Water and the Circular Economy
Project Round
Project Number
8TR8 - 001
Research Organisation
NERA Economic Consulting

Economic Framework for Estimating Long Run Marginal Costs

The Challenge

Setting prices for water and wastewater services by reference to long-run marginal cost (LRMC) principles is not a new concept in Victoria. However, water businesses have not so far adopted a common approach to estimating LRMC. The primary objective of this project was therefore to assist Victorian water businesses to estimate the LRMC associated with their services and to inform the application of LRMC principles in pricing decisions.

The Project

The term ‘marginal cost’ refers to the incremental cost of producing an additional unit of output or, alternatively, the cost that can be avoided by reducing output by a specified amount. Economic theory suggests that the essential, resource allocation function of markets works best when prices are set equal to the marginal cost of production.  When prices reflect marginal cost, allocative efficiency will be maximised, since the incremental benefit to society from the consumption of an additional unit of a good or service is equated with the incremental cost of providing that good or service, i.e. society would not be ‘better off’ from more or less consumption.

In the context of the water and wastewater sector, setting prices equal to marginal cost can encourage:

  • the sustainable use of water by consumers, since water consumption decisions will ensure that the marginal benefit of an additional kL of water reflects the costs of providing that water;
  • efficient procurement and provision decisions by water businesses, e.g. when to pump and when to use stored water; and
  • efficient investment decisions, eg, investment in transportation, treatment and storage infrastructure, water efficient appliances, etc.

While economic theory suggests that pricing by reference to marginal cost is more likely to deliver an efficient market outcome, this principle is often difficult to achieve in practice. Careful consideration must be given to the nature of the output and its associated costs that need to be estimated.

One such consideration is whether prices are more appropriately set by reference to short-run marginal cost (SRMC) or LRMC. This usually depends on the institutional circumstances of the market and the demand and supply implications of departing from either cost measure.  In the water industry, pricing by reference to LRMC is often preferred since, in practice, the realisation of any significant benefits under SRMC pricing arrangements would require customers to be able to respond to short-run changes in price.  Significant technological developments and associated investment in metering infrastructure would be required for any such benefits to be available in the Victorian water industry.

Further, water and wastewater sectors are generally highly capital intensive and characterised by ‘lumpy’ investment in new capacity.  At any one time, most water and/or wastewater systems operate with some spare capacity such that the system is capable of serving additional demand at relatively low or zero cost.  Given this, marginal costs are generally measured on the basis of the change in the per unit costs of supply associated with permanent step changes in forecast demand that require some level of additional capital investment.

Additionally, many water and wastewater services are multi-dimensional and it is not always practicable to set prices for all dimensions of service, even if it were possible to estimate a marginal cost for each individually. For instance, given that consumption of the residential sewerage service for each individual residence is not currently measured, it would simply not be practicable to charge for residential sewerage services on the basis of the marginal cost of either volume or pollution load.

There are two broad approaches that can be used to estimate the LRMC in the water and wastewater sector, namely:

  • the perturbation approach (also known as the ‘Turvey’ approach); and
  • the average incremental cost approach (AIC).

The two approaches involve similar steps, but differ in the way in which they measure the effect of changes in demand on operating and capital costs. Further, each of these approaches embodies differing levels of estimation complexity as well as differing levels of suitability under ‘lumpy’ and more periodic expenditure.

The Outcome

Before estimating any LRMC, one must first be clear about what needs to be measured. Therefore, NERA recommends that the critical first steps to estimating LRMC are to:

  • identify the relevant dimensions of each service provided.
  • determine whether it is practicable to measure how much of each service dimension is ‘consumed’ and whether customers are able to make independent decisions as to how much of a particular service dimension is ‘consumed’.
  • assess the likelihood and magnitude of customer response to price signals for each measurable service dimension.

In the case of a generic, vertically integrated water business, the framework developed by NERA showed that water volumes along with trade waste volumes and/or pollution load are likely to be the only service dimensions for which it is appropriate to estimate LRMC for application in tariff structure decisions.

NERA recommends that:

  1. For the functional levels of service dimensions characterised by a relatively ‘lumpy’ profile of capital expenditure, LRMC should be estimated according to the Turvey approach; and
  2. For the functional levels of service dimensions characterised by a relatively ‘smooth’ profile of capital expenditure, LRMC should be estimated according to the AIC approach.

The framework developed by NERA also included a set of common assumptions/procedures to assist businesses in the application of each of these approaches.

Overall, since any estimate of LRMC is based on a forward looking set of assumptions, NERA strongly recommends that sensitivity and/or scenario testing be conducted to ascertain the extent to which changes in these assumptions affect the estimated LRMC. This allows for the effect of the various uncertainties in LRMC assumptions to be examined, and so to convey the range of uncertainty attributed to any particular estimate.

Supporting documents