4.1. Chapter 3 considered which segments of the market are not likely to see effective competition over the period from 1997. Price control will be required in these areas. Chapters 5 and 6 on network and retail price caps deal with specific issues relating to the individual type of price cap being considered. In addition to these, there are a number of issues relating to price cap regulation generally, on which decisions have to be taken.
4.2. Instead of a combination of separate retail and network price caps, BT's prices might be regulated by means of a global price cap which would include the prices of interconnection services as well as retail services. Within this cap, BT would be free to set the relative prices of all services subject to some constraints, including a limit on the overall level of charges. Proponents of the global price cap have argued that it would promote efficient pricing and reduce the incentive on BT to behave anti-competitively.
4.3. However, the efficiency properties of the global price cap depend on applying the correct weights to the various services in the basket, a difficult task. In addition, a global price cap would give BT a considerable amount of freedom over relative prices and there is a danger that this would be used in ways which would inhibit the development of effective competition, even assuming that the standard requirement that BT Retail should face the same charges as other operators would be imposed. BT would, for example, be able to increase interconnection charges and reduce retail prices within the cap and this would subject BT's competitors to a margin squeeze.
4.4 It has been suggested that this problem could be addressed by requiring that all prices lie between incremental costs and stand-alone costs. This would however be equivalent to applying the Efficient Component Pricing Rule (ECPR), which Oftel considered in its December 1994 consultative document on interconnection. It was acknowledged that the ECPR had some useful properties in limiting inefficient entry and allowing recovery of common costs, but the view was taken that these advantages were outweighed by the limitations it would impose on competition and by the uncertainty and lack of transparency it would create. It seems likely that BT would take full advantage of the permitted latitude to squeeze the margin between wholesale and retail prices down to the minimum permitted. This would deter the growth of competition.
4.5. A global price cap would have two further undesirable features. It would not sit easily with Oftel's preferred approach which is to withdraw services from the price cap where effective competition makes this possible. It would also have the effect of including competitive and monopolistic services in the same price control basket which is generally regarded as undesirable: reductions in prices of competitive services may be offset by rises in prices of monopoly services.
Oftel is not inclined to consider further the introduction of a global price cap but would welcome comments.
4.6. The view taken when BT's price cap was last reviewed was that the price control should continue to be based on the RPI rather than any alternative index. However, two disadvantages have been pointed out: RPI does not closely reflect BT's costs and use of RPI involves feedback effects in that BT's prices are themselves part of RPI.
4.7. The main alternative to the RPI would be an index of BT's costs such as labour and supplies. This has the advantage that it is not affected by factors unrelated to telecommunications. The main drawback is that it would remove any incentive on BT to control the prices of the items included in the index because BT itself influences the cost of these inputs. In addition, such an index would be less transparent than the RPI which has the advantage of being widely known and used, not least by the public itself. An index of economy-wide costs might be less influenced by BT but the advantage of greater relevance to telecoms would be lost.
4.8. One of the perceived drawbacks of the RPI is inclusion of irrelevant items such as mortgage payments. This could be addressed by using one of the available variants of the RPI which excludes mortgage payments or by use of the GDP deflator. However, the benefits of doing so would be small compared to the loss of the RPI's advantages of simplicity and widespread acceptance as a measure of inflation. It seems unlikely that feedback effects are significant.
Therefore Oftel's view is that it remains appropriate for the price control to be based on the RPI. However, Oftel would welcome comments on the possible alternatives. These could involve the use of RPI in addition to some other index.
4.9. Price control, in general, is appropriate only to those services where there is no effective competition to BT. However, this approach needs to be considered more carefully where new services are concerned.
4.10. In some cases where innovative services are newly introduced by BT, BT may not face effective competition. It may thus be in a position, at least in the short-term, to exploit the customer. On the other hand, such profits act both as a reward for the introduction of new services which benefit the customer and as a signal for competitors to enter the market. In such cases, control of prices may stifle entry and discourage innovation, to the detriment of the customer in the longer term. For this reason, Oftel believes that new services should not be included in the price cap. Other mechanisms will be used to deal with unfair cross-subsidy or other anti-competitive behaviour which might be identified in relation to such services.
Oftel would be grateful for comments on the view that new services should not be price capped and on how new services should be defined for this purpose.
4.11. Whilst it may be inappropriate to cap the prices of new services, in a situation where there are considerable common and joint costs between BT's basic services and its enhanced activities, customers of BT's basic services might reasonably expect to share some of the benefits of BT's exploitation of its economies of scope. However, it is important that the mechanism for sharing these benefits does not lead to a weakening of the incentive on BT to attempt to achieve them. Oftel believes that the price control should be set on the basis of a reasonable forecast of the benefits which BT might derive from shared use of common assets.
Oftel would be grateful for comments on how far it is appropriate to allow, in setting the price cap, for the benefits to BT from economies of scope derived from the introduction of new services.
4.12. A number of proposals have recently been made for explicit 'profit sharing' as an alternative to pure RPI-X regulation. Profit sharing can take a number of forms. Their common feature is that they aim to give customers a greater and/or earlier share of any unanticipated efficiency gains than does RPI-X regulation. To an extent, profit sharing is intermediate between RPI-X regulation and profit control. The relative merits of those two approaches were discussed in Chapter 1 (paragraphs 1.33 - 1.37).
4.13. Oftel has considered the case for introducing a measure of explicit profit sharing. On the one hand, the present system of RPI-X regulation is likely to encourage greater productive efficiency (ie production at minimum cost). On the other hand, profit sharing may result in prices being closer to costs, which represents a gain in allocative efficiency (because resources are directed to their most efficient uses when prices reflect costs of production). There may also be certain presentational advantages in allowing consumers to share explicitly in the profits of BT although these might be less apparent if lower than expected profits resulted in price rises. Proponents of profit sharing argue that it could have real advantages, including perhaps a lower cost of capital, if it led to a perception that the regulatory regime was likely to be more stable as a result.
4.14. Profit sharing would not be a substitute for price capping. As the history of regulation earlier in this century shows, profit sharing schemes will be undermined unless the terms are periodically revised to take account of changes in costs. However, with profit sharing it might be possible to increase the period between regulatory reviews.
4.15. But profit sharing has disadvantages. In particular, it introduces a need, at least under certain schemes, to set the threshold level of profit above which gains are shared and to determine the proportion of profits to be returned to customers. There would also be a need to monitor the level of profit against the threshold. This will make regulation more detailed, complex and uncertain. Some proposed profit sharing mechanisms are likely to restrict the methods of raising finance open to BT and this, perversely, could increase its cost of capital. It would also be necessary to determine the means of sharing profits between groups of customers and this might be controversial.
4.16. It is important to remember that RPI-X regulation is in fact a form of profit sharing. The benefits of increased efficiency within a price cap period are redistributed to customers when X is reset for the new control period. Recent work has shown that more than half of the net present value of unanticipated cost savings are passed to customers under RPI-X regulation. Other work has suggested that there may already be too great a degree of profit sharing in the sense that the reduction in incentives to reduce costs (due to the way in which X is set to yield an acceptable rate of return) outweighs the benefit to consumers from the reflection in prices of a larger share of profits. What may be more important than the percentage share is the size of the total benefit to be shared. And another mechanism by which profits already benefit consumers, at least as taxpayers, is by payments of corporation tax.
4.17. Oftel's view is that the clear and demonstrable benefits of the price cap since 1984 (see Annex B) indicate its value as both a control and an incentive device. On this basis Oftel firmly believes that the long-term gains from greater productive efficiency under price capping will outweigh the short-term loss, within the price cap period, of allocative efficiency. In particular, setting thresholds for profits is likely to encourage 'gaming' by the regulated company in order to avoid distributing profits to shareholders. This could give rise to some quite marked disincentive effects when profits approach the threshold. In addition, the gains in regulatory certainty may also be small or even negative. There have been no instances of the regulator intervening between reviews to tighten BT's price cap, whilst the possibility of intervention would not completely be eliminated even under profit sharing (especially if the control period were lengthened). In addition, the level of profit to be shared and the mechanism of sharing it would introduce uncertainty into BT price levels and affect the whole market price structure.
4.18. A further and the crucial consideration is that, in the UK telecoms market, there is the growing prospect of strong competition to BT. This is a key difference from other markets where recently privatised companies are dominant and where competition is at a much earlier stage of development. In the context of telecommunications, profit sharing (in which excess profits would be redistributed to customers in the form of prices reductions) would increase the uncertainty for competitors to BT since BT's prices have a major influence on their business. As competition takes off, it will ensure that customers enjoy the benefits of greater efficiency and lower costs, reducing the need for price controls and profit sharing.
4.19. There is little evidence available as to the effects of profit sharing on efficiency, and not surprisingly this evidence relates almost entirely to US experience. The evidence is somewhat ambiguous but does tend to suggest that price capping yields greater efficiency gains than some other forms of regulation, including profit sharing.
Oftel is not persuaded that the present system of regulation, which includes a substantial degree of implicit profit sharing, should be modified to introduce an explicit profit sharing element. However, Oftel is eager to have further comments on the value of profit sharing.
4.20. Whilst Oftel is not persuaded of the merit of profit sharing as an integral element of the price cap, nonetheless it remains concerned at the possible need to cater for the specific impact of exogenous factors. By definition, these are outside BT's control yet they can have a major effect on BT's profits. Thus it is arguable that increased profits resulting from exogenous changes should be redistributed to customers as they do not reflect any enhanced efficiency on the part of BT and this would not reduce the incentive on BT to lower costs which it could control. Of the exogenous factors which could potentially affect profitability, unanticipated growth in Gross Domestic Product (GDP) has particular significance for performance in the telecoms industry.
4.21. When X is set, account is taken of the expected impact of economic growth on BT's profits over the period of the price cap. However, BT's cost structure is such that profitability is very sensitive to the level of BT's output and unanticipated volume growth can lead to profits significantly above the levels expected in the price control review.
4.22. The issue of concern here is the impact of the rate of growth of the economy (that is the increase in GDP) on the growth of traffic volumes rather than other factors affecting volume growth which are to some extent under BT's control. There may be a case for ensuring that when economic growth exceeds expectations and results in significant additional profitability, this might be shared with customers: such an arrangement, because it is beyond BT's control, would not undermine the incentive attributes of the price cap. To implement such an adjustment would not be straightforward and careful consideration would need to be given as to how it would work beyond simple adjustments to reflect changes greater than predicted changes in GDP. The main difficulty is that this would require reliable knowledge of the relationship between GDP and telecoms growth, which is not available. Account would also need to be taken of whether the adjustment should be made for lower than as well as higher than anticipated GDP growth.
Oftel does not favour making an allowance for unanticipated growth in GDP or other exogenous factors in the price cap. However, comments on the desirability and practicality of an adjustment to reflect unanticipated growth would be welcome.
4.23. The duration of the price cap is a further issue which has important implications for both the distributional effects of price control and the underlying incentive properties. The period of the price cap needs to balance two opposing pressures. On the one hand, the longer the period between reviews, the longer the time in which BT may benefit from efficiency gains and thus the greater the incentive to reduce costs. A very short interval between reviews would make price control very similar to profit control, with the poor incentive properties already discussed in Chapter 1.
4.24. On the other hand, a longer interval between reviews adds to the uncertainty surrounding the profits which BT will actually achieve. If the effect is to make BT's profits appear riskier, then its cost of capital could be increased.
4.25. BT's initial price cap was set for five years but subsequent price cap periods were reduced to four years. Experience suggests that, with a four year cap, the work of resetting X comes round very quickly and there may be only 2 years' data on performance under the existing cap when work on the new one begins. For this reason, and because of its greater incentives for efficiency, Oftel believes that there could be a case for returning to a cap of five years duration.
Views on the appropriate duration of the price cap would be welcome.
4.26. Oftel proposes to introduce a 'roll-over' provision into the next price control arrangements so that, if further controls are needed from 2001/2 and it does not prove possible to agree them with BT before the existing controls expire, the existing price control would continue in force in the interim until a new control is agreed. This has the advantage that reviews can be carried out later with better information. This raises some practical issues, for example, whether a new control should operate retrospectively if it is agreed after the formal end of the old control and, if so, how.
Oftel would welcome comments on the introduction of 'roll-over'.
4.27. It is necessary to consider the precise mechanics of how the individual elements of network and retail price cap baskets would operate. In particular, consideration needs to be given to the way in which the constituent services in the price cap baskets would be weighted in the overall price change calculation. In the past the basket weights for BT's price controls have been set equal to the proportions of basket revenue accruing to the relevant services in the prior year to that in which the price change takes place.
4.28 This approach can be adapted to any pattern of BT's tariffs which might follow the lifting of the RPI+2% constraint on line rentals. As now, compliance with the RPI-X formula would be assessed against a reference tariff which would need to be defined. The value of discounts against the reference tariff to customers on packages which were outside the basket would need to be calculated in a manner similar to that in which BT's current discounts are assessed. These issues are discussed further in Chapter 6 (paragraphs 6.30 - 6.37).
4.29. An alternative method would be to base the weights on the proportion of revenues derived from the service in question in the current year. This is equivalent to a constraint on average revenue, and has been adopted in some regulated industries. One drawback with this approach is that forecasts of the relevant weights are required, with the complication that there is a need for retrospective adjustment.
Oftel's preference is to continue with the prior year revenue weighting system for a network price cap but views on the appropriate method of weighting the constituents of the price control basket would be welcome.
4.30. If BT reduces its prices more than is required in a particular price control year then it is allowed to count the excess price reductions towards its obligations in the following year. The current controls limit BT's ability to carry over price reductions into the fourth year of a price control, depending on the extent to which BT has used this facility in previous years.
4.31. One possibility is to prevent excessive price reductions from contributing to the following price control basket reductions. However, this would constrain BT's flexibility in implementing price reductions. A good example of this is the introduction of Per Second Pricing in June 1995. At that time BT was required to make price reductions of around £100m by 31 July. The Per Second Pricing reductions exceeded £300m. Without the facility to carry over reductions into the following price control year it is likely that BT would either have postponed Per Second Pricing or have introduced it with much lower price reductions. Oftel agreed with BT that customers would benefit if Per Second Pricing was introduced as early as practicable and with large price reductions to ensure that nearly all calls would be cheaper under Per Second Pricing than previously.
Oftel believes that it is appropriate to make provision for carry over in future price control rules and that present arrangements should therefore continue.
4.32. One of the features of price cap regulation is that profits may diverge from the level expected at the time when X was set. Any such divergence may be taken into account when X is reset in the next price control review. In principle, one way in which this could be done is by a one-off adjustment to prices which would bring the firm's expected rate of return to an acceptable level in the first year of the new cap. A second option would be to set X so that, taking the price control period as a whole, the firm was expected to earn, on average, an acceptable rate of return, but without a sharp adjustment to prices.
4.33. A third approach would be for the new control to be set so that the expected rate of return reaches an acceptable level by the end of the price control period. This approximates more closely than the other two methods to the workings of a competitive market in which excess profits are gradually eroded as rivals improve their own efficiency. It also avoids discontinuities in prices over time and leads to a more stable and predictable background against which investment and other decisions may be taken, by both suppliers and customers in the telecoms market. This is particularly important for telecoms as there are now many players besides BT. This approach also has greater incentives for efficiency as it allows the firm to retain the benefits of cost reductions for longer. The key difference between price control and rate of return control, in terms of their incentive properties, arises from the longer regulatory lag in the former. This means that cost reductions feed into price reductions only after a period during which the firm receives the benefit of increased efficiency. One-off adjustments to prices would reduce the effective regulatory lag, and hence the incentives to reduce costs.
4.34. In addition, one-off adjustments would create a particular distortion to the incentives on the firm near the end of a price control period. If the gains from increased efficiency were always taken away in the first period of the new cap, there would be little incentive to improve efficiency towards the end of a control period. The second option would also weaken incentives, particularly as in some years the rate of return might have to be below an acceptable level to reduce the previous years' excess; it would be quite close to rate of return control.
Oftel believes that the best approach is to set X so that the expected rate of return reaches an acceptable level by the end of the price control period. It welcomes views.
4.35. Another key consideration in setting a price control is the level of investment that is implicit in the financial forecasts underlying the price cap. Clearly, some method is required of ensuring that the company actually delivers on this part of the deal. However, if cost minimisation and efficiency are the aims of the price cap, it is not necessarily appropriate to require that the forecast amounts of investment are undertaken. Investment is an input rather than an output, and efficiency-promoting regulation should not focus on inputs. What is required is an assurance that the outputs which were implicit in the calculation are delivered. To this end, it will be necessary to assess what the measurable benefits of the investment programme are likely to be, in terms of quality of service and other benefits to customers.
4.36. In the past, there has been a general acceptance of the need to upgrade many parts of BT's network and therefore to allow investment costs relating to network upgrading into the calculations underlying price control. It is not clear that this is an appropriate approach for a price cap covering the period beyond 1997, although some further upgrading of the network may be required in order to allow all customers access to all the supplementary services available from a digital line. However, there must be a concern that, in the current stage of network development, much of any upgrade proposed by BT might relate to services other than basic telephony, for example to BT's plans for broadband switched mass-market services.
4.37. Accordingly, Oftel intends to take a critical view of BT's plans for investment in the forthcoming review, and, in particular, to press BT to justify its plans by giving explicit details of the consequences for customers of network services of alternative levels of investment. Oftel intends to ensure that basic telephony customers, either at wholesale or retail level, pay only for network upgrades which are necessary to provide them with the appropriate level of service. Upgrades which relate to BT's plans in other markets will not be included in BT's investment plans for the purposes of setting the price controls.
4.38. One of the merits of adopting a forward looking approach to setting interconnection charges (see Chapter 5 paragraph 5.6) is that it might make it easier to ensure that new investment allowed for in BT's prices relates to basic network services. A forward looking approach to pricing decisions implies that accounting information will be looked at on a Current Cost Accounting (CCA) basis. If a firm is experiencing no increase in the volume of its activities then on average, the level of its gross capital expenditure will be equal to the CCA depreciation, if asset lives and prices are properly specified. If a relationship between volume growth and required asset base can be estimated, the level of capital expenditure necessary to sustain the existing activities can be computed. Following this approach, it would be for BT to justify any additional capital expenditure over and above that identified as necessary. BT would have to demonstrate that the expenditure was necessary to bring identifiable benefits to basic voice telephony customers through an enhancement of services supplied by BT Network.
4.39. The 'outputs' considered above related to investment in the network. However 'output' could also cover broader concepts including the way BT interfaces with its customers and its sensitivity and responsiveness in meeting their needs. This reflects on BT's investment in staff management and training. Whilst clearly important, it is not so obvious that this relates directly, as with network investment and operational activities, to major outlays of expenditure. However, if it were thought that BT's service to customers was lacking and that emerging competition was not offering the appropriate spur, it would be possible, though extremely complex and difficult, to identify specific expenditures associated with staff training, customer servicing and management development and monitor them against customer performance. Whether the shortfall might trigger a reassessment of X is less clear given the relative size of the financial resources involved. It may therefore be appropriate to look for other mechanisms outside the price cap to ensure that output is maintained at the level predicted in the calculations leading to the price control.
The views of respondents on the best way of ensuring that BT achieves the outputs implicit in the price cap are invited.
Return to the start of this chapter
Return to this document's table of contents