A.1This Annex discusses the future treatment of certain secondary issues and parameters relating to the Oftel monitoring formula not dealt with in Section V of the main text.
A.2 A difficulty with basing the required rate of return on actual returns is that they may include some excess or deficient returns. Accordingly, Oftel will also review analytic approaches to estimating service providers' cost of capital.
A.3 A number of methods exist for estimating a company's cost of capital. The most widely used is the Capital Asset Pricing Model (CAPM). According to the CAPM, the cost of equity is the sum of the risk-free rate and a risk premium which depends on the riskiness of the firm relative to the riskiness of the stock market as a whole. The underlying idea is that investors can diversify risks which are not correlated with the stock market as a whole by buying shares in a large number of companies. Risks which are correlated with the market cannot be diversified away and so form a component of the cost of equity.
A.4 The basic CAPM may be expressed by the following formula:
rE = rF + B(rM - rF )
where rE is the required return on equity, rF is the risk free rate of interest, rM is the average rate of return on the stock market as a whole and B, the firm's "beta coefficient", measures the correlation between the firm's share prices and those on the market as a whole. A firm whose (correlated) returns fluctuate more widely than those of the market as a whole will have a beta greater than one, while a company whose returns are less volatile will have a beta less than one.
A.5 All the parameters of the CAPM may be estimated empirically, although there is no universal agreement on the appropriate methodology. The recent Oftel Consultative Document "Pricing of Telecommunication Services from 1997" suggests that the nominal risk free rate currently lies between 7.1% and 8.8% net of taxes (depending, inter alia, on the marginal income tax rate which investors are assumed to face), whilst the equity risk premium (rM - rF) lies between 4% and 6%.
A.6 The most difficult parameter to estimate accurately in the present case is beta. Estimation relies on share price data. However, many service providers are not quoted companies or are part of larger groups whose share prices reflect the performance of a number of diverse activities. One possibility in these circumstances would be to use a sector or company average beta. Beta for the telecommunications sector as a whole has been estimated at 1.08, although this must have been heavily influenced by BT's beta which is substantially lower. Estimates of the cost of capital published in connection with the flotation of Orange and the proposed flotation (now lapsed) of People's Phone suggested a beta of 1.3 and costs of capital of 14 - 15%. There may therefore be grounds for thinking that an appropriate beta for service providers would be somewhat greater than 1.08. Service provision is a competitive activity and this is likely to increase volatility and hence beta. It is also often regarded as a risky business although some of this greater risk may be diversifiable. In addition, the generally low fixed costs in service provision tend to argue for a somewhat lower beta. Oftel provisionally concludes that an equity beta between 1 and 1.4 is appropriate.
A.7 In order to arrive at an estimate of the cost of capital, the cost of equity must be combined with the cost of debt, taking account of the tax advantages of the latter. This requires knowledge of gearing ratios which again may be difficult to obtain for non-quoted companies. In the circumstances, the simplest assumption to make is that of 100% equity finance.
Oftel proposes to estimate service providers' cost of capital using the CAPM and the values shown above, and would welcome comments on this approach.
A.8 Data problems appear to preclude use of the Dividend Growth Model to estimate the cost of equity. In addition, Oftel sees little scope for estimating individual rates of return of different service providers.
A.9 The May 1994 measures required that airtime profits should be calculated net of bonuses in respect of subscribers above a base level of 50,000. This was because it was felt that airtime bonuses disproportionately favoured TSPs who tended to be significantly larger than independent service providers. It is not obvious that the networks' costs per customer fall as the number of subscribers per service provider rises. Thus there do not appear to be any strong cost-based grounds for the offering of such bonuses. Arguments that bonuses should only be excluded if there is evidence of cross-subsidy neglect this point. However, it may be that 50,000 is no longer the appropriate base figure, because of growth in both tied and independent service providers.
Oftel believes it is appropriate to continue to exclude airtime bonuses which benefit large service providers and proposes to do so. However, it would welcome comment on the level above which they should be excluded from the calculation of cross-subsidy.
A.10 Another concern is the correct measure of cost to use. Fully-allocated cost ("FAC") is the benchmark currently used by Oftel, whereas economic theory would suggest that cross-subsidy exists only where prices are below incremental cost. However, where there are significant barriers to entry and exit, a price-floor of incremental cost may give an established firm with market power the ability to exclude rivals. This would be detrimental to competition. As we are concerned with the problem of long-run loss-making by TSPs, any measure based on incremental costs should reflect long-run incremental costs, and these are likely to be close to FAC.
A.11 Where, therefore, predation is prevented by prohibition of "unfair cross-subsidy", fully allocated cost is a good and readily practicable proxy for incremental costs. In these circumstances Oftel believes that it should continue to define cross-subsidy with respect to fully allocated costs.
Oftel believes that it is appropriate to use fully-allocated costs in order to allow independent service providers to compete on a level playing field with the TSPs.
A.12 At present, cross-subsidy is measured according to each service provider's own costs and revenues. It is arguable that it would be more appropriate to use the revenues obtained and costs incurred by one of representative efficiency. In a competitive market, a less efficient service provider would make losses until it either improved its efficiency or exited the market. Such losses would not be regarded as unfair cross-subsidy.
A.13 In practice it would be very difficult to construct cost and revenue figures for a TSP of representative efficiency. Therefore the appropriate response is likely to be to allow the Director General to exercise some discretion in deciding where an unfair cross-subsidy exists, taking account of whether any failure to comply with the formula was the result of inefficiency. This would represent no more than a continuation of current practice and could also be applied to other issues such as alleged "seasonality" in results.
Oftel does not intend to define costs as those of an efficient operator, but will take these costs into account when considering whether a failure to comply represents unfair cross-subsidy.
A.14 It has been argued that some costs should be spread over a greater number of years than is currently allowed for in the monitoring formula. For example, there may be costs associated with start-up operations which are incurred at the outset but which bring in benefits which are reflected in profits earned in later periods. The question then is whether these costs should be amortised over a number of quarters.
A.15 Oftel is not attracted by this idea. The TSPs to which the formula applies are established operations and so it is hard to advance this argument in terms of start-up costs. Indeed Oftel is concerned that such a measure could be exploited by the established operators by setting up ostensibly new service providers for this purpose. The treatment of MOTO and Orange, in respect of whom this argument might be stronger, is discussed above.