MD 159

TO THE MANAGING DIRECTORS OF ALL
WATER AND SEWERAGE COMPANIES
AND WATER ONLY COMPANIES

11 February 2000

LRMC AND THE REGULATORY FRAMEWORK

Previous MD letters (particularly MD148) have explained the importance of long run marginal cost (LRMC) for water pricing and for efficient investment planning. The value of LRMC depends in part upon the quality of companies' estimates, which should be subject to a regular process of re-evaluation. In order to facilitate that re-evaluation, this letter updates Ofwat's advice on:

• the role of LRMC within the regulatory framework;
• Ofwat's approach to LRMC in a number of important policy areas, principally charging and competition policy;
• some of the practical challenges posed by applying LRMC to prices, and some potential solutions;
• the process for updating LRMC estimates (companies are invited to review their estimates and, if they wish, to submit revised numbers by the end of July 2000).

• a general review of companies' numbers;
• Ofwat's response to the outstanding issues related to LRMC estimation, as identified by companies.

Objectives

LRMC is relevant to my duties to promote economy and efficiency, and to facilitate competition. It is a tool to encourage better, more consistent decision-making across a number of areas. Its role is to ensure that customers, water companies and potential new entrants take proper account of the costs of their consumption and investment decisions. In so doing, LRMC can help to achieve:

• more cost-effective use of existing assets;
• least cost investment in new assets to balance supply and demand;
• conditions to ensure that entry into the market for water services is beneficial to customers.

Ofwat's approach to LRMC

LRMC is relevant to decision-making about investment plans, particularly with regard to:

• Ofwat's Paying for Growth methodology;
• the economics of leakage control.

• terms for bulk supplies and inset appointments;
• issues of tariff structure, including large user tariffs and Special Agreements;
• terms for common carriage agreements.

Since companies will be expected to apply LRMC consistently, it is in their interests to ensure that it is estimated robustly. My office will seek to ensure consistency in companies' use of LRMC through the approaches discussed below. Other companies, together with customers and potential new entrants, should also be able to challenge LRMC estimates by exposing inconsistencies between those estimates and the applications to which they are put.

Paying for growth

As long as prices are not below robust estimates of LRMC, it should be possible to finance the incremental cost of meeting growth in demand from the extra revenues generated by that growth. This formed the basis of Ofwat's approach to assessing growth-related costs as part of the 1999 Periodic Review. Allowance in price limits for growth-related expenditure was made to the extent that it could be financed from additional revenues.

Economics of leakage control

Companies should choose the least cost combination of options to balance future supply and demand. Those options should then form the basis of companies' LRMC estimates. Consequently, if the costs of leakage control are less than a company's estimated LRMC, it could reduce the cost of balancing supply and demand by reducing leakage further. This will be reflected in mandatory leakage targets. If the costs of leakage control are above LRMC, companies should be able to demonstrate that they have pursued opportunities to sell 'surplus' water.

Demand management

Incentives for economy in the use of water are most effective when prices reflect LRMC (which is only possible when customers are measured). If prices were above LRMC, customers would conserve water when it would be less costly for a company to invest to balance supply and demand. If prices were below LRMC, a company would have to invest to balance supply and demand when it would be less costly for customers to conserve water. Prices that reflect LRMC can help to achieve an economic balance between companies investing in new supplies and customers conserving water.

Bulk supplies and inset appointments

Companies with a supply/demand imbalance may find it more economical to seek a bulk supply from another company than to invest in their own new resources. Differences between neighbouring companies' LRMC estimates should highlight opportunities for bulk supplies. A company expecting marginal costs to rise should be ready to seek a bulk supply, perhaps for a limited period. A company expecting marginal costs to fall should be prepared for bulk supply requests from other companies, including its competitors, and indeed should seek out opportunities to provide bulk supplies.

In the first instance, I look to companies to seek or provide bulk supplies through their own negotiations. Where companies are unable to agree, I have powers to determine terms for them. Ministers plan to introduce legislation later this year that will give the Environment Agency powers to compel an undertaker to seek a bulk supply from another company. As now, I would then have powers to determine financial terms if the parties concerned were unable to agree. In determining terms, I will make particular reference to LRMC, but with due regard also to:

• expected changes in marginal costs over time;
• the need to maintain incentives for companies to make bulk supplies;
• preventing undue discrimination.

Large user tariffs

Tariffs for customers consuming at least 250Ml/a are now outside the tariff basket. The emerging competitive market for supplies for large users should now be the primary driver for these tariffs. In this new environment, my role is to:

• prevent undue discrimination and undue preference;
• prevent abuses of a dominant market position and other anti-competitive behaviour;
• ensure that tariffs are structured to send appropriate price signals.

As far as practicable, average prices for large users should be determined by a robust allocation of accounting costs. This is the principal means by which abuses of market power and undue discrimination are assessed.

LRMC is relevant to the balance between volumetric rates and fixed/standing charges. Volumetric rates should reflect LRMC as closely as possible to provide appropriate incentives to promote economy in the use of water. It would also ensure that, when their demand falls, large users enjoy bill reductions that reflect their suppliers' cost savings.

Common carriage

The LRMC of supply networks is relevant to determining charges for access to companies' networks. In MD154 I said that I would expect each company to charge entrants as it would charge itself. In this way, incumbents should be able to recover reasonable network costs and capital maintenance charges. I said that charges might be based on average costs, where appropriate, or LRMC. A low 'bulk transport' component of LRMC, for example, might indicate that access to this part of the network could be provided at low cost.

I will consider any complaints relating to access pricing under the Competition Act 1998. The context of each complaint will determine the most appropriate cost concept to be used as the basis for determining reasonable terms.

Practical issues of application

Price limits are determined at Periodic Reviews on the basis that companies need to recover sufficient revenue to finance their total costs, including a reasonable return. On this basis, average tariff basket prices should equate approximately to average accounting costs. For customers outside the tariff basket, competitive and regulatory pressures should together ensure that average prices equate approximately to average accounting costs.

LRMC could be above, below or equal to average accounting costs. On the one hand, the regulatory capital value is around one tenth of the replacement cost of companies' assets. On the other hand, companies may be able to meet additional demands at low cost as a result of past investment (this is explored in the annex). Consequently, companies need to find ways of reconciling LRMC pricing with total cost recovery. There are two main approaches:

i. Set volumetric rates as close as possible to robust estimates of LRMC and allow fixed/standing charges to bridge the gap with average accounting costs.
ii. Set volumetric rates closer to LRMC for some customers, leaving other customers to contribute more or less to total revenues than the total costs their demands impose upon the system.

The first approach should also be applied carefully. A mechanical translation of LRMC estimates into volumetric rates could result in very high standing charges (where LRMC is well below average accounting costs) or free tranches of water (where LRMC is well above average accounting costs). Either case could lead to individual customers paying significantly more or less than justified by an allocation of accounting costs. That could constitute undue discrimination. The balance between volumetric rates and standing charges should be informed by reference to LRMC, therefore, provided this avoids unacceptable distributional effects.

Companies also need to take account of the range of uncertainty surrounding their LRMC estimates when applying them to tariffs. As the attached annex suggests, all LRMC estimates should be subject to sensitivity analysis. Similarly, companies should develop their understanding of the time path of marginal costs. This would enable them to avoid excessive discontinuities in pricing policy over time, while retaining appropriate signals about the long run value of water.

Next steps

In this letter I have explained why LRMC is important and how Ofwat will seek to ensure that companies apply it consistently. I have explained that it is in companies' interests to ensure that their own and their competitors' LRMC estimates are robust. Ownership of the numbers rests with the companies but they should be in the public domain and open to challenge. My role is to:

• provide guidance, without being overly prescriptive, so that companies follow a consistent methodology for estimating LRMC;
• ensure that LRMC is applied consistently.

I will continue to apply to regulatory decision making the information on LRMC available to me. Companies may wish, therefore, to review their previous LRMC work in the light of this letter and its Annex. Should this review prompt a company to submit revised LRMC estimates, my Office would be glad to receive this information not later than 31 July 2000. Moreover, since LRMC should be subject to continuous review, estimates will be collected as part of the June Returns from 2001 onwards.

I C R Byatt

ANNEX

Technical annex – calculation of LRMC

The purpose of this annex is to:

• provide some general observations on the range of estimates published in the 1999-2000 Report on tariff structure and charges; and
• address outstanding issues regarding the calculation of long run marginal cost.

Estimates of the LRMC of steady demand, published in table 27 of the 1999-2000 Report on tariff structure and charges, range from 14p/m³ to 377p/m³. Wessex Water, which lies at the top end of the range, has revised down its estimate to reflect assumptions made in the Periodic Review. In any case, this wide range disguises the fact that most estimates are clustered around 30p/m³ to 50p/m³.

Some of the estimates run counter to expectations based on Ofwat's understanding of the supply/demand balance positions of various companies. For example, some companies with an ample margin of supply over demand have produced relatively high estimates while some companies facing supply constraints have produced relatively low estimates.

There are various explanations for these counter-intuitive results. In some cases, estimates are unexpectedly high because companies have used average accounting costs as a proxy for some or all of the components of LRMC. This is motivated by an awareness of the problem of reconciling LRMC pricing and total cost recovery. In effect, however, it prejudges the best way to address that problem. Ofwat would prefer each company to estimate LRMC robustly and to tackle the pricing problem in light of that estimate. This makes more sense than working backwards from the pricing problem to the estimate.

Similarly, there are various explanations for unexpectedly low LRMC estimates. One of the dangers of forward-looking estimates of LRMC is that they may be subject to errors of exclusion. This is particularly likely for those companies whose LRMC estimates are based on future schemes that have not been evaluated in detail. Companies run the risk when under-estimating LRMC that low volumetric prices, based on these estimates, will generate increased demand. In turn, such companies would find that they cannot finance the costs of this increased demand from incremental revenues.

Calculation of LRMC

Following the publication of MD123 and the Ofwat seminar held on 30 April 1997, companies were invited to submit estimates of LRMC in October 1997. Since then, estimates have been updated by the companies for submission as part of Periodic Review Information Requirement E, and for submission in companies' draft Business Plans in 1999.

Guidance on the calculation of LRMC was updated in the reporting requirements for companies' Supply/Demand Balance Submissions (1998) and again in the Guidelines for Business Plans (1999).

Some companies have reported difficulty in interpreting this guidance. The 1999-2000 Report on tariff structure and charges recognised that there are some outstanding methodological issues relating to the calculation of LRMC. This annex addresses those issues that have caused particular difficulty, namely:

1. the treatment of sunk costs and excess capacity;
2. calculation of LRMC when there is no foreseeable need for new resources;
3. the relevance of hypothetical increments and decrements to demand;
4. application of efficiency assumptions;
5. allocation of LRMC across its components (resources, treatment etc).

The London Economics paper Water Pricing: The Importance of Long Run Marginal Costs (attached to MD123) took the view that LRMC refers to marginal costs calculated for a system that is optimally configured. Issues of "lumpy investment" and consequent spare capacity do not arise in these idealised circumstances where LRMC matches short run marginal costs (SRMC)¹. In practice, however, companies cannot configure their systems with the benefit of perfect foresight. London Economics preferred to describe long run costs at the margin as "long run incremental costs" in order to distinguish this practical situation from the abstract definition of LRMC.

¹ The long run is defined as the period over which all inputs can be varied. In the short run, at least one input – usually capital equipment – is fixed.

Ofwat has tended not to make such a distinction in terminology. When referring to LRMC, however, Ofwat expects companies' calculations to be based on the world as it is. LRMC is a forward-looking concept, representing the incremental costs, on a continuing basis, which a company would incur as a result of changes in its level of output. Those costs include both additional operating costs and the cost of bringing forward future capital investment. Costs already incurred ("sunk") to create existing capacity should not be included, so LRMC will tend to be lower when there is excess capacity.

It would be surprising, and indeed uneconomic, if companies generally had excess capacity. Ofwat expects companies to plan, using economic criteria, for a degree of 'spare' capacity for security of supply purposes. Investment in security of supply is economic where the cost of additional investment equals the cost of not investing (including the potential cost of emergency supply measures). Excess capacity – that is, spare capacity in excess of this economically derived security margin - might arise for a time as a result of 'lumpy' investment, but Ofwat would not expect this to be the norm.

The relevance of sunk costs is not necessarily confined to excess capacity, as defined above. A company with no excess capacity may find that marginal costs are relatively low as a result of past investment. Raising a dam, for example, may carry a lower unit cost than building a new dam. Equally, additional investment in leakage control may bear a low unit cost once the set-up costs of establishing DMAs have been incurred. These examples might be thought of as another form of spare capacity, not least because, like conventional spare capacity, the economies offered cannot be exploited indefinitely. There is, for example, a limit to how far a dam can be raised.

LRMC and future resource needs

For companies facing growing demand, and with a future investment strategy to meet those demands, calculating LRMC is relatively straightforward. The reporting guidance for Periodic Review Information Requirement E suggested that LRMC could be calculated as the present value unit cost of the next representative scheme, including consequent costs from source to tap. Identifying a "representative" scheme is highly subjective, and some companies have preferred to avoid this problem by calculating the present value unit cost of their respective future resource strategies. The numerator in such a calculation is the sum of the present values of future capital and operating costs over an appropriate investment horizon, while the denominator is the present value of the associated contribution to maintaining the supply demand/balance. Since all costs and outputs are forward looking, this calculation makes allowances for the effects of spare capacity and sunk costs. Estimates calculated in this way are consistent with Ofwat's approach.

Where future demand is forecast to match future supply the present value unit cost of a future investment strategy cannot be calculated. To calculate LRMC under such circumstances is more difficult. It is necessary to hypothesise an increment (and/or decrement) to demand. Calculations of LRMC may be sensitive to the size of the increment or decrement assumed, so it is particularly important to undertake a sensitivity analysis. Of course, all estimates of LRMC should be subject to sensitivity analysis. Even when demand is forecast to increase, so that it is not strictly necessary to hypothesise an increment to demand, it is useful to assess the sensitivity of LRMC estimates to errors in forecast demands and costs.

Hypothetical increments and decrements

Some companies have argued that LRMC should be calculated as the difference in the present value of future costs attributable to an increment of demand additional to a company's central forecast of demand. That is, the company would first derive an investment strategy based on its best estimate of future demands. It would then reconfigure its strategy to meet the same forecast demand plus an increment. LRMC would be calculated as the change in the present value of costs divided by the change in the present value of output (the demand increment). Several variations on this theme have been suggested, including an average of LRMCs calculated on the basis of increments and decrements to the central demand forecast.

This kind of approach (usually attributed to Ralph Turvey) is common in the academic literature, but it raises questions about the relevance of hypothetical increments or decrements when demand is, in any case, forecast to increase or decrease. The previous section described the calculation of the present value unit cost of an investment strategy. That calculation was essentially the same as the one attributed to Turvey; only the inputs differed. The "increment" of demand was the central forecast of demand growth, and the change in costs was calculated as the cost of meeting forecast demands relative to the cost of meeting an unchanged level of demand in the future. The "Turvey" approach, in effect, calculated the change in costs from an increment on an increment. This might be useful as part of a sensitivity analysis, but it is not clear that this is an appropriate basis for calculating a central estimate of LRMC.

Application of efficiency assumptions

Following the publication of final determinations, some companies have recognised that the efficiency assumptions implicit in price limits are relevant to the calculation of LRMC. In order to deliver normal (or above-normal) rates of return to shareholders, companies will need to achieve assumed efficiency savings, and this applies equally to new investment as well as to the use of existing assets.

Allocation of LRMC across its components

The components of LRMC (resources, treatment, bulk transport and local distribution) are relevant to a distinction between the:

• LRMC of producing water ('resources' LRMC for non-potable water; 'resources' and 'treatment' for potable water).
• LRMC of delivering water ('bulk transport' for large users; 'bulk transport' and 'local distribution' for other customers).

Some companies have experienced difficulty in allocating total LRMC between its components. These difficulties are particularly apparent when supply/demand balance strategies are dominated by leakage control, which provides additional treated water for local delivery. The reporting guidance for the 1999 Business Plans suggested that the costs of leakage control, together with metering and other demand management measures, should, as far as possible, be allocated in the same proportions as the alternative supply-side schemes for which they are substitutes. For example, if a company chose leakage control as an alternative to augmenting resources, the costs of leakage control could be allocated to the resources component of LRMC. Equally, if a company chose leakage control to alleviate constraints on distribution capacity, then the costs of leakage control should be allocated to the distribution component of LRMC.

Some companies have calculated the components of LRMC by making different assumptions about demand in each case. For example, resources LRMC might be calculated on the basis of a 1Ml/d increment to demand while treatment LRMC is based on a 2 Ml/d increment. The motivation for this approach was that different increments to demand were required to derive non-zero estimates of LRMC for different components. Total LRMC, derived as the sum of these components, was therefore based on inconsistent assumptions about incremental demand. Ofwat recommends that companies explore the sensitivity of total LRMC and its components to different demand assumptions, but that any given assumption should be applied consistently across all components.

Top of Page