect current market data. Only by making extreme assumptions about, for example, discount rates can prices lower than those proposed by the RA be derived. Industry and government can have confidence that the proposed spectrum prices are fair and reasonable.

Download:

The Word 6.0 version of the Whole document (excluding Appendix A) (252 KB)

Excel 5.0 version of Appendix A-1 (31 KB)

Excel 5.0 version of Appendix A-2 (55 KB)

Excel 5.0 version of Appendix A-3 (35 KB)

Excel 5.0 version of Appendix A-4 (18 KB)

or download the self-extracting zipped versions of

The Whole document (excluding Appendix A) in Word 6.0 (97 KB) and;

Appendix A in Excel 5.0 (50 KB)

Top