Annex D

Current obligations imposed on SMP operators

Obligation to interconnect with Schedule 2 Public Operators

D.1 A licence condition requires BT and Kingston to offer to enter into interconnect agreements with other operators for the provision of interconnection services. It differs from the more general obligation placed on non-SMP operators that only requires them to negotiate interconnect agreements.

Requirement to meet requests for access other than from Schedule 2 Public Operators

D.2 A licence condition requires BT and Kingston to provide access to their systems to any other organisation running telecoms systems that is not a Schedule 2 Public Operator.

Requirement not to unduly discriminate

D.3 A licence condition requires BT and Kingston not to unduly discriminate.

Obligation to have cost oriented charges for interconnection services

D.4 A licence condition requires BT’s and Kingston’s interconnect charges to be cost-orientated (reasonably derived from the costs of providing the service based on a forward-looking incremental cost approach).

Network charge control regime

D.5 The Network Charge Control regime imposes charge controls on BT’s interconnection services. There are two categories of interconnection services that are subject to charge controls. The first is for services in markets in which there is already some competitive pressure that is likely to increase (known as ‘prospectively competitive services’). Such services are subject to safeguard controls set at RPI+0 per cent, which means that the price of the services is not allowed to increase by more than the rate of inflation. The BT services that are subject to a safeguard control are inter-tandem conveyance and inter-tandem transit.

D.6 The second category is for services in markets in which there is little or no competition (‘non-competitive services’). These services are subject to RPI-X per cent controls. The value of ‘X’ is set on the basis of a number of factors and the services are controlled in one of six separate charge control baskets to which the charge control relates. The baskets relevant to this market review are:

• call origination;
• interconnect specific services;
• single transit and local-tandem conveyance;
• flat rate internet access call origination at the local exchange; and
• flat rate internet access call origination at the tandem exchange.

Requirement to publish a Reference Interconnection Offer

D.7 A licence condition requires BT and Kingston to publish a Reference Interconnect Offer (RIO). It must be published every six months and include:

• a full list of the services to be supplied to Schedule 2 Public Operators including the charge for each service and details of terms and conditions; and
• a description of interconnection services to be supplied, broken down into components according to market needs.

Network Charge Change Notice regime

D.8 A licence condition requires BT and Kingston to notify changes to the charges they make for interconnection services. The period of notification depends on the service in question. BT is required to give 90 days’ notice of a change to a charge or the structure of the charge for non-competitive interconnection services and 28 days for prospectively competitive services and new services. For competitive services, BT is required to give one day’s notice before changing a charge or the structure of it. The Director may consent to a reduction of the notice period, if he considers that there has been a change in the level of competition in the relevant market or that such a change would benefit the development of competition in the long-term.

Requirement to send individual agreements to the Director and publish them

D.9 A licence condition requires BT and Kingston to send, on entering into an interconnection agreement, a copy of the agreement to the Director. BT and Kingston are obliged to publish the full text of the interconnection agreement unless the Director has determined that confidential information can be excluded.

Requirement to have cost accounting systems and accounting separation

D.10 Licence conditions require BT and Kingston to maintain cost accounting systems to demonstrate that their interconnection charges have been fairly and properly calculated and keep separate accounts for services provided to themselves and services provided to others.

Wholesale quality of service obligations

D.11 A licence condition requires BT to publish annually a set of statistics about its performance against a target performance. The performance targets are agreed between the Director and BT.

Prohibition on cross subsidies

D.12 A licence condition provides a power for the Director to direct BT and Kingston to take steps to cease cross subsidising parts of their business from profits made in other parts of their business. The condition also ensures that BT and Kingston keep records of material transfers between different parts of their business.

Code of practice on the confidentiality of customer information

D.13 A licence condition on BT ensures that information it gains from its wholesale business relating to the customers of other operators is not used to its own advantage.

Requirement to provide carrier pre-selection

D.14 A licence condition requires BT and Kingston to provide carrier pre-selection (CPS).

D.15 The Director has issued a number of determinations relating to BT's CPS offering:

• Charge for Per Operator Set-Up Facilities. This is the one-off charge made by BT to each operator wishing to become a CPS operator. The current charge was determined in Final Determination on costs and charges for Permanent Carrier Pre-selection, January 2001;

• Charge for On-Going Per Operator Activities. This is the annual charge made by BT to each CPS operator. The current charge was also determined in Final Determination on costs and charges for Permanent Carrier Pre-selection, January 2001;

• Charge for Per Customer Line Set-Up Facilities. This is the charge levied by BT for each new CPS order. It is variable and dependent on, for example, the type of order and how far an order progresses through BT's systems (so that an order which is rejected immediately will incur a smaller charge than one that is successfully completed). The current charges are set out in Final Determination on costs and charges for the provision of permanent carrier pre selection, September 2002; and

• The CPS surcharge BT applies to its wholesale call origination tariffs. This is the surcharge levied by BT on its wholesale call origination tariffs in order to recover the system set-up costs incurred by BT in the provision of CPS. The current surcharge and further details are set out in Final Determination of surcharges for the provision by BT of carrier pre-selection facilities, February 2002.

Requirement to provide wholesale line rental

D.16 A licence condition requires BT to provide wholesale line rental (WLR) services. WLR services are subject to a no undue discrimination requirement, requirements to notify changes to charges and a charge control.

Obligations relating to the Number Translation Services regime

D.17 In January 1996, Oftel determined a formula for the financial arrangements that should apply to Number Translation Services (NTS), where the call is originated on one operator’s network and terminated on another. The formula is:

Originating operator (ONO) retains: P - D + C

Terminating operator (TNO) receives: D – C.

P is the actual retail price charged by the originating operator to the customer;

C is the pence per minute charge for conveyance

D is the Deemed Retail Price

D.18 Oftel has since issued several determinations on various aspects of the regime. The principal areas that these determinations relate to are:

• the use of INCA and the Network Charge Differential (NCD) for calculating conveyance charges;

• the relevant retail costs that BT can recover for NTS calls (the retail uplift)

• the ability of operators and service providers to set the retail charges for calls to NTS numbers; and

• the ability for operators to opt out of BT's retail discounts.

Flat Rate Internet Access Call Origination (FRIACO)

D.19 A direction requires BT to offer wholesale Flat Rate Internet Access Call Origination (FRIACO) at the local exchange (DLE FRIACO).

D.20 A further direction requires BT to make ST FRIACO available from its tandem exchanges. The Direction contained a two-stage solution. During Stage 1, which ran from 26 February 2001 to 1 February 2003, there was a limit on the volume of ST FRIACO ports that each operator could purchase at any given tandem switch and BT was able to require traffic re-arrangements at tandem exchanges with insufficient capacity. From the start of Stage 2 on 1 February 2003, these restrictions were lifted, but the additional cost of any additional capacity necessary for Stage 2 must be borne by all operators who purchase ST FRIACO.

Annex E

Quantified charge control cost benefit analysis

This annex is available as a pdf document. Please click here to download

Annex F

NTS Retail uplift charge control methodology

Introduction

F.1 Option 3 in Chapter 14 proposes that a charge control is applied to the NTS retail uplift charge. This annex explains the proposed methodologies for the three alternative approaches, options 3A, 3B and 3C.

F.2 The charge control is based on an RPI-X model, ie this means that BT is allowed to increase the retail uplift charge by the inflation rate, minus an X factor. This annex seeks to calculate this X factor. The X factor represents cost saving achieved by BT over time, as a result of technical progress and economies of scale incurred by BT due to volume growth. Economies of scale imply that increases in volumes of NTS calls results in a reduction in unit (per minute) costs, because retail costs do not increase at the same rate as volumes. This economies of scale effect is represented in the cost volume ratio (CVR) used in the calculation. For example, if the CVR is 0.25, then if volumes rise by 100 per cent, total costs rise by 25 per cent and unit costs fall by 37.5 per cent. Use of this CVR in the retail uplift charge derivation is explained in annex G.

F.3 Oftel is consulting on three options for the charge control. Option 3A builds upon the retail uplift methodology set out in annex G, and projects it forward. Option 3B begins with the same retail uplift methodology, but phases into BT’s up to date Fully Allocated Cost (FAC) using a glide path. Option 3C uses the BT FAC allocation for the charge control base charge and target charge. The charge control is set to run for 4 years, from 25 July 2003 to 24 July 2007. This requires forecasts of data for five years ie from 2001/02 (actual volumes and base year cost data) to 2006/07, so that the charge can run up to 24 July 2007.

Option 3A – Retail uplift charge control fully based on current retail uplift charge methodology (1994/1995 cost allocation)

F.4 There are three steps in calculating the value of X based on this methodology which are set out below:

• Step 1: Establishing a base year;

• Step 2: Volume forecasts, assumptions for price reductions and efficiency savings over the controlled period; and

• Step 3: Projecting forward the retail costs.

Step 1: Establishing a base year

F.5 In order to devise an RPI-X charge control, it is necessary to establish a base year of cost. The base year data provides the foundation for forecast cost changes to be applied. The base year is the most recent year that actual volume data is available and is year zero of the charge control. The base year cost data is the allowable cost figure under the existing approach. In this case, it is the year 2001/02, that gives an average retail uplift figure of 0.19ppm as set out in annex G. In order to project the costs forward, it is necessary to project forward volumes and revenues along with assumptions about efficiency gains.

Step 2: Volume forecasts, assumptions for price reductions and efficiency savings over the controlled period

Volume forecasts

F.6 Oftel has forecasted BT’s originating metered NTS call volumes as follows, and the explanation of how they were derived is given below:

Table F.1: Originating NTS call volume forecasts for BT

F.7 Prior to this year, traffic data submitted to Oftel’s regular market information programme did not distinguish NTS ‘freephone’ traffic from FRIACO traffic. Consequently, it was difficult to determine recent growth trends in NTS volumes. As part of this review, however, BT has submitted separate information on FRIACO and NTS freephone volumes for the years 1999/2000 and 2001/02. Oftel is also continuing to collect this information via its market information program.

F.8 As Figure F.1 below shows, the large growth in non-geographic volumes reported by the market information over the period 1997-2000 was due to a significant increase in NTS traffic, reflecting the growth of pay-as-you-go Internet access. The introduction of unmetered Internet access has, however, led to a decline in NTS volumes over the last couple of years. There is also evidence of a slowdown in overall Internet PSTN traffic growth in recent months.

Figure F.1: PSTN Internet and NTS volumes

F.9 On this basis, it seems unlikely that BT originated NTS call volumes will increase over the next four years. On the other hand it seems unlikely that volumes will decrease at the same rate as in recent quarters. Oftel’s market research shows, for example, that the proportion of Internet subscribers using pay-as-you-go access has stabilised at around 30 per cent over the last year.

F.10 Oftel has thus forecast future NTS volumes on the basis that a very small downward linear trend will continue until 2006/07. The precise forecast methodology is based on a combination of linear extrapolation and exponential smoothing of data over the last 5 quarters. Essentially this predicts future values using the current trend but places more weight on the more recent observations. A five quarter period was chosen to remove any seasonal effects.

Efficiency

F.11 It is Oftel’s view that retail activities (except bad debt) are subject to technical progress so costs fall over time. This view is supported by actual data from 1994/5 to 2001/02, where Oftel estimated that the efficiency savings were 6.5 per cent p.a. ie BT’s real retail costs fell by 6.5 per cent on average each year excluding volume effects (annex G explains how this was calculated). It is necessary to take a view on the efficiency savings factor that must be applied in future years. Oftel believes that that it is reasonable to apply this historic efficiency factor to the charge control; Oftel assumes that real retail costs will continue to fall by 6.5 per cent p.a. over the four years covered by the charge control. However, Oftel have undertaken a sensitivity analysis on this figure.

Step 3: Projecting forward the retail costs

Treatment of retail costs except bad debt

F.12 The first stage of calculating X is to forecast the change in the NTS retail costs (except bad debt) over the controlled years in order to calculate a target cost for 2006/07. This is computed by undertaking two adjustments:

1. Incorporating the NTS volume increase over time in the cost projection via the CVR

F.13 The relevant formula is:

Formula A

C_t = C_t-1(1+CVR*Δvol(t, t-1))

Where:

C_t = costs in current year

C_t-1 = costs in the year before current year

Δvol(t, t-1) = per cent change in volume between current year and year before current year.

The above formula calculates the cost for year t, using the previous year’s cost figure.

2. Incorporating efficiency savings via the efficiency factor

F.14 In order to incorporate an efficiency effect, equation (a) must be multiplied by the following:

(1+e) where e is the annual reduction in real unit costs at constant volumes (estimated to be 6.5 per cent).

Treatment of bad debt

Revenue projections

F.15 Oftel considers that it is appropriate to separate out bad debt costs written off from all other operating retail costs because they track revenues rather than volumes ie a cost revenue ratio ('CRR') is appropriate in forecasting bad debt costs. As explained in the retail uplift methodology (annex G) Oftel takes the view that a CRR of 1 is a reasonable assumption because there are not likely to be any economies of scale in bad debts incurred. In the current methodology, bad debt is separated out in the 1994/5 baseline, and projected forward to 2001/02.

F.16 In order to forecast bad debt costs, it is necessary to forecast revenue growth over the controlled years. Since revenue = price*volumes, it is therefore necessary to forecast price changes over the relevant years in addition to the volume forecast above. Oftel has used the price change forecasts based on the forecasts for Local and National calls in the model underlying the 2001/02 Retail Price Control review. Special Local Rate and Special National Rate NTS prices are linked to Local and National price movements respectively, so Oftel considers that these price forecasts are a good approximation. Oftel is therefore proposing that NTS call prices are likely to decrease by 10 per cent per year in real terms. This forecast assumes a degree of re-balancing. Re-balancing implies that retail call prices will tend towards cost over time, whilst line rental will correspondingly increase over time.

F.17 Revenue must be projected forward to allow for projection of bad debt costs. Since revenue is a function of volumes and prices, it is forecasted using the following formula:

Formula B

R_t = R_t-1 ((1+Δvol (t, t-1))*(1+Δprice (t, t-1)))

Where R = revenues, using the same notation as above

F.18 The baseline revenue figure is extracted from confidential Market Information Statistics. A projected set of revenue figures allows for the bad debt costs to be forecasted using the cost revenue ratio.

Formula C

BD_t = BD_t-1*(1+Δrev (t, t-1)*CRR)

Where BD = bad debt, and using the same notation as above.

Computing X

F.19 For each of the controlled years, we now have the projected retail costs, bad debt costs and volumes. However, the cost figures for the base year and that for the target year of 2006/07, are used to determine the value of X. In order to work out the forecasted pence per minute retail uplift charge for the target year, the costs are added together and divided up by the forecasted NTS volume:

Formula D

Forecasted retail costs in 2006/07 + Forecasted bad debt costs in 2006/07

Forecasted volumes in 2006/7

In order to calculate X, the following formula is applied:

-X =

where P_t = the unit cost (RU charge) in the base year (2001/02)

P_t+5 = target year unit cost (2006/07) of 0.18ppm

F.20 This gives a value of X of 1.3 per cent, ie a charge control of RPI-1.3 per cent.

Sensitivity analysis

F.21 Oftel has undertaken a sensitivity analysis on each of the parameter values which are set out in the table below.

Table F.2

The results above indicate that the price assumption and efficiency assumptions are particularly sensitive. A smaller reduction in prices results in a negative value of X; i.e. unit costs are increasing the charge control would be RPI+0.3 per cent.

Option 3B: Retail uplift charge control with glide path from the current methodology (1994/1995 cost allocation) to up to date BT FAC

F.22 For option 3B, Oftel proposes to migrate the retail uplift charge based on the approach set out in Annex G, to the BT FAC approach of cost allocation, i.e. BT’s FAC costs in its current Financial Statements, over the period of the charge control, using a glide path.

F.23 The base year charge for the charge control would remain the same, i.e. it would be the same base charge proposed in Annex G and used for Option 3A. The target charge would be derived from 2001/02 BT FAC costs. This would allow a smooth adjustment or glide path from the current methodology to the BT FAC allocation for the retail uplift. In summary, the methodology for this variant of the charge control would be the same as that in the first option, but with a different target due to a different cost allocation.

F.24 There are five steps in calculating the value of X based on this methodology which are set out below:

• Step 1: Calculating a 2001/02 BT FAC cost allocation in order to calculate the target;

• Step 2: Separation of bad debt;

• Step 3: Forecasting geographic call volumes;

• Step 4: Forecasting the target cost allocation; and

• Step 5: Devising the "glide path" for NTS calls.

Step 1: Calculating a 2001/02 BT FAC cost allocation

F.25 This step is necessary in order to calculate the target cost figure. It is derived using cost allocations for OLO terminated NTS calls. This information has been provided to Oftel by BT. This is retail cost information for BT-OLO special local rate, special national rate and PRS calls. Normally, Oftel would add on a return on capital employed. But in this case, the capital employed is negative, so Oftel have not added a return.

Two adjustments are made to this cost allocation:

Inclusion of retail (origination) costs for BT terminated NTS calls

F.26 This adjustment is necessary to reflect the fact that retail costs for the BT terminated non-geographic calls are not present in these particular cost categories. This adjustment is made using confidential information on BT terminated NTS volume data. It must be noted that retail costs associated with termination of non-geographic calls are excluded from this cost allocation. The adjustment is undertaken by calculating the unit cost of NTS calls, and multiplying that figure by BT terminated NTS volumes. This resulting cost figure is added onto the initial cost allocation.

A. Unit cost = cost BTtoOLO NTS calls/volumes of BTtoOLO NTS calls

B. Unit cost * BTtoBT NTS calls = Cost for BTtoBT calls

Cost allocation = BTtoBT NTS cost allocn + BTtoOLO NTS cost alln

Removal of PRS bad debt

F.27 Oftel has removed bad debt associated with PRS calls because this is present in a separate PRS bad debt surcharge.

F.28 This gives rise to a final cost allocation of £135.8m.

Exclusion of irrelevant costs

F.29 The BT FAC methodology may contain costs that are not relevant to the retail uplift charge. (This was not the case for the 1994/5 cost baseline (as in Option 3A), because in that case, there was an allowance for all costs that were considered to be relevant for NTS calls, given that they were common between charge control services (geographic calls) and NTS calls.)

F.30 Costs that are not likely to be relevant to NTS calls are those which are neither incremental to NTS calls, nor common between retail services for NTS calls and other call types. They are likely to be incremental to retail services for calls other than NTS.

F.31 Oftel is in the process of obtaining information from BT to allow an assessment of any retail costs that are not likely to be relevant to NTS calls. However, for the purposes of this proposal, Oftel has not made any exclusions from the cost allocation, because it appears that significant exclusions have already taken place from the separate retail cost allocation for NTS. This separate allocation for retail costs to NTS calls was not available prior to December 2002. The proportion of marketing costs alone in the retail cost allocation is fairly small, although the proportion of marketing and sales costs together is significantly higher. Oftel will revise the cost allocation if necessary when further information has been obtained from BT.

Step 2: Separation of bad debt

F.32 Based on the same reasoning as explained in option 1, Oftel has separated bad debt costs out of the cost allocation, and inflated bad debt costs with a CRR. Oftel has estimated the bad debt percentage of total up to date costs in the following way. A breakdown of NTS retail costs from BT indicates the proportion of total costs accounted for by finance and billing in 2001/02 is 22 per cent. Historic data provided to Oftel by BT indicated that bad debt accounted for 67 per cent of finance and billing costs. Oftel will be updating this figure from BT in due course. This further indicates that the proportion of total retail costs accounted for by bad debt is 15 per cent.

Step 3: Forecasting geographic call volumes

F.33 This step is undertaken in the same way as option 3A.

Step 4: Forecasting the target cost allocation

F.34 The target non bad debt cost allocation is forecasted in the same way as in Option 1, i.e. using a CVR of 0.25, and an efficiency savings factor of 6.5 per cent. The bad debt costs are also forecasted in the same way, using the CRR of 1, and no efficiency savings. This gives rise to a total cost allocation figure for 2006/07, adjusted for economies of scale due to volume effects and efficiency savings.

Step 5: Deriving the 'glide path' for NTS calls

F.35 In order to derive the glide path from the starting year costs to the target year costs, it is necessary to convert the target year data into a unit (ppm) costs. The 2006/07 forecasted cost is divided by the relevant forecasted volume figure in order to obtain the unit cost, which is 0.30ppm. We now have the unit cost for the starting year (which is the current 2001/02 retail uplift charge) and the unit cost in the target year. The glide path consists of the calculation of X using these two cost figures. X is therefore calculated in the following way:

-X =

where P_t = the unit cost of NTS retail costs (current retail uplift charge) in the base year (2001/02) = 0.19ppm

P_t+5 = target year FAC unit cost (2006/07) = 0.30ppm

F.36 This gives a value of X of -10.2 per cent ie the price cap is RPI+10.2 per cent

Sensitivity analysis

Table F.3

Option 3C - Retail uplift charge control with base charge and target charge based on BT FAC allocation

F.37 For this option, the base charge for the charge control uses the FAC approach. The target charge is also based on the FAC approach and is the same as options 3B and 3C. This option completely departs from the cost allocation in 1994/5 and uses an assumed FAC retail uplift charge in 2001/02. Step 1 in Option 3B explains how the base year (2001/02) cost allocation is derived. This cost allocation is divided by NTS volumes in for that year:

£135.8m /(48,015 million minutes)*100 = 0.28ppm

F.38 Application of the formula gives rise to an X value of -1.9 per cent ie a charge control of RPI+1.9 per cent.

Sensitivity analysis

Table F.4

Comments on the magnitude and sign of X in each option

F.39 The growth in unit costs is a function of growth in total costs less growth in volumes:

Growth unit costs = f[(growth in total costs) – [growth in volumes)]

F.40 The impact of volume growth on total costs is complicated. It is clear though that falls in volumes reduce operating expenditure due to the link to the cost volume relationship, and falls in volume (and price) reduce revenues, which are fundamental to the fall in bad debts.

F.41 In option 3A, unit costs are decreasing. This is because there is a change in total costs of -30 per cent, and change in volumes of –25 per cent. Hence by definition unit costs are decreasing.

F.42 Option 3B gives rise to higher retail uplift charges over the controlled years, because the cost base is changing from a 1994/5 allocation to a 2001/02 FAC allocation. The latter allocation has a greater amount of retail costs allocated to NTS calls. Therefore, over the controlled years, the charge will increase and a positive value of X in the RPI-X formula will result.

F.43 In option 3C, the unit retail uplift charge increases over the controlled years, because reduction in total costs is outweighed by the reduction in volumes. Volumes have reduced by 25 per cent, whilst total costs have reduced by 18 per cent. This gives a positive increase in unit costs.

 Annex G

NTS Retail uplift base charge – Options 3A and 3B

Introduction

G.1 In chapter 14, Oftel sets out its proposal for a charge control on the NTS retail uplift. This annex sets out Oftel's proposed methodology for the base charge used for charge control options 3A and 3B.

G.2 Oftel proposes to use the same methodology, including the using same baseline year, for the charge control base charge as has been used in its directions for the years 1999/2000 and 2000/01. In summary, the retail uplift is calculated on the basis of the sum of recoverable costs in the baseline year plus the change in costs incurred as a result of volume growth over the intervening years, inflation, and efficiency changes by BT.

Scope of recoverable costs

G.3 The retail uplift allows BT to recover a properly attributed portion of the retail costs that it incurs in order to satisfy the requirement for the charge to be LRIC plus an appropriate mark-up. BT has three main categories of direct retail costs. These are customer service, finance and billing and marketing and sales. These cost categories are supported by indirect retail costs, such as computer equipment and accommodation. Oftel considers that it is appropriate to include all three categories in the retail uplift.

G.4 Customer service and finance and billing costs are likely, to a certain extent, to be incremental to NTS services. The majority of marketing and sales costs are likely to be common between geographic calls, which were regulated under the 1996 Price Control Review and NTS services. These are likely to be general call stimulation campaigns that affect all calls, not just geographic calls.

G.5 Oftel has identified BT’s relevant retail costs (common between geographic and NTS calls) as:

• finance and billing costs, 66.9 per cent which are due to retail bad debt, derived from an average figure over the period 1997/98 to 2000/01 from information provided by BT;

• customer service;

• marketing – aimed at getting more people connected to BT in the UK;

• marketing – aimed at increasing call revenue;

• billing enquiries;

• fault report;

• complaints; and

• indirect retail costs.

Recoverable costs in baseline year

G.6 The recoverable costs for the retail uplift in the baseline year 1994/95 are derived from the total allocation of relevant retail costs. The total allocation of relevant retail costs is given in the local and national call category of the Retail Systems Business (RSB) for 1994/5.

G.7 These relevant retail costs are common between services included in the 1996 Price Control Review ('PCR') and NTS calls. When the 1996 PCR was set, there was an implicit allocation of BTs retail costs between geographic and NTS calls on the basis of relative volumes. Therefore, in order to estimate the baseline figure for NTS retail costs, the unit retail cost figure is obtained by dividing the total retail costs (i.e. those attributable to PCR and NTS services), by total volumes (Local, National and NTS Volumes) and multiplying this by BT’s NTS call volumes in 1994/95. BT terminated NTS costs are present in the 'Other' category of the RSB. The baseline is therefore suitably adjusted to allow for retail costs (for call origination not termination) of these calls. This is undertaken using confidential volume data on BT terminated NTS calls.

G.8 During the time of the 1996 price control review, a lack of consideration of the price control in the retail uplift methodology would result in BT over-recovering retail costs overall, by them being recovered in the price control and then again in the retail uplift charge. This would happen because the enormous growth in NTS volumes would, under BT’s FAC methodology of cost allocation, lead to a re-allocation of costs towards NTS calls. This would result in a charge that is not cost orientated, because the combinatorial test would not be satisfied. The combinatorial test requires that the revenue from any combination of services covers the common costs among those services as well as the incremental cost of each service. If the test is satisfied for all groups of services, it avoids under or over recovery of common costs.

G.9 An allowance for Return on Capital Employed (ROCE) is added onto the baseline figure for total retail costs to allow BT to earn a reasonable return for providing retail services. This is the cost of capital multiplied by the mean capital employed.

G.10 During its review of the price cap on calls to mobiles in 1998, the Monopolies and Mergers Commission (MMC) revised the distribution of the retail costs of calls. In doing so, it reduced the costs that were attributed to calls to mobiles, for the year 1997/98. These costs were then re-allocated across local and national calls (including BT to Operator NTS calls), giving an increase in total costs of 9.7 per cent. Since these costs are common to NTS calls as well, the baseline of the NTS retail uplift calculated above has been increased by 9.7 per cent.

Projecting baseline costs to 25 July 2003

G.11 As the baseline costs are derived from 1994/95 costs, Oftel has to estimate how the attributable costs would change as a result of efficiency changes by BT, inflation and the huge increase in NTS call volumes over the intervening period, which, because of economies of scale, led to a reduction in unit costs. In particular, largely due to the take up of dial-up internet access using NTS numbers, the volumes of NTS calls increased by 2,558 per cent during the period from 1994/1995 to 1999/2000. In order to update the 1994/5 baseline to derive a retail uplift charge applicable from April 2001 to March 2002, and a further update from April 2002 to March 2003, Oftel has used BT originated NTS volume data provided by BT, which excludes all the volumes for unmetered services. Specifically, for the charge applied from April 2001, Oftel uses volume data for the financial year 2000/01. For the charge applied from April 2002, Oftel has used volume data for the financial year 2001/02. The relevant years’ costs and volume data are also used to update the efficiency part of the calculation, as discussed below.

G.12 Since the volume of calls has increased, it is reasonable to expect that this causes costs to be higher than would otherwise be the case, i.e. a positive cost volume relationship (CVR). However, Oftel does not believe that costs would have increased on a one to one ratio with volume (ie: so that a 100 per cent increase in volume would lead to a 100 per cent increase in costs, giving a CVR of 1).

G.13 Oftel recognise that retail costs in this sector exhibit substantial economies of scale. On the basis of the work performed and assumptions made in the PCR, Oftel has adopted the view that a volume increase of 100 per cent would entail an increase in costs of 25 per cent. This results in a CVR of 0.25. Oftel believes that this is a reasonable value for all costs except for bad debt which it treats slightly differently.

G.14 Oftel believes that costs for bad debt more closely track revenue movements than volume movements. Accordingly, Oftel considers that a cost revenue relationship ("CRR") rather than a CVR should be applied to bad debt. Oftel takes the view that the CRR is 1 (ie: costs of bad debt increased directly with revenues on a 1 to 1 ratio). An adjustment is made for freephone services because there is no bad debt for freephone. PRS calls are high value calls, implying that they incur a greater level of bad debt compared to normal NTS calls. Oftel has removed all this 'excess' bad debt from the retail uplift calculation, and applied a surcharge to PRS calls, to ensure allow BT to recover this excess bad debt from PRS calls only.

G.15 Oftel finally makes adjustments for inflation and efficiency (independent of volume changes) recognising that BT’s Retail Systems Business ('RSB') has become more efficient since 1994/95 so that its retail costs had reduced. These efficiency savings can be attributed to technical progress in retail activities ie the discovery of lower cost methods of undertaking the same retail functions. More generally, this is why one can observe that, although volumes of all call cateogories have increased over the time period, total retail costs have reduced, because the efficiency effect that reduces costs outweighs the volume and revenue induced cost increases. Oftel has calculated an annual efficiency factor for retail costs for Local, National and NTS calls using cost and volume data for these calls.

G.16 In order to determine the efficiency adjustment, the following formula is applied:

c = c (1 + CVR*% Δ in call volumes)(1+x)

(1+x)=

where:

x = efficiency savings ie year on year gain in underlying efficiency net of volume effects.

C_t = costs at end of relevant period

C_t-1 – costs at beginning of relevant period

G.17 This calculation gives an average efficiency savings factor of roughly six per cent (it varies slightly for different years) ie retail costs, independent of volume changes decrease by six per cent per annum. This efficiency savings factor is applied to all calls except bad debt. Oftel takes the view that there is limited scope for additional efficiency savings in bad debt expensed.

G.18 Oftel considers that this methodology enables the calculation of charges which are meet the requirement for charges to be LRIC plus an appropriate mark up. The charge covers any costs incremental to NTS and a proportion of common costs, allocated initially by volume (ie in the baseline calculation).

Initial conclusion

G.19 Based on the methodology described above, the base year charge for charge control options 3A and 3B is 0.19ppm.

Annex H

Bad debt surcharge for calls to Premium Rate Services

H.1 This annex explains the methodology used to calculate the PRS bad debt surcharge. Note that the Director has recently consulted on the PRS bad debt surcharge under the current regulatory framework and is still considering some issues that have been raised in response to that consultation.

H.3 With the PRS bad debt surcharge there are two cost types which have been treated in this way. These are the extra bad debt expense and the extra financing costs of working capital associated with an average PRS call, over and above the fixed amount recovered within the NTS retail uplift for a local or national rate call.

H.4 The extra bad debt expense can be further subdivided between two distinct factors:

• PRS calls on average have a significantly higher pence per minute net retail price than local or national rate calls - the 'price factor’. For example, if a person made 100 minutes worth of local calls charged at 4ppm, and the person defaulted on paying, the bad debt would be £4. However, if a person made 100 minutes of PRS calls charged at 50ppm and defaulted, the bad debt would be £50; and

• PRS calls, as compared with local or national rate calls, may experience a different level of bad debts expressed as a percentage of relevant turnover due to the fact that the customers who make these calls are more (or less) likely to default on payment – the 'incidence factor’.

The ‘price factor’

H.5 To calculate the ‘price factor’ of the cost of bad debt for PRS calls, it is

necessary to remove the standard bad debt allowance from the overall PRS bad debt cost figure. The standard bad debt refers to that incurred by any general call type:- it excludes the excess bad debt incurred by a PRS call. This adjustment is necessary because the bad debt figure includes the excess PRS bad debt and standard NTS bad debt (which is covered by the retail uplift charge). This adjustment will therefore eliminate double counting of standard bad debt, which should only feature in the retail uplift charge (see diagram below).

Figure H.1

H.6 The standard bad debt is derived by determining the proportion of the retail uplift pence per minute figure that comprises bad debt. This figure is then expressed as an average percentage of the PRS retail call price. This percentage figure is subtracted from the percentage ratio of bad debt to turnover as follows:

x% = PRS bad debt cost figure (£s) expressed as a percentage of total PRS retail revenue (£s)

y% = standard bad debt in retail uplift (pence per minute) expressed as a percentage of the PRS retail price (pence per minute)

z% (the bulk of the PRS bad debt surcharge) is the difference between x% and y% ie x% minus y%

Extra working capital associated with PRS calls

H.7 Oftel understands that on average, BT receives call revenues from retail customers three months in arrears and makes a payment to the terminating operator, known as the POLO, two months after they have been incurred. This leaves a month in which BT must finance its debtors after it has paid its creditors. This implies that BT is incurring an opportunity cost due to this timing factor. The greater the POLO and the shorter the timing gap, the lower the liability. So, for example, if the POLO was 100 per cent of the retail price ie all retail revenue was payable to the OLO and this was paid after three months, BT would incur no liability because its outpayments would exactly match its receipts both in terms of timing and value. If the POLO was a very small percentage of the retail price and had to be paid immediately, BT would incur maximum liability of three months before it received the revenue. In other words, BT is incurring a higher cost of working capital – this is the finance that BT must have in place in order to finance its own costs and payments to other operators before it receives payments from its retail customers.

H.8 An allowance has to be made for BT to recover this extra cost. Using information provided by BT about the POLO as a percentage of the retail price, Oftel has derived the net capital employed associated with the call. This figure is adjusted by the working capital from a standard call type in order to prevent BT from over recovering costs.

H.9 The following BT information (with some adjustments made to reflect current

circumstances) is used to calculate the extra working capital associated with PRS calls (actual numbers are, in most cases, confidential):

• PRS call revenues and volumes for 2000/01;

• Average retail PRS price (derived from above);

• SNR (special national rate) and SLR (special local rate) call revenues and volumes for 2000/01;
• Average retail price for NTS Local and National calls;

• Net retail price (ie average PRS price minus average NTS Local and National price);

• Proportion of net retail PRS price (as defined above) made up by POLO;
• Proportion of net retail SNR and SLR retail price made up by POLO;

• Billing cycle = quarterly (every three months for customers not paying by direct debit);

• Bill preparation time = three days;

• Bill payment date = one month after receipt (for customers not paying by direct debit);

• Interconnection payment cycle = one month; and
• Percentage of BT’s customers on direct debit or Monthly Payment Plans (source: 2001/02 Price Control Review data)

H.10 The following calculations show the methodology used to calculate the extra working capital associated with PRS calls. It is worth noting that the extra cost associated with financing working capital represents a minimal proportion of the overall PRS bad debt surcharge. The bulk of the surcharge is caused by the calculation of ‘z per cent’ as outlined in paragraph H.6 above.

A) The ‘income debtor’ (the retail customer) is calculated as follows:

Table H.1: Calculation of income debtor

The total income debtor is applied to the retail price to obtain a debtor figure as follows:

Let x = % of customers on direct debit or monthly payment plan

x*0.6 + (1-x)*2.6 = y

y/12 x net retail price (as defined in paragraph 4.19) = (a) pence

B) The ‘POLO creditor’ (i.e. the creditor balance arising from payments owed to the terminating operator) is calculated as follows:

Table F.2: Calculation of POLO creditor

The POLO creditor is therefore 2/12 x (proportion of net retail price made up by POLO) = (b) pence

C) Net capital employed per call minute = (a) – (b) = (c) pence

This figure is multiplied by PRS call volumes to obtain the additional working capital for PRS calls ie:

(c) pence x PRS call volumes = £d

D) To calculate BT’s opportunity cost, a cost of capital of 13.5 per cent (this is BT’s cost of capital as used in the Director’s statement of February 2001 entitled "Network Charge and Retail Price Controls from 2001") is applied to C) above:

13.5% x £d = £e

E) This figure is divided by retail revenues for PRS bad debt to obtain a percentage figure:

£e / retail revenues = f%

H.11 The two percentage figures obtained from the calculations set out in paragraphs 4.16 and 4.20 are added together to obtain the level of the PRS bad debt surcharge based on the higher price of these calls :

The ‘incidence factor’

H.12 Information provided by BT suggests that the incidence of bad debt, as a percentage of turnover net of discounts, is not materially different when contrasting PRS calls with Local and National rate calls. Therefore the ‘incidence factor’ has no impact on the calculation of the PRS bad debt surcharge.

Initial Conclusion

H.13 The Director's initial view is that the PRS bad debt surcharge should be 2.05 per cent.

Annex I

List of questions

Chapter 4 Wholesale fixed narrowband exchange line services

Question 4.1: Do you agree with the market definitions?

Question 4.2: Is there evidence that might support alternative market definitions?

Question 4.3: Do you agree with the SMP criteria used?

Question 4.4: Do you agree on the assessment of SMP?

Question 4.5: Do you have any comments on future developments that might affect these assessments?

Chapter 5 Call origination

Question 5.1: Do you agree with the market definitions?

Question 5.2: Is there evidence that might support alternative market definitions?

Question 5.3: Do you agree with the SMP criteria used?

Question 5.4: Do you agree on the assessment of SMP?

Question 5.5: Do you have any comments on future developments that might affect these assessments?

Chapter 6 Local-tandem conveyance and transit

Question 6.1: Do you agree with the market definition?

Question 6.2: Is there evidence that might support an alternative market definition?

Question 6.3: Do you agree with the SMP criteria used?

Question 6.4: Do you agree on the assessment of SMP?

Question 6.5: Do you have any comments on future developments that might affect this assessment?

Chapter 7 Wholesale transit services

Question 7.1: Do you agree with the market definitions?

Question 7.2: Is there evidence that might support alternative market definitions?

Question 7.3: Do you agree with the SMP criteria used?

Question 7.4: Do you agree on the assessment of SMP?

Question 7.5: Do you have any comments on future developments that might affect these assessments?

Chapter 8 Interconnection circuits

Question 8.1: Do you agree that it is necessary and appropriate to regulate interconnection circuits as part of a package of remedies to address BT and Kingston’s market power in the assessed markets?

Question 8.2: Do you agree that it is necessary to regulate each type of interconnection circuit?

Chapter 9 Approach to remedies

Question 9.1: Do you agree that in general terms, ex-ante regulation is justified in the markets in this review where SMP is proposed?

Chapter 10 General remedies

Question 10.1: Do you agree that a requirement on BT and Kingston to provide network access on reasonable request should be imposed in the markets proposed?

Question 10.2: Do you agree that a requirement not to unduly discriminate should be imposed on BT and Kingston in the markets proposed?

Question 10.3: Do you agree that the basis of charges should be LRIC plus appropriate mark-up and allowance for return on capital employed for the markets proposed?

Question 10.4: Do you agree that no ex-ante pricing rule is necessary for ISDN30?

Question 10.5: Do you agree that a charge control in the form of RPI-X should be imposed on BT in the markets proposed?

Question 10.6: Do you agree that the values of X should be as set out in the proposed condition (ie the current values of X)?

Question 10.7: Do you agree that Kingston should not be subject to a charge control?

Question 10.8: Do you agree a requirement to publish a reference offer should be imposed on BT and Kingston in the markets proposed?

Question 10.9: Do you have any views on what a reference offer should contain?

Question 10.10: Do you agree that changes to charges should be notified in advance - 90 days notice in markets where SMP is persistent and 28 days notice in markets where competition is more developed?

Question 10.11: Do you agree that a requirement to notify technical information should be imposed on BT and Kingston in the markets proposed?

Question 10.12: Do you agree that 90 days is a reasonable period for notification of new or changed technical information in the markets proposed?

Question 10.13: Do you agree that a requirement to consult on interfaces is no longer appropriate?

Question 10.14: Do you agree that a requirement to publish a set of KPIs, subject to the detail being agreed after consultation, should be imposed on BT in the markets proposed?

Questions 10.15: Do you have any views on the key areas where KPIs might be required in the markets proposed?

Question 10.16: Do you have evidence and comments on the extent to which the SOR process has problems in the proposed markets? As noted above, Oftel has received earlier representations with respect to the SOR process between operators and BT. Stakeholders are also asked to comment on the SOR process between Kingston and operators.

Question 10.17: Do you have comments on the outline proposals for regulation of the SOR process set out above?

Chapter 11 Carrier pre-selection

Question 11.1: Do you have any comments on the proposed CPS condition set out in Annex A?

Question 11.2: Do you have any comments on the proposed direction set out in Annex B?

Chapter 12 Indirect access

Question 12.1: Do you have any comments on the proposed IA condition set out in Annex A?

Chapter 13 Wholesale line rental

Question 13.1: Do you agree that BT should be required to provide a fit for purpose WLR product for analogue, business ISDN2 and ISDN30?

Question 13.2: Do you have a view on whether the proposed conditions for Option 1 or Option II should be imposed on BT?

Question 13.3: Do you have any comments on the proposed provisions for cost recovery?

Question 13.4: Do you have any comments on the draft directions set out in Annex C?

Chapter 14 NTS call origination

Question 14.1: Do you agree that NTS call origination should be specifically regulated?

Question 14.2: Do you have any comments on the range of options that Oftel has considered?

Question 14.3 Do you agree that a cost oriented retention is an appropriate remedy for promoting competition in downstream markets?

Question 14.4: Do you consider that a charge control for the retail uplift as opposed to annually setting charges would be beneficial?

Question 14.5: What are your views on the relative merits of the methodologies considered for the charge control?

Question 14.5: In relation to Kingston, do you consider that the current unregulated arrangements are preferable to the introduction of regulation?

Chapter 15 Flat Rate Internet Access Call Origination (FRIACO)

Question 15.1: Do you have any comments on the range of options considered?

Question 15.2: Are there any additional advantages or disadvantages of the options that Oftel has considered?

Question 15.3: Do you agree with the Director's initial conclusion on the appropriate regulatory option for both BT and Kingston?

Chapter 16 Cost accounting and accounting separation

Question 16.1: Do you agree that cost accounting and accounting separation requirements should be imposed in the markets proposed?

Annex J

Glossary

ADSL (Asymmetric Digital Subscriber Line): a digital technology that allows the use of a copper line to send a large quantity of data in one direction and a lesser quantity in the other.

Analogue: the direct representation of a waveform, as opposed to digital, which is a binary coded representation.

Barriers to entry: an additional cost which must be borne by entrants but not by firms already in the industry; or other factors, which enable an incumbent to maintain prices above the competitive level without inducing entry.

Broadband: a service or connection allowing a considerable amount of information to be conveyed. Generally defined as a bandwidth greater than 128kbit/s.

BT: British Telecommunications plc.

Cable modem: a cable modem is a device that enables a consumer to access the Internet via a cable line

Channels: an exchange line consists of one or a number of channels, each having a bandwidth of 64 kbit/s.

Combinatorial test: a test to be applied on a combination of services where there are common costs between services. The revenue from any combination of services would need to cover the common costs between the services as well as the incremental cost of each service.

Communications provider: a person who provides an Electronic Communications Network or provides an Electronic Communications Service.

CVR (Cost Volume Ratio): the relationship between movements in volumes of say call minutes with the underlying cost of provision

Dedicated port: A connection, typically on a Network Access Server or telephone exchange, which is reserved for use by a single (wholesale) customer, for example an ISP.

Dial-up Internet access: Internet access that uses a dial-up connection over an analogue or ISDN telephone line.

Digital: the binary coded representation of a waveform, as opposed to analogue, which is the direct representation of a waveform.

Digital Local Exchange (DLE) and Local exchange: the telephone exchange to which customers are directly connected, often via a remote concentrator unit.

Direct Access: the situation where a customer is directly connected to a telecommunications operator’s network by a fixed link.

DMSU (Digital Main Switching Unit): a tandem exchange primarily used for connecting calls between DLEs.

DSL (Digital Subscriber Line): a family of technologies generically referred to as DSL, or xDSL, capable of transforming ordinary phone lines (also known as "twisted copper pairs") into high-speed digital lines, capable of supporting advanced services such as fast Internet access and video-on-demand. ADSL (Asymmetric Digital Subscriber Line), HDSL (High data rate Digital Subscriber Line) and VDSL (Very high data rate Digital Subscriber Line) are all variants of xDSL.

DSLAM (Digital Subscriber Loop Access Multiplexer): apparatus sited in the same exchange building as is used to terminate DSL enabled copper loops, which comprises a bank of DSL modems and a multiplexer which combines many customer lines into one data path.

Exchange line: the telephone line that connects the customers’ network terminating point to the local exchange.

Fully Allocated Costs (FAC): an accounting method for attributing all the costs of the company to defined activities such as products and services. Typically this method would follow the principle of cost causality.

FRIACO (Flat Rate Internet Access Call Origination): the provision of Flat Rate Internet Access Call Origination via a wholesale unmetered Internet access product from BT.

DLE FRIACO: digital local exchange FRIACO. The provision of Flat Rate Internet Access Call Origination via a wholesale unmetered Internet access product from BT at the local exchange.

ST FRIACO: Single Tandem FRIACO. The provision of Flat Rate Internet Access Call Origination via a wholesale unmetered Internet access product from BT at the tandem exchange.

HDSL (High data rate Digital Subscriber Line): one of the earliest forms of DSL services to be widely used. It is symmetrical, offering the same data rates upstream and downstream. The maximum data rate is however lower than that for ADSL.

Hull Area: the area defined as the 'Licensed Area' in the licence granted on 30 November 1987 by the Secretary of State under section 7 of the Telecommunications Act 1984 to Kingston upon Hull City Council and Kingston Communications (Hull) plc.

Indirect Access: where a customer establishes a connection with a particular operator’s network by dialling a short code to switch through the network on which his exchange line terminates. Such calls are usually billed by the Indirect Access operator.