5.1 In order to roll forward the outputs of the incremental cost work to the start date for the new controls, and in order to calculate the appropriate value of X for the control period, Oftel has constructed a financial model. This model is based on the outputs of the incremental cost work for 1994/95, 1995/96 and 1996/7, along with data supplied by BT. Forecasts of the level of BT's costs of providing network services at 1 August 1997 and during the period 1997-2001 are produced by the model. The forecasts depend upon a number of assumptions about key variables, notably:
5.2 Oftel has based its forecasts on broadly the same assumptions that it used for setting the retail price controls. This case is based on recent past trends, in which call volumes have been growing at a rate of 7.5% per annum over the last 2 to 3 years.
5.3 Oftel has argued that there could be scope for even higher call growth, perhaps as high as 9% annually, in future. BT has disputed this, arguing that such levels of growth are outside the feasible range for the next few years. Recently reported market information tends to contradict BT's view of what is feasible, however. In 1995/96 BT's inland retail (local and national) call volumes increased by 6% over 1994/95 levels, to be compared with a market growth rate for inland calls of around 8-9%. The equivalent growth in volumes over BT's inland conveyance network was 8.5-9%.
5.4 Oftel intends to evaluate the effects on the values of initial charges and X for the Network Charge Control of increases in market call volumes of between 7% and 10% per annum. Oftel will be interested to receive comments from respondents to this document on the appropriate assumption to make about market call volume growth.
5.5 All commentators expect BT's share of retail services to decline over the next price control period as competition increases. The outside forecasts which Oftel has seen forecast BT's market share in 2001 to be in a range between 60% and 70% for calls (in total) and between 70% and 80% for access lines. Oftel is considering the way in which changes of market share at the retail level and increasing infrastructure competition at the network level are likely to impact on BT's market share for network services.
5.6 The volume growth rates over BT's network might be higher or lower than the retail market growth rates for local, national and international calls discussed in the preceding section, given the growth in demand for other call types and BT's network market share.
5.7 This will depend, first, on whether market growth of other call types, notably of calls to and from mobiles and special rate (NTS) services, exceeds that for local, national and international calls. On the face of it this seems likely, which will tend to increase volumes over BT's network relative to those discussed in the preceding section.
5.8 Offsetting this will be the extent to which BT Network is likely to lose market share. This will differ by type of network service. For instance, BT is likely to lose a higher proportion of intertandem conveyance than local exchange segments and local-tandem conveyance, because intertandem conveyance is likely to be more competitive (as reflected in its price control treatment - see Chapter 2). Oftel will need to take these offsetting effects into account when deciding on a range of likely outcomes for volume growth over different parts of BT's network.
5.9 In setting the retail price controls, Oftel based the proposed value of X on assumptions about BT's efficiency which would enable it to match the most efficient operator of comparable systems by the year 2001. While benchmarking comparisons are very useful in assessing the scope for improvements in the efficiency of an individual operator, Oftel believes that a more detailed study of BT's operational efficiency is required for the setting of the Network Charge Control. A consultancy study of BT's operational efficiency in providing network services will be starting very shortly. The results of this study, along with the outputs of the incremental costs work, will be key factors in informing Oftel's assessment of achievable efficiency gains, and hence the value of X for the Network Charge Control.
5.10 The achievable rate of unit cost reduction is estimated separately for capital and operating costs. In the case of capital costs, some underlying assumptions are derived outside the financial model. In particular, inputs to the financial model have been provided by the work of the Incremental Cost Steering and Working Groups. This has included an assessment of a range of factors which relate to the efficiency of BT's capital assets. The Working Group has made assessments of utilisation rates of capacity compared with what was regarded by members of the Group as an achievable level. Oftel will take this information into account in assessing the level of inefficiency in BT's use of capital when setting the Network Charge Control. Assumptions about BT's volume growth and the projections of investment, depreciation and disposals of assets also impact upon the productivity of capital inputs and these factors are also incorporated in the financial model.
5.11 To the extent that BT's actually incurred costs incorporate inefficiencies in the utilisation of its capital assets, these will be squeezed out over time. The Network Charge Control will be set so as to allow BT to recover its actual relevant costs at the start of the period. BT's charges will be moved down to a level of efficiently incurred costs by the end of the period.
5.12 The study of BT's operational efficiency will be an important element in the assessment of BT's scope for reducing its unit operating costs. For the moment, the best evidence available to Oftel is based on the Benchmarking study, in which BT was compared with Local Exchange Carriers (LECs) in the United States, which was performed by consultants for Oftel for the retail price control review. This study suggested that the best performing Local Exchange Carriers are improving their efficiency, as measured by movements in real unit operating costs, by around 3% per annum. This includes reduction in costs resulting from increased volumes and economies of scale. This rate of cost reduction is therefore a benchmark for what is achievable if BT's volume growth matches that of the comparator group.
5.13 The benchmarking suggested that the gap between the unit operating costs of BT and the best performing Local Exchange Carriers was of the order of 5% in 1995/6. This efficiency gap was assessed by comparing the operating costs per unit of output of BT with the comparator group, after making appropriate adjustments for environmental factors. The resulting efficiency gap was further adjusted - downwards - by a statistical technique designed to make an appropriate allowance for observation and data errors.
5.14 The provisional targets for reductions in real unit operating costs are based on the findings of the benchmarking study which underlay the retail price control proposals, along with the most up to date outputs of the incremental costs work. As the results of the operational efficiency study become available, Oftel will move towards a firmer view about the appropriate target level for cost reductions.
5.15 Oftel has considered the cost and asset volume relationships forgrowth in the volume of calls over the network. Oftel considers that it is probable that reasonable values for both of these parameters lie in the region of 0.2 to 0.3. (That is, a 1% increase in calls would lead to an increase in costs and assets of between 0.2% and 0.3%). Oftel would welcome comments from respondents on the appropriate values for these parameters, and would particularly welcome any evidence respondents are able to make available to support their views.
5.16 It should be noted that the appropriate values for these parameters might be somewhat lower than those derived from a study of the long run incremental costs of providing different volumes of output. This is because what is required is an estimate of how costs and assets will move with volumes within the period of the price control.
5.17 Following the methodology adopted in the Retail Price Control review would lead to Oftel targeting a return on capital, on a nominal post tax basis, in the range 8-9% in its calculations of BT's projected financial performance.
5.18 Oftel has also computed its estimate of the difference between the post tax and pre tax costs of capital which is implied by the forecast of BT's tax payments. The effective tax rate is in the range 27%-29%.
5.19 Oftel is also considering, in the light of recent market developments, whether it would be appropriate to consider BT's cost of capital from a global perspective. As BT becomes a more globally oriented company, its cost of capital should perhaps be assessed by reference to returns across stock markets across the world, rather than within the UK. The parameters of the models which Oftel has used, including the Capital Asset Pricing Model (CAPM), could be adjusted to take account of global factors. In the case of the CAPM, it would be appropriate to incorporate a global equity risk premium and a value of beta that related the volatility of BT's returns to the volatility of global stock markets. Oftel will be considering, during the period of the Network Charge Control Review, whether, and if so how, to take account of these factors in deriving BT's cost of capital.
5.20 In line with the adoption of Long Run Incremental Costs as the appropriate costing methodology for the services covered by the Network Charge Control, Oftel has adopted an asset valuation methodology based on replacement cost. The starting point for the charges is derived from the incremental costs work, in which assets are valued on a forward looking basis.
5.21 An important determinant of the way in which costs will move over time is the change in value of the capital assets employed in the network. It is also important to consider how the costs of these capital assets are recovered over time, in terms of the appropriate asset lives and depreciation profiles.
5.22 The underlying principle, in the top down and bottom up incremental cost models, and in the Oftel financial model, is that assets should be valued on the basis of their Modern Equivalent Asset value. That is, assets are valued according to the lowest cost method of providing the services delivered by the assets in question. The lowest cost assets will generally incorporate the latest proven technology available.
5.23 The bottom up model calculates the recovery of capital costs on the basis of principles of economic depreciation. This is a methodology by which an asset is depreciated according to its earning power over its life, with the end of the asset's life coming when the earning power falls to zero. While this is conceptually the correct way to value assets and recover capital costs, the methodology requires a number of assumptions (eg about the future movements in asset prices and maintenance costs) in order to e implemented. These assumptions are difficult to forecast with confidence.
5.24 The top down incremental cost model uses straight line depreciation, with some allowance for holding gains and losses as assets change in price over time. This might approximate to the profile of capital charge recovery implied by economic depreciation and is rather simpler to implement. Whilst this type of accounting depreciation differs from the estimated economic depreciation for the bottom up model, the analysis conducted in the reconciliation exercise found no evidence of systematic bias. The financial model is based on a similar approach, with holding gains and losses incorporated in line with the assumptions made about changes in asset valuation. It is clearly important that similar assumptions about asset lives and asset price changes are made in the Incremental Cost Model outputs as are made in the financial model.
5.25 It should be noted that, while falling real asset prices will tend to generate high values of X (ie large reductions in unit costs), the need to recover the holding losses implied by such movements in prices also generates a higher level for the initial charges. That is, if asset prices are forecast to fall in real terms, the depreciation charge in the base year is set at a higher level in order to compensate the owners of the assets for the holding losses that are expected to occur.
5.26 Oftel is considering the capital costs involved in providing network services at a level of disaggregation where up to six different categories of assets are identified. These categories will include: local switching, main switching, duct and transmission. Oftel would be interested to receive the views of respondents about how they would expect the real prices of these assets to move over the period of the price cap.
5.27 Oftel has incorporated BT's projected capital expenditures into its financial projections. It has, however, made certain adjustments to these to reflect the fact that BT's nonregulated businesses may be the beneficiaries of some of this expenditure.
5.28 In Annex F there is an example of the printout of a simplified version of the financial model, along with a flow diagram setting out the key relationships and interactions within the model. The key assumptions (and parameter values) which determine the outputs of the model are, as set out in the earlier part of this Chapter:
5.29 The model is sensitive to changes in assumptions about these factors in the following way:
5.30 A 1% p.a. increase in labour productivity growth leads to an increase in the value of X of about 0.2.
5.31 The sensitivity to changes in network volume growth depends upon the parameter values for the asset volume and cost volume relationships. With asset and cost volume relationships set at a value of 0.25, each 1% p.a. increase in volume leads to an increase in the value of X o about 0.7. With both asset and cost volume relationships set at 0.5, each 1% increase in network volume leads to an increase in the value of X of about 0.4.
5.32 The impact of changing the parameter value for the asset volume and cost volume relationships depends upon the volume growth assumed in the particular run. At low volume growth, changes to these parameter values have little effect. However, at high volumes there is a considerable impact. At 10% volume growth, an increase in the asset volume relationship of 0.1 (eg increasing it from 0.25 to 0.35) would decrease the value of X by about 0.6. This relationship was estimated on the basis of an assumption that the cost volume relationship was unaltered.
5.33 At 10% volume growth an increase in the cost volume relationship of 0.1 leads to a decrease in the value of X of about 0.4, assuming the asset volume relationship is left constant.
5.34 Real asset price changes feed through into capital costs on a oneforone basis. That is, an increase in the rate of asset price decline of 1% p.a. would tend to reduce real unit capital costs by 1% p.a. Capital costs account for 55-65% of BT's relevant costs, so the effect of a 1% p.a. fall in capital input prices would be to increase X by around 0.6. However, as noted earlier, this effect is offset by the increase in the initial starting price which is implied by such an assumption, as the holding loss associated with asset price reductions would form part of the depreciation charge for setting initial prices.