7.1 This chapter sets out Oftel's proposals about how the price controls for the period after 1997 should be set. The chapter is divided into three sections. They deal in turn with the cost of capital; the regulatory asset base and the windfall gain issue identified in the first consultative document, and the financial model used by Oftel. Indicative ranges of X are specified for retail and network caps and a quantification is set out of the effect of varying assumptions on the computed values of X.
7.2 In response to the proposals set out in the first consultative document, Oftel received a number of comments about the issue of how to estimate the cost of capital for BT in order to set the parameters of the price control for the forthcoming period. While consumer groups were generally content with the approach proposed, a number of respondents from the telecommunications industry and from the investment community were critical of the proposals.
7.3 Some respondents criticised Oftel's intention to set price control limits in a way that constrained returns to equal an estimate of the cost of capital. It was suggested that, in the telecommunications market from 1997, such an approach to regulation would be inappropriate, and more reliance should be placed on market forces to control prices. The commentators from the telecommunications industry and the financial community also suggested that the higher level of actual and prospective competition in the telecommunications market meant that the cost of capital for BT should be set at a level rather greater than the levels allowed by other regulators and the MMC in relation to utilities not subject to the same degree of competition.
7.4 Specific criticisms were also made of Oftel's choice of parameters for computing BT's Weighted Average Cost of Capital by use of the Capital Asset Pricing Model. Respondents argued that the values of Beta suggested by Oftel were too low, as was the equity premium proposed by Oftel. It was noted that Oftel's suggested range for the equity premium was lower than the historic average. Some respondents, notably BT, criticised the taxation adjustments proposed by Oftel if it were necessary to estimate a pre-tax Weighted Average Cost of Capital.
7.5 Oftel has carefully considered the responses to its proposals for computing BT's cost of capital. In relation to the argument that it is inappropriate to set prices in a way that limits BT's profitability in an increasingly competitive market, Oftel's view is that this issue is dealt with through the coverage and specification of the price control baskets. These issues are considered in the chapters on Retail Price and Network Charge Caps. Where competition is strong or developing strongly, so that producers are forced to improve their efficiency and pass the benefits on to customers, Oftel agrees with the assessment that it is inappropriate to regulate prices tightly. However, if the judgement has been taken that services are not sufficiently competitive for this to be the case, then Oftel believes that those services should be price capped. Oftel believes that the purpose of price caps is to protect consumers from the exercise of market power, which may manifest itself through the making of excessive profits or through a failure to improve efficiency.
7.6 Oftel has some sympathy with the argument that a competitive telecommunications market may raise the cost of capital above that allowed for utility businesses. However, the extent of the extra riskiness in telecommunications should be captured through the value of Beta for BT in the Capital Asset Pricing Model formula, and Oftel's view is that no further adjustment for risk is appropriate.
7.7 Having considered the comments made by respondents in relation to the setting of parameters for the Capital Asset Pricing Model, Oftel's view on the appropriate values of these parameters is as follows.
7.8 In relation to the equity premium, Oftel is persuaded that the weight of evidence and the views of market participants would support the choice of an equity premium towards the top end of the range of 4-6% set out in the first consultative document.
7.9 The London Business School Risk Measurement Service (LBS RMS) estimate of Beta for BT's whole business is 0.82. In the first consultative document Oftel suggested that figure should be adjusted to reflect Oftel's expectation that the riskiness of the network business would be rather lower than that for BT as a whole. Oftel suggested that on this basis the Beta for BT's network business could be as low as 0.6. Respondents suggested that the adjustment to the LBS RMS figure to reflect lower risks for network activities should result in a Beta for these activities near the top end of the range proposed by Oftel.
7.10 Oftel accepts that a more rigorous approach to this calculation would be appropriate, and has therefore computed an estimate of Beta for the network and retail businesses. This has been achieved by identifying particular parts of BT's activities which are likely to have different, and identifiable, Betas, such as Cellnet and the investment in MCI. BT's average Beta (ie the Beta for the group as a whole estimated by the LBS RMS) has then been decomposed into a number of components, including those for the network and retail businesses. No further adjustment has been made to the Beta for the network business, on the basis that BT is facing prospective competition in all of network, access and retail activities. The resulting level of Beta for these activities is in the region of 0.75.
7.11 Some respondents criticised the tax adjustment formula used by Oftel to convert a post-tax into a pre-tax cost of capital. In particular, BT set out a number of detailed arguments, in which it was asserted that the Oftel adjustment was inconsistent with the values proposed for the other parameters of the formula. As Oftel stated in the first consultative document, it considers that the appropriate method for estimating a cost of capital and applying it for the purposes of computing price controls is to use a post-tax cost of capital and to assess post-tax returns in relation to this. Such an approach is conceptually appropriate, since estimates of what investors require are measured after tax, rather than before tax. This approach has the added advantage of removing the need for debates about the appropriate adjustments. The pre-tax cost of capital estimates set out by Oftel in the first document were calculated for the purpose of comparison with estimates of pre-tax costs of capital for other companies.
7.12 The computation of after-tax returns requires detailed knowledge of the tax treatment of the company's assets, and Oftel is not yet in a position to make the appropriate calculations. When the price controls are finally set it is expected that this will be the basis on which the calculation will be done, but for the time being Oftel proposes to estimate BT's tax liabilities using a conventional formula.
7.13 The tax formula referred to above is:
WACC pre-tax = WACC post-tax
(1-Tc)
with the value for Tc being BT's average or effective corporate tax rate over the last five years, rather than the statutory corporate tax rate of 33%. Use of the statutory rate is inappropriate because it is necessary to adjust the rate for the capital allowances and tax lags available under current UK tax law. Oftel's view is that the pre-tax cost of capital calculation, on this basis, would produce results not dissimilar to those set out in the first consultative document, and rather below the estimates proposed by BT, because BT proposed to use the full corporate tax rate rather than the actual tax rate in its calculation.
7.14 As a result of its further consideration in the light of responses to the first consultative document, Oftel considers it would be appropriate to apply a cost of capital towards the upper end of the range (9.2% - 13.4%) shown in the first document.
7.15 Oftel proposed in the first consultative document a move from an HCA basis of asset valuation to a CCA valuation, which would be consistent with the basing of charges on forward looking long run incremental costs. Oftel noted that, because this would involve an upward revaluation in BT's overall asset base, there was the prospect that windfall gains to shareholders could result.
7.16 The responses to this proposition suggested that the issues had either not been sufficiently well explained or not generally well understood. In particular, it was suggested by some respondents that if all that was proposed were a change to the unit of measurement of asset values and profits, then no windfall gain would result. Oftel accepts that, in a competitive market, it would be quite true that a change of accounting base would be irrelevant and would not create any windfall gains or losses for shareholders. Prices would be unaffected by such a change, and so there would be no economic effect.
7.17 However, Oftel faces the task of regulating an industry which is not fully competitive, where there are considerable sunk costs and assets, and in which prices have been set historically below a level consistent with the provision for the replacement of existing assets, and therefore below the costs facing new entrants using existing technology. It is necessary in these circumstances for the regulator, in determining the future allowable level of prices, to identify an asset base on which an efficient operator could make a reasonable return. If the asset base is to be uprated to a level above that on which prices had previously been set, because for competitive reasons it is desirable for prices to reflect the costs facing new entrants, then prices would be higher than if this uprating had not taken place. Shareholders would then enjoy windfall gains on the existing assets of incumbent operators .
7.18 It is Oftel's view that, while it would be desirable for prices to reflect the costs of new entrants, it would not be appropriate for such a move to result in windfall gains to shareholders. Oftel therefore proposed in the first document a number of alternative methods of ensuring that these windfall gains did not materialise. Oftel favours an approach by which any movement in prices towards the costs of new entrants would be gradual and justified by the additional investment made by BT's shareholders. This would be achieved if existing assets were depreciated at a proportion of the full CCA rate, in a way which related the depreciation charge on old assets to the current regulatory value, which is regarded by Oftel as being broadly equal to the HCA valuation of BT's assets.
7.19 If this approach were adopted, capital expenditure would normally exceed depreciation and the asset base would rise over time towards full replacement cost. Prices would ultimately be set at new entrant pricing levels. In the interim, prices would reflect the fact that less than a full depreciation charge on existing financial capital was allowed. In cashflow terms, this would lead to lower cashflows in earlier years being offset by higher flows in later years.
7.20 Apart from the adjustments to deal with windfall gains which could arise from a move from HCA to CCA asset valuations, the principles of Current Cost Accounting have been applied in the projections of costs after the base year.
7.21 The principles set out above have been applied in the assessment of the value of the asset base for BT's Access Business. The effect is to reduce the value of X by around 0.25, and to close the gap between the regulatory asset base and the CCA value of BT's assets by 2001.
7.22 In order to establish a framework within which the parameters of the new price control arrangements could be set, Oftel has developed a financial model of BT. This model enables Oftel to calculate the expected impact of various alternative price control regimes under a range of alternative assumptions.
7.23 The model consists of a number of modules. First, there is a calculation of the expected level of retail demand for telecommunication services from BT and other operators. This is derived from estimates of the overall growth in the market for these services which have been derived from recent experience, together with underlying assumptions about market developments, the performance of the economy and the effects of expected real price reductions. These forecasts have been
combined with a calculation of the expected market shares of BT and other operators.
7.24 Call volumes have been growing at a rate of 7.5% per annum over the last 2 to 3 years and Oftel believes there is scope for even higher growth, perhaps as high as 10% annually, in future. This reflects the prospects for continuing price reductions and general economic growth, the expansion of new services and the entry into the market of new operators. By contrast, BT forecast that call volumes would grow by only about 4% per annum, significantly less than recent trends would suggest. Oftel believes that it is hard to justify such a pessimistic outlook in view of these factors. Other forecasts seen by Oftel tended to be at the upper end of this range.
7.25 Growth in lines will be lower than growth in call volumes, with numbers of lines probably not increasing significantly faster than recent trends indicate. BT forecast that growth in lines would fall from the 3.5% of the last two years to 2% per annum. Again Oftel believes that this is likely to be too pessimistic even allowing for the possibility of future migration into ISDN.
7.26 Given this diversity of view, Oftel has modelled the impact of a number of different market growth scenarios. These correspond broadly to the rates of growth forecast by BT, to a continuation of recent trends and to higher rates of call volume growth of 9% per annum combined with growth in line numbers of 3.5% per annum. Oftel's modelling suggests that an across-the-board increase of 1 percentage point in the annual growth rates of calls and lines increases X by about 0.75.
7.27 The market shares are estimated through an assessment of competitors' expected discounts strategies, which are related to the competitors' costs and BT's assumed tariffing for each service. The market share losses for BT are separately identified in the case of direct, indirect and international simple resale operators.
7.28 The calculation of BT's loss of market share to directly connected operators is based on a comparison of customers' bills on the basis of BT's tariffs and the competitors' tariffs. This is performed separately for a number of customer types (three types of residential and four types of business customers). The bill relativities are combined with projections of the proportion of each customer type which will have choice of access suppliers, along with an assumed relationship between relative prices and the propensity of customers to switch suppliers (a form of cross elasticity of demand). The availability of number portability is assumed to increase the effective cross elasticity of demand between BT and competitors over time as number portability should reduce the costs to customers of switching suppliers.
7.29 In the case of market share lost to indirect and international simple resale operators, estimates are made of the availability of such services and the cross price elasticity of demand. In all cases the cross elasticity is assumed to vary with the level of discount offered by the competitors with respect to BT.
7.30 All commentators expect BT's share to decline over the next price control period as competition increases. Forecasts from commentators who are more optimistic about the prospects of competitors (notably BT) suggest that BT's share of total call minutes could fall below 65% and its share of lines to below 75% by 2001. Others were closer to 70% for calls and 80% for access. Work by NERA on behalf of a group of BT's competitors suggested that BT's share of call minutes would fall to around 65% by 2001 whilst its share of local calls and access combined would be just over 70%. Although NERA did not make this explicit, this is broadly consistent with a share of lines alone of about 80%.
7.31 It is clear that there can be no certainty about how far BT's competitors will succeed in taking customers away from it and it is therefore necessary to consider the implications of BT retaining a variety of different shares of the market. In its modelling, Oftel has used BT shares of total call minutes ranging, in 2001, from just above 60% to just above 70% (clearly, shares within this vary by call type) with BT's share of lines being roughly ten percentage points higher than its share of call minutes. However, there is likely to be some inverse correlation between market growth and BT's market share as competitors are likely to find it easier to obtain business in a faster-growing market. Thus Oftel considers it unlikely that a "high market growth - high market share" scenario will be realised.
7.32 The second module of the model consists of forecasts of demand and costs for BT's Network Account. The demands at the retail level for services of BT and other operators are converted into demand for BT's network services through estimates of usage factors and routing factors. The usage factors are those which appear in the Financial Statements of BT's Business and relate to the usage of network components by BT's retail services. The routing factors are those which appear in the Element Based Costing matrix, which underlies the calculation of interconnection charges under the current standard charges regime.
7.33 The starting point for modelling the costs of BT's network services is today's level of costs on a current cost basis. Projections of these costs are then made on the basis of productivity adjustments, along with forecasts of movements in prices of inputs and an estimate of how changes in volume cause costs to change. The productivity adjustments are informed by the Efficiency Study conducted by NERA for Oftel and the results of the work concluded by the Steering and Working Groups convened by Oftel to estimate long run incremental costs. This calculation yields both a starting point for network changes and the value of X to be applied to the baskets of General Network services and Call Termination services.
7.34 On the basis of the study of BT's efficiency carried out for Oftel by NERA, Oftel believes that it is reasonable to expect BT's real unit costs to fall by 4.5% per annum over the next price control period. This is made up of an underlying 3.5% improvement which was estimated to be the achievable performance by an already efficient operator, and a closing of the 10% efficiency gap between BT and the most efficient comparable operators identified by NERA. Oftel's modelling suggests that a 1 percentage point increase in the rate of reduction of BT's real unit costs increase X by about 1 point.
7.35 Two approaches to investment are considered in developing these forecasts. One method is to incorporate BT's own forecast investment with any adjustments considered appropriate. The alternative is to use projections of capital expenditure derived within the model. In this approach, the capital expenditure within the model is set on the basis that the capital expenditures should be equal to the Current Cost Accounting (Operating Capability Maintenance) depreciation charge. If asset lives are correctly specified, this should on average, over a run of years, be the required capital expenditure to maintain the ability of the firm to sustain the current level of output. It is also necessary to allow for the effect of additional volume of sales on investment and assets through a relationship between the volume of sales and asset requirements.
7.36 The costs of Access are modelled in much the same way as those of Network, except that adjustments are made to the asset base to avoid an immediate movement from an HCA basis of valuation to a CCA basis generating a windfall gain to BT shareholders. This adjustment was described in the earlier section dealing with the asset base.
7.37 Retail costs are projected on the same basis as Network Costs, taking the existing CCA levels, applying productivity and input price adjustments along with allowances for increases in volumes through a cost volume relationship.
7.38 When the modules are integrated into a single model, it is possible to compute prices, volumes, unit costs and asset movements for BT's network, retail and access activities. It is then possible to derive appropriate values of price control parameters which are consistent with the assumptions made, which yield a rate of return equivalent to the cost of capital. The effect of the proposed levels for the value of X on the cashflows generated by the regulated business is to reduce them over the period from about £4.2bn per annum to around £3bn per annum.
7.39 BT's cost of capital is also of considerable importance given the way in which X is set. Oftel's modelling suggests that each 1% increase in the assumed the cost of capital reduces X by 0.5.
7.40 In considering the effect on BT of alternative price control parameters, it is worth noting the current position in terms of the profitability and cost generative capability of the services which it is proposed to regulate. For the proposed "broad basket" of access plus local, national and international calls, BT's return on capital (on a historic cost basis) was 24%. This calculation excludes BT's costs of redundancy and restructuring. These costs amounted to £335m in 1994/95. Oftel's calculations of BT's costs which underlie the values of X for the wide basket retail price control include an allowance of £63m per annum for redundancy and restructuring.
7.41 This return on capital is rather higher than the cost of capital suggested in this document. However, the adjustments proposed by Oftel, by which the Asset base will rise over time towards the CCA value, mean that the proposed price controls are based on more than a simple calculation of price reductions necessary to reduce this rate of return to the cost of capital proposed. While it is true that the current level of profits exceeds Oftel's view of BT's cost of capital, and a reduction in this profit level is implicit in the proposed controls, account should be taken of the effects of the adjustments to the Asset base in estimating the effects of Oftel's proposals on BT's profits.
7.42 The revenues generated by the "broad basket" services in 1995 were £8.2bn, compared with BT's group turnover of £13.9bn in the same period. The operating cashflow from the "broad basket" services was around £4.2bn, which implies that, for those services, cash retained after operating costs amounts to 51% of revenue. A summary of the financial performance of BT Group and of the services for which price controls are proposed is set out in the table below.
Table 7.1 1994/5 Financial Performance of BT Group and Proposed Price Controlled Services
£bn BT Group Oftel's Proposed Broad Basket Revenue 13.9 8.2 Operating Profit 3.6 2.9 EBITDA* 5.9 4.2 Capital Employed 16.4 12.0 ROCE 22% 24% EBITDA/Revenue 42% 51% Redundancy 0.8 0.3 /restructuring Reported profit 2.8 2.6 Debt 20% * Earnings before Interest, Tax, Depreciation, Amortization.
7.43 As was explained in the earlier sections of this chapter Oftel has considered the impact of a number of factors on the profitability of BT's regulated retail business. Oftel has modelled the effects of alternative assumptions about these factors and has derived ranges within which it is expected that the retail price controls would be set. The ranges vary according to which set of services is subject to price controls, and the range in each case is determined by the assumptions made about market growth, BT'S market share, BT's efficiency improvements, the cost of capital and the asset base.
7.44 Clearly the composition of the price control basket is one such factor. National and international calls are significantly more profitable than local calls and, in particular, access. Thus a basket which includes only access and local calls will have a lower X than one which includes national and international calls as well. Oftel's modelling suggests that removing international calls from the basket reduces X by about 2, whilst removing national calls as well reduces X by a further 4.5 to 5.
7.45 The table below sets out for consultation a range of values of X for the 2 baskets Oftel has proposed - the minimum basket (local calls and access) and the preferred, broad basket. Oftel has examined a range of possible values for the main variables which affect X and the values of X they produce. Its analysis will be further refined over the next few weeks and, in the light of this and responses to this consultation, Oftel will reach a final view on the value of X for statutory consultation at the end of May. The figures presented here should not be taken as precise values of X for a particular value of the key variable. Rather, they show the centre of gravity of a number of values for X produced for different values of the key variables. Nevertheless Oftel considers that the values of X set out here probably span the likely range.
7.46 BT's volume growth (the product of forecast market share and growth in market volume) is the variable having the most impact on X. It is the most difficult to get right. The table sets out 3 alternatives:
7.47 Taking central assumptions for efficiency and for cost of capital as explained below, the values of X on the alternative scenarios are:
Table 7.2 Alternative values of X for retail price control
BT Volume Growth
Basket Accelerated Historic "BT"
Minimum 3 1 -1.5
Discounts in
Broad 9 7 5
Discounts out
7.48 Underlying each scenario are common assumptions about BT's efficiency, the cost of capital and the rate base to which it should be applied. Specifically, BT's real unit costs are assumed to fall by 4.5% per annum. This is at the centre of the range which Oftel believes is reasonable, based on the study of BT's efficiency carried out for it by NERA. In each case, X has been set to achieve a rate of return approximately equal to 12.5% in the last year of the price control. This is towards the upper end of the range of cost of capital estimates presented by Oftel in the first consultative document. The rate base is assumed to move from HCA towards CCA over the period of the price cap, as described in paragraph 7.15 - 7.21.
7.49 It is clear from the table that the composition of the price control basket has a major effect on X. National and international calls are more profitable than local calls and especially access. Thus the Minimum Basket, which includes only access and local calls, will have a lower X than one which includes national and international calls as well because there is less supernormal profit to be eroded over time.
7.46 The appropriate level of X in the network charge control depends upon similar factors to those determining the parameters of the retail price control, except that the relevant level of demand is the whole market's demand for BT's network services rather than the demand for BT's retail services. A further consideration is that, since it is intended that the starting level of charges for the network services should be set equal to BT's costs, there would be no supernormal profit at the start of the period. This will lead to the proposed value of X being equal to the projected reduction in BT's real unit costs in providing the services covered by the network charge cap over the period of operation of the cap.
7.47 It is proposed that the starting point for the network charge cap should be based on the forecast level of BT's incurred costs for the first year of the cap's operation. The projected reduction in real unit costs will be based on the assumption that BT achieves the level of costs of an efficient operator by the end of the control period. This level of costs will be assessed by reference to the outputs of the work which has been, and is continuing to be, undertaken to estimate the long run incremental costs of an efficient operator.
7.48 The reduction in BT's real unit costs in providing network services will depend, to some extent, on the volume of traffic using BT's network. Oftel has therefore estimated the effect of each of the three alternative market growth scenarios set out earlier on BT's real unit costs in providing network services. The table below sets out the values of X for the Network Charge cap which result from alternative assumptions about volume growth and BT'S productivity improvement. In the light of these calculations, Oftel considers that it is probable that the value of X for the Network Charge cap will lie in the range 3 to 6.
Table 7.3 Network price cap - Possible values of X
Productivity growth
Market Growth High Medium Low
Accelerated 6 5 4
Historic 5.5 4.5 3.4
'BT' 4 3 2
Click here to go to next chapter