Contents

Executive summary

Chapter 1 – Introduction

Chapter 2 – Costs and benefits

Chapter 3 – Policy issues

Chapter 4 – Conclusions

Chapter 5 – Consultation

Annex A – Calculating the effect of 2PC

Annex B – Sensitivity analysis

Executive summary

S.1 This consultation presents the initial findings of Oftel's policy review of two part charging (2PC). It considers the introduction of a 2PC system for all successful interconnection calls.

S.2 Oftel has discussed the question of 2PC, at a general level, on a number of occasions:

• Pricing of calls to the Internet: possible initiatives to bring about more appropriate and flexible tariffs, November 1999, see www.oftel.gov.uk/publications/1999/info_super/oifp1199.htm;
• Oftel's statement on the relationship between interconnection charges and retail prices for NTS, December 1999, see www.oftel.gov.uk/publications/1999/pricing/nts1299.htm; and
• Price control review, March 2000, see www.oftel.gov.uk/publications/pricing/pcr0300.htm.

However, each time the subject has been raised, very few operators have shown interest in furthering the discussions. As a result, BT has rejected requests for two-part charging from individual operators on the grounds that a broad industry consensus needs to be gained before changing existing charging practice.

S.3 Oftel has therefore decided to stimulate debate by undertaking this policy review to assess the likely costs and benefits of implementing a system of two-part charging, and to consider its potential impact on industry and the interconnection regime. Since the size of these costs and benefits rests on various assumptions, Oftel has set the analysis in the context of a wider regulatory option appraisal. As such, this review also investigates the practicalities of implementation, including some of the less tangible costs and likely knock-on effects to other parts of the regulatory framework.

S.4 There are two key elements that make up the interconnection charges for calls. They are the cost of setting up a call (call set-up) and the cost of maintaining the connection for the duration of the call (call duration).

S.5 Under the current arrangements, BT's interconnection charges are charged for on a pence per minute (ppm) basis without a call set-up charge. The effect of this averaging is that the charge for shorter than average calls is lower than it would be under a system of 2PC and the charge for longer than average calls is higher than under a system of 2PC. A move towards 2PC would introduce a separate charge for the set up of a call as well as a ppm charge for call duration and could result in wholesale charges for each individual call that more efficiently align cost recovery to its cause.

S.6 If 2PC is considered appropriate for certain types of calls, Oftel believes that the same principles should be applied to all calls as it would be impractical to maintain different charging regimes by call type. If as a result of this consultation 2PC is favoured, this would mean a universal change to 2PC for all BT's interconnection charges and the withdrawal of the current arrangements.

S.7 Oftel's analysis shows that, over a period of five years, there could be a welfare benefit in excess of £5 million from 2PC. The size of the benefit depends on the assumptions in the analysis, in particular the sizes of the set-up and duration charges under 2PC. Oftel has undertaken a sensitivity analysis of these assumptions.

S.8 Based on the estimates Oftel has at this stage, the total cost to BT and other operators involved in implementing 2PC, were it to be required, could be in excess of £12 million over a period of five years. It is likely that the individual costs faced by different operators would vary substantially depending, for example, on the current state of each operator's wholesale billing system, its scale and complexity. Oftel would welcome more detailed cost information as part of the consultation.

S.9 At this stage, when the costs and benefits are compared, there does not appear to be an overwhelming economic benefit associated with the implementation of 2PC. In addition, examining the call set-up and call duration charges adopted by other European countries, it seems unlikely that the level of variability in the ratio between the two costs reflects real differences in network design. It is more likely that this reflects the absence of uniformity in the treatment of cost allocation. On this basis, Oftel sees no clear methodology to determine an appropriate and meaningful split between call set-up and call duration which would lead to maximum benefits.

S.10 It is also worth noting that at a time of large regulatory change, the implementation of 2PC would undoubtedly consume scarce regulatory resources in Oftel/Ofcom, BT and in OLOs, in addition to the more direct operational cost of adjustment. In the absence of any overwhelming economic benefit it is Oftel's initial view that wholesale changes to the current interconnection arrangements would be disproportionate. In order to implement the change, given the intangible and non-quantified costs, it is Oftel's initial view that the benefits should significantly exceed the costs.

S.11Oftel seeks stakeholders' views on the practicality of changing to a new system, and the likely costs involved. The deadline for comments is 25 April 2003. Details about the consultation process are at the end of this document, see chapter 5.

S.12 Oftel hopes to publish a statement by July 2003. If as a result of this consultation, Oftel concludes that a system of 2PC is favoured, Oftel will also publish a programme of implementation in accordance with the new European Directives (see timetable at paragraph 4.4).

Chapter 1

Introduction

Interconnection charging arrangements

1.1 There are two key elements that make up the interconnection charge for a call. They are:

• the cost of setting up a call (call set-up); and
• the separate cost of maintaining the connection for the duration of the call (call duration).

1.2 Under the current arrangements, BT's interconnection charges are charged for on a pence per minute (ppm) basis without a charge for call set-up. The ppm charge is calculated to take account of the costs of call set-up and average call duration. These costs also include the costs of unsuccessful calls.

1.3 Under this system, the costs of originating calls of average duration are fully met. However, the effect of this averaging is that the charge for shorter than average calls is probably lower than under a system of 2PC and the charge for longer than average calls is probably higher than under a system of 2PC. This effect evens out over all calls, implying that BT recovers its costs appropriately. But where operators purchase interconnection for calls that are predominantly longer or shorter than the average duration, the charges could be either higher or lower by BT for interconnection than if charged under a system of 2PC.

Scope of the review

Successful calls

1.4 Strict cost causation suggests that the charge for interconnection calls should be per call attempt, not per successful call. However, the economic efficiency benefits achieved when a price reflects costs derive from the consumer facing an appropriate price signal to modify their behaviour. Unsuccessful call attempts impose a cost on the network provider, so charging for them signals to consumers that they should reduce their consumption – however, the consumer is likely to have limited ability to reduce the number of unsuccessful call attempts. This is because consumers generally would not know, when initiating a call, whether it would be successful or unsuccessful. This argument assumes that the costs of 2PC would be passed on to consumers.

1.5 Oftel does not favour charging for unsuccessful call attempts on the basis that callers are not in a position to regulate their unsuccessful call attempts, even with the full information on cost available to them (MMC's public interest argument – Cellnet and Vodafone, December 1998). It is unclear, therefore, that charging for unsuccessful calls would promote economic efficiency and result in consumer benefit. On these grounds, this document considers 2PC for successful calls, with no separate change for unsuccessful calls: costs incurred in unsuccessful calls should be recovered from the generality of successful calls.

All call types

1.6 This review was launched in response to representations made by operators carrying long-duration number translation service (NTS) Internet calls. As discussed above, the current averaged charging structure means that the charge for longer than average duration calls is probably higher than under a system of 2PC. Oftel estimates that the average length of metered Internet calls is 11.5 minutes and that operators terminating calls lasting longer than around 3.7 minutes (the 'average') could benefit from a 2PC regime. 2PC could reduce costs incurred by operators terminating calls to the Internet.

1.7 As explained in paragraph 1.3, the current system of charging leads to appropriate cost recovery at the aggregate level by BT. If BT is to continue to recover its overall costs then a reduction in the charge for long duration calls must be accompanied by an increase in the charge for short-duration calls. As such, there is no objective reason why the necessary cost-recovery should be restricted to a subset of short-duration calls (eg NTS voice), or to call origination, rather than to all calls including call termination.

1.8 In conclusion, if 2PC is considered appropriate for certain types of interconnection calls, Oftel believes that the same principles should be applied to all interconnection calls, excluding FRIACO. (The FRIACO charge is priced in relation to the cost of an average circuit and not on a pence per minute (ppm) basis. A change in the ppm charge for metered calls does not therefore imply that the FRIACO charge should also change). If as a result of this consultation 2PC is favoured, this, in Oftel's initial view, would mean a universal change to 2PC for all interconnection charges (BT’s wholesale call origination and call termination) and the withdrawal of the current arrangements.

The interconnection regime

1.9 Interconnection charges are controlled by the network charge control and charges for each interconnection service should fall between the long run incremental cost of provision (LRIC) and the stand-alone cost of provision and pass the relevant combinatorial tests.

1.10 Under the current interconnection regime, BT must allow other licensed operators to interconnect their telecommunications systems with it on non-discriminatory terms. The Interconnection Directive (ICD) requires that tariffs be sufficiently unbundled so that the operator requesting interconnection is not required to pay for anything not strictly related to the service requested. In practice, this includes an amount for common costs.

1.11 Services generally should only be made up of the elements that are necessary for the service in question, and operators should only be required to pay for costs that they have caused to be incurred. However the nature of interconnection is such that many charges are derived from a set of averaged costs - for example single tandem call charges are irrespective of distance and geography, which can vary greatly among individual calls.

1.12 Since the current charging arrangements and a system of 2PC both ensure an appropriate recovery of costs, including common costs, whether or not 2PC should be mandated under the current regime is a question of general policy, depending on the overall costs and benefits involved, rather than being related to any individual economic market. The market reviews (see below) have yet to establish the details of the new regulatory regime, however Oftel would expect that the issue of 2PC would similarly be considered a matter of policy.

Oftel's market reviews

1.13 The new EU Directives establish a new framework for the regulation of electronic communications networks, services and associated facilities throughout Europe. These Directives entered into force on 24 April 2002. Member States have 15 months in which to transpose the provisions into national legislation and to bring all regulation into line with the new Directives. The Directives must be implemented in all Member States at midnight of the 24 July 2003, 15 months plus one day after entry into force.

1.14 The new package of Directives consists of:

• the Framework Directive (FD);
• the Authorisation Directive (AD);
• the Universal Service Directive (USD);
• the Directive on Privacy and Electronic Communications (PEC);
• the Access and Interconnection Directive (AID); and
• the Radio Spectrum Decision (RSD).

1.15 The FD sets the overall framework, including objectives and principles, which Oftel must take into account when making regulatory decisions. In addition, the FD sets out the principle that, in most cases, market reviews must be carried out before regulation is imposed and that regulation is only to be imposed where the market is not effectively competitive ie where at least one operator has significant market power (SMP). A summary of all of the new Directives can be found on Oftel’s website at www.oftel.gov.uk.

1.16 Oftel will be consulting on its market reviews during the first half of 2003. In order to implement 2PC under the new regime, Oftel would need to wait until the draft conditions outlined in the market review consultations have been finalised in the market review statements.

1.17 If 2PC is favoured, any consultation on the detail of its implementation would be published after 25 July 2003. A draft implementation programme for 2PC is set out later in this review (see chapter 4).

Chapter 2

Benefits and costs

2.1 As explored below, a move towards 2PC could result in wholesale charges for each individual call that more efficiently align cost recovery to its cause (the economic principle of cost causation). However, while 2PC has a sound theoretical basis, it is not clear whether in practice it could be easily implemented and whether the implementation costs would outweigh the benefits.

2.2 Oftel's six principles of cost recovery provide a useful framework within which to discuss the relevant issues (Oftel suggested and the MMC adopted six principles of cost recovery in its report on telephone number portability, November 1995).

Six principles of cost recovery

(i) Cost causation

2.3 This principle requires that costs should be recovered from those whose actions cause the costs to be incurred at the margin.

2.4 From a viewpoint of pure economic efficiency, 2PC more accurately reflects the principle of cost causation, since the costs incurred are attributed to the factors that cause them. If this more granular approach to pricing at the wholesale level is reflected in retail prices, there would be benefits to consumers.

(ii) Distribution of benefits

2.5 2PC would benefit consumers of longer than average calls (assuming that the lower costs would be passed on to consumers). This provides a possible argument for these consumers bearing a larger share of the implementation costs. However this is unlikely to occur in practice, given the practical difficulties an operator would be likely to face if it tried to recover these costs only from customers of certain (long duration) call types. Therefore the distribution of benefits argument is of limited relevance for 2PC.

(iii) Effective competition

2.6 This principle requires that the mechanism for cost recovery should not undermine or weaken the pressure for effective competition. Under 2PC, providers of solely voice traffic are likely to find their costs increasing (since these are generally shorter duration calls), and providers of data traffic are likely to find their costs decreasing (since these are generally longer duration calls). However, voice and data providers are not in close competition with each other (consistent with Oftel market definitions) so competition is unlikely to be materially affected as a result of a move to 2PC. To the extent that short and long calls are in the same market, 2PC promotes competition on the merits by reflecting underlying costs.

(iv) Cost minimisation

2.7 The mechanism for cost recovery should ensure that there are strong incentives to minimise costs. Incentives to minimise interconnection costs generally are not altered by the introduction of 2PC.

(v) Practicality

2.8 The mechanism for cost recovery needs to be practical and relatively easy to implement. While 2PC has a sound theoretical basis, it is not clear whether in practice it could be easily implemented and what the wider impact would be. The likely practicalities of implementing 2PC are therefore considered in chapter 3.

(vi) Reciprocity

2.9 Finally, where services are provided reciprocally, charges should also be reciprocal. This principle is not relevant to call origination, but would generally apply for call termination for normal geographic calls, which is also considered by this review. However this principle applies under the current system of charging so there would be no change. Nonetheless, comments from respondents on the need for 2PC to be implemented on a reciprocal basis between BT and OLOs are particularly welcome.

Conclusions

2.10 Having considered all six of the cost recovery principles, it is clear that the key issues in the case of 2PC are associated with cost causation and the practicality and costs of implementation. On this basis, this document compares the likely costs and welfare benefits of implementing 2PC and considers the practicalities of implementation.

2.11 The following paragraphs set out Oftel's analysis of the likely benefits and direct costs of introducing 2PC arrangements. However, since the size of costs and benefits rests on various assumptions, it is not on its own definitive. It is therefore necessary to set the analysis in the context of a wider regulatory option appraisal. To this end, the practicalities of implementation, including some of the less tangible costs and the likely knock-on effects to other parts of the regulatory framework, are considered in chapter 3.

Calculating the welfare benefits of 2PC

2.12 In order to assess the possible welfare benefits of 2PC, Oftel has analysed data on average call duration together with the processing costs of call set-up and call duration provided by BT and partly present in the 2001-02 financial statements. This has enabled Oftel to estimate the actual cost savings for different types of call:

• local calls;
• national calls;
• fixed calls to mobiles;
• local non-geographic calls;
• national non-geographic calls; and
• 'other', excluding FRIACO (see 1.8).

2.13 Oftel's benefit analysis is based on the following calculations, set out as five steps below. Annex A explains in greater detail how these calculations were made, and the information and data on which they are based.

Step 1: Calculating the unit set-up cost and duration cost of each network component of a call

2.14 Oftel obtained information from BT on the cost of call set-up (pence per call) and duration (pence per minute) on a fully allocated cost basis for the various network elements of a call. Most of this information is present in the Financial Statements. BT has informed Oftel, at a general level, how it divided costs between the relevant existing network components (see annex A). During the consultation period Oftel will continue to discuss with BT the data sources used and how they impact on the attribution of costs.

Step 2: Calculating average call duration

2.15 It was then necessary to calculate the average duration of each call type. This was calculated using residential duration information from the price control review, originally from BT. This information was suitably adjusted so that it was consistent with the cost and volume information used in the benefits analysis. The average durations were then compared to the average duration of all calls across BT’s network.

Step 3: Calculating the costs of different call types under 2PC

2.16 The costs of different call types under 2PC was calculated using usage factors mainly for end to end call types provided by BT for the Network Charge Control modelling. These factors enable mapping between network components and call types, since the proportion of each component used by a call depends on the type of call. By applying the usage factors to each network component and summing the costs for the call type, the 2PC and averaged unit cost (ie the current arrangement) were calculated.

Step 4: Calculating the benefits of 2PC

2.17 Once the costs and average durations had been calculated, the possible welfare effects could be determined. The following paragraphs seek to provide an intuitive explanation of the possible welfare effects.

2.18 The welfare effects of 2PC manifest themselves slightly differently on relatively long and short calls. The welfare effects are measured at the retail level, i.e. using retail prices. The welfare analysis therefore implicitly assumes that the costs under 2PC are reflected in retail prices (see paragraph 3.8).

(i) Long calls

2.19 Under the current charging arrangements, calls that are longer than the overall average duration are inefficiently priced higher than the level they should be by the operator. Under 2PC, this price is brought down to cost (see note one below), which is likely to result in an increase in benefits to consumers (measured by consumer surplus) (see note two below). Two effects bring about this increase in consumer surplus:

• a transfer from the producer to the consumer; and
• a reduction in 'dead-weight loss' (ie the loss of welfare to society due to prices not reflecting costs).

2.20 In terms of this benefits analysis, only the reduction in the 'dead-weight' loss is actually measured, for reasons explained below.

2.21 The benefits analysis can be explained diagrammatically.

Figure 2.1: Possible welfare effects resulting from the introduction of 2PC to calls that are longer than the overall average duration

P1 =retail price of calls under average 1PC
P2 = retail price of calls under 2PC
Q1 = quantity of calls under 1PC
Q2 = quantity of calls under 2PC

2.22 In this situation, the average retail price of a longer duration call (eg a national call), decreases after the introduction of 2PC. The reduction in the network charges due to the introduction of 2PC is assumed to result in a reduction in the retail price, as shown above. It is this percentage reduction in the retail price that causes the welfare effects to come about.

2.23 Under the current arrangements, the consumer surplus is equal to area A, and the producer surplus (see note three below) is equal to area B. Area C is the dead-weight loss - it belongs to neither the producer nor the consumer and is therefore inefficiently lost due to the pricing above cost.

2.24 Under 2PC, the producer surplus is transferred to the consumer (area B), as is the dead-weight loss (area C). The consumer surplus is now denoted by A+B+C. The gain to society (which in this case is a gain to the consumer) is denoted by area C. This is the area that is being measured in the welfare analysis (see note four below).

(ii) Short calls

Figure 2.2: possible welfare effects resulting from the introduction of 2PC to calls that are shorter than the overall average duration

Referring to fig 2.2:

P1 =retail price of calls under average 1PC
P2 = retail price of calls under 2PC
Q1 = quantity of calls under 1PC
Q2 = quantity of calls under 2PC

2.25 In figure 2.2, the average retail price of a shorter than average duration call type is likely to increase (eg a local call).

2.26 2PC results in a price increasing from P₁ to P₂. However, since the new greater price now reflects the average cost, this is economically efficient compared to 1PC. This is because, under the current arrangements, prices are inefficiently low since they are below cost - and effectively 'subsidised' by longer call charges. As such, the producer incurs the cost of this subsidy, which is B + C + D.

2.27 A move to 2PC results in two things:

• a transfer from consumer surplus to producer surplus, as measured by areas B + C. This is reflected in the fact that consumers now have to pay a higher retail price; and
• a gain to the producer of area D, because it no longer has to incur this area of cost because the prices are now reflective of costs. Hence, area D is the welfare gain to the producer given a shift to 2PC. For the cost benefit Analysis, area D is calculated to determine the welfare gain for short calls.

2.28 To conclude, only the 'dead-weight' losses are measured for the welfare analysis, and not the transfers between consumers and producer. This is because, in aggregate, the transfers between consumers and producers should net out to zero (if both are given an equal weight when assessing welfare).

Step 5: Computing the welfare gains

2.29 Using the demand function, duration information, cost information and elasticity (see Annex A), it is possible to estimate the welfare gains explained above for short and long call types for the year 2001-02. This information can be extrapolated forward using an appropriate discount rate to calculate the benefits of 2PC over a number of years. Table 2.3 illustrates the results from this model.

Table 2.3: Benefits of 2PC on each call category over five years (central case)

Source: Oftel, note that discount rate used is 3.5% (see note five below)

2.30 The size of the benefit depends on the assumptions in the analysis, in particular the sizes of the set-up and duration charges under 2PC and the elasticities of demand of the different call types. The reported figure of £5 million reflects the central case used in the sensitivity analysis.

2.31 Oftel has undertaken three sensitivity analyses based on elasticity, average call duration and the cost split between set-up and duration. Annex B sets out these analyses. In summary, the sensitivity analysis results in a range of total cumulated discounted benefit of £2,995,758 to £8,377,153 based a margin of variation of parameters considered reasonable by Oftel.

2.32 This information set out in table 2.3 shows that, over a period of five years, there could be a welfare benefit in excess of £5 million.

Calculating the costs of 2PC

2.33 As discussed earlier, the benefits associated with 2PC must be weighed against the costs. The following paragraphs consider the direct implementation costs BT and other operators could face.

Direct costs incurred by BT

2.34 BT has commented that the implementation costs associated with 2PC are likely to be significant and is carrying out a feasibility study of the potential costs it would face. BT has given Oftel an estimate of the potential costs involved (see below). Oftel expects BT to provide further details when responding to this consultation.

2.35 In order to bill for calls under the current charging system, BT's wholesale billing system must first gather data on call volumes, disaggregated by call type. It must then rate this call volume data, which involves multiplying each set of call volume data by an appropriate unit charge. A move to 2PC would require the billing system to collect and rate volume data on call set-up as well as on call duration.

2.36 Oftel understands that BT's wholesale billing system already collects data on call attempts, and that this data is included in the bills passed to BT's wholesale customers. It is not clear, however, whether this data is of billable quality.

2.37 If the existing volume data on call attempts can be used for billing purposes, then it is likely that BT's direct costs associated with implementing 2PC would be relatively low. This is because the main changes required are likely to be confined to the rating modules within the billing system. If, however, the volume data on call attempts which is already collected is not sufficiently accurate, then more major changes might be required, and these are likely to extend to other systems associated with the handling of switch call data records (CDRs) and not just the wholesale billing system itself.

2.39 BT estimates that the costs it would incur could be as high as £2.5m in the first year of implementation and £500,000 a year thereafter. Over a period of five years this would total £4.5 million (undiscounted).

Direct costs incurred by industry

2.40 In many ways, the cost to industry of implementing 2PC would mirror BT's costs. For example, OLOs paying BT interconnection charges would need to ensure that their own billing systems were capable of collecting and rating data on successful call-attempts, in order to be able to validate BT's bills. It is likely that the costs would vary substantially between different operators depending, for example, on the current state of each operator's billing system, its scale and complexity and individual operator's decisions such as whether to introduce 2PC at a retail level.

2.41 Oftel has not attempted to collect this information on an operator by operator basis before issuing this consultation document, and would welcome detailed comments as part of the consultation.

2.42 At this stage, based on the magnitude of BT's potential costs, Oftel estimates that the total cost to all other operators, spread across both the larger and smaller market players, would be in excess of £8 million over a period of five years (undiscounted).

Q. Please can operators provide information on the magnitude and types of costs that they would expect to incur over a five year period? This information will form a key input to Oftel's final decision on this issue. In particular:

• What are the costs of changes to billing systems, and associated CDR-handling?
• What are the costs of changes to pricing systems?
• What are the costs of changes to accounting systems?

2.43 Based on the information available, Oftel estimates that the total costs to industry (BT – £4.5 million and other operators – £8 million) involved in implementing 2PC, were it to be required, could be in excess of £12 million over a period of five years (discounted, see table 2.4).

Conclusions

2.44 Despite the potential economic benefits of 2PC calculated earlier, an estimate of the direct costs associated with its practical implementation shows that they are likely to outweigh the benefits. Oftel estimates that the total net cost over a period of five years resulting from the implementation of 2PC would be in excess of £7 million, as shown in table 2.4.

2.45 The table below shows that the net present value is -£7 million.

Table 2.4: Estimate of net benefit

Source: Oftel estimate, benefits and costs are discounted at 3.5 per cent.

2.46 As highlighted earlier in this chapter, Oftel has not attempted to collect information on costs from operators on an individual basis before issuing this consultation document. Oftel would therefore welcome detailed comments on costs, as part of the consultation. This information will form a key input to Oftel's final decision on this issue.

Notes:

1. For the purposes of the analysis, it is assumed that the price is brought right down to cost. In reality, the retail price is nearer cost rather than actually at cost.

2. The welfare that consumers gain from consuming a certain number of units of a good.

3. The supernormal profits that producers gain when prices are above average costs.

4. This is based on the assumption that consumer and producer surplus should be given the same weight when assessing welfare. Typically Oftel would tend to place greater weight on consumer surplus. But in this analysis such a judgement is not important since (by construction) aggregate producer surplus across all call types is zero under the current arrangements and under 2PC, because the focus is on the effect of changing the structure of charges.

5. The discount rate is used to compare costs and benefits arising at different times. This is important because this proposal involves streams of costs and benefits that arise over a few years. Without discounting, the assessment process is distorted. The 3.5 per cent figure equates to the social rate of time preference (SRTP). HM-Treasury recently published this new discount rate in its Green Book, 16 Jan 2003.

Chapter 3

Policy issues

Introduction

3.1 Oftel accepts the theoretical benefits associated with two-part charging due to charges more closely reflecting costs. However Oftel also acknowledges there are likely to be significant costs associated with the implementation of a new charging arrangement. Chapter 2 shows that these costs may outweigh the benefits. In addition to these direct costs, there are a number of less quantifiable issues, which also need to be taken into consideration. In general terms, these include:

• the process of determining a two-part charge;
• the relationship between unmetered and metered Internet access;
• passing the benefits to consumers; and
• the regulatory, commercial and operational resources needed to adjust the interconnection regime and maintain consistency and transparency of charging across all regulated services.

(i) Process of determining a two-part charge

3.2 The division of costs between call set-up and call duration is unlikely to be straightforward. This view is supported by experience elsewhere in Europe. The table below shows the call set-up and PPM charges established in other countries that have adopted 2PC. The data below is taken from the Interconnect Atlas provided by Analysys (www.analysys.com/atlas/series/Call_origination.asp). It shows that the ratio between the call set-up charge and the duration charge for single tandem call origination under 2PC arrangements varies widely from 40 per cent (France) to over 400 per cent (Finland). Even if Finland is excluded, as an extreme case, there is a high degree of variability in this ratio. Based on its understanding of the relevant networks, Oftel believes it is unlikely that this variation reflects real differences in network design between the different countries listed here and is more likely to reflect the absence of uniformity in the treatment of costs. Uncertainty surrounding the most appropriate way to split costs between call set up and call duration leads to uncertainty in the magnitude of benefits that would be obtained from a move to 2PC. See annex B for sensitivity calculations on the magnitude of call set up and call duration charges.

Figure 3.1: Charging arrangements in other European countries

*UK data added by Oftel based on 0.401 ppc call set up and 0.189 ppm call duration charges, converted at exchange rate of 1GBP = 1.50872 EUR as at 23/01/03.

(ii) The relationship between metered and unmetered Internet access

3.3 As set out in chapter 2, benefits associated with FRIACO volumes are not included in the analysis undertaken by this review. This is because the FRIACO charge depends on the cost of an average circuit and does not depend on the ppm charges. Because it was not derived from the ppm charge for metered calls, a change in the ppm charge for metered calls does not imply that the FRIACO charge should also change.

3.4 Indeed this review is initially in response to representations made by operators carrying long-duration metered Internet calls. Since this type of call is, on average, much longer than a voice call, operators terminating a considerable amount of metered Internet traffic may find an interconnection regime based on 2PC favourable.

3.5 Certainly, the increase in the number of homes and businesses now online (42 per cent of homes in August 2002 compared to 25 per cent in May 2000) has significantly boosted the volume of Internet traffic being carried over BT's network. However, Oftel's market information shows that the significant growth in Internet volumes over the last two years has been in unmetered (ie FRIACO-based) traffic rather than metered traffic. As such, Oftel does not expect metered access to increase significantly in the future and does not consider that any future benefits are not being captured by this review's cost benefit analysis, which is based on current volumes. Indeed if metered Internet access volumes decline, the future benefits would be less than estimated.

Figure 3.2: PSTN dial-up Internet volumes

Source: Oftel estimate

Q. Do you consider that there will be a continued demand for long duration pence per minute Internet calls?

Q. What might be the impact on FRIACO of reduced charges for long-duration NTS calls?

(iii) Passing the benefits to consumers

3.6 When calculating the benefits in chapter 2, Oftel made the assumption that the introduction of 2PC at the wholesale level would be reflected at the retail level. The economic benefits of 2PC depend on consumers facing prices that accurately reflect costs. The extent to which operators pass on the benefits to consumers will depend, inter alia, on the competitive pressures that the operators face.

3.7 It appears likely to Oftel that competitive pressure in Internet call termination will be sufficient to drive cost savings on long duration calls through to the retail level (though this may not be the case where interconnection is based on 0845 number ranges which are, in most cases locked to local rate tariffs. This is due to the barriers ISPs may face if they choose to switch to 0844 number ranges in order to gain greater flexibility and the ability to offer their customers lower retail prices.) It also appears likely to Oftel that the higher costs of short duration voice calls will be reflected at the retail level.

Q. If 2PC is introduced at the wholesale level are operators likely to introduce 2PC at the retail level? Why?

Q. Are there other mechanisms that operators may use to reflect differences in wholesale costs for long and short duration calls at the retail level?

3.8 It is also worth considering the strong correlation between length of call and vulnerability of consumer. 78 per cent of households with an annual income over £30,000 access the Internet at home, compared to 23 per cent with an income under £17,500. This suggests that those on lower incomes are likely to be making calls of shorter than average duration. This is supported by a lower average quarterly telecoms spend by households in the lower income group – £64 compared to £104 by higher income households (Oftel's residential survey, Q11 November 2002). However, it is not clear whether this lower spend is the result of consumers making fewer calls, cheaper calls or shorter calls and whether it is a result of either choice or cost.

3.9 If the assumption that price changes caused by 2PC would be passed on at a retail level is right, it is possible that by introducing the change in charging, a mechanism will have been put in place that transfers welfare benefits from consumers with lower income levels to the comparatively better off.

Q. What might be the impact on consumers if price changes were passed through?

(iv) Regulatory resources

3.10 Oftel believes that if 2PC is adopted, then it must be adopted for all call types except FRIACO, for reasons explained in paragraph 1.8. A move towards 2PC would therefore require significant regulatory resource to adjust charges and charge controls across most of the interconnection regime. At a time of large regulatory change, the implementation of 2PC would undoubtedly consume scarce regulatory resources in Oftel/Ofcom, BT and in OLOs.

3.11 Given the scale of the changes introduced by the new EU Directives and the additional changes that will be introduced by the Communications Bill, Oftel feels that considerable weight must be given to the costs of the regulatory effort needed to introduce 2PC.

Q. How significant is the cost of regulatory and industry resources to implement 2PC?

Chapter 4

Conclusions

4.1 Without doubt, there are clear welfare benefits associated with 2PC, as set out in chapter 1. However, the direct costs to industry are also likely to be significant. Based on available information, Oftel's initial analysis suggests that no overwhelming net economic benefit can be associated with the implementation of 2PC. However, Oftel has not attempted to collect information on implementation costs from operators on an individual basis before issuing this review, and would welcome detailed comments on this as part of the consultation.

4.2 While 2PC has a sound theoretical basis and has been adopted by other countries, apportioning costs between call set-up and call duration does not appear to be a straightforward task. Examining the two part charges adopted by other countries, it seems unlikely that the high degree of variability in the ratio between the two costs reflects real differences in network design and, more likely, reveals the absence of uniformity in the treatment of cost allocation. Uncertainty about cost allocations leads to uncertainty about the magnitude of the benefits that will be obtained.

4.3 In addition, at a time of significant regulatory change, the implementation of 2PC would undoubtedly consume scarce regulatory resources in Oftel/Ofcom and across industry. In the absence of any overwhelming economic benefit, it is Oftel's initial view that wholesale changes to the current interconnection arrangements to implement a system of 2PC appear to be disproportionate.

4.4 However, if as a result of this consultation, Oftel concludes that a system of2PC is favoured, Oftel will publish a statement together with a programme of implementation under the new regulatory regime. Below is a draft implementation programme that outlines the likely time scales involved.

  Chapter 5

Consultation

5.1 The Director General seeks the views of interested parties on the proposals contained in this consultation document. Please supply your comments by 25 April 2003. All non-confidential responses will be published. The Director General will then publish a statement setting out his conclusions

5.2 Comments should be made in writing and where possible sent by e-mail to lucy.rhodes@oftel.gov.uk. However, copies may also be posted or faxed to the address below. If any stakeholders are unable to supply their comments in one of these ways, please use the contact details below to discuss alternatives.

Lucy Rhodes
Oftel
50 Ludgate Hill
London
EC4M 7JJ

tel: 020 7634 8836
fax: 020 7634 8738
e-mail: lucy.rhodes@oftel.gov.uk

5.3 Confidential responses should not be sent via the Internet. Responses will be published on Oftel's website in the Publications section under Responses to Oftel consultations except where respondents indicate that the response, or part of it, is confidential. Appointments to view written comments in Oftel's Research and Information Unit must be made in advance (see contact details below). Respondents are therefore asked to separate out any confidential material into a clearly marked annex. In the interests of transparency, respondents are requested to avoid confidential markings wherever possible.

Internet

5.4 This document is available on Oftel's web site at www.oftel.gov.uk. Oftel has a free e-mail based mailing list to help people stay informed about the work that Oftel is doing. Each time an Oftel document is published and placed on Oftel's website at www.oftel.gov.uk, subscribers to the list receive an e-mail informing them about the document. To register, please go to the What's New section of the website.

Alternative formats

5.5 Copies of this consultation document are available on disk. Accessible formats such as large print, Braille and audiocassette can be made available on request. Please contact the Oftel Research and Information Unit on 020 7634 8761 or by e-mail at infocent@oftel.gov.uk for more information.

The consultation criteria

5.6 Oftel considers that this document meets the Cabinet Office code of practice on written consultation documents. The code is reproduced below for convenience. If you have any comments or complaints about this consultation process please contact:

Oftel Co-ordinator for the code of practice:

Rob Jex
Oftel
50 Ludgate Hill
London
EC4M 7JJ

tel: 020 7634 5350
fax: 020 7634 8943
e-mail: rob.jex@oftel.gov.uk

Code of practice

5.7 The code of practice states that:

• timing of consultation should be built into the planning process for a policy (including legislation) or service from the start, so that it has the best prospect of improving the proposals concerned, and so that sufficient time is left for it at each stage;
• it should be clear who is being consulted, about what questions, in what timescale and for what purpose;
• a consultation document should be as simple and concise as possible. It should include a summary, in two main pages at most, of the main questions it seeks views on. It should make it as easy as possible for readers to respond, make contact or complain;
• documents should be made widely available, with the fullest use of electronic means (though not to the exclusion of others), and effectively drawn to the attention of all interested groups and individuals;
• sufficient time should be allowed for considered responses from all groups with an interest;
• responses should be carefully and open-mindedly analysed, and the results made widely available, with an account of the views expressed, and reasons for decisions finally taken; and
• departments should monitor and evaluate consultations, designating a consultation co-ordinator that will endure that all lessons are used.

Annex A

Calculating the effect of 2PC

A.1 This annex sets out the effect of moving to a 2PC arrangement for the following call types:

• local calls;
• national calls;
• fixed calls to mobiles;
• local non geographic calls;
• national non-geographic calls; and
• 'other' excluding FRIACO

A.2 The objective is to calculate the welfare effects achieved for the different call types through the introduction of 2PC. Benefits associated with FRIACO volumes are not included in the analysis. The FRIACO charge depends on the cost of an average circuit and does not depend on the ppm charges. A change in the ppm charge for metered calls does not therefore imply that the FRIACO charge should also change.

Step 1: Calculating the unit cost of each call component under 2PC

A.3 In order to calculate the effect of 2PC Oftel obtained information on the cost of call set up (pence per call) and duration (pence per minute) for the various network components of a call. Table 1 sets out the information that was provided by BT on set up and duration costs. It is based on costs disclosed in BT’s CCA financial statements for the year ending 31 March 2002.

Table A1: Set up and conveyance costs 2001-02

Source: BT financial statement 2001-02.

A.4 BT divided up the set-up costs and conveyance costs by using its equipment manufacturers' cost allocation matrix. This meant mapping each network element shown on the exchange contracts (ie pricing element schedule items) onto various main cost driver categories (eg call set-up, call duration, access etc) according to the manufacturers' guidance. As seen above, only the Local exchange concentrator, local exchange processor and main and digital junction switching call components are split between set up and duration. The rest of the components are wholly allocated to call duration.

A.5 BT has informed Oftel that it divided the relevant existing network components on the following basis:

A.6 Oftel intends to ask BT to explain what the equipment manufacturers' cost allocation matrix is and how it impacts the attribution of costs between call set up and duration.

A.7 This information was used to work out the unit set-up cost and conveyance cost for each network component of a call. The set up part of the call was worked out by dividing up the set-up cost per call component by the number of successful calls going through that element of the network. For the conveyance element of the call, this was done by dividing conveyance cost by the volume of calls (in minutes) going through that particular component. For both the set up and duration unit costs, an allowance for the cost of capital was included. This was derived by adding on 13.5 per cent of the mean capital employed for each network element and dividing it up by call volumes to obtain the unit cost. BT’s cost of capital is 13.5 per cent.

A.8 So, for example, to work out the unit cost for the local exchange concentrator under two part charging, using the figures in table A1:

Unit CCA call set up costs:
= (£70m/104283m)*100
= 0.067 pence per call

Unit CCA call set up MCE:
= (£240m/104283m)*100*13.5%
= 0.031pence per call<