Transmission and distribution pricing

advertisement
Transmission and distribution pricing
ABARE submission to the NECA review
The central objective of transmission and distribution services pricing is the
achievement of efficient outcomes in terms of network use, operation and
maintenance, and investment (NECA 1997). Designing a satisfactory pricing structure
for network services which achieves this objective is one of the most complex tasks of
regulatory authorities. Yet, it is important to get the prices right as these have
implications not only for efficiency in the use of the network system, but also for
investment decisions on grid expansion, generator and customer location. In addition,
in a vertically related electricity system the terms of network access also impinge on
the effectiveness of competition at the production and retail ends of the electricity
supply chain.
The focus of ABARE’s submission is on economic efficiency in network pricing. In
particular, the objectives in this submission are:
•
to provide a framework for achieving economic efficiency through network
pricing arrangements; and
•
to address the following specific questions raised in the National Electricity
Code Administrator’s issues paper (NECA 1997):
–
–
–
–
How should network charges be structured in order to reflect the different
cost components of network service provision?
Does the code provide sufficient locational signals through the wholesale
market arrangements and the network pricing arrangements?
How can the network pricing arrangements encourage economic bypass,
while ensuring that uneconomic bypass does not occur?
What is the potential for contracts to be introduced for aspects of network
service?
1. Framework for efficient network pricing
An efficient pricing regime needs to reflect the costs incurred in providing network
services. Transmission and distribution networks are characterised by high capital
1
intensities. Therefore, a high proportion of network costs is fixed and attributable to
the capital costs of establishing network infrastructures. The variable costs associated
with electricity transmission and distribution systems are relatively low and consist
essentially of energy losses when capacity of the network is not constrained and of
energy losses and congestion costs (as measured by the divergence of electricity spot
market prices across regions) when capacity is constrained.
Electricity networks are also subject to significant economies of scale, which implies
that, for a given geographic area, transmission and distribution systems are natural
monopolies. The high fixed costs and low variable costs have important implications
for the pricing of network services if optimal network use and investment is to be
delivered.
1.1 Efficient network use
Efficiency in network use requires that network services be priced at levels reflecting
the costs to society of supplying those services. Setting prices equal to short run
marginal cost (SRMC) is a possible reference benchmark as it equates the cost of
using an additional unit of access to the value of that unit of access to network users.
SRMC includes only the costs directly associated with the provision of an additional
unit of access and, therefore, does not cover capital costs and ongoing maintenance
costs. This outcome reflects the sunk nature of fixed costs, such as infrastructure
costs, which have no bearing on the efficient level of network utilisation for a given
level of capacity.
However, one of the major criticisms of marginal cost pricing in highly capital
intensive industries, such as electricity transmission, is that marginal cost is typically
below average cost when the network is not capacity constrained. Therefore, SRMC
pricing is likely to lead to a shortfall in revenue for network service providers. In
addition, when capacity is constrained, SRMC would not reflect the additional capital
cost incurred in adjusting capacity to provide an extra unit of access services. The
implementation of SRMC pricing also requires regulators to have access to
information on costs.
There is a number of pricing options which allow the regulator to balance the need for
efficient pricing, as embodied in SRMC pricing, against the need to finance
investment (Berg and Tschirhart 1988; Armstrong, Cowan and Vickers 1994).
Ramsey pricing, for example, provides a means to cover the regulated firm's costs by
a mark-up over marginal cost which is inversely proportional to the demand
2
elasticities of different user groups (Berg and Tschirhart 1988). Another option is a
two-part tariff with variable prices set at marginal cost and an access fee to cover any
financial deficit. The major objective of these pricing regimes is to allocate fixed costs
in a manner that minimises the distortions caused by any deviation from marginal cost
pricing.
1.2 Efficient network investment
The design of network pricing regulation has significant implications for investment
decisions. Long run marginal cost (LRMC), which includes the capital cost of
augmenting the network and the cost of operating the extended network relative to the
cost of operating the facility at existing capacity, provides better signals for long term
investment decisions than SRMC. Investment in capacity is optimal if a user’s
willingness to pay for an additional unit of network services in peak periods is greater
than the LRMC of expanding the network.
However, LRMC pricing distorts prices away from their optimal levels in the short
run when capacity is not constrained. It is therefore generally accepted that it is
difficult to design a pricing regime that allocates existing capacity efficiently and also
provides the correct incentives to expand the network (Hinchy and Low 1993; King
1995). Even if the focus was on providing the right economic signals for investment
through LRMC pricing at least in periods when capacity constraints are binding, there
are several factors which are likely to distort investment decisions.
First, there are both positive and negative externalities inherent in electricity networks
because networks are typically shared assets. For example, the upgrading of a line
may not benefit only users who are directly affected by the expansion but it could also
provide indirect benefits to all users in terms of enhanced system security. To the
extent that the marginal benefits experienced by all users exceed the marginal benefit
of an individual investor at that individual’s optimal investment quantity, this could
result in a lower level of investment than is socially desirable (Bushnell and Stoft
1997).
In addition, transmission and distribution systems are natural monopolies. A network
owner assessing the optimal capacity for a new or expanded line would invest only up
to the point where the marginal revenue produced by the additional capacity is equal
to the marginal cost. Therefore, monopolistic behavior would also lead to
underinvestment in network capacity.
3
For these reasons, it may be difficult to achieve ‘first best’ outcomes for network
investment through market based incentives only. This suggests that some central
coordination of grid expansion would be required for procedural oversight and
assessment of the market investment process. In particular, the oversight would need
to ensure that investment decisions provide a cost effective way of meeting additional
electricity demand after equal consideration of network augmentation, additional
generation and demand side management options. Nonetheless, the role of the price
mechanism in terms of providing clear signals about the need for additional capacity
should not be overlooked.
1.3 Efficient location of new generation and load
In order to promote efficient location of new generation and load, network prices
should be location dependent, thus reflecting the location dependent costs that new
generating facilities or end use loads impose on transmission and distribution systems.
This pricing regime would need to reflect the positive correlation between the costs of
transmission and distribution and the distance between generation and load.
In many ways a pricing regime which delivers efficient use of the existing
infrastructure would also send the appropriate price signals for efficient location of
new supply and demand facilities. This requires a pricing system which reflects
marginal energy losses and also congestion costs when capacity is binding. If fixed
costs are to be covered, the provision of locational pricing signals needs to involve the
allocation of these costs on a location dependent basis.
2. Network pricing issues
The key feature of an efficient network pricing regime is a high degree of cost
reflectivity, with marginal cost as the appropriate cost parameter rather than average
cost. However, it is also clear from the previous section that, in practice, an access
regime based strictly on marginal cost pricing may be made difficult by a number of
considerations. In the following section comments are provided on specific questions
raised in the NECA issues paper and some guidelines are provided on how these
should be addressed to achieve efficient outcomes.
2.1 Costs of network services
How should network charges be structured in order to reflect the different cost
components of network service provision?
4
The pricing methodology proposed in the National Electricity Code for covering fixed
costs is a combination of cost reflective pricing (defined as a pattern of pricing which
reflects the differentials in cost related to distance or the value of the assets used) and
postage stamp pricing (which is essentially average cost pricing). Common costs that
cannot be attributed to particular user groups are allocated on the basis of network
use. The code also provides for connection costs to be covered through fixed charges
to network users. Within this broad framework network charges can be designed to
include one or more of the following components: a fixed charge, a demand
component and an energy component. Variable network costs including energy losses
and congestion rents are transacted through the wholesale spot market.
The point of departure for an access regime which is to provide the right signals for
efficient network use should be SRMC pricing, reflecting energy losses when capacity
is not constrained and also congestion cost when capacity is constrained. However,
given that SRMC is lower than average cost when capacity is not constrained, this
approach may not deliver a sustainable revenue stream to network owners which
includes a reasonable rate of return on investment, which is one of the objectives of
the pricing regime as specified in the code. In this case, the distortions caused by
average cost pricing can be reduced by adopting a nonlinear pricing regime such as a
two-part tariff.
A two-part tariff enables prices to approach SRMC while allowing network owners to
cover fixed costs of service provision through an access fee. If the fixed charge does
not encourage some network users to exit the market then nonlinear tariffs are clearly
superior to uniform tariffs in terms of efficiency (Berg and Tschirhart 1988). In
general, this situation would apply to electricity networks as demand for access
services from most network users is unlikely to be sensitive to the access fee given
that the demand for electricity is price inelastic.
However, some users who have the option of bypassing the incumbent network
service providers may be driven away if the access fee is perceived to be too high. The
efficient nonlinear tariffs would then require that network owners be in a position to
identify different user groups and tailor the tariffs to particular groups according to
their elasticity of market participation with respect to prices and access fees.
Alternatively, network service providers could be encouraged to offer self-selecting
two-part tariffs in which consumers are provided with a range of tariffs to choose
from (Brown and Sibley 1986). In essence, self-selecting tariffs allow users to select
5
the tariff that best fits their own demand with the objective of preventing a reduction
in the market participation rate.
A structure for network prices which aims to minimise distortions could therefore
consist of an energy based charge which reflects the SRMC of network use and a
fixed access fee which ideally should be completely insensitive to use. The access fee
would cover both common costs, such as power system security, as well as costs of
capital that are directly attributable to particular users. In practice, however, the fixed
access fee is often determined on the basis of some measure of maximum demand, a
measure which is not fully independent of use. A key consideration for a demand
based charge is that it should reflect the costs that particular users impose on the
system. This implies that the demand charge should be based on users’ demand during
the time interval in which systemwide demand is at its peak rather than based on
maximum demand irrespective of time. The latter approach does not account for spare
capacity in the network and attempts to cover fixed cost in of!f-peak periods may
deter the utilisation of this capacity.
The concept of marginal cost is an important prerequisite for any efficient pricing
regime. It is clear that the feasibility of an optimal tariff structure depends highly on
the availability of information on demand and cost profiles. Asymmetric information
can be an important issue for network pricing because regulators typically have less
information than network service providers. While access to demand and cost
information may be difficult to obtain, prices determined without this information are
likely to be inconsistent with economic efficiency (Baumol and Sidak 1994). This
highlights the need for regulators to reduce information asymmetries through
independent studies and yardstick comparisons where these are feasible.
2.2 Locational signals
Does the code provide sufficient locational signals through the wholesale market
arrangements and the network pricing arrangements?
Locational pricing signals in the national electricity market are provided primarily
through loss factors in the determination of merit order dispatch for generators and the
calculation of wholesale electricity prices, and through the cost reflective method
proposed in the code for pricing the use of transmission services.
6
The key characteristic of an efficient pricing regime is that prices should reflect the
underlying costs. The existing arrangements for network pricing, while providing
some price signals for the location of new load and generation, are likely to distort
network use and long term investment decisions for the following reasons:
•
Network losses
The pricing of energy losses should be such as to provide signals about the costs
of losses. The relevant cost parameters are marginal losses as these provide clear
signals about the extent and timing of any congestion. The code provides for
three types of losses:
–
–
–
interregional loss factors based on marginal losses;
intraregional loss factors based on the average of marginal losses; and
distribution loss factors based on average losses.
Energy losses across regions in the national electricity market are to be based on
incremental losses. This can be expected to provide appropriate price signals for
investment in major interconnectors and the location of new generation and load
across regions. However, the averaging of marginal transmission losses and of
total distribution losses within regions is likely to weaken the signalling function
of prices particularly given that regions in the national electricity market are
broadly defined. To the extent that marginal losses are substantially higher than
average losses, this would not provide the incentive for network users to
minimise losses through locational and cyclical load adjustment. The distortions
are likely to be more significant in periods when the differential between
marginal and average losses is relatively high such as when capacity is
constrained or when there i!s a large amount of spare capacity.
The averaging of intraregional transmission and distribution losses are justified
in the code by the complexity involved in calculating dynamic marginal losses
at each connection point within a region. However, the merits of reducing the
degree of averaging and moving toward full marginal cost pricing need to be
reconsidered in the light of technological advances which would make the
adoption of marginal pricing feasible across all segments of the network.
•
The postage stamp component
It is proposed in the code that 50 per cent of network costs be covered through
cost reflective prices which are location dependent with the balance to be
covered on a postage stamp basis. Postage stamp pricing involves a uniform
7
price per unit of capacity used irrespective of distance. The postage stamp
component is justified on the basis that full cost reflective pricing, as defined in
the code, would oversignal locational costs (Gawler 1997). The oversignalling
of locational costs would arise because the concept of cost reflective pricing as
proposed in the code is based on average cost pricing rather than marginal cost
pricing. Consequently, given that incremental cost is typically lower than
average cost, a pricing method based on the latter would clearly inflate location
based costs. A further justification for the postage stamp component i!s related
to government social policy objectives regarding rural development.
The common service rate resulting from the averaging of transmission and
distribution costs would clearly lead to distortions in terms of network use and
investment choices between additional network capacity, generation and
demand side options. In particular, by reducing the degree of cost reflectivity
through cross-subsidies, this approach would induce responses by market
participants which do not reflect the underlying costs of network service
provision. For example, the reduction in the degree of cost reflectivity could
potentially provide incentives for inefficient entry and the duplication of
network assets.
It is therefore important that alternatives for achieving social policy objectives
be canvassed as these objectives are often in conflict with economic efficiency
goals. At the very least, the extent of network pricing cross-subsidies should be
identified and funded explicitly through community service obligations.
•
The allocation of transmission charges
Another feature of the proposed pricing regime which is likely to distort
locational decisions is that generators are to pay network charges only for their
‘shallow’ connection costs (that is, the direct costs of connecting a new
generator to the network and not for the costs of any subsequent augmentation
elsewhere in the network). Fundamentally, customers pay for the bulk of use of
system and common charges.
This allocation of transmission costs can be expected to exacerbate the
distortions resulting from the postage stamp component of network pricing by
removing any incentive for generators to minimise transmission costs through
locational decisions. To the extent that the location of new generation facilities
is not regulated, this allocation would have the effect of creating additional
8
demand for transmission services, resulting in higher costs to consumers,
without any consideration of the benefits involved in locating generation close
to load centres. Moreover, the advantage of cogeneration in terms of savings in
transmission costs would be reduced.
If efficient outcomes in terms of location of generators are to be achieved, it is
critical that the costs of transmission which are affected by the location of
generation facilities be internalised as much as possible. The implication is that
generators should be allocated transmission charges which are fully cost
reflective and location dependent.
2.3 Network bypass
How can the network pricing arrangements encourage economic bypass, while
ensuring that uneconomic bypass does not occur?
Network bypass occurs when a network user invests in new capacity which duplicates
parts of the existing network infrastructure. The threat of bypass can be expected to
result in a degree of contestability for the services provided by incumbent network
service providers. In particular, the option to bypass ensures that incumbent network
owners cannot charge users more than it would cost to duplicate the service
themselves. The costs that a user would incur by bypassing a segment of the network
include the stand alone cost of establishing the necessary infrastructure in addition to
the variable costs attached to the level of network use (NECA 1997). The code does
not explicitly promote or restrict network bypass; however, there is a case for the code
to encourage efficient bypass which reduces total network cost.
Bypass is privately beneficial if the per unit cost of transmission or distribution along
the newly built segment of the network is less than the network price charged by the
incumbent network owner. However, socially efficient bypass may not coincide with
private incentives to bypass when externalities are present or when network prices are
not reflective of the costs of service provision. For example, a user’s private decision
to bypass is not likely to take into account the costs which are external to that user.
These may include the cost of stranded assets and the higher cost burden borne by the
remaining captive customers. In contrast, bypass is socially efficient or economic if
the cost to society of meeting the demand for network services is lower with a bypass
than without.
9
The likelihood of inefficient bypass increases if network prices are greater than the
cost of network service provision. One example of where network prices may not
reflect the cost of service provision is when a network owner is in a position to
exercise market power to earn monopoly rents. Equally, prices may not reflect costs if
network prices include some degree of cross-subsidisation for social policy reasons.
An example of this situation is the use of postage stamp pricing which results in users
paying an average of the cost of services over large segments of the network. Through
postage stamp pricing, users of low cost segments of the network will face network
charges which are higher than the cost of service provision.
The divergence between private and social incentives for bypass may be reduced by:
•
removing cross-subsidies inherent in average cost pricing;
•
allowing negotiation to reduce prices to a more cost reflective level for those
users who threaten to bypass; and
•
regulating against socially inefficient bypass.
To the extent that it may not be feasible to remove cross-subsidies, allowing different
network prices to different users may represent a more viable means of reducing
inefficient bypass. If users threaten to bypass in response to prices which do not
reflect the costs of service provision then a system of negotiation between these users
and the incumbent network owner may reduce inefficient bypass (Groom 1996).
Faced with the threat of bypass, an incumbent network service provider would be
willing to reduce its network price to a more cost reflective level and, hence, the
negotiated outcome would make both parties better off than if the bypass had gone
ahead. However, this market based approach for dealing with non cost reflective
prices does not provide a solution to the problem of external costs which are imposed
on stakeholders other than the negotiating parties. Similarly, if the incumbent network
owner is in a position to exercise market power over the captive users,! the
negotiation approach may lead to welfare losses as the incumbent attempts to recover
the revenue forgone through higher charges to the remaining users.
The scope for inefficient bypass could thus be reduced through a regulatory system
which allows bypass only when it can be shown that the net public benefit of bypass is
positive. This will require that regulators be able to identify the costs and benefits of
bypass proposals on a case by case basis.
Measures which remove any divergence between the cost and price of network
services clearly provides a first best approach for reducing the potential for inefficient
10
bypass. However, the presence of externalities provides a role for a regulatory body to
assess the social consequences of any private decision to bypass.
11
2.4 Contracts for network capacity
What is the potential for contracts to be introduced for aspects of network service?
Under the existing arrangements for network pricing, annual tariffs are used to
allocate costs to network users. However, the pressure for capacity contracts is
increasing as network users request differential standards of network services such as
firm access and standby service (NECA 1997).
One of the key advantages of capacity contracts over tariffs is that tradable property
rights assigned to network capacity provide a means of allowing buyers and sellers to
establish a market-clearing price for the use of network capacity. However, property
right structures which would lead to an efficient allocation of network capacity need
to be exclusive in the sense that all benefits and costs accrued as a result of owning
and using network capacity are captured exclusively by the owner of the property
right, transferable and enforceable (Tietenberg 1992).
While some categories of network services satisfy these conditions, other types of
network services are less amenable to a system of network access rights. In particular,
when network capacity rights cannot be well defined the use of existing infrastructure
and new investment resulting from a system of capacity rights are likely to be
distorted.
Capacity rights could be applied to new major interconnectors which are independent
of the meshed transmission and distribution networks. The auctioning of these rights
would provide a means of introducing competition in the provision of additional
network capacity.
In addition to using access rights for new interconnectors, a system of transferable
transmission rights could also be adopted for the provision of firm or uninterrupted
access to network users. This would allow the development of a market based
mechanism for constrained transmission capacity with holders of these rights able to
resell their rights in a short term secondary market.
There are several important issues involved with a market based system for allocating
constrained capacity. A major issue is the potential for the exercise of market power
in the access rights market. If the market for firm access rights is not competitive,
holders of transmission capacity rights may be in a position to raise the price of firm
access above the competitive benchmark by increasing transmission constraints. For
12
example, large generation portfolios could behave strategically by creating congestion
costs for rival generators through their bidding strategies. Therefore, it will be
necessary to monitor market outcomes if the scope for the exercise of market power is
assessed to be significant.
Conclusion
The key focus of the NECA review is on the allocation of network costs to network
users in a manner which is consistent with economic efficiency objectives. An
efficient network pricing regime is characterised by a high degree of cost reflectivity,
with marginal cost as the benchmark. The various considerations that may reduce the
feasibility of marginal cost pricing call for pricing arrangements that minimise the
welfare losses associated with deviation from marginal cost.
However, any pricing regime designed to deliver efficient outcomes would require
that regulators have access to information on demand and cost profiles of network
services. It is virtually impossible to formulate network prices which would provide
appropriate signals for network use and investment without information on demand
elasticities, market participation elasticities, short and long run marginal costs, the
extent of cross-subsidies and capacity constraints. Regulators can generally be
expected to have less information on these variables than network owners themselves.
It will therefore be important that this asymmetry be addressed through independent
studies and comparative assessments where possible.
References
Armstrong, M., Cowan, S. and Vickers, J. 1994, Regulatory Reform: Economic
Analysis and British Experience, MIT, London.
Baumol, W. and Sidak, J. 1994, Toward Competition in Local Telephony, MIT,
London.
Berg, S. and Tschirhart, J. 1988, Natural Monopoly Regulation, Cambridge
University Press, England.
Brown, S. and Sibley, D. 1986, The Theory of Public Utility Pricing, Cambridge
University Press, England.
13
Bushnell, J. and Stoft, S. 1997, ‘Improving private incentives for electric grid
investment’, Resource and Energy Economics, vol. 19, no. 1, pp. 87–108.
Gawler, R. 1997, Electricity transmission issues, Paper presented at the IIR
Conference, ‘The Electricity Spot Market’, Melbourne, 12–13 February.
Groom, E. 1996, Implementation issues, presented at the ACCC’s National Electricity
Market network pricing forum, Canberra.
Hinchy, M. and Low, J. 1993, Electricity Transmission and Bulk Pricing under
Deregulation, ABARE Research Report 93.14, Canberra.
King, S. 1995, Access Pricing, Discussion Paper for the Government Pricing Tribunal
of New South Wales, Research Paper no. 3, Sydney, February.
National Electricity Code Administrator (NECA) 1997, Transmission and
Distribution Pricing Review: Issues in Network Pricing, Sydney, December.
Tietenberg, T. 1992, Environmental and Natural Resource Economics, third edition,
HarperCollins, New York.
14
15
Download