Home > Consultations > Consultation Documents > Mobile call termination > Statement > Chapter 6

6.1 As set out in Chapter 5, consistent with the December consultation, Ofcom has reached the conclusion that, given the finding of SMP for each operator in the relevant market, direct controls (through a charge control) should be imposed on the charges for terminating mobile voice calls on the 2G mobile networks of Vodafone, O2, Orange and T-Mobile. This chapter sets out in more detail the proposed level and structure for these controls.

6.2 In order to impose a charge control it is necessary to identify:

1. the 'efficient charge' level that these charges should be brought down to by the end of the control period; and
   
   
2. how these charges should be brought down to the level of this efficient charge.

The efficient charge level

6.3 Ofcom's decision regarding regulatory remedies, including the charge control, reflects considerations of economic efficiency and the intention to maximise benefits to end-users. Ofcom refers to the level of wholesale termination charges which it believes best achieves these objectives as the 'efficient charge' level.

6.4 As proposed in the May and December consultations, Ofcom has set the target charge on the basis of long run incremental cost (LRIC) plus a mark-up for common costs, based on the equal proportionate mark-up (EPMU) approach, and a network externality surcharge.

LRIC

6.5 Ofcom is of the view that the most appropriate and economically efficient basis for regulatory charge controls is forward-looking LRIC. The LRIC of voice termination is the additional cost an MNO incurs to provide termination. This can also be seen as the cost that the firm would avoid if it decided not to provide voice termination, taking a long-run perspective. LRIC based charges correspond more closely to the charges that would prevail in an effectively competitive market than accounting-based measures of cost. It is a fundamental goal of price regulation to mimic the effects of a competitive market and this consideration underpins the use of LRIC. Further details and references regarding the use of LRIC in a regulatory context are provided in Annex C.

6.6 More generally, further details concerning the implementation of LRIC and the mark-up for common costs, as well as a discussion of the responses to the December consultation regarding the calculation of the LRIC+ efficient charge level can be found in Annex C.

Economic depreciation

6.7 As stated in Annex E of the May consultation and paragraph 6.7 of the December consultation, the depreciation approach selected by Ofcom for the LRIC model is economic depreciation (for further details of the conceptual underpinnings, see Calls to mobile: economic depreciation, September 2001^(-22-). For a discussion of the cost path over time using economic depreciation in the LRIC model for key assets, see Additional Information Concerning Oftel's LRIC Model, 12 February 2002^(-23-)). This matches the cost of equipment to its actual and forecast usage over the long term. As a consequence, there is relatively little depreciation in years where utilisation is low and relatively high depreciation in years of full, or almost full, equipment utilisation. By contrast, the usual accounting method takes the actual price paid for equipment (or its replacement cost) and divides by the expected equipment life to reach a depreciation charge for the year (thus adopting a straight-line depreciation profile). The timing of cost recovery under economic depreciation varies from that under such accounting depreciation. Between 2001 and 2006 the use of economic depreciation results in a higher per minute cost of terminating calls whilst in years prior to 2001, economic depreciation would have resulted in lower costs compared to an equivalent calculation based on accounting straight-line depreciation.

EPMUs for recovery of common costs

6.8 Ofcom considers it appropriate for regulated services to contribute towards the recovery of relevant common costs through a mark-up in addition to LRIC to allow for full cost recovery. Ofcom believes that it is appropriate for these costs to be recovered by an EPMU. In the May and December consultations, it was considered whether the efficient charge level should be set in accordance with Ramsey principles, that is, whether the mark-up for the recovery of common costs should be set on the basis of demand conditions. In theory, Ramsey prices minimise the loss in economic efficiency introduced by the departure from marginal cost pricing due to the presence of common costs. However, Ofcom has concluded that the derivation of Ramsey prices, or more generally of welfare-optimal prices, raises complex conceptual and practical issues which do not allow for sufficiently reliable optimal prices to be estimated. Ofcom believes that EPMU achieves a more appropriate balance between practicality and efficiency than the Ramsey methodology. These issues are discussed in detail in Annex K of the December consultation and paragraphs 5.19-5.33 of Chapter 5 in the May consultation.

Responses to the December consultation

6.9 T-Mobile (paragraphs II.10-17 of its response) disagrees with a number of the reasons on the basis of which Ofcom has decided not to employ a Ramsey-methodology. The specific points raised by T-Mobile are summarised and listed below.

1. T-Mobile considers that Ofcom's argument that the retail market is imperfectly competitive is contradicted by the empirical evidence and by Ofcom's own conclusion in the review of the retail mobile market that this market is effectively competitive.
   
   
2. T-Mobile considers that the MNOs' overall pricing structure is not Ramsey-based, but even if it was T Mobile argues that it would still be welfare enhancing to set those prices that were being regulated on the basis of broad relative elasticities rather than to ignore significant differences in these elasticities completely.
   
   
3. T-Mobile rejects Ofcom's claims that the MNOs' ability to price discriminate and offer multi-part tariffs enables them to recover common costs from infra-marginal subscribers and limits the required mark-up on termination. T Mobile claims that if MNOs had this ability, they would be using it now.
   
   
4. T-Mobile argues that there cannot be more practical difficulties in identifying an estimate of Ramsey-based termination charges than in estimating EPMU, since Ofcom's own consultant has developed a model that estimates Ramsey prices. It also adds that, given the large sums at stake in this regulation, there can be no justification for basing mark-ups on implied assumptions about the relative super-elasticities of termination and origination that are outside the range of the empirical estimates.
   
   
5. T-Mobile contests Ofcom's reliance on the past performance of the mobile market to justify cutting rates. It also notes that since the 24 July 2003 cut in termination charges, there are indications that market growth is slowing down at a penetration rate significantly below that of some other European countries.
   
   
6. T-Mobile suggests there are inconsistencies in Ofcom's justification for not setting Ramsey-based termination charges. T-Mobile claims that Ofcom rejects the models proposed by Vodafone, Orange, O2 and T-Mobile on the grounds that these are over-simplified, but then adopts EPMU, which is the most simplified approach of them all. Moreover, T-Mobile argues that Ofcom appears content to use the Ramsey models to help estimate the externality surcharge.
   
   
7. T-Mobile argues that Ofcom does not calculate the EPMU for common costs correctly because it allocates administration and customer acquisition, retention and service (CARS) costs, which represent the vast bulk of the common costs, to retail services.

6.10 Ofcom has addressed the first of T-Mobile's points in paragraphs 4.38-4.44 and in K.9-K.15 of the December consultation. In summary, Ofcom does not consider that a finding of 'no SMP' in the retail access and outgoing calls market is equivalent to a finding that MNOs will fully pass-through all excess profits earned in termination markets, and therefore set Ramsey-based prices in the retail market. Ofcom therefore continues to believe that there is a strong risk that setting Ramsey-based termination charges would not maximise social welfare.

6.11 In response to point (b), as explained previously (see paragraphs 5.19-5.25 of the May consultation and K.34-K.35 of the December consultation), Ramsey-based prices are a set of prices for a group of services which maximise economic efficiency, given that the presence of common costs across these services does not allow the adoption of marginal cost pricing (since the firm(s) would not break even if it priced all services at marginal cost). Hence, Ramsey-based prices allow for the recovery of common costs across all the services to which they are common and are based on the relative demand conditions for all these services. If some services are excluded, along with their marginal costs, they would be assumed not to contribute to the recovery of these costs and this would generate upwardly biased estimates of the mark-ups for the services included in the model. Such a set of prices would, thus, be sub-optimal because of this error of omission and would not maximise welfare. Hence, even if the termination charge was set on the basis of the Ramsey principle by Ofcom, the overall set of mobile prices would be efficient only if MNOs set Ramsey-based prices for the remaining services. Ofcom is of the view that MNOs do not have the incentive to set Ramsey-based retail prices. The reasoning behind this view has been discussed in paragraphs 5.26-5.30 of the May consultation and K.9-K.15 of the December consultation.

6.12 Regarding the ability to price discriminate, Ofcom has previously suggested (see paragraphs K.18-K.20 of the December consultation) that linear prices are not the most efficient set of prices that could be achieved in the mobile markets. This claim is supported by economic theory and the fact that MNOs do employ multi-part tariffs. However, Ofcom has never maintained that perfect price discrimination is necessarily feasible (see Ofcom's analysis of the externality mark-up in Annex G of the December consultation), but simply that even if full price discrimination is not possible, non-linear pricing is. Simple linear Ramsey pricing models (submitted by or on behalf of the MNOs) fail to take this into account and raise questions as to the claimed efficiency properties of these models.

6.13 Ofcom has previously rejected the claim in point (d) and in so doing, exposed the conceptual and practical reasons why it considers that Ramsey pricing is not the appropriate methodology for setting termination charges. This discussion is set out in Annex K of the December consultation. Ofcom has also explained why it believes that EPMU is the appropriate methodology to be used in this case.

6.14 Regarding T-Mobile's claim that Ofcom has relied on the past performance of the mobile market, in the December consultation (paragraph K.46) Ofcom stated that

"evidence from the history of the mobile market does not support the claim that EPMU represents inappropriate regulation. Since 1998 termination charges (for Vodafone and O2) have been regulated on the basis of Fully Allocated Costs (plus an externality mark-up), which is very close to setting charges on LRIC plus EMPU, and the mobile market has thrived (i.e. penetration rate and level of usage have increased dramatically)".

6.15 However, Ofcom has never relied on past performance of the mobile market to calculate the target charge, and thus set the charge controls, but has based it on what Ofcom considers to be sound economic principles and careful estimates of costs in current conditions.

6.16 In response to point (f), Ofcom does not consider its reasoning is inconsistent. Ofcom has previously stated that it is not aware of any model of efficient pricing which it believes is sufficiently reliable to develop 'optimal' prices for fixed-to-mobile calls and mobile termination charges. The weaknesses of the currently available models primarily relate to deficiencies in capturing all relevant market features and the extensive informational requirements that underpin them, as well as in their sensitivity to changes in this information. This generates doubts on the reliability and robustness of the results derived from these models. In this context, Ofcom considers that EPMU, as a basis on which to recover common costs, strikes a reasonable balance between practicality and efficiency. Ofcom's views on this issue are presented in further detail in Annex K of the December consultation (see in particular paragraphs K.42-K.43). In relation to the calculation of the externality surcharge, Ofcom is aware of the difficulty of robust quantification. It has used the best information available, which includes the use of Ramsey-based models, to derive a range of estimates for the surcharge. But Ofcom continues to recognise that all of the models used have deficiencies, which it has taken into account in its interpretation of the estimates and in making its judgement of a reasonable surcharge.

6.17 T-Mobile's final point (g) has also been raised by T-Mobile in another part of its response and is dealt with in detail in paragraphs C.101-C.105 of this document.

6.18 Also in response to the December consultation, Vodafone (paragraph 1.58 of its response) claims that Ofcom's arguments (in paragraphs K.42-K.44 of the December consultation) for not relying on the elasticity estimates provided by the MNOs to set termination charges merely shows that there is a range of uncertainty around these estimates, and that this is not enough to maintain that EMPU is more efficient than the Ramsey methodology.

6.19 Ofcom has not maintained that the EPMU methodology is theoretically more efficient than Ramsey pricing principles. Ofcom considers that, given the limited size of the common costs and the difficulties of setting efficient mark-ups, the use of an EPMU for common costs and a mark-up for the un-internalised network externality achieves a more appropriate balance between practicality and efficiency than the Ramsey methodology. Annex K of the December consultation sets out further details of Ofcom's view on these issues.

Network externality surcharge

6.20 In the May and December consultations, Ofcom proposed that it would be appropriate to allow MNOs to add an additional mark-up on cost when setting charges for mobile termination services. This mark-up (or surcharge) was designed to ensure that MNOs account for the external benefits that callers to and from mobile telephones receive from the addition of new subscribers to the network, and the maintenance of existing subscribers on the network. An outline of the approach in the December consultation, responses to the December consultation, and Ofcom's comments and conclusions can be found in Annex D.

The structure of the charge controls

6.21 The previous section addressed how the level of the efficient charge should be derived. This section describes Ofcom's approach for reducing current termination charges to this level and the structure of the control, specifically:

• the control periods;
• the treatment of calls from fixed networks and off-net mobile-to-mobile calls;
• the calculation of compliance with the control;
• the treatment of ported numbers; and
• the treatment of combined 900/1800MHz vs 1800MHz operators.

The control periods

6.22 Ofcom considers that the charge control regime should last until 31 March 2006, as proposed in the May and December consultations. Ofcom does not currently believe it would be appropriate to extend the period of regulation past this date without undertaking a subsequent market review.

6.23 The December consultation proposed an implementation date of 1 April 2004. Taking into consideration the publication date of this document, it is appropriate to revise the implementation date for the charge control proposed.

6.24 In considering the appropriate date from which the charge control should apply, Ofcom notes that a major motivation for a reduction in termination charges is that consumers should benefit from lower retail prices for calls to mobiles. Ofcom therefore believes that it is desirable to ensure that changes in retail fixed-to-mobile prices, by BT and other fixed operators, can occur at approximately the same time as changes in mobile termination rates. This will allow consumers to benefit at the same time as mobile operators reduce their prices. An implementation date of 1 September 2004 allows a reasonable period of time to achieve this objective, since it gives mobile operators 28 days (consistent with the industry standard and previous regulation) to revise and notify BT and other operators of their new charges^(-24-), and a further 2 months for BT to effect retail price changes.

6.25 The charge control will therefore apply to the two periods:

• 1 September 2004 to 31 March 2005; and
• 1 April 2005 to 31 March 2006 (2005/06).

Responses to the December consultation

6.26 In section 3.4.3 of its response, Orange states its surprise that, as in the charge control proposed in September 2001, the current proposals still consider a charge control period until March 2006 as appropriate, despite the duration of the control period having fallen from four to two years. Orange concludes that due consideration has not been given to the appropriate period over which the controls should apply and over which termination charges should be reduced to the efficient charge level.

6.27 Contrary to Orange's conclusion, Ofcom has given careful consideration to the appropriate period of the charge control. As stated in paragraph H.2 of the December consultation and reiterated in paragraph 6.23 above, Ofcom does not believe it would be appropriate to extend the period of regulation past 31 March 2006 without undertaking a subsequent market review. Ofcom considers it important to review the nature and extent of regulation in this market earlier rather than later and therefore believes that a longer charge control period would not be appropriate.

6.28 T-Mobile expresses its belief, in paragraphs II.82-II.84 of its response, that if it chooses, it should be able to comply with the charge control by making a single change to its tariffs at the start of each control period. However, given BT's requirement for two months' notice prior to price changes, T-Mobile argues that this is not possible since the relevant call volume data would not be available at the point at which T-Mobile would need to make its pricing decisions.

6.29 As discussed in paragraph 6.24 above, Ofcom has concluded an implementation date of 1 September 2004 is appropriate. This date provides sufficient time to ensure that T-Mobile's practical concerns are addressed.

Calls from fixed networks and off-net calls

6.30 As proposed in the May and December consultations, Ofcom has decided to impose two separate sets of controls:

1. one on the charges for terminating voice calls from fixed phones on 2G networks; and
   
   
2. one on the charges for terminating off-net mobile-to-mobile voice calls on 2G networks.

6.31 Ofcom has set the level of these two controls to be the same. The LRIC of termination does not differ depending on where the call originates. The efficient charge level, and in particular the network externality surcharge, has been set primarily by reference to termination of fixed-to-mobile calls, but this is also the appropriate level to act as a safeguard control for the termination of off-net mobile-to-mobile calls (see the discussion on bilateral agreement in Chapter 5 of the December consultation for further details).

6.32 As in the previous proposals, Ofcom has decided that the control on each MNO should be placed only on the weighted average of the current three time-of-day charges (day, evening, and weekend) as MNOs should be free to vary these charges provided the overall charge control is met.

Responses to the December consultation

6.33 Orange raises two arguments against the imposition of two controls in section 3.2.1 of its response:

• Two controls is "somewhat otiose" given the interaction with the condition prohibiting undue discrimination; and
• It is not practically possible for Orange to comply with both charge control conditions, as it does not separately record mobile voice termination services by reference to the originating network.

6.34 Orange correctly asserts that the December consultation proposal for two charge control conditions is to address the possibility of an MNO otherwise being able to set a higher than average charge for one category of operator and lower than average charge for another category, whilst still meeting an average charge required by a single charge control. This was explained in paragraph 6.23 of the December consultation. However, Ofcom does not accept Orange's argument that the imposition of two controls is "somewhat otiose". As discussed in paragraph 6.23 of the December consultation, fixed-to-mobile and off-net mobile-to-mobile charges need not be identical, but should not permit the loading of the majority of charges onto one type of call. The two separate charge controls provide the necessary specificity to ensure such loading is avoided. Ofcom therefore considers two separate sets of charge controls necessary.

6.35 In terms of an inability to meet two separate controls, as set out in paragraph 6.24 of the December consultation, consent for compliance with a charge control to change from traffic-specific to total traffic volumes would be expected to be given for the period requested where an MNO is unable to identify the origin of the calls it terminates on its network. This would avoid the potential problem identified by Orange.

Compliance with the control

6.36 As proposed in Chapter 7 of the May consultation and Chapter 6 of the December consultation, Ofcom has decided to place a charge control on the average of the charges levied by each of the four MNOs (i.e. daytime, evening and weekend charges) for terminating voice calls on their 2G networks, weighted by the relative call volumes in the previous year. This charge control will bring the weighted average charge down to the efficient charge level by 2005/06. Ofcom's charge controls require that, during each period of the control, the average charge set by the regulated MNO (the Average Interconnection Charge or 'AIC') does not exceed the charge with which the operator is required to comply (the Target Average Charge or 'TAC').

6.37 Annex I of the December consultation and Annex H of the May consultation contain proposals on the specific form of the calculation of weighted average charges. As operators set different termination charges for different times of the day or week, a weighting mechanism must be used to determine the AIC. Ofcom also proposed that the TAC for each operator should be weighted on a consistent basis with the AIC, rather than the TAC being unweighted as per the previous charge control. This weighting of the TAC is designed to prevent unintended changes in traffic mix from distorting the impact of the charge control.

6.38 An adjustment mechanism was also proposed to ensure that the TAC is not distorted by over- or under-shooting in the previous year.

6.39 In the December consultation the conditions MC3, MC4, MD3 and MD4 specified the TAC for the second period (T₂) in terms of the Controlling Percentage^(-25-) ( RP₁ - X₂), the Adjusted Base Target Charge (B₂) and the Adjustment Percentage (A₂) for the second period, and the AIC (I₁) for the first period:

and p₁(i), v₁(i) are the time of day charges in the first period and time of day volume shares in the first period^(-26-), respectively.

6.40 Given the decision to set the TAC in the first period equal to the efficient charge level for 2005/06 (see paragraph 6.81 below), X₂ is effectively equal to RPI₁and the Controlling Percentage term is no longer required. Noting that the AIC is defined in terms of volume shares in the previous year, the remaining terms can be restated to give:

6.41 As noted in paragraph 6.87 below, Ofcom believes that it is simplest to specify the TAC for the final period in relation to the target average charge in the first period. This can be achieved by setting T₂ equal to T₁ multiplied by a Weights Adjustment Factor^(-27-) (W₂) to account for changes in time of day weights. As demonstrated above, this formulation is equivalent to that proposed in the December consultation with a control for the second period of RPI-RPI.

Responses to the December consultation

6.42 Vodafone and Orange both re-iterated their opposition to the proposed methodology:

The formula generates unjustifiable arbitrary gains / losses that cannot be objectively justified as being representative of a MNO's underlying costs. It is our understanding that the objective of the proposed price control is to drive a MNO's charges to the regulator's estimate of its reasonably incurred costs. The proposed TAC formula undermines this objective (Orange, p.30)

Vodafone repeats its view that a change to the manner in which the price cap is implemented is unnecessary. The impact of Oftel's proposals will be to reward 'overshooting' of the price cap by adjusting upwards the TAC in the following year. (Vodafone, paragraph 1.61)

6.43 Ofcom has explored in some detail the benefits of the proposed methodology in the previous consultations. Ofcom rejects Orange's interpretation that it creates arbitrary or unjustifiable gains or losses - it is designed specifically to address the arbitrary gains and losses that occur under the approach incorporated in the previous control. Orange's response does not provide an explanation of how, in the absence of an adjustment to the TAC, the arbitrary gains or losses under the previous approach would be accounted for. Indeed, under Orange's favoured approach, an MNO may find it could comply with the TAC in a completely illusory way, i.e. not via price changes, but solely through a change in traffic profile. Alternatively, if the weights move in the opposite direction (towards the daytime), an (unadjusted) charge control will necessitate extremely large price reductions. Thus, Ofcom considers the objective of reducing charges to an efficient level would be met in a more appropriate manner under the proposed approach.

6.44 Ofcom does not accept Vodafone's view of the effects of the proposed operation of the TAC formula with regards to 'overshooting'. As noted in paragraph A6.29 of the Review of the Charge Control on Calls to Mobiles, published on 26 September 2001, the adjustment factor here is intended to ensure that the TAC this year is not distorted by under- or over-shooting last year, i.e. that the target is calculated as if there had been neither under- nor over-shoot (see the expression for the Weights Adjustment Factor in paragraph 6.40 above). The adjustment factor does not account for adjustments that should be made to the target in order to compensate (either) customers, for an overshoot in the prior year, or the regulated operator, for an undershoot.

6.45 Ofcom also does not accept that the treatment between an overshoot and overshoot are unreasonable (Vodafone, paragraph 1.62). Both overshooting and undershooting require Ofcom's intervention: in the first case a decision regarding the appropriate remedy for failure to comply with the charge control, and in the second to provide consent for recovery of an undershoot in the subsequent period. This reflects Ofcom's main objective to ensure that the charge control ceiling is not exceeded.

Treatment of ported numbers

6.46 Number portability is the facility which allows subscribers of publicly available telephone services (including mobile services) to change their service provider whilst keeping their existing telephone number. Its purpose is to foster consumer choice and effective competition by enabling subscribers to switch between providers without the costs and inconvenience of changing telephone number. Mobile number portability was introduced in the UK in January 1999. The current commercial arrangements and its implications are described in Annex E.

6.47 In the May consultation Ofcom expressed the view that the level of porting of mobile numbers has become significant enough to warrant proper consideration of how they should be treated in the charge controls (see paragraphs 7.33 and 7.34 of the May consultation). Having examined the issue further, Ofcom noted that including ported-in minutes and then allowing the MNOs to request their exclusion (as proposed in the May consultation) could result in an undesirable outcome. Thus in the December consultation Ofcom modified its proposal to address this concern and suggested excluding call minutes to ported-in numbers from the charge controls (see paragraphs 6.26 to 6.29 and Annex J of the December consultation). However, Ofcom also proposed that it would include these call minutes in the controls if a concern arose that the MNOs might be reducing the effectiveness of the charge controls by setting excessive termination charges for calls to ported-in numbers.

6.48 Ofcom has not changed its view and intends to implement the proposal put forward in the December consultation. Given the current charging arrangements, Ofcom considers its December proposal to be the most appropriate treatment of calls to ported numbers. Whilst call minutes to ported numbers are not going to be included in the charge controls, this does not prevent Ofcom including these minutes in the control if the MNOs are found to be manipulating the situation and setting excessive termination charges for calls to ported-in numbers.

6.49 Further details regarding ported numbers and the responses to the December consultation are discussed in Annex E.

The controls for the combined 900/1800MHz and the 1800MHz operators

6.50 Ofcom has considered whether there should be different target charges for each operator or whether they should all be subject to the same charge control. As proposed in paragraph 7.36 of the May consultation and paragraph 6.30 of the December consultation, Ofcom believes that the efficient charge level for combined 900/1800MHz and 1800MHz operators should be different and, thus, that the controls on these two types of operators should be set at different levels. However, Ofcom believes that Vodafone and O2 (the two combined 900/1800MHz operators) should have the same target charges as each other, as should the two 1800MHz operators (Orange and T-Mobile), since operators of the same operator-type face the same cost conditions.

6.51 As to the magnitude of the difference in efficient charge levels for different types of operators and the underlying reason for this, Ofcom's view has not changed from that stated in the May and December consultations, that at current traffic levels, neither operator type has a significant cost advantage over the other on an accounting basis (see paragraph 6.31 of the December consultation). Whilst a minor adjustment has been made regarding the inter-operator differential following amendments to the LRIC model output (see paragraph C.87) the net difference in efficient charge levels remains essentially unchanged from the December consultation and reflects the difference in LRIC derived from the use of economic depreciation to obtain the path of costs over time. A full discussion of these issues including responses to the December consultation is presented in paragraphs C.71-C.87 of Annex C.

6.52 Given the difference in efficient charge levels for the two types of operators, as stated in the May and December consultations, Ofcom believes that it is appropriate to set the target average charge in the first period as an absolute target in pence per minute to allow the charges of the four operators to be aligned. As explained in paragraph H.3 of the December consultation, this allows the target average charge for each type of operator to be set at the same amount above the efficient charge ensuring that the difference between the target average charges for the two types of operators equals the difference in the level of the efficient charge for each type of operator. Thus one type of operator would not have an advantage over the other which potentially might result in a distortion in retail competition. It also means that the same target average charge can be set for Orange and T-Mobile, reflecting the identical efficient costs that they incur as 1800MHz operators^(-28-).

Summary

6.53 In summary, as discussed above, Ofcom has concluded that:

1. the charge controls should apply until 31 March 2006 and operate over two periods: in both the first period (1 September 2004 to 31 March 2005) and the final period (1 April 2005 to 31 March 2006) the target average charge should be set as a specified figure^(-29-);
   
   
2. there should be two separate sets of controls for termination of fixed-to-mobile and off-net mobile-to-mobile calls;
   
   
3. the weights in each charge control should be based on the volumes of minutes of the relevant traffic experienced by each MNO during the previous year;
   
   
4. call minutes to ported-in mobile numbers should be excluded from the weights and therefore from the controls; and
   
   
5. since the efficient charge levels for the combined 900/1800MHz and the 1800MHz operators are different, the controls on these two types of operators should be set at different levels.

The specific controls

6.54 This section describes the detailed specification of the charge controls and how it has been derived.

6.55 Tables 1 and 2 below summarise the target average charges for fixed-to-mobile 2G voice termination and off-net mobile-to-mobile 2G voice termination respectively. The target average charges for fixed-to-mobile voice termination are identical to those for off-net mobile-to-mobile voice termination.

Table 1: Target average charge for 2G fixed-to-mobile voice termination

Pence per minute (nominal)	900/1800MHz operators (Vodafone, O2)	1800MHz operators (Orange, T-Mobile)
Charge in first period (1 Sep 04 - 31 Mar 05)	5.63	6.31
Charge in final period (2005/06)*	5.63	6.31

*subject to changes in time of day weights

 

Table 2: Target average charge for 2G off-net mobile-to-mobile voice termination

Pence per minute (nominal)	900/1800MHz operators (Vodafone, O2)	1800MHz operators (Orange, T-Mobile)
Charge in first period (1 Sep 04 - 31 Mar 05)	5.63	6.31
Charge in final period (2005/06)*	5.63	6.31

* - subject to changes in time of day weights

6.56 The derivation of the efficient charge level and the appropriate target average charges in the two periods of the control is discussed below.

Derivation of the efficient charge

6.57 The efficient charge level for 2005/06 is composed of the LRIC for voice call termination plus a mark-up for common costs, based on the equal proportionate mark-up (EPMU) approach, and a further mark-up for the network externality.

6.58 The LRIC for voice call termination is calculated from a LRIC model developed by Oftel and published in April 2002^(-30-) based on the costs of a reasonably efficient 2G mobile operator in the UK^(-31-). In its review of the charges for calls to mobiles, the CC agreed with the general principles underlying the model methodology and that the model is a suitable starting point for the assessment of costs (see paragraph 2.287 of the CC report and paragraphs C.4-C.6 of Annex C for further references).

6.59 In the light of further information made available during the CC inquiry of 2002 and responses to both the May and December consultations, Ofcom has considered a number of issues and potential adjustments to the output of the April 2002 model. These issues are discussed in detail in Annex F of the December consultation and subsequent responses are discussed in Annex C, which covers:

1. Cost of capital;
   
   
2. Amendments to the LRIC model output;
   
   
3. Comparison with MNO data;
   
   
4. Network common costs; and
   
   
5. Non-network common costs.

6.60 Following responses to the December consultation, Ofcom has also revised its estimates of the level of the network externality surcharge as discussed in paragraphs 6.70-6.72 below.

Cost of capital

6.61 Ofcom takes the view that the appropriate cost of capital in the context of this market review is the cost of capital for a reasonably efficient 2G mobile operator in the UK. Ofcom has considered a number of methodologies in forming its view about the cost of capital, but believes that the main emphasis should be on the use of the Capital Asset Pricing Model (CAPM). Ofcom has undertaken a fresh analysis of each of the components of the CAPM used to derive an estimate for the cost of capital in the light of more recent information and after consideration of comments received in response to the December consultation. On this basis, Ofcom estimates the pre-tax real cost of capital to be in the range of 9.8% to 14% with a mid-point of 12%, which is a small decrease from the 12.25% that was proposed in the December consultation. All other factors remaining unchanged, this results in a small decrease in the economic cost of termination for 2005/06 of about 0.03ppm (in real 2000/01 terms), which is less than 1%. Further details of the derivation of this range and a discussion of responses to the December consultation are provided in Annex B.

Amendments to the LRIC model output

6.62 Ofcom is currently undertaking a review of the annual administration fees paid by MNOs for their 2G spectrum allocation which may result in revision to the fees from 2005/06. This has resulted in an amendment to the assumed input costs to the LRIC model regarding 2G spectrum pricing. Ofcom has made further amendments to the LRIC model where appropriate in order to address responses to the December consultation. The issues raised are discussed in detail in paragraphs C.7-C.22 of Annex C. In summary, Ofcom has amended the model calculation with reference to:

• administered incentive pricing for 2G spectrum; and
• the treatment of equipment that declines in quantity.

6.63 In paragraphs C.23-C.39 of Annex C, Ofcom also addresses comments regarding the lifetime of assets used in the model, the asset prices used in the model after 2010, and the impact of uncertainties in a dynamic market such as migration to 3G.

6.64 The overall impact of these amendments is an increase in the economic cost of termination in 2005/06 of about 0.06ppm (in real 2000/01 terms), or about 1.5%.

Comparison with MNO data

6.65 As stated in Annex E of the May consultation and Annex F of the December consultation, in order to address concerns over the accuracy of the LRIC model, Ofcom has undertaken a comparison between the outputs of the model and actual cost accounting data from the MNOs. Ofcom has derived adjustments to be applied to the output of the LRIC model following the methodology proposed by the CC in its inquiry.

6.66 As described in paragraphs C.41-C.93 of Annex C, Ofcom has given detailed consideration to the responses to the December consultation regarding the appropriateness of Ofcom's approach. In particular, Ofcom has considered the following issues:

• the appropriate level of gross book value (GBV) averaged across the four MNOs;
• the appropriate 'data adjustment' factor (given that the MNOs' submitted information reflects both voice and data services whilst the LRIC model considers a voice-only network);
• the varying proportions over time of capital and operating costs that contribute towards the total economic cost;
• the appropriateness of a reconciliation of the model results with the MNOs' cost information conducted at a total level vs at a termination-specific level;
• the comparability of coverage and quality between combined 900/1800MHz and 1800MHz operators in 2001 and subsequent investment;
• the magnitude of the differential in economic cost between combined 900/1800MHz and 1800MHz networks; and
• the appropriate basis of cost recovery (economic depreciation vs accounting depreciation).

6.67 In summary, as a result of the amendments made to the LRIC model output noted in paragraph 6.62 above, Ofcom finds that an upward adjustment of 38.7% should be applied to the capital costs and a downwards adjustment of 8.5% should be applied to the operating costs in the LRIC model, to reconcile the model's output with the actual costs incurred as reported by the MNOs. These percentage adjustments compare with a capital adjustment of +35.6% and operating adjustment of -14.9% considered in the December consultation (see paragraph 6.45 of the December consultation). Overall, the net adjustments to the LRIC model figures following comparison with the MNOs' data increase the results for the 2005/06 economic cost by 0.33ppm and 0.14ppm (in real 2000/01 terms) for combined 900/1800MHz and 1800MHz operators respectively. These adjustments are approximately 0.2ppm higher than the adjustments proposed in the December consultation.

Network common costs

6.68 Consistent with the approach described in paragraph 6.4 above, the LRIC model incorporates an EPMU for network common costs. Responses to the December consultation include concerns regarding the definition of the minimum coverage network common costs and the allocation of these common costs. These concerns are discussed in paragraphs C.96-C.100 of Annex C. Ofcom continues to believe that both the calculation of the magnitude of network common costs and their allocation is reasonable for the reasons set out in Annex C.

Non-network common costs

6.69 As in paragraph 7.48 of the May consultation and paragraph 6.48 of the December consultation, Ofcom has also included a common cost mark-up for the recovery of non-network administrative costs that should be recovered across all areas of the business, including both retail and network services. Responses to the December consultation stated the view that an unreasonable proportion of these costs were recovered from retail services since the retail activities category included all relevant cost elements, but the costs included in the network category were not similarly complete as the cost of capital tied up in the capital base was missing. Ofcom agrees that a better measure of the capital component of the network cost is the sum of network depreciation and the cost of capital associated with the network assets and, accordingly, has revised the non-network common cost mark-up to 0.41ppm (in real 2000/01 terms) from the figure of 0.33ppm proposed in the December consultation. Details of this calculation, discussion of other responses regarding non-network common costs, and Ofcom's view of these issues are provided in paragraphs C.101-C.116 of Annex C.

The economically efficient network externality surcharge

6.70 Ofcom considers it appropriate to add a further mark-up (an 'externality surcharge') to the LRIC of termination and EPMU for common cost recovery, which reflects the value of the network externality.

6.71 In both Annex F of the May consultation and Annex G of the December consultation, caution was expressed regarding the estimation of a surcharge, noting that the conceptual and practical estimation obstacles were formidable. A judgement was made on the basis of a range of estimates produced by different models of behaviour in wholesale and retail mobile markets. Each of these estimates provided a relevant, although incomplete, perspective on the efficient surcharge. Ofcom maintains a similar approach to the calculation of the appropriate surcharge in this Statement.

6.72 Ofcom considers that, broadly speaking, the estimates used in the previous consultations remain relevant to the decision about an appropriate externality surcharge. However, following responses to the December consultation, Ofcom has refined its methodology resulting in an upwards revision to two of these estimates. Ofcom therefore believes it would be appropriate to allow an additional 0.1ppm for the externality surcharge. This takes the appropriate surcharge to 0.5ppm. An outline of Ofcom's approach, responses to the December consultation, and Ofcom's view of the responses are discussed in Annex D.

Summary of the efficient charge

6.73 Taking account of the factors raised above, Ofcom has determined the efficient charge level for 2005/06 to be 5.00ppm and 5.60ppm (in real 2000/01 terms) for combined 900/1800MHz and 1800MHz operators respectively, as shown in the table below^(-32-). These figures are the target average charges for the final year of the control and are approximately 0.4ppm higher than those proposed in the December consultation.

 

Table 3: Efficient charge level (LRIC + common cost mark-up + network externality mark-up)

Pence per minute (real 2000/01)	2001/02	2002/03	2003/04	2004/05	2005/06
900/1800MHz operators
LRIC+ mark-up for common costs	6.03	5.58	4.72	4.43	4.50
Network externality mark-up	0.50	0.50	0.50	0.50	0.50
Efficient charge	6.53	6.08	5.22	4.93	5.00
1800MHz operators
LRIC+ mark-up for common costs	7.19	6.58	5.40	5.01	5.10
Network externality mark-up	0.50	0.50	0.50	0.50	0.50
Efficient charge	7.69	7.08	5.90	5.51	5.60

 

Path of reductions to the efficient charge

6.74 Having determined the level of the target charge in the final period (2005/06), the appropriate level of the target charge in the first period (1 September 2004 to 31 March 2005) must be determined, and the method for specifying the charge control for the final period (2005/06). Therefore this section considers:

• the starting charge;
• the level of the target average charge for the first period; and
• the method for specifying the control for 2005/06.

Starting charge

6.75 In order to assess how quickly termination rates should be reduced to the efficient charge level at the end of the control period (2005/06) it is relevant to establish the level of current termination rates. Ofcom's derivation of current termination charges is set out in paragraphs H.15-H.17 of the December consultation and summarised in nominal and real 2000/01 terms in the table below.

 

Table 4: Starting charges in nominal and real 2000/01 terms

	900/1800MHz operators	1800MHz operators
Charge in 2003/04 (nominal ppm)	8.04	9.47
Charge in 2003/04 (real 2000/01 ppm)	7.53	8.88

 

Target average charge for first period

6.76 The purpose of the charge control is to set charges at the efficient level by the end of the control period. The control period lasts until 31 March 2006 and the methodology described above determines the efficient charge level for the period 2005/06 (1 April 2005 to 31 March 2006). The target average charge for the final period 2005/06 is therefore set to equal this efficient charge level. Given the implementation date for the charge control of 1 September 2004, it is necessary to determine the path of charge reductions to the efficient charge level by specifying the target average charge for the first period (1 September 2004 to 31 March 2005).

6.77 In determining the appropriate level of the target charge in the first period, Ofcom has given careful consideration to balancing two objectives:

• reductions should be achieved sufficiently quickly in order to deliver substantial benefits to consumers; and
• reductions should allow sufficient time for operators and customers to adjust to new levels and structures of mobile charges.

6.78 The first point seeks to ensure that consumers benefit through lower fixed-to-mobile call charges. The second point notes that benefits to callers to mobiles should not be at the expense of unacceptable disruption to the mobile sector, the industry and consumers more generally. In practice, any delays in implementing the charge control results in shifting the balance away from the first objective because consumers benefit less quickly from the price cuts.

6.79 In the May consultation the level of the target charge for 2004/05 was based on calculating the size of three equal real percentage reductions to take the starting charge down to the efficient charge level in 2005/06, and applying two such reductions to obtain the 2004/05 target charge. The December consultation followed the same approach but applied an additional adjustment to the 2004/05 target charge following the principle of maintaining a given balance between the two objectives above.

6.80 However, an implementation date of 1 September 2004 means that maintaining the same balance would result in setting a target charge for the first period below the efficient charge level. Ofcom believes that setting such a target charge would be unreasonable and have undesirable consequences. It follows that the lowest reasonable target charge that can be set will result in a reduced benefit to consumers in comparison to the previous proposals.

6.81 Ofcom has considered this issue carefully in light of representations in response to the December consultation. Consistent with the objectives in paragraph 6.77 above (and stated in previous consultations^(-33-)), Ofcom has determined that it is appropriate for the charge control to move straight to the efficient charge in 2004/05, so that the target average charge in the first period is set at the efficient charge level for 2005/06.

6.82 Ofcom's objective is not only to deliver benefits to consumers sufficiently quickly but also to consider the potential effect of reductions in termination charges on operators, the industry and consumers more generally. Ofcom believes that the charge control allows sufficient time for preparation and adjustment, both in terms of financial planning and adjustment to retail prices. With regards to financial planning, Ofcom considers that the industry has been made sufficiently aware of regulatory intention in this area to plan for and accommodate changes in termination revenues since the publication of the market review in September 2001. With regards to adjustment to retail prices and the potential for associated disruption, Ofcom does not believe that a one-off reduction to the efficient charge level is likely to lead to excessive disruption or damaging consequences to mobile subscribers. The magnitude of the reduction is similar to the initial reduction proposed in the December consultation and not much greater than the compound effect of the CC's recommendation of two real reductions of 15% to occur in 2003/04.

6.83 From Table 3 above, it is apparent that the efficient charge level in 2004/05 is actually lower than the efficient charge in 2005/06. This is due to the revision in assumed 2G spectrum pricing from 2005/06 onwards as discussed in paragraphs C.12-C.16 of Annex C. However, Ofcom believes it would be undesirable to set a target average charge for the first period which is lower than the target average charge for the final period of 2005/06 as this would potentially have a disruptive effect on consumer prices. Therefore Ofcom has decided to set the target average charge for the first period equal to the efficient charge level in 2005/06.

6.84 Taking the efficient charges for 2005/06 of 5.00ppm and 5.60ppm in real 2000/01 terms results in nominal target average charges of 5.63ppm and 6.31ppm for combined 900/1800MHz and 1800MHz operators respectively after inflating using compounded RPI^(-34-).

Specification of control for 2005/06

6.85 By setting the target average charge for the first period to be equal to the efficient charge level for 2005/06 as an absolute target in pence per minute, this ensures that the difference in target average charges for the two types of operators is equal to the difference in efficient charges. This meets the objective stated in paragraph 6.52 above.

6.86 In the December consultation the charge control for the final period (2005/06) was specified as a RPI-X control to reflect the required reduction from the 2004/05 charges necessary to reach the efficient charge level for 2005/06.

6.87 Given that Ofcom has now set the target average charge for the first period at the efficient charge level for 2005/06, the RPI-X specification of the charge control for 2005/06 is no longer necessary. It is simplest to specify the target average charge for 2005/06 to be the same absolute target in pence per minute as specified for the first period (subject to changes in time of day weights)^(-35-): that is 5.63ppm for combined 900/1800MHz operators and 6.31ppm for 1800MHz operator.

Responses to the December consultation

Starting charge

6.88 In paragraph A.3 of its response, Vodafone argues that the incorrect value of RPI has been used in setting the starting charge. Whilst a pro-rata RPI adjustment was used to derive the RPI-15% reduction on 24 July 2003, Vodafone argues that if this rate persists through to 31 March 2004 then the full annual RPI should have been used and thus Vodafone has been under-recovering termination revenue for the period from 25 July 2003. Furthermore, Vodafone asserts in paragraphs A.8-A.9 of its response that this RPI error in deriving the appropriate starting charge is perpetuated in Table 3 of Annex H of the December consultation since an inconsistent value of RPI is then used to convert from nominal to real.

6.89 In the context of this market review, the relevant question is not how the reduction on 24 July 2003 was calculated but the current level of termination charges as these dictate the starting point for the charge control^(-36-). In any case, Ofcom has set the target average charge in both control periods with reference to the efficient charge level for 2005/06 and not in relation to the starting charge (see paragraphs 6.81 and 6.73 above).

6.90 With regards to Vodafone's second point, Ofcom notes that this is no longer relevant given that Ofcom has chosen to specify the charge control exclusively with reference to the efficient charge level for 2005/06^(-37-).

Adjustment for delay and 'retrospection'

6.91 In its response, Vodafone (paragraphs 1.16-1.43) suggests that the adjustment to the target average charge in 2004/05 proposed in the December consultation (see paragraph 6.79 above) had the effect of setting "the regulated charge at a lower level than would otherwise be appropriate to offset customers' 'overpayments' in an earlier period". Vodafone suggests that the intention is to achieve a retrospective effect from a forward-looking control and believes this proposal is unlawful and unreasonable. Similarly, O2 (pages 4-5), Orange (section 3.4), and T-Mobile (paragraphs I.12-I.15) argue that the proposed charge control is 'retrospective' or 'backward-looking'.

6.92 The purpose of the charge control is to set charges at the efficient level by the end of the charge control period. Consequently it is necessary to determine the path of charge reductions over the course of the control period to reach this efficient charge level. The implementation date of the charge control affects the path of reductions to reach the efficient level. It is legitimate for Ofcom in considering that path to seek to balance the interests set out in paragraph 6.77 above. This is necessarily a forward-looking exercise, and as set out in paragraphs 5.7-5.11 of Chapter 5, Ofcom does not accept an argument that the resulting charge controls are retrospective and therefore does not accept that in setting such controls Ofcom is acting ultra vires.

Impact on MNOs

6.93 Vodafone claims that operators have "reaped no financial advantage" (paragraph 1.33 of its response) from higher termination rates because these higher rates have been competed away in the retail market. This view is shared by O2.

6.94 In section 3.4.3 of its response, Orange disagrees with the view expressed in paragraph H.9 of the December consultation that MNOs have had sufficient opportunity to anticipate the reduction in future termination revenues or to accommodate these changes in their financial planning. Orange believes that re-balancing of revenues between incoming and outgoing calls cannot take place before reductions in termination charges due to the competitive pressures in the retail market which result in any potential excesses from call termination being competed away.

6.95 Ofcom notes that the claim by Vodafone and O2 would hold only if the 'waterbed' effect is complete which Ofcom does not accept (see paragraph 4.34 of Chapter 4). But in any case, Ofcom's rationale for its choice of target charges arises from a concern regarding the impact of delay on callers to mobiles.

6.96 In response to Orange's point, Ofcom has not commented upon the precise date for implementing changes to mobile outgoing call prices. The magnitude and timing of any changes to (unregulated) mobile retail charges is for MNOs to decide. As already noted in paragraph 6.82 above, the magnitude of the reduction in termination rates to occur in 2004/05 is similar to that proposed in the December consultation. Ofcom believes that MNOs have had sufficient time to consider the impact of this reduction and determine how to set their retail charges appropriately.

Impact on consumers

6.97 Regarding the benefit to callers to mobiles, in paragraphs 1.38-1.52 of its response, Vodafone states that it is not possible that fixed customers would have seen a reduction in charges from 1 January 2004 given the necessary processes such as the time required to notify BT prior to pass-through of reductions.

6.98 Vodafone also adds that callers to mobiles would not see the full benefits of this adjustment due to the fact that BT is not required to fully pass-through reductions in termination rates to callers from fixed to mobile phones. Orange also raises its concern regarding pass-through obligations on BT in section 3.4.2 of its response.

6.99 In contrast, on page 4 of its response, UKCTA argues that whilst it agrees with the approach described in the December consultation, it nevertheless believes that the proposed compensation for delay does not provide adequate relief to other sectors of the economy which are paying inefficient subsidies to the mobile sector.

6.100 Whilst Ofcom acknowledges Vodafone's point that a reduction in mobile termination charges may not pass through to callers to mobiles instantaneously, this is not the key consideration for the purposes of calculating the adjustment for delay. Under the neutral assumption that the time required for reductions in termination rates to pass through to retail prices remains unchanged whether the implementation date for the charge control is 1 January 2004 or 1 April 2004, the key observation is that a three month delay in implementation of the charge control will translate into a three month period during which consumers are likely to pay higher prices for calls to mobiles. However, as stated in paragraph 6.24 above, Ofcom has now set an implementation date of 1 September 2004 to enable consumers to benefit at approximately the same time that mobile operators reduce their prices, taking account of the concerns raised by industry that pass-through might not be achievable without both sufficient notice for MNOs to adjust and advise third parties of termination charges, and a sufficient period for FNOs to adjust retail rates to reflect the new charges.

6.101 The concern raised by Vodafone and Orange regarding BT's requirement to pass through the reductions in termination rates to callers from fixed-to-mobile phones is addressed in paragraphs 5.16-5.20.

6.102 In response to UKCTA's concerns, Ofcom has given consideration to the interests of different parties in balancing the objectives noted in paragraph 6.77 above. The main focus of these objectives is the interests of consumers, however, Ofcom has also given due consideration to the impact of the charge control on operators, consistent with its duty in section 47(2) of the Act to set a proportionate charge control condition. In particular, given the constraints of setting a target charge which is no lower than the efficient charge level, and setting a reasonable implementation date taking account of practical considerations so that reductions in termination and retail prices can occur together, Ofcom believes that the charge control set out in this document strikes an appropriate balance. Whilst UKCTA expresses its concern that the charge control does not formally require MNOs to reduce their rates until just before 31 March 2005 (the end of the first period), if MNOs do not reduce their rates early in the first control period, given how much higher current charges are above the target charge, MNOs would have to set their charges significantly below the efficient charge level in order to comply with the target average charge for the period. Having regard to Ofcom's requirement that MNOs must notify interconnecting operators of their effective charges on 1 September 2004 within 28 days of when the conditions come into force (see conditions MC6 and MD6), Ofcom's expectation is that MNOs are likely to implement charge reductions at, or close to, the beginning of the first control period.

Detailed calculation of adjustment for delay

6.103 With regards to the detailed calculation of the adjustment for delay, Vodafone submits in paragraph A.10 of its response that the effect of differential volumes between 2003/04 and 2004/05 should be taken into account. Vodafone also states its belief that it is more appropriate to calculate Table 3 in Annex H of the December consultation entirely in nominal terms rather than undertaking translations between nominal and real rates.

6.104 These concerns are no longer relevant to the charge control specified in this document since Ofcom's decision to set the target average charge in the first period at the efficient charge level no longer involves the calculation of an adjustment for delay^(-38-).

Cost benefit analysis of regulation

6.105 Ofcom's approach is that regulatory intervention is to be considered appropriate only when there is a reasonable expectation that its benefits will exceed its costs. In paragraphs 7.58-7.63 of the May consultation and Annex L of the December consultation, an assessment of the net benefits to be gained from the regulation of termination charges was provided. Ofcom is continuing to use the same approach - the figures below are an updated version given the changes in LRIC noted above.

6.106 As noted in previous consultations, models of economically efficient pricing are in principle well suited to deriving estimates of the welfare gains from regulation. Ofcom's reservations about the relevance and practicality of deriving Ramsey prices means that such estimates should not be regarded as precise, and consequently Ofcom does not use these models to derive 'optimal' prices for fixed-to-mobile calls or mobile termination. Rather, Ofcom uses these models to provide an indication of the direction and broad magnitude of the effect of regulation.

6.107 The approach to the assessment of the appropriate level of voice termination charges involves an estimation of the set of charges that maximises the welfare of consumers, subject to ensuring that MNOs are able to earn a reasonable return on their investment. This assessment takes account of the benefits to all users, including those calling mobiles as well as the mobile customers themselves.

6.108 In particular, the assessment undertaken involves a comparison between two scenarios:

• a scenario in which termination charges are brought down via the charge control to the efficient charge and where other mobile prices are assumed to be set on a Ramsey basis ('constrained Ramsey'); and
• an unregulated scenario in which MNOs set excessive termination charges, but are assumed to make no supernormal profits overall (i.e. make sufficient revenues to cover costs, including the cost of capital, for retail and wholesale mobile services in total) ('zero-profit unregulated').

6.109 The comparison uses the relevant versions of the model produced by Dr Rohlfs and described in his paper A model of prices and costs of mobile network operators, 22 May 2002^(-39-). Dr Rohlfs' model is based on four services: mobile subscription, mobile-originated usage other than off-net, fixed-to-mobile usage and off-net mobile-to-mobile usage.

6.110 The comparison indicates that there are likely to be large gains from regulation. This is due to the highly inefficient structure of charges under the 'unregulated' scenario. That is, compared to taking an approach purely based on efficiency considerations, fixed-to-mobile and off-net mobile-to-mobile charges are priced considerably higher; and mobile-originated calls and subscription are priced considerably lower, than would be appropriate.

6.111 A summary of the results is as follows. The expected welfare gain of (GBP) £222.5m is the per-quarter gain in 2005/06 (expressed in 2000/01 real terms). This is shown in Table 5 below. Expressed in current (2004/05) terms, the total benefit delivered over the course of the charge control is (GBP) £1,449m. This has been calculated in line with the methodology used in the December consultation (see Table 3, Annex L).

 

Table 5: Summary of outputs of cost-benefit analysis (real 2000/01 terms)

		"Unregulated" scenario	"Constrained Ramsey" scenario
Subscription price		(GBP) £16.66	(GBP) £16.02
Mobile-originated usage price		5.1ppm	9.3ppm
Fixed-to-mobile usage price		26.5ppm	7.1ppm
Off-net mobile-to-mobile price		17.2ppm	14.7ppm
Mobile termination charge		24.8ppm	5.3ppm
Compared to calibration prices and quantities	Change in consumer surplus (per quarter)	(GBP) £.332.8m	(GBP) £555.3m
	Change in producer surplus (per quarter)	-(GBP) £495.8m	-(GBP) £495.8m
	Change in total surplus (per quarter)	-(GBP) £163.0m	(GBP) £59.5m
Welfare gain from regulation (per quarter)		(GBP) £222.5m

 

Responses to the December consultation

6.112 Vodafone, Orange and T-Mobile all comment on the reasonableness of the 'unregulated' termination charge, as used in Ofcom's comparison, and state that if a lower charge is used, the welfare gains would not be sufficient to justify a price control.

...the true gain (based on Frontier Economics' model, and the inputs proposed by the CC) is only (GBP) £4.7 million per quarter...if one moves from 10.5ppm to the supposed optimal charge; and the true gain is only (GBP) £63 million per quarter if one moves from 17ppm to the supposed optimal charge. (Vodafone, 1.59)

The base case unregulated scenario is fundamentally flawed and consequently the Director's reliance on the welfare analysis to justify the proposed price control obligation is without substance...due to (this deficiency) and the failure of the cost-benefit analysis to consider the dynamic effects of proposed regulation, the case for intervention is marginal at best. (Orange, p.7)

The Director's welfare model can also be fairly criticised because it is predicated on a Ramsey pricing structure in the retail and wholesale mobile markets. Given the Director's conclusions in respect of applying Ramsey pricing in his 'fair charge' calculations, the use of a Ramsey model can hardly be considered to be reflective of the Director's view of the "prevailing characteristics of the retail and wholesale mobile markets" (Orange, p.7)

6.113 The reasonableness of the unregulated charge is discussed in paragraphs 4.3-4.18 of this document. While Ofcom believes that the monopoly charge is likely to be in the range of 20-25ppm, the result of substantial welfare gains is robust to any plausible termination charge. As noted in paragraph 4.16, Vodafone has previously indicated that charges in the absence of regulation would be likely to be as high as 17-20ppm. If the unregulated charge were at this level, regulation would continue to provide substantial welfare gains as illustrated by Vodafone's own figure of (GBP) £63m per quarter. There are also gains from regulation at lower termination charges (including reducing current termination charges); however, Ofcom does not consider such charges to be plausible, for the reasons expressed in paragraphs 4.3-4.18.

6.114 As to the reference to 'dynamic effects' of the proposed regulation, Ofcom has commented at paragraph 4.29 that it does not believe that substantive detrimental impact on dynamic efficiency is likely. This issue is further addressed in paragraphs L.19-L.23 of the December consultation, including a discussion of the effect of regulation on incentives to invest.

6.115 Orange makes a further point about the use of a scenario which uses a Ramsey pricing outcome. Ofcom believes this point is substantively addressed in paragraphs L.25-L.26 of the December consultation. To summarise, the Ramsey pricing model is not the only model of price-setting behaviour that could be assumed in the 'regulated' scenario. However, it is not critical to the size of the gains from regulation whether this model is used or not. Efficiency gains in welfare models are driven by either or both of two effects: (i) a general move of prices closer to cost (involving reductions in excess profit); and / or (ii) a closer alignment of specific prices to cost. Ofcom's welfare analysis compares two scenarios which are, by definition, zero-profit outcomes. This means that the first of the effects identified is not included in this case. Rather, the efficiency gains from regulation are driven solely by the second of these effects - in particular, by reducing the welfare losses associated with an excessive termination charge. That is, the critical driver of the gains from regulation is the size of the 'unregulated' price of termination. The structure of the other prices (set in the retail market for access and outgoing calls) is largely inconsequential to the result of welfare gains. Hence, once the termination charge is assumed fixed by regulation, it makes relatively little difference whether a Ramsey model, or another model of retail competition, is used.

6.116 T-Mobile makes a number of further claims about the welfare analysis. In particular it claims that the cost benefit analysis is "propped up" by an assumption of imperfect competition and excess profits, that Ofcom "effectively rejects" its own welfare analysis, and that the effect of price controls on subscriber numbers (in particular, pre-pay subscribers) predicted by the cost-benefit analysis is "inconsistent with both economic theory and the available evidence". (T-Mobile, paragraphs II.24-II.28 and II.32-II.34)

6.117 Ofcom believes that these claims are either inaccurate or that they have been addressed satisfactorily in previous consultations:

• the estimate of welfare gains from regulation in no way relies on an assumption of excess profits or of imperfect competition. As noted above, no profits are earned in total under either the regulated or unregulated scenarios in Ofcom's analysis. This point was addressed in paragraph L.34 of the December consultation.
• Paragraph L.29 of the December consultation addresses why weaknesses in welfare models means that they are inappropriate for setting regulated charges. T-Mobile's further points about the welfare costs of choosing an under- vs an over-estimate of the optimal mark-up are discussed in paragraphs K.44-K.45 of the December consultation.
• Paragraphs L.16-L.18 of the December consultation address the effect of price controls on pre-pay subscribers. In summary, the welfare model used makes a number of simplifying assumptions (e.g. that all subscribers pay the same price for subscription), in part due to the complexity associated with modelling existing retail tariffs. However, the model's result that it would be more efficient for the average subscription charge to rise from existing levels does not mean that charges to all subscribers must rise equally. The model does not capture price discrimination, and T-Mobile's claim that ongoing subscription charges for pre-pay users will have to be introduced to recover costs does not follow.
• T-Mobile further claims that the treatment of pre-pay subscribers is unreasonable, in that Oftel had effectively assumed that low value subscribers had a perfectly inelastic demand. However, the point initially made in the December consultation was that pre pay subscribers incur depreciation costs in holding a handset. These ongoing costs form part of the subscription cost, from the point of view of the pre-pay subscriber. Therefore, it is not correct to say that these users currently pay no 'ongoing' charge (as modelled), and that even if (ongoing or upfront) charges to these users were to rise, it is not plausible to suggest that there would a "massive fall" in the number of pre-pay subscribers, as claimed by T-Mobile (p.93 of its response to the May consultation). The assumed elasticity of demand for subscription (-0.3) provides a reasonable estimate of the effect on subscriber numbers from a price increase. Ofcom therefore believes the treatment of pre-pay subscribers in the welfare model is reasonable.

Conclusion on the cost-benefit analysis

6.118 Ofcom considers that the cost-benefit analysis shows that consumers will, on the whole, be made substantially better off by the regulation of mobile termination charges.

---

Footnotes:

22. - see http://www.ofcom.org.uk/static/archive/oftel/publications/mobile/depr0901.htm

23. - see http://www.ofcom.org.uk/static/archive/oftel/publications/mobile/ctm_2002/
lric_more120202.pdf

24. - Conditions MC6 and MD6 have been amended to require the mobile operators to notify wholesale purchasers of the mobile call termination charges that will be in effect on 1 September 2004 no later than 28 days after the condition comes into force.

25. - The charge control uses a controlling percentage which is defined in terms of changes in RPI in the previous year. More precisely, the charge control years proposed in the December consultation run from April to March; the volume weights are taken from the volume shares in April to March of the previous year; and the change in RPI is the change in the latest available calendar year, January to December.

26. - Similarly, v₀(i) are the time of day volume shares in the year before the first period.

27. - The charge control specified in this document has a shorter first period, lasting 7 months in duration. The charge control conditions therefore specify the Weights Adjustment Factor explicitly in terms the Average Revenue for the first period divided by the AIC for the first period of the charge control, consistent with the expression in paragraph 6.40.

28. - Currently Orange and T-Mobile have different average charges.

29. - subject to changes in time of day weights for the final period target average charge (see paragraphs 6.36-6.41)

30. - see http://www.ofcom.org.uk/static/archive/oftel/publications/mobile/ctm_2002/
april02_model.zip

31. - An updated version incorporating the changes discussed in the December consultation and in Annex C will be available shortly on Ofcom's website.

32. - The figures of 5.00ppm and 5.60ppm (in real 2000/01 terms) are equivalent to the nominal figures of 5.63ppm and 6.31ppm presented in Tables 1 and 2.

33. - See paragraph 7.23 of the May consultation and paragraph H.5 of the December consultation.

34. - Assuming the value of RPI used for 2005/06 is the same as the value for 2004/05 (2.8%).

35. - See paragraphs 6.36-6.41

36. - Nevertheless, Ofcom does not accept that Vodafone has been under-recovering termination revenue since the use of a pro-rata RPI figure in setting the reduction on 24 July 2003 reflects the implementation of a one-off cut to take effect before 25 July 2003. The fact that this charge level has continued to be in effect in accordance with the Continuation Notices given to the four MNOs on 23 July 2003 did not affect the calculation of that one-off cut.

37. - Even if it was relevant, Ofcom does not agree that there is an error in the nominal to real conversion since the purpose of the calculation is to determine the current charges in real 2000/01 terms consistent with the start of the charge control - 1 April 2004 in the case of the December proposals, and not 24 July 2003.

38. - Nevertheless, Ofcom notes that Vodafone's first concern was commented upon in footnote 81 of paragraph H.22 of the December consultation, and with regards to Vodafone's second concern, the approach taken and consulted upon in the May and December consultations considered a target charge for 2004/05 based on applying two of three equal real percentage reductions: performing the same calculation based on nominal percentage reductions does not necessarily generate the same result.

39. - See http://www.ofcom.org.uk/static/archive/oftel/publications/mobile/ctm_2002/
main_report.pdf

next section

Back to top