5.1 The formula requires data on a number of parameters in order to be made operational. These will be discussed in turn since each raises difficult measurement issues.
5.2 Monthly airtime profit is measured from data supplied by the TSPs in their quarterly monitoring returns. The value of a new subscriber is estimated as the ratio of the total profit on airtime sales earned in a quarter to the average subscriber base in that quarter. Thus there is an implicit assumption that new subscribers are expected to generate the same level of airtime profits as the "installed base" of subscribers.
5.3 It has been suggested that this implicit assumption is not valid. In particular, it has been argued that the increasing proportion of new customers on residential, as opposed to business, tariffs means that the formula as currently applied overstates the economic value of new subscribers, because residential users will typically generate lower airtime profits than business users. A similar argument could be applied to cases where the charging schemes for new and existing subscribers are different. It is argued that for these reasons cases of cross-subsidy could be overlooked.
5.4 The measured net cost of subscriber acquisitions is indeed the cost of those gross new acquisitions made in the relevant quarter. This will not be distorted by differences between new and existing subscribers. However, it has been pointed out that acquisition costs may differ according to the type of subscriber. For example, the net cost of acquiring subscribers to digital service is higher than for analogue subscribers. If the percentage of digital connections is increasing yet average earnings reflect a large base of analogue subscribers, then there could be a distortion in the measurement of cross-subsidy.
5.5 Oftel agrees with those who have argued that the current method of applying the monitoring formula may be misleading where the subscriber mix is changing rapidly. A decision then has to be taken on what is the best practical method of allowing for the changing subscriber mix. Oftel has considered a number of alternatives.
5.6 The challenge is to estimate as accurately as possible the airtime profits to be expected from each quarter's new subscribers. One possibility would be to use an explicit forecast of the expected profits. However, it would be difficult to make forecasts with sufficient accuracy and there is a danger that incentives could be created to manipulate the forecasts. For both these reasons the use of forecasts would require provision for retrospective adjustment. This could add considerably to the complexity of the monitoring process.
5.7 Moreover, any forecasts are likely to rely heavily on past data. There may therefore be little gain over the use of past data in the formula itself. However, this past data should be as relevant as possible to the new subscribers currently joining the networks.
5.8 One way of increasing the relevance of the airtime profits data would be to reduce the age of the subscriber base on which the calculation is made. For example, the economic value of new subscribers could be calculated from the average airtime profits of customers in place for less than a year or perhaps six months. This would have the effect of stripping out progressively the earlier acquisitions on higher tariffs and would allow both for tariff developments and changes in the customer mix.
5.9 Oftel believes that this is a promising idea. However, the relevance of the airtime profits of customers in place for less than six months or a year to the profits to be expected over the lifetime of an average subscriber may be reduced by two factors. Firstly, usage may not be constant but may, for example, increase substantially over this lifetime. If so, this method could underestimate the economic value of a new subscriber. Secondly, churn may be particularly high within the first few quarters from connection. Customers who join then rapidly leave a network may have atypical usage patterns and thus, again, may not be a very precise guide to average subscribers.
5.10 These drawbacks could be overcome by increasing the threshold for inclusion in the subscriber base for monitoring purposes to something closer to the average subscriber life, currently assumed to be 35 months. However, this begins to erode the main advantage of this approach, that it quickly reflects changes in the subscriber mix.
5.11 Nonetheless, Oftel believes that this is an idea worth pursuing and would be interested in views on the ability of service providers to provide data in the form required, particularly if subscribers are to be progressively removed from the base for monitoring as their time on the network reaches the prescribed threshold. If this were thought too complex, monitoring could be confined to subscribers joining after a particular date (January 1994, for example). The start date would then have to be periodically updated so that the monitoring returns would not decline in relevance over time.
5.12 An alternative would be to start from the proposition that airtime profits on new subscribers differ from the average for subscribers as a whole because of the increasing proportion of subscribers on residential tariffs and taking digital service. The main differences in tariffs and usage patterns are between business and residential subscribers and between digital and analogue service. Thus one approach would be to require separate monitoring returns for business and residential subscribers and, perhaps, within each for digital and analogue subscribers.
5.13 Oftel believes that this is also a potentially useful approach. However, it has the disadvantage that it does not directly address the question of changing tariffs as, within each group, all subscribers would be averaged together. The extent to which this is a problem depends on how far tariff changes affect existing as well as new subscribers. It could also increase the complexity of monitoring as up to four separate returns would be required from each service provider and again it is unclear whether service providers have the data in the form required. However, if this last obstacle were overcome, consideration could be given to use of a finer disaggregation by tariff type. This would better address the issue of changing tariffs although at the cost of increased complexity.
5.14 All the above options for dealing with changes in the subscriber mix would entail more work for service providers in producing the monitoring returns. It is clearly essential that any proposal adopted should be practicable and Oftel has no wish to impose undue burdens on service providers. In deciding which proposal, if any, to adopt in the light of responses to the Consultation, Oftel will take into account any extra monitoring effort required. However, Oftel believes that it is essential that the monitoring formula is relevant to the subscribers being acquired by service providers now. Therefore, Oftel believes it should be prepared to require the necessary monitoring systems to be put in place if responses to this Consultation suggest that this is necessary and desirable.
Oftel believes that it would be desirable to modify the monitoring formula the better to reflect differences between new and existing subscribers. This could be done on the basis of the length of time that customers have been in place, or by tariff type. Comments are invited on the feasibility and desirability of these and any other alternatives.
5.15 The average subscriber life, n, is a key determinant of the economic value of a subscriber. Various lives were put to Oftel during the investigation which led up to the May 1994 Directions. Most tended to be in the range 30 to 40 months. This is clearly consistent with the value of 35 months chosen.
5.16 The figure of 35 months was calculated arithmetically from empirical evidence on the percentage gross reduction in the number of subscribers suffered on average by service providers per month, known as the rate of "churn". Oftel estimated this at 2% per month. Then, a service provider would expect to have lost one half of any given initial stock of new subscribers after a period of 35 months. This period was taken to be the expected average subscriber life.
5.17 Such data are likely to change over time as market conditions alter, and new calculations are required.
Oftel proposes to re-estimate average subscriber lives, based on recent evidence and (possibly) broken down by tariff type. Comments are sought on this proposal.
5.18 It has been submitted to Oftel at various times that particular service providers' subscribers have lives which are longer than average. These differences may reflect variations in the subscriber mix referred to above. However, the performance of individual calculations for each service provider would impose considerable burdens on Oftel.
Oftel does not propose to compute subscriber lives individually for service providers.
5.19 The other key determinant of economic value is the required rate of return, I. In the May 1994 measures this was set at 2% per month or, allowing for compounding, 27% per annum in nominal terms.
5.20 This figure was based on actual returns then being earned by Cellnet and Vodafone, although it was set at a level some way below them. It was felt that a reasonable profit for service providers must have regard to the profitability of the industry as a whole.
5.21 It was expected that industry profits would decline over time, partly because of increasing competition from new entrants. For this reason, the May 1994 statement made it clear that Oftel would review the required rate of return in the light of experience. As part of the current exercise Oftel will collect financial data from a range of service providers.
Oftel will seek to collect data on rates of return currently being earned by network operators and service providers.
5.22 Connected with the issue of the required rate of return is the issue of how best to estimate service providers' cost of capital. For further discussion of this question see Annex A, which also discusses certain other secondary issues and parameters relating to the Oftel formula.
5.23 Oftel has become concerned that some behaviour on the part of the network operators may be anti-competitive but does not fall within the scope of the formula. An example of this is the making of incentive payments direct to dealers for connections to the network. The Oftel formula is in principle intended to cover all the financial flows arising from tied service provision. When it was introduced incentive payments and connection bonuses, even if financed in part by network operators, were in the main channelled to dealers via service providers. It therefore appeared sufficient to apply the monitoring arrangements only to TSPs themselves. However, incentive payments made direct to a dealer by a network operator, and which are contingent on the dealer connecting the customer through an airtime contract with that operator's TSP, are an equally important element in the financing of tied service provision. The monitoring system will not therefore achieve its objectives if they are excluded from it.
5.24 Further, the provisions in the mobile licences prohibiting undue preference and undue discrimination apply to all such incentive payments, A direct relationship between a network and a dealer is not, of itself, necessarily anti-competitive, but it has the potential to be so if it effectively ties the retailer to a particular TSP. For example, if the network provides financial support to the retailer on the basis, whether explicit or implicit, that the retailer gives preference to a TSP in arranging for a customer to sign an airtime contract then the network's actions will distort competition between service providers. In addition, given the size of the TSPs' market share a distortion may arise because only TSPs are perceived as able to offer a nationwide service to the high street multiples that also operate nationwide. Oftel is currently investigating whether some of the incentive payments made direct by network operators may be discriminatory. It is proposed to continue these investigations parallel with this review.
Oftel intends to ensure that discount and other incentive payments made direct by network operators to dealers are brought within the formula and the monitoring arrangements, and would welcome comments on how best to integrate them into the quarterly returns.