Draft Determination

on the Proposed
Access Undertaking
of AGL Gas
Companies

Submission

by

London Economics,
Joshua Gans

& John Kay

October 1996
Draft Determination on the Proposed
Access Undertaking of AGL Gas
Companies

Submission

by

London Economics,

Joshua Gans

& John Kay

October 1996
Contents Page

1. Introduction 1

2. Terms of reference 2

3. Roles and objective of regulation 4

3.1 Regulation to facilitate bargaining 5
3.2 Regulation to facilitate competitive market outcomes 5
3.3 Regulatory threats 6
3.4 Minimising adverse incentives 6
3.5 Regulation in the NSW gas market 7
3.6 Summary 7

4. The design of the overall price control 8

4.1 Options for regulation 8
4.2 The length of the price review 11
4.3 Summary 12

5. Calculating allowable revenue - practical and theoretical

considerations 13
5.1 Valuation of assets 14
5.2 The cost of capital 16
5.3 Operating expenditure 17
5.4 Capital expenditure 18
5.5 Depreciation 18
5.6 Summary 19

6. Structure of access charges and the promotion of competition

20
6.1 Barriers to entry 20
6.2 The structure of prices 22
6.3 Administered charges versus price negotiations 23
6.4 The definition of the cross-subsidy 24
6.5 Summary 24

7. AGL’s incentives 26
7.1 Regulatory transparency 27
7.2 Potential for adverse incentives 28
7.3 Summary 30

London Economics i
October 1996
Contents Page

Notes 32
Annex 1: Asset valuation 34
Annex 2: Calculation of allowable revenues 36

London Economics ii
October 1996
Executive summary

Executive summary
1. London Economics has been asked by BHP Petroleum to provide an
independent review of the proposed gas access price regulation
arrangements in NSW, as outlined in the Independent Pricing and
Regulatory Tribunal’s (IPART’s) Draft Determination of September
1996.

2. This submission summarises London Economics’ findings in this

respect.
We comment on two central aspects of network regulation:

§ imposing constraints on the tariffs that may be charged by a

monopoly network owner to protect users and final customers;
and

§ structuring tariffs to ensure that effective competition can take

place in the presence of a vertically integrated network owner.

The role of regulation

3. Substantial parts of the gas industry - the transmission and the

distribution network are essentially natural monopolies. The role of
regulation is to set in place the conditions for maximising the value
created by the industry. To theextent that the regulator is able to make
credible threats, this may not require ongoing regulatory intervention.

4. The effectiveness of regulation is likely to be limited by a regulated

utility’s incentive to exploit its informational advantage
in relation to the
regulator in order to increase revenues and profits.

Overall price control

5. The range of options for a regulatory framework range from rate of

return regulation to price cap and average revenue
regulation. All forms
of regulation, including the regulatory approach that IPART is
considering, suffer from adverse incentives by regulated utilities to
exploit the informationasymmetry that is present between the regulator
and the regulated company.

6. The length of time between reviews of the price control reflects a trade-off
between providing efficiency incentives to the regulated company and
transferring efficiency gains to customers. This balance
is a key factor in
determining the sustainability and legitimacy of the regulatory regime.
IPART’s recommendation of a three-year review process is probably
appropriate in this context.

London Economics iii

October 1996
Executive summary

Allowable revenues

7. The calculation of allowable revenues derived from the overall price

control relies on the appropriate definition and measurement of a
number of key components. In practice, this has presented significant
problems in the history of UK utility regulation. These issues are
independent of the design of the overall price control.

8. IPART has not addressed a number of central issues that are critical in
determining allowable revenues.

9. Where the valuation of assets is concerned, the proposed approach:

§ is essentially circular;

§ fails to address the question ofan appropriate opening value for

assets and associateddistributional issues between shareholders
and customers; and

§ does not comment on how asset values will be rolled forward

which is likely to affect AGL’s capital expenditure incentives.

10. The cost of capital discussion in the Draft

Determinationappears to focus
on AGL’s actual financial structure. This is likely to present AGL with
adverse regulatory incentives to adjust its capital structure to increase its
cost of capital.

11. IPART do not comment on how appropriate levels of operating

expenditureshould be determined. Without further investigation, the
Regulator has no means of determining appropriate levels of operating
(and capital) expenditure, and what efficiencies AGL is capable of. AGL
should be required to undertake independent yardstick competition
analyses or detailedefficiency studies as a means of assessing potential
cost improvements.

12. Determination and measurement of capital expenditureand depreciation

play an important role in determining incentives by the regulated
company to over- or underinvest:

§ the circular approach to asset valuation could discourage

investment that should optimally occur.

§ In the absence of clear regulatory guidelines for the depreciation

approach, there are likely to be credibility problems and AGL will
again have an incentive to distort its short and long-term
decisions.

London Economics iv
October 1996
Executive summary

The structure of access charges

13. AGL’s incentives to erect barriers to entry to undermine competition

have
only been addressed in passing. We are particularly concerned with
AGL’s apparent ability to:

§ undertake strategic price discrimination;

§ cross-subsidise different network activities; and

§ charge predatory prices.

14. These concerns are directly linked to AGL’s vertically integrated structure
and its role as a competitor in the gas supply market.

15. It appears that AGL’s proposed Reference Tariffs are structured

according to a standard ‘two-part tariff’approach. However, it is likely
that a marginal cost pricing approach will over-recoverAGL’s allowed
revenues.

16. We have concerns that the emphasis on negotiated as opposed to

administered charges leaves AGL with the ability to
price discriminate in
order to undermine competition. Ultimately, the only feasible route to
fair competition may be to unbundle charges for gas trading services
from
those for the gas transportation business.

17. While IPART’s definition of what constitutes a cross-subsidy based

is on
avoidable costs, their definition of these costs appears to deviate from
common economic usage.

AGL’s incentive structure

18. The UK experience illustrates that

the benefits from increased regulatory
commitment and accountability may be limited if this undermines the
exercise of informed judgement. However, uncertainty regarding the
terms of access to AGL’s network may discourage investment by other
industry participants.

19. More generally, we have some sympathy with IPART’s concerns about
adverse incentives on the part of AGL to manipulate a single driving
determinant of asset valuation and revenue requirements, such as the
return on rate base. However, the proposedlack of certainty in the price
regulation process does not imply that the incentives to AGL to
manipulate future reviews are eliminated or even diminished.

London Economics v
October 1996
Executive summary

20. Revenue determination based on a basket of financial indicatorsdoes not

remove adverse incentive effects. Indeed, it offers AGL insurance against
any potential risk of losses from these actions. This is the opposite of
good incentive creation in regulation which would reduce the firm’s
insurance against uncertainty in order to ensure that it internalisesthe
risk of poor decision-making.

21. AGL may take actions such that strong regulatory steps will harm its
financial indicators without reducing its anticipated revenues in the
absence of such steps. Such actions include signing take or pay contracts
that are sensitive to redetermined revenues or undertaking risky
investments.

22. More generally, the implications of the regulatory regime for remaining
gas industry participants have not been considered. In effect, IPART has
proposed to put in place a framework characterised by a substantial
degree of uncertainty regarding the basis on which prices will be reviewed
at the end of the next review period.

Conclusions

23. The issues raised in this submission are in manyways endemic to access
price regulation. The concerns raised here outline the need for a careful
review of the structure and particulars of the industry at hand, especially
in any initial determination. Today, regulatorsplay an important role in
setting the agenda for other future determinations and thus, special
attention must be paid to the signals sent by current determinations.

London Economics vi
October 1996
Section 1 Introduction

1. Introduction
The Independent Pricing and Regulatory Tribunal’s (IPART’s) Draft
Determinationof September 19961 describes AGL’s proposed undertaking
to provide access to third parties for the use
of its infrastructureassets, and
IPART’s preliminary findings in this regard. This submission by London
Economics on behalf of BHP Petroleum reviews and comments on the
approach proposed by IPART.

The IPART Draft Determination covers a broad range of issues and

concerns. In this paper, drawing onthe UK experience to date, we focus on
those that we believe to be the most important:

§ Section 3 of this paper reviews the roles and objectives of regulation.

§ The trade-offs associated with different forms of regulation and the

length of the review period are discussed in Section 4.

§ Section 5 reviews essential considerations for determining an overall

price cap. The manner in which each of these inputs derived
is is likely
to have profound implications for the operation of the proposed
regulatory regime.

§ Section 6 addresses the structure of relative prices and key concerns

for
the development of competition in NSW.

§ Section 7, finally, considers how the specific form of

regulation is likely
to influence AGL’s incentives.

This document is supported by two Annexes:

§ Annex 1 contains an illustrativecalculation linking together key items

in the regulatory formula.

§ Annex 2 provides an overview of asset valuation techniques.

London Economics 1
October 1996
Section 2 Terms of reference

2. Terms of reference
IPART’s Draft Determination

IPART’s Draft Determination comes in the context of a developing

regulatory structure for gas in NSW. Prior to the introduction of a formal
regulatory framework in 1990,Boards of Inquiry administered a system of
‘reasonable profit’ to determine gas tariff prices.2 After this time, the Gas
Council of NSW placed AGL Gas’ five companies under separate CPI-X
price control formulae. Following a request by AGL in 1993, the price
control formula was replaced with a Combined Maximum Average Price
(CMAP). The formula set the average tariff charge based on the costs of
production, government fees, profit and CPI-X for the industry as a whole.

IPART highlight the importance of competition:

It needs to be stressed that

the overridingobjectivein providingaccessto
distribution networks is to facilitate competition in upstream and
downstream markets. Access to these networks is a means to an end;
it is not an end in itself. (Foreword, page i)

The Draft Determination for access price regulation

in NSW gas distribution
is based on the following key features:

§ a form of average revenue regulation where maximum prices are set

with reference to the forecast ‘sustainable revenue’ stream of AGL;

§ ‘sustainable’ revenue is determined with reference to a basket of

financial indicators;

§ there are no limitations on contract negotiations;

§ periodic review of price caps; and

§ transitional provisions.

IPART reject strict reliance on the ‘return

on rate base’ principle as a driving
determinant of revenue requirement of the network. Rather, the
determination of the level ofrevenue regarded as ‘sustainable’ is made with
reference to a basket of financial indicators.

London Economics 2
October 1996
Section 2 Terms of reference

These indicators include the rate of return on assets, but also:

§ the nature of contracts signed;

§ the cash position of AGL;

§ risk measures based on debt and equity positions;

§ historical cost valuations of AGL assets; and

§ operating cost measures.

The Draft Determination also makes reference to a number of other

important issues, including:

§ the definition of cross-subsidies;

§ the appropriate rate of return, given AGL’s business risks and

financial
structure;

§ the role of past user contributions to network infrastructure;

§ the definition and structure of reference tariffs;

§ as well as non-price issues, including access principles, reference price

services and key terms and conditions of supply.

London Economics’ terms of reference

This submission comments on the two central aspects of network regulation:

§ imposing constraints on the tariffs that may be charged bymonopoly

a
network owner to protect users and final customers; and

§ given the objective of achieving competition in the supply of gas,

structuring tariffs to ensure that effective competition can take place
in
the presence of a vertically integrated network owner.

London Economics 3
October 1996
Section 3 Roles and objective of regulation

3. Roles and objective of regulation

Substantial parts of the gas industry - the transmission and distribution
network - are essentially natural monopolies. Investmentin infrastructureis
substantial with material and sunk costs. The characteristics of the
infrastructureare such that duplication is uneconomic, and competition in
infrastructure provision only tends to occur where there are substantial
volumes of trade or tariff distortions encouraging by-pass by customers.

Regulation has a positiveeconomic role in industries such as gas because it

can potentially ensure that the weighted sum of consumer and producer
welfare is maximised. That is, regulators can take actions
which ensure that
value created in the industry is as large as possible. Without regulation,
different parties have incentives to take actions which maximise their own
value at the expense of others. Monopoly power is the archetypal such
action, involving one firm choosing prices in order to maximise its own
revenues. This notonly reduces the value captured by consumers, but also
the total amount of value created - creating so-called dead-weight losses.

In the case of access to gas distribution, we are concerned with the value
created by the industry today, as well as in the future. This is determined by
the decisions of:

§ gas producers regarding what quantity of gas to supply and whether

to invest in expanding their production facilities;

§ gas distributors as to how they operate pipelines,

the technologies they
choose, and the capacity they invest in; and

§ final consumers and users as to how much gas they choose to use.

Deciding the optimal combination of these decision is a difficult

problem. It
is certainly beyond the role of any regulator to monitor and
ensure that these
decisions are made appropriately. However, under two theoretical
benchmarks, appropriate decision-making can be achieved even when the
relevant parties behave in a unregulated and non-cooperative fashion:

§ When the parties (producers, distributor and consumers) are

identifiable and there are no significant
transactionscosts, constructive
bargaining can take place and they will reach agreements that
maximise value created. Their relative bargaining strengths will then
determine how this value is divided between them.

§ When markets involve a large number of participants who cannot

efficiently engage inface-to-face negotiations, perfect competition can,
theoretically,ensure that their individual, self-interesteddecisions will
be based on price signals that accurately reflect the costs and benefits
of particular arrangements.

London Economics 4
October 1996
Section 3 Roles and objective of regulation

Thus, rather than intervene directly in decision-making, therole of regulation

is to set in place the conditions for favourable
decentralised outcomes. That
is, the regulator can take actions that facilitate efficient bargaining,
competitive outcomes or some combination of the two. We deal with each
of these in turn.

3.1 Regulation to facilitate bargaining

The efficient bargaining arrangement is an appropriate benchmark when
there are a small number of parties negotiating. The relationship between
gas producers and a distributor is one such example.

However, efficient bargaining outcomes may be difficult to achieve. Two

broad reasons involve information asymmetries and investment hold-up:

§ Information asymmetries occur when one party does not know the value
of the relationshipto the other. This can lead to inefficient bargaining
breakdown as each party tries to negotiate better
a deal for themselves
at the risk of causing the other party to withdraw from further
negotiations.3

§ Investment hold-up occurs when bargaining over price and other

elements can take place after key investments have been made. In this
environment, parties may distort their investment decisions to improve
their later bargaining position, or because of a fear that they will not
earn a sufficient rate of return after bargaining has taken place.
This is
sometimes termed ‘hold-up,’ referring to a situation where one party
4
delays trade in order to renegotiate contract terms.

When bargaining is not efficient, regulation can facilitate value creation by

providing compulsory arbitration of disputes. This canreduce the incidence
of bargaining breakdown and secure the returns to earlier investments.

3.2 Regulation to facilitate competitive market outcomes

In markets with a large number of participants, face-to-face bargaining is
not practical, nor is it likelyto be efficient under these circumstances. Here,
competitive pressure can ensure that prices reflect the costs of production
and hence can form an appropriate basis for value enhancing decisions. The
retail market for gas is an example of an environment with too many
participants to support face-to-face bargaining.

London Economics 5
October 1996
Section 3 Roles and objective of regulation

However, when competitive pressures are imperfect (i.e., there are

monopolistic tendencies), prices no longer send signals which result in
efficient consumption. This, in turn,alters the decisions of participants in a
way that reduces value created. The role of regulation in this circumstance
is to provide mechanisms by which the correct prices will be generated and
all market participants will be able to base their decisions on appropriate
information.5

3.3 Regulatory threats

The above discussion has implied that regulation takes an active role in
facilitating value creation. However, it
is possible that this positive role can
be achieved without ongoing and explicit intervention. If the regulator can
be seen as a potential player, this may constrain undesirable behaviour.

For example, without regulation, a monopolist network owner will set

transport charges too high, resulting in dead weight losses. If there is
effective regulation, prices will be brought down, but regulation is a costly
activity. If the regulator can make a credible commitment to intervene if
prices are observed to be too high, a monopolist fearing this situation may
limit price rises so as not to provoke a regulatory response. While not as
effective as active regulation, this mechanism has the benefit that it avoids
regulatory costs.6

However, the threat of regulatory action must be credible. If a regulator is

perceived to have observed a breach in conduct and fails to take action, the
threat of regulatory intervention and associated costs to the utility loses its
significance.

Beliefs about what a regulator might do can have a profound influence on

the behaviour of market participants and this is the subject of our discussion
of AGL’s incentives in Section 7. While this may lead to positive outcomes,
it might also create adverse incentives (see below) to engage in behaviour
designed to influence anticipatory regulatory actions.

3.4 Minimising adverse incentives

There are limits on the ability of regulation to facilitate value creation.
These
limits are directlyrelated to the information asymmetry that exists between
the regulator and the regulated utility. AGL has superior information to the
regulator about existing and efficient cost structures and levels, investment
programmes, technological choices and associated costs, and future trends
in demand for its services. This issue is also recognised by IPART. 7

London Economics 6
October 1996
Section 3 Roles and objective of regulation

There is therefore a fundamental conflict of interest arising from the

fact that
the information needed for effective regulation comes from the regulated
utility itself, which is interested in securing flexibility to raise prices.
Moreover, different forms of regulation themselves generate adverse
incentives. These incentives involve interested parties attempting to ‘game
the system’.8

Thus, the role of regulation is to facilitate welfare enhancing outcomes

subject to the constraint of minimising adverse incentives. In some cases, the
costs of such adverse incentives or gaming can be so great as to render
virtually impossible the facilitating role of regulation.

3.5 Regulation in the NSW gas market

The NSW gas market iscurrently structured around AGL which acts as the
sole, vertically integrated distributor/supplier of gas to all final customers.
Thus, while the producer-distributor relationship might be characterised by
face-to-face negotiations, the distributor-user relationshipis more of a mass
market. What is the role of regulation in such a mixed trading environment?

To see the critical issue here, suppose that the regulator

effectively facilitates
efficient agreements among the producersand AGL. Those agreements will
be based on their interests alone and are likely to be the same
as if producers
and distributor acted as a pure monopolist or group of oligopolists in the
consumer/user market.

This is clearlynot a desirable outcome. Therefore, regulation involves more

than facilitating efficient agreements among producers and AGL. It must
ensure that the final prices resulting from that agreement are reflective of
costs so that users are able to make decisions based on appropriate
information. Regulatory rules must balance the interestsof producers, AGL
and consumers in facilitating agreements, and effective regulation involves
limiting the set of agreements reached between upstream suppliers so that
they reflect the interests of final users.

3.6 Summary
As a vertically integrated gas distributor / supplier, AGL is in a strong
position to exert market power and to reduce valuecreated by the NSW gas
industry. The role of regulation is to put in place a framework that
facilitates efficient industry outcomes by balancing the interests
of upstream
producers (existing and new), AGL and final customers by playing an active
role and the use of regulatory threats. However, the effectiveness of
regulation is limited by the degree to which a regulated utility
can exploit its
information advantage to undermine regulatory objectives.

London Economics 7
October 1996
Section 4 The design of the overall price control

4. The design of the overall price control

This Section considers alternative forms of utility regulation which set the
context for the approach that IPART is contemplating. We consider the
practical and informational requirementsfor implementation, the incentives
for efficient operation and investment and then comment on the frequency of
the regulatory review.

4.1 Options for regulation

The tension between confiscating profits and inducing productive efficiency
is at the heart of the regulation debate. It is encapsulated in the contrast
between two types of regulatory contract or price rules:

§ Cost plus regulation, or rate of return regulation comprises a profits

confiscationrule that aims to achieve allocative efficiency by relating
price to reported marginal or average cost.

§ Fixed price regulation comprises a rule that allows

the public utility to
be the residual claimant to the profits achieved by lowest cost
productive efficiency.

The fact that we observe neither of these two regimes practised for very long
is due to another consideration coming into play, which is that the regime
must be sustainable. Public utilities supply essential services; moreover,
they tend to be the sole supplier of that service in a particular region or
nation. These characteristics mean that the public do not expect to be
exploited, either through the inefficiency or bad management of the utility
induced by one system of regulation, or through the accumulation of
excessive profits by shareholders induced by the other system.

4.1.1 Rate of return regulation

Rate of return regulation aims to give the regulated firm an incentive to keep
prices low. Without regulation, the firm will choose prices so as to maximise
profits. Therefore, by limiting its profits, it can indirectly be forced to keep
prices low.

In practice, regulation of this form sets the rate of return the firm can earn
off its capital assets. So long as this rate of return is determined
appropriately, it will give the firm the incentive to invest optimally in its
capacity. This has several advantages:

§ the firm is free to price to respond to changing cost and demand

conditions; and

§ the nature of future investment is flexible.

London Economics 8
October 1996
Section 4 The design of the overall price control

However, there are several (well-known) limitations to rate of return

regulation:

§ given that the rate of return is fixed (and profits will remain
unchanged), the firm has little incentive to operate to minimise costs;

§ in an access environment, setting the rate of return for the monopolist

removes its incentive to ensure that network users enter the marketin a
cost minimising way;9 and

§ the firm’s incentives to invest in capacity depend

not only on economic
conditions, but also on how the rate of return is measured.

There are means in which these incentive problems can be addressed,

although each of these, in turn, require substantial information on the part
of
the regulator:

§ if excess investment is a concern, a credible commitment to revalue

assets and not consider assets that have been developed
with excessive
cost will mitigate against this;

§ if the optimal technological choice cannot be easily observed, the

use of
alternative measures including benchmarking with best practice
elsewhere and a revaluation that takes best practice into account can
provide countervailing incentives;10 and

§ if there is a possibility of by-pass, this might provide the firm with

sufficient incentives to invest properly.

These measures may mitigate adverse incentive effects. However, the

discussion in Section 5 highlights that a range of fundamental theoretical
and measurement issues need to be addressed in the process.

More generally, adverse incentives arelikely to be reduced if the regulator is

able to issue credible threats to sanction
inappropriate behaviour. However,
such mechanisms will have a positive influence if they are explicit and the
regulator is able to commit to carrying them out in the future. We return to
this issue in Section 7.

4.1.2 Price caps

Price cap regulation determines a maximum price that the network owner
can charge in any given year.In practice this involves determining a rule by
which an initial maximum price isset, and subsequently adjusting the price
to grow at some rate below inflation (i.e. CPI-X). Intheory at least, the X is
set to reflect productivity growth in the industry and is reviewed
periodically.

London Economics 9
October 1996
Section 4 The design of the overall price control

Price cap regulation tends to providebetter incentives for the regulated firm
to operate efficiently. This is because the firm can capture in profits the
benefits of any productivity improvements until the next review period.
However, in practice, afirm considering lowering costs might hesitate to do
so prior to a regulatory review, in orderto ensure that the X is set low upon
review. This ratchet effect undermines the incentive benefits of price caps.

Another potential disadvantage of price caps on individual prices is that

these can limit the utility’s ability to negotiate terms with other market
participants. Subject to competition policy considerations (see Section 6),
this can be addressed by allowing renegotiation to take place on the basis
that other market participants are at least as well off under alternative
arrangementsas they are if they purchase the service at the price cap level.
Such negotiations are permitted under the IPART Draft Determination.

The effectiveness of price cap regulation is limited as a result of the

information asymmetry problem we have referred toabove and concerns the
estimation of potential productivity improvements - this is discussed in
more detail in Section 5. It should be noted that the methodology that has
emerged in the UK, which identifies a strong role for the recovery of a fair
rate of return, is quite different to the early expectations of how CPI-X
regulation would work. In the 1980sit was envisaged that X would simply
represent an estimate of future productivity growth - it is clear that
in fact X
captures far more than that simple statistic.

4.1.3 Average revenue regulation

Average revenue regulation (sometimes called the revenue yield approach) is
a form of average price regulation, where total revenue divided by total
output is constrained to be no greater than a price cap and some
adjustment
factor.11 IPART is considering a form of total revenue regulation where
AGL’s total revenue stream from the provision of network services is fixed,
and this revenue must be raised from network users.

Average revenue regulation allows the regulated firm price rebalancing

within the confines of the overall restriction. Within a vertically integrated
structure, this raises competition policy concerns (see Section 6).

Average revenue regulation can only be applied when products are

commensurable, for example, when the same output is delivered to different
geographic areas. Under average revenue regulation, as it is carried out in
the UK, TransCo’s allowed revenue depends on the total number of therms it
transports. TransCo therefore has incentives to expand the volume of gas
transported while underestimating future sales prospect. In order to
address this, the UK regulator is currently considering linking allowed
revenue to three different elements - a capacity element and different
commodity elements for large and smaller loads.

London Economics 10
October 1996
Section 4 The design of the overall price control

4.1.4 Sliding scale regulation

Sliding scale regulation aims to create positive
incentives for regulated firms
to implement cost savings by sharing the resulting benefits with customers.
Various forms of sliding scale regulation have been introduced in the US
where state regulators have developed indices linked to either utility prices
or revenues. Most plans specify ranges within which there is no sharing of
risks (a so-called‘deadband’), broader ranges within which sharing in some
proportion is to occur and limits beyond which automatic reviews and / or
adjustments are required.

The most substantial group of issues to be tackled in implementing sliding

scale regulation is concerned with the measurement of profits (or profit
equivalents). Theseproblems concern both the base level (however defined)
and verifying reported figures. Given the measurement and accounting
difficulties raised by this, it hasbeen suggested that the differences between
operating cost & revenue projections and outturns are shared, because these
are easily measurable, cashbased concepts. In the absence of unambiguous
guidelines, sliding scale regulation may aggravate the problem of gaming
between regulator and regulated company, and lead to a significant increase
in the intrusiveness of regulation. 12

4.2 The length of the price review

The design of the price control has significant implications for the efficiency
and sustainability of any regulatory regime. Even under the most
meticulously conducted price review, it is inevitable that the outcomes for
operating and capital expenditure will diverge from their expected values.
Moreover, given the informational advantage of the firm, they are likely to
differ in only one direction. Consequently, the important issue, for both the
efficiency of the overall price control, and its sustainability, is how
frequently
unanticipated gains are shared with customers.

Early proponents of price cap regulation would argue that price controls
should only be reset infrequently, if at all. Advocates of this form of
regulation such as Starkey and Van Pelt(1995) do, however, draw attention
to the key premise underlying this form of regulation, namely, that

increasedprofits for the firm will be viewed by regulators and their

constituencyas somethingother than a failureof regulationitself. If this
premise is false then the regulators will be under constant pressure to
13
recontract when the firm reports higher profits.

London Economics 11
October 1996
Section 4 The design of the overall price control

It is the understanding that this premise is generally false that gives rise to
regulatory mechanisms that depart from this very high powered incentive
contract. At the other end ofthe scale, complete clawback of profits due to
unanticipated operatingand capital expenditure efficiency gains is likely to
have an unacceptably adverse impact on incentives. The important
judgement to be made is where to settle between these two extremes.

Having decided where to settle, the instruments should be agreed:

§ Is the regime to be a fixedprice control with reviews every three years,

or is a price formula adjustment mechanism to be adopted over a
rather longer period?

§ Furthermore, once the transitional

period is over and the regulator has
gained experience, would it be possible to extend the regulatory lag?

The experience of the UK is that its regulatory system has

been weakened by
not appearing to include customers in the gains that were accruing to
shareholders and managers of the regulated businesses in the first few years
after privatisation. By the time the gains began to flow to customers
following the first quinquennial reviews, the system had ceasedto command
broad public support and has become unsustainable.

Thus, the recommendation inIPART’s Draft Determinationto review prices

after three years isa welcome acknowledgment that all stakeholders should
be included in the benefits of incentive regulation.

4.3 Summary
It is clear that the design of the price control can have significant
implications for the efficiency and sustainability of the regulatory regime.
Even under the most meticulously conducted price review, it is inevitable
that the outcomes for operating and capital expenditure will diverge from
their expected values. Moreover, given the informational advantage of the
firm, they are likely to differ in only one direction. Consequently, the
important issue, forboth the efficiency of the regime, and its sustainability,
is how frequently unanticipated gains are shared with customers.

London Economics 12
October 1996
Section 5 Calculating allowable revenue -
practical and theoretical considerations

5. Calculating allowable revenue - practical

and theoretical considerations
Put at its simplest, the job of a regulator is to determine allowable revenue
for a period of time, that is to determine the price of the regulated output for
a period of time. The number of inputs into this process is small. They are:
§ an opening asset valuation;

§ a cost of capital;

§ expected values for operating expenditure;

§ expected values for capital expenditure; and

§ expected values for depreciation.

The approaches by which allowablerevenue can be calculated is reported in

Annex 2. Energy regulators in the UK have adopted the
cash-flow approach
to setting prices.

While the derivation of allowed revenue is straightforward, the measurement

of each variable raises quite different practical and
theoretical considerations
depending on which variable is being considered. It appears that these
issues have not been considered by IPART in sufficient depth:

§ The measurement of the asset value and the cost of capital raise
distributional, commitment and credibility issues that are largely
invariant to the design of the regulatory constraint.

§ The measurement of the operating and capital expenditure and

depreciation highlight the problemof asymmetric information between
the regulator and the regulated company. How each variable is
calculated, especially operating expenditure, will have key implications
for the incentive properties of the regulatory system.

Before considering each ‘building block’ in turn it is important to recognise

that the approach to price setting which has evolved in the UK is quite
different to what is being proposed for AGL. The proposed IPART
approach is to calculate revenue requirements based on a range of
indicators, and then to ensure that the present value of the net revenue
stream is reasonable, that is, it lies between historic and replacement cost
asset values. While we recognise the need for regulatory discretion, this
approach to price setting is circular and potentially capricious, as we
discuss in the next section.

London Economics 13
October 1996
Section 5 Calculating allowable revenue -
practical and theoretical considerations

5.1 Valuation of assets

The opening asset valuation should be the fixed point in any system of
regulation. An asset valuation that is based purely on the present value of
net revenues is circular and leaves the shareholder at risk of capital
expropriation, which will require compensation through a higher rate of
return. Moreover, such an approach sets no clear standard for judging the
efficiency of tariffs, or its distributional implications. Different approaches
for asset valuation are set out in Annex 2 to this document.

The cost of a marginal investmentis based on the actual cost of purchasing

the asset and the opportunity cost of capital. Applied to a network, an
efficient set of prices would recover:

§ the opportunity cost of capital on the replacement cost of the assets;

and

§ the consumption of the assets (or full replacement cost depreciation).

There may be good distributional reasons why regulators may be unwilling

to allow the firm to recover such costs in their entirety, but the Draft
Determination does not touch on these.

It appears that the only constraint on IPART’s discretion on setting

allowable revenues is that the present value of those revenues should lie
between historic and replacement cost values. An indication that this
approach to price setting is circular is summarised by two quotes on two
consecutive pages of the Draft Determination:

The Tribunal wishes to emphasisethat the rate of return on the

regulatory asset base is only one indicator used
testto
the reasonableness
of the sustainable revenue stream and the resulting prices (page 19).

Note that the sustainable revenue stream has been used to determine the
asset value, not that the asset value has been used to determinethe
revenue stream (page 20, emphasis added).

These two remarks appear to be circular. The opening value of the assets,
and the basis for rolling the asset base forward, should be immutable.
IPART should reconsider the issue of asset valuation.

London Economics 14
October 1996
Section 5 Calculating allowable revenue -
practical and theoretical considerations

5.1.1 Customers’ versus shareholders’ interests

There are two obvious bases for deriving the opening value of assets:

§ the flotation / market value of the company (or a variant thereof); or

§ the replacement costs of the assets.

If a company is valued for less than the replacement costs of the

assets, then
allowing shareholders to receive a return on the entire replacement costs of
the assets will hand them a windfall gain. Thus, basing the opening asset
value on the market value preserves the existing distributional settlement
between shareholders and customers.

A separate question is how the asset base is rolled forward over time. itIfis
a regulatory objective that regulated prices should be efficient in the long
run, then a regulator may start from the market valuation of the company
and put in an automatic mechanism such that as assets are replaced the
regulatory value of the company tends towards thereplacement costs of the
assets - this is essentially the approach the Monopolies and Mergers
Commission took in the case of British Gas.

If, however, the market valuation of the assets is so far below the
replacement cost then it may not be feasible to reduce the gap because it
would involve an unacceptable redistributionof income from customers to
shareholders. This is the viewtaken by the water regulator in the UK - new
investment earns a full rate of return but even when all assets have been
replaced there will bea wedge between the regulatory and replacement cost
values.14

What is important is not necessarily the particular method chosen, but that a
method is chosen so that all stakeholders understand the distributional
settlement, and so that shareholders can be assured that their assets will not
be expropriated. The reluctance of IPART to specify an asset value will not
be emulated by commentators, customer groups and shareholders, eager to
establish the size of the slices of the cake accruing to particular groups.

The treatment of the opening asset value, and the manner in which it is rolled
forward requires a rule. The treatmentof the other inputs into the process
requires guidelines, as we now discuss.

London Economics 15
October 1996
Section 5 Calculating allowable revenue -
practical and theoretical considerations

5.2 The cost of capital

IPART’s cost of capital calculation applies a Capital Asset Pricing Model
(CAPM) approach to what appears to be AGL’s future capital structure:

In future years, the

capitalstructurewill changeto reflectthe applicationof
the Gas Customers Reserve Account (GCRA), and this will have a
corresponding effect on AGL’s WACC. (page 19)

We have not verified the precise nature of the AGL’s cost of capital
calculations. However, we would emphasise the following principles:

§ The calculation ofthe cost of capital should be open and transparent,

revealing data sources, measurement techniques, sample sizes and
other relevant information.

§ When establishing the cost of capital for a regulated company the

following guidelines should be followed:

1. Gearing (or leverage) of the company should be based on the

optimal level, rather than the actual level adopted by the
company. If the regulator permits the cost of capital to reflect
actual gearing levels, then the company may have an incentive to
choose a sub-optimal position at the price review which is then
adjusted post-review to allow the company to benefit.15

2. The cost of equity should be calculated on the basis of the

CAPM
with the correct adjustment for the equity beta to take into
account the optimal, rather than actual gearing. A range of beta
values, based on the standard error of the beta estimate, should
be considered. Where no actual beta value can be found,
comparative beta estimates should be considered - from other
domestic utilities and international gas companies - with
adjustments made to take into account:

⇒ differences in regulatory regimes; and

⇒ expected differences in non-diversifiable risk (arising from

differences in customer base etc).

3. Finally, the cost ofdebt should be measured as the risk-free rate

from the cost of equity calculation, plus a premium. The
premium should either be based on the observed premium for
bonds from this company over and above the equivalent
maturity government bond redemption yields, or a comparison
of comparator premia - derived from the same sources as the
comparator beta values.

London Economics 16
October 1996
Section 5 Calculating allowable revenue -
practical and theoretical considerations

5.3 Operating expenditure

In determining expectations of operating expenditure, capital expenditure
and depreciation we begin to consider the problem of asymmetric
information inherent in regulation. Quite simply, the regulator has no way
of knowing what operating expenditure efficiencies the firm is capable of.
The manner in which the regulator forms his or her expectations has
important implications for the efficiency of the regulatory regime.

IPART have not beenexplicit in how forecast efficiencies have been derived,
although the Tribunal and AGL are considering a forward looking revenue
path expressed as “CPI-13, 10, 7” (page 15).

There are a number of ways in which a regulator can form expectations:

1. extrapolation of past performance;

2. productivity elsewhere in the economy;

3. productivity elsewhere in the industry - yardstick competition; and

4. conduct detailed efficiency studies of the firm’s activities.

Extrapolation of past performance can be expected to have poor incentive

properties unless the firm is so myopicthat it does not realise that straining
to achieve a productivity rate now will simply result in
that rate forming the
basis for future price reductions.

Option 2. is likely to give unsustainable outcomes - depending on which

sector of the economy is chosen the firm will receive an undemanding price
regime, leading to protest from customers, or an excessively demanding
regime leading to financial distress for the firm.

In multi-firm industries, yardstick competition is likely to be an effective

way of both encouraging productive efficiency and reducing prices to
customers - essentially the regulator sets prices on the
basis of average costs
in the industry (oreven lowest cost in the industry) which encourages firms
to reduce costs which will result in lower prices. The difficulty with this
approach is that firms are not directly comparable, but there are a number of
techniques available to estimate efficiency, after taking into account firm
specific characteristics.

If the firm is a single monopolist, and there are no obvious comparators,

then in order to establish the scope for future efficiency gains the regulator
may need to send engineers and management consultants into the firm to
conduct efficiency studies of the firm’s processes and audit the forecasts of
future operating expenditure. On the basis of this information, the regulator
may be better equipped to understand what productivity gains can be
achieved.

London Economics 17
October 1996
Section 5 Calculating allowable revenue -
practical and theoretical considerations

Guidelines in this area should be minimal and should essentially relate to

best regulatory practice. Revelation of the basis on which operating
expenditure targets aremade may precipitate a strategic response from the
firm resulting in a sub-optimal long run outcome.

5.4 Capital expenditure

Capital expenditure allowances do not impact on allowable revenue by as
much as operating expenditure and depreciation, because only the financing
cost of capital expenditure (i.e. depreciation and the required
rate of return)
is covered by allowable revenue. 16

What is important is that the design of the regulatory regime itself should
not impart perverse incentives to over- or underinvest.Underinvestmentcan
occur when regulators cannot commit to remuneratingpast investment. In
this context, a circular approach to asset valuation will tend to discourage
investment that should optimally occur. This is because, under the
proposed IPART methodology, the asset value derived bytaking an NPV of
future revenues is only likely to be equal to the money shareholders have
actually invested by chance. This reinforces the need to fixthe opening asset
value and the manner in which it is rolled forward.

Overinvestment occurs if, given a consistent approach to asset values, the

firm expects the regulator to allow a rate of return greater than the true cost
of capital more often than he/she allows a rate of return that is lower than
the true cost ofcapital over the entire life of the asset. Since the chances of
this happening are not that great, whilst the downside is potentially quite
large, the incentives to over-invest in the classic Averch-Johnson sense are not
that serious. However, clawback of unspent capital expenditure, like
clawback of extraordinary profits resulting from unanticipated efficiency
gains, can be expected to be distortionary.In the UK, Ofgas have proposed
to clawback the revenue associated with unspent capital expenditure from
the previous period - this completely removes the incentive to economise on
investment plans, and tilts the balance strongly in favour of sub-optimal
over-investment.17

5.5 Depreciation
The issue of depreciation is strongly linked with both the asymmetric
information problem and the asset valuation problem discussed above.

Depreciation can either be calculated on a replacement cost or renewals

basis. In the former case the regulator must take a view on the extent to
which the replacement cost value of the assets will change from year to year.
If depreciation is calculated on a renewals basis, the regulator must take a
longer term view of capital expenditure requirements and set the
depreciation profile to ensurecash-flow neutrality over the period - this has

London Economics 18
October 1996
Section 5 Calculating allowable revenue -
practical and theoretical considerations

the advantage of avoiding the large gaps that can develop between capital
expenditure and depreciation.

As for the asset valuequestion, guidelines should be established which give

a clear indication of the methodology the regulator proposes to adopt -
otherwise credibility problems about the remuneration of past investment
will emerge. For example, a regulator could choose replacement cost
depreciation when capital expenditure is high relative to depreciation but
then switch to a renewals approach when the major investment programme
has finished. This would result in temporarily low prices but would deter
any new investment.

5.6 Summary
This Section has outlined how the calculation of each element of the
allowable revenue calculation can affect behaviour. These issues have not
been sufficiently addressed in the design of the proposed NSW access
regime:

§ IPART’s proposed treatment ofthe opening asset value appears to be

circular and the manner in which it
will be rolled forward has not been
clarified.

§ The calculation of AGL’s proposed cost of capital appears to have

been based on AGL’s current and forecast capital structure. Given that
this approach is likely to present AGLwith adverse gaming incentives,
the cost of capital should be calculated independently of AGL’s
position.

§ It is not clear how IPART intends to form a view regarding appropriate

levels of operating and capital expenditure.Yardstick and engineering
comparisons represent one approach for overcoming the information
asymmetry of the Regulator vis a vis AGL. At the same time, AGL
should retain sufficient incentives to invest appropriately.

§ Where depreciation is concerned, IPART should establish guidelines

which give a clear indication of the methodology the regulator
proposes to adopt in order to address credibility problems about the
remuneration of past investment.

London Economics 19
October 1996
Section 6 Structure of access charges and the
promotion of competition

6. Structure of access charges and the

promotion of competition
The NSW gas market is essentiallystructured around AGL which currently
acts as the sole, vertically integrated distributor/supplier of gas to final
customers. AGL has contracted with the upstream Cooper Basin producers
for gas supplies to all customer classes.

The previous discussion of the role of regulation has taken the structure of
the NSW gas industry as given. However, there are numerous issues that
arise because of the vertically integrated structure of the industry. Any
serious reforms aimed at producing competitive outcomes in the gas
industry have to address the question of the overall industry structure.

6.1 Barriers to entry

Concerns with insufficient competitive entry in vertically related markets
focus on the possibility of strategicbehaviour by firms to create barriersto
entry. The implication of this threat for
price regulation is that the ability of
the network access provider to engage in the following should be limited:

1. Strategicdiscrimination: placing itself at a cost advantage

relative to its downstream competitors by implicitly charging
access prices to itself lower than those charged to other firms.

2. Cross-subsidy: subsidisingunduly low margins downstream,

for example, with unduly high network access prices.

3. Predation:setting retail and network access prices at a level

which will prevent other equally or more efficient firms from
earning sufficient margins to enable them to remain
competitive, thus deterring entry or inducing exit from
upstream and downstream markets.

London Economics 20
October 1996
Section 6 Structure of access charges and the
promotion of competition

In order to achieve these aims there are two general principles of access
pricing, frequently referred to as the Baumol-Willigconditions,which are
designed to deal with economies of scale (where costs per unit fall
as output
increases) and scope (where total costs fall when activities are carried out
together rather than separately). The conditions may be stated as follows:

1. No price, or set of prices, shouldexceed the stand-alone costs

(SAC) of providing the service or services, where stand-alone
costs are determined as the costs an efficient competitor
would incur in providing just that service or group of services.

2. No price, or set of prices, should be less than the incremental

(or avoidable) costs of providing the serviceor services, where
incremental costs are the additional costs incurred by the
monopolist in providing just that service or group of services.

These are designed to mimic the constraints of contestable

marketsand should
ensure that a network access monopolist is unable to earn more than a
competitive rate of return, engage in predation or cross-subsidisation and
that inefficient bypass is deterred.

However these conditions do not select a singlepricing scheme and hence an

analysis of individual schemes is still necessary. IPART has set out the
following basic criteria which a network access pricing regime should satisfy:

1. charges should permit the infrastructureowner to recover its network

costs and to earn a “reasonable level of profit”;

2. charges should not price off the network those customers who could
cover the avoidable costs of their use of the network;

3. charges should preclude inefficient network by-pass; and

4. charges should facilitate the development of efficient upstream and

downstream competition.

How the first of the above objectives can be achieved has already been
considered in Section 4. In this Section we consider the latter three goals. It is
intended that the Regulator will determine Reference Tariffs for network
access services which are compatible with these aims. However, these
Reference Tariffs are not to preclude a customer and a network service
provider from entering into individual negotiations to determine an
alternative price which reflects cost and market conditions. Within the
framework of access pricing principles, negotiation is
viewed by IPART as a
vital element.

London Economics 21
October 1996
Section 6 Structure of access charges and the
promotion of competition

6.2 The structure of prices

First, however, what structure should prices take? The structure and
derivation of the proposed Reference Tariffs has not been clarified in the
Draft Determination.

A standard discussion of network pricing begins from the premise that

efficient marginal cost prices are insufficient to recover the costs of the
business. In fact, the opposite is often the case. Utility regulators rarely
allow the recovery of the full replacement cost depreciation or allow the firm
to earn a full return on theentire replacementcost assetvaluation,in particular
when assets are sunk. Indeed, IPART state that it will permit AGLearn to a
return on assets somewhere between historical cost and replacement cost
valuation. Joshua,can we put in a sentence here to say that it is not really
appropriate to try and recover replacement costs on sunk assets, specially if
they have been valued at historical cost in the past?

The question of the basis on which existing network assets are valued then
has wider implications for the optimalstructure of charges. Whether or not
marginal cost charging will lead to cost over- or underrecovery in
the specific
case of AGL depends on three factors which work in opposite direction:

§ In the presence ofeconomies of scale and scope, marginal cost pricing

will lead to cost under-recovery.

§ In the presence of substantial excess capacity, marginal cost pricing

may lead to cost under-recovery.

§ However, marginal costs are forwardlooking and are derived within a

replacement cost framework. Marginal costpricing will then be greater
than average costs andlead to cost over-recovery, if large parts of the
network are valued on a historical cost basis, rather than at
replacement cost.

The net effect of these three forces is an empirical matter and cannot be
determined ex ante.

Thus, average cost allowed to be recovered by the regulator may be less than
marginal cost. This is also likely to be the case forAGL, since the regulatory
asset base has been set at less than the replacement cost value. Thus, unless
prices are hopelessly out of line with costs it is unlikely that points (2) and
18
(3) of IPART’s criteria will be violated.

London Economics 22
October 1996
Section 6 Structure of access charges and the
promotion of competition

6.3 Administered charges versus price negotiations

As a vertically integrated player, AGL is in a position to exert significant
market power on potential competitors in order to deter entry in
downstream markets. In the absence of vertical integration,we would agree
that a regime of negotiated charges (subject to an
overall revenue constraint)
19
would be superior to the regulator administering network prices.

It should be remembered, however, that bargaining and negotiation can be

costly and can also result in inefficiencies. It is crucial, therefore, that the
framework within which negotiations take place is one designed to maximise
the likelihood that efficient agreements will be reached. In short, a general
message of modern bargaining theory is that bargaining costs (i.e. delays,
break-downs in negotiations, impasses, failure to reach efficient agreements
etc.) increase with the amount of private information held by the parties.
Efficient and effective outcomes are more likely to occur, the greater the
amount of information shared between the parties. From the standpoint of
encouraging economic efficiency, therefore, the regulator’s task should be to
ensure that all information concerning the network service provider’s costs is
made known to those entering into negotiations with it.

These considerations are consistent with the recommendations in the Hilmer

report relating to access to “Essential Facilities” and reflect our comments on
the role of regulation in Section 3.1:

To facilitate negotiation of appropriate access agreements

once a facilityhas
been declared,the owner of the facility should be required to provide
relevantcost or other data to the party entitledto seek accessand, if need
be, to the arbitrator. (Hilmer Report, page 256)

However, if the network service provider is vertically integrated either

upstream or downstream - and both are the case for AGL, the conclusion
that negotiated charges are preferable to administered charges becomes far
less obvious. Much depends on the protection afforded to actual and
potential competitors in the verticallyrelated markets by competition policy
and the regulatory regime. Weakcompetition rules or ineffectual regulation
deters optimal entry, whereas pro-competitor regulation (as opposed to pro-
competition regulation) can result in a backlash from the incumbent - both
Oftel and Ofgas have had a poorrelationship with, respectively BT and BG,
as a result of their efforts to promote competition.

Ultimately, unless competition rules are flexible and speedy dispute

resolutions are possible, the only feasible route to fair competition
may be to
further unbundle the trading activity from the transportationbusiness - this
has been the result in the gas and telecommunications markets in the UK.
Only then, irrespective of the efficiency of actual prices charged, will all
parties be assured that they are being charged on the same basis as all the
players in the market.

London Economics 23
October 1996
Section 6 Structure of access charges and the
promotion of competition

6.4 The definition of the cross-subsidy

While the IPART Draft Determination recognises that the existence of
different rates to different market segments does not in itself amount to
cross subsidisation, and that a subsidy exists if the rate does not cover
avoidable costs, their definition of such costs appears to deviate from
common economic usage.

Avoidablecosts in the tariffmarketincludethe operating,maintenance

and connection costs of supplying the tariff market and the opportunity
cost of utilising assets to supplythe marketwherethe assets and
fundsemployedin those assets could be used more profitably
elsewhere. (page 17, emphasis added)

Most commonly, avoidable costs are defined as those costs that would not
be incurred if a user or group of users did not receive the service.20 This
definition excludes joint costs among users, and the inclusion in IPART’s
Draft Determination of opportunity cost of funds aspart of avoidable costs
is inconsistentwith standard definitions of avoidable costs. If an asset has
been purchased and is effectively sunk,then if some users do not utilise the
asset, those costs remain unchanged and are therefore not avoidable.21,22,23

London Economics 24
October 1996
Section 6 Structure of access charges and the
promotion of competition

6.5 Summary
This Section isconcerned with the structure of Reference Tariffs which AGL
is proposing to post for potential competitorsand focuses on three separate
issues:

§ the structure of these tariffs;

§ AGL’s ability and incentives to price discriminate in order to

undermine competition; and

§ the definition of the cross subsidy.

The basis on which these Reference Tariffs are

likely to be derived is unclear.
Given the tension between an asset valuation approach below replacement
costs and the forward looking nature of marginal costs, charges below
marginal costs may be appropriate.

More importantly, as a vertically integrated supplier, AGL has strong

incentives to price discriminate against potential entrants.

Finally, while IPART propose to define cross-subsidies on the basis of

avoidable costs, the definition of thesecosts that has been used in the Draft
Determination requires clarification.

London Economics 25
October 1996
Section 7 AGL’s incentives

7. AGL’s incentives
This Section comments in detail to the proposed approach for determining
AGL’s revenues. IPARTproposes to regulate AGL’s maximum prices with
reference to a forecast ‘sustainable revenue’ stream. Sustainablerevenue is
determined with regard to a range of publicly available financial indicators.

The basic rationale for using a range of financial indicators is to prevent AGL
from taking actionsdesigned to manipulated any single indicator. In effect,
IPART is proposing to play what game theorist’s term a “mixed strategy.”
Such a strategy involves placing weights on each possible method of revenue
determination in such a way that AGLcannot game any single method. The
logic here is akin to a tennis player mixing up shots to an opponent’s left and
right side. This approach prevents an opponent from becoming prepared for
any particular shot. However, the success of a “mixed strategy” depends on
just that: it mustbe completelymixed. Providing certainty on any particular
dimension can undermine the basis for the entire approach. 24 Below we show

that IPART’s proposed approach, while correct in spirit, falls short of its
desired goal.

IPART claim that:

The use of indicatorsbasedon publiclyavailableor easily accessible

information reduces the
problemsof informationasymmetryprevalent
in regulatoryregimesaroundthe world. The Tribunalhas taken the
view that reliance
on any singleindicatormay distortthe Regulatory
framework by encouraging inappropriate behaviour. (page 10)

However, it is difficult to agree with either of these claims:

§ The majority of financial indicators used must come from AGL, and
hence there is scope (and indeed an incentive for) misreporting.IPART
do not explain how this information will be obtained.

§ As we shall discuss in more detail below, there are actions that AGL
can take that can distort each of these indicators in a way that
improves their average revenue over time and in subsequent reviews.
Thus, even if these indicators can be
measured accurately, there are no
provisions in the proposal that prevents their manipulation.

The expectation that no such manipulation will occur appears to rest on the
notion that inappropriate actions can only affect single indicators. Such an
implicit assumption would support the rationale for a mixed strategy
approach to regulation. Hence, it is argued that AGL would not have
adverse incentives since such actions are costly and each single indicator
receives a small weight in the basket. However, this does not imply that AGL
will not take such actions. Since a number of the financial indicators listed
are, in fact, closely linked, IPART has in effect reverted to the use of, perhaps
not one, but a very limited range of indicators of AGL’s financial status.

London Economics 26
October 1996
Section 7 AGL’s incentives

In addition, IPART’s intention to rely on a number of financial indicators has

implications for remaining gas industry participants which have not been
considered. In effect, IPART has put in place a regime characterised by a
substantial degree of uncertainty regarding the basis on which prices will be
reviewed at the end of the next review period. If this increased uncertainty
gave AGL incentives to reduce access charges, this would not present a
difficulty for other participants. But to the
extent that it fails to achieve this
end, IPART’s proposed approach leaves other market participants with little
indication of what is likely to be an appropriate price path for transportation
charges. This undermines their ability to take appropriate decisions.

7.1 Regulatory transparency

The question of regulatory commitment and accountability is receiving
increasing attention in the UK. One effect of increased judicial or public
(MMC) review of the regulatory process has been that it has increased the
reluctance of regulators to provide detailed rationale for their decisions.
It is
easier to mount legal challenges to the steps of an argument than to the
simple exercise of a general discretion whichstatute undoubtedly confers on
the regulator.25

However, in many instances the sought for clarity and transparency of

regulatory procedures and formulae, is largely illusory. The UK experience
has shown that many regulatory decisions can only be made with the
exercise of informed judgement. It is certainly possible to construct
formulae which would, in whole or in part, provide answers to these
questions by arithmetic rules, rather than regulatory discretion.
At the same
time, there isa risk that the operation of these formulae would be arbitrary
and unfair.

More importantly, regulated companies would almost certainly attempt to

manipulate the parameters of these formulae. The demand for greater
precision and less discretion in regulation is therefore often a demand for a
combination of the more extensive use of these formulae together with a
much larger role for the Courts in interpreting such statutory requirements
as “the proper financing of functions” and “the promotion of effective
competition”. The practical issues is therefore, whether decisions on such
issues as the market consequences of price discrimination or the
reasonableness of efficiency targets, are better made by regulators, or by
judges.

However, there are good reasons for thinking that these considerations do
not apply in AGL’s case. Our primary concern is that, far from removing
AGL’s incentives to game, the proposed system of financial indicators
aggravates the adverse incentives. We comment in greater depth on this
issue below. Secondly, AGL acts as a key infrastructureprovider and link
between upstream producers and final customers. The degree of uncertainty
surrounding AGL’s overall price cap and hence the likely level of access

London Economics 27
October 1996
Section 7 AGL’s incentives

charges over the medium-term planning horizon is unlikely to encourage

investment by potential competitors and hence benefits to final customers.

7.2 Potential for adverse incentives

IPART’s proposed approach provides AGL with significant adverse
incentives. Thisbecomes apparent if regulatory decision-making is thought
of in two stages. In the initial or regulatory determination stage, the
regulator chooses a mechanism and approach for regulation. In the future
or regulatory redetermination stage, the regulator reviews the
price caps and
other decisions in light of new information.

Between these stages,the regulated utility chooses its own behaviour based
on knowledge of the initial determination and its anticipation of the
regulator’s future behaviour. The regulated firm seeks to maximise its
profits over the entire time frame.As such, the regulated firm will consider
very carefully how their actions at the end of stage one, affect the Regulator’s
decisions in stage two.

This essential model captures most of the concerns regarding adverse

incentives. Consider these examples:

§ Suppose the regulator announces a policy of rate of return regulation

in Stage 1 with an intention to review pricesin Stage 2, again based on
rate of return using historic costs as a measure of the rate base.
Anticipating the Stage 2 continuation of the policy, the firm knows that
if it can expand its rate base it will make more profits. Hence, its
anticipation of a continued policy of rate of return regulation
adversely
affects the firm’s investment behaviour.

§ Alternatively, suppose that the regulator intends to use rate of return

regulation, but that this time the investment chosen by the firm in after
Stage 1 is sunk for the entire life of the industry. In Stage 2, the
regulator will maximise social welfare by using scrap value as its
measure of the rate base as this is most likely to lead to optimal
pricing. Anticipating this, the firmwill realise that it will be unable to
cover the costs of its investment and hence, will choose to under-invest.
Even if the regulatormakes a promise not to use scrap value in Stage
2, unless this promise is credible, the firm will fear a scrap value
approach and curtail its investment.

The difference between these two conclusions rests

on what the firm believes
the regulator will do. IPART correctly point out that committing to any
particular method of price redetermination can lead to adverse incentives.
However, the question is: does not committing to a particular method of
future redetermination reduce those incentives?

London Economics 28
October 1996
Section 7 AGL’s incentives

7.2.1 Implications of IPART’s proposed regulatory approach

IPART’s proposals can also be viewed as part of a two-stage process. In
Stage 1, IPART set out an initial price determination and also the method for
redetermination. The redeterminedmethod involves reviewing the price cap
on the basis of a basket of financial indicators. What can AGL anticipate
about how the Stage 2 redeterminationwill actually occur as oppose to the
announced criterion. In other words, what has IPART committed to do in
Stage 2?

IPART claim to have committed to have a redetermination based on the

indicators. All of those indicators have something in common: they reflect
the financial viability of AGL. IPART is concerned for the continued viability
of AGL and have stated their “responsibility not to jeopardise the financial
integrity of the infrastructure owner.” This promise not to redetermine
prices in such away as to raise the risk of bankruptcy of AGL is a credible
one - it is not in the interest of producers or final users see
to the distribution
network cease operations. Thus, while AGL cannot anticipate the precise
price redetermination, it can be assured that any redeterminationthat harms
its financial integrity will not be carried out.

This anticipation is likely to adversely affect AGL’s revenues. In its

investment expenditures and otheractions, such as contractual negotiations
and terms, AGL will often face a trade-off between:

§ the expected revenues from these actions; and

§ the volatility of those revenues.

No profit maximising firm will choose a path that raises volatility without
resulting in a higher expected revenue, as this imposes additional risks
without greater return. In the normal operation of its business, AGL will
then choose its actions to balance risk and return. However, IPART’s
credible promise not to jeopardise the firm’s financial integrity alters this
balance, adversely, in favour of greater return and greater risk.

To see this, note that the costs of greater risks AGL

to are losses in the event
of lower than anticipated demand. However, given IPART’s promise, AGL
knows that, if it is at risk of financial harm, IPART will favourably review
revenues. Even if demand is high at the time of the redetermination, the
basket of financial indicators will demonstrate that risk analysis for the firm
is not conducive to price decreases and may favour anincrease. At Stage 2,
IPART cannot refuse to take into account these risks as they are real by that
time.

London Economics 29
October 1996
Section 7 AGL’s incentives

7.2.2 Illustrative examples

Consider some examples of actions which AGLmay take that unfavourably
harm its risk assessment without reducing its anticipated revenues (in
anticipation of IPART’s promise):

§ Agreeing to price sensitive contracts. The terms and conditions of

contracts are listed among the financial indicators that IPART will take
into account in price redetermination.If AGL signs a series of take or
pay contracts that are sensitive to redetermined revenues or reference
tariffs, then it is forcedto reduce contracted prices accordingly. AGL,
however, knows that no unfavourable redetermination will likely occur,
because IPART will be concerned about the adverse affect this will
have on financial viability.

§ Risky investments. If AGL invests in a pipelineto attract business from

another state or location this is risky because it may simultaneously
face new competition. Nonetheless, such investments can allow it to
earn higher average revenues. Such investments alter the risk profile of
the firm. Thus, on redetermination, revenues will have to be
maintained so as not to risk AGL’s financial integrity. This is of
particular concern if the gas distributor’s activities are not ring fenced.

§ Inefficientgrowth of the tariff market. AGL might decide to engage in

inefficient growth of the residentialmarket (there is some evidence to
suggest that thishas been AGL’s policy in the past). It could do this
by setting low connection charges to consumers and offering initial
discounts on usage. The effect of this would be similar to expanding
the rate base of AGL. Inefficient expansion can affect AGL’s financial
viability in the short-term with a view to raising future Reference
Tariffs to already connected users (who face high switching costs).

§ Accounting difficulties. AGL’s method of accounting for profits and

costs could be manipulated in such a way as to make them look at
risk.26 Most accounting procedures are designed toeliminate reporting
of too little risk.

7.3 Summary
We have some sympathy with IPART’s concerns about adverse incentives on
the part of AGL to manipulate a single driving determinant of asset
valuation and revenue requirements,such as return on rate base. However,
IPART’s basic approach of a lack of transparency in its price regulation
process does not necessarily imply that the incentives
to AGL to manipulate
future reviews are eliminated or even diminished.

London Economics 30
October 1996
Section 7 AGL’s incentives

Revenue determination based on a basket of financial indicators does not

remove adverse incentive effects. Indeed, it offers AGL insurance against
any potential risk of losses from these actions. This
is precisely the opposite
of good incentive creation in regulation which would reduce the firm’s
insurance against uncertainty in order to ensure that it internalisesrisk
the of
poor decision-making.

8. Conclusions
The issues raised in this submission are in many ways endemic to access
price regulation. Standardforms of price regulation carry with them issues
of sensible asset valuation and the risk of adverse
incentives. Regulators are
constrained in that they mustpay attention to the financial consequences of
their determinations, buteven this might be the subject of gaming. Indeed,
the problems regulators face in this regard are similar to concerns about
“soft budget constraints” facing state owned firms in the former Communist
countries of Eastern Europe.27

This makes the task of regulation in the post-Hilmer environment quite

difficult. In particular, it brings to the forefront the notion that general
prescriptive approaches to regulation are unlikely to be optimal.The details
of the access dispute at hand have to be considered and assessments must
be made regarding whether ways of alleviating the problems of information
asymmetry -- specific to the industry -- are available. In the case of gas,
benchmarking and tough reviews of investment decisions have emerged as
regulatory solutions in both the UK and the US.

Finally, it should be emphasised that regulators today play an important

role in setting the agenda, not only for future
determinationsin the industry,
but also determinations in future access disputes. Thus, it is critical thatthe
economic objectives of regulation be adhered and referred to in access
determinations. Such an approach sends a signal to others of the criteria for
regulation and hence, can strengthen the actual effectiveness of regulation in
creating value.

London Economics 31
October 1996
Notes

Notes

1
Independent Pricing and Regulatory Tribunal of New South Wales, Draft
Determination on the Proposed Access Undertaking of AGL Gas Companies
(NSW)
Limited, September 1996.

2
Gas Council of NSW, “Four Year Review of the Price Control Formula the
in
Tariff Gas Market,” November 1994.

3
John McMillan, Games, Strategiesand Managers,Oxford University Press:
Oxford, 1992.

4
Oliver E. Williamson, The Economic Institutionsof Capitalism,Free Press:
New York, 1985, Oliver Hart, Firms, Contracts and Financial Structure,
Oxford University Press: Oxford, 1995.

5
Paul Milgrom and John Roberts, Economics,Organizationsand Management,
Prentice-Hall: New York, 1992.

6
Another example of implied behaviour by regulators is the threat of
compulsory arbitration. As Stephen P. King and Rodney Maddock, “A
Strategic Analysis of Part IIIA of the Trade Practices Act,” mimeo., ANU,
1996, show in relation to the Competition Policy Reform Act in Australia,
this threat may not only ensure that thereis no bargaining breakdown, but
that (under some conditions) the agreement reached will partly reflect the
interests of consumers. The benefit of this is the avoidance legal
of and court
costs and the need for a dedicated regulatory institution in many industries.

7
The use of indicators based on publicly available or easily accessible
information reduces the problems of information asymmetry prevalent in
regulatory regimes around the world.

8
J-J. Laffont and Jean Tirole, A Theory of Incentives in Procurement and
Regulation,MIT Press: Cambridge (MA), 1993; Mark Armstrong, Simon
Cowan and John Vickers, RegulatoryReform,MIT Press: Cambridge (MA).,
1994.

9
Stephen P. King, “Who Took the Competition Out of National Competition
Policy: Access Pricing Under Rate-of-Return Regulation,” mimeo.,
Australian
National University, 1996.

10
Note that it is notnecessary that alternative systems be strictly comparable.
If the regulator can, over time, observe productivity improvements
elsewhere then this is evidence that there ought to be such improvements in
the local industry as well.

11
Average revenue regulation is applied to British Gas (BG) in the UK.

12
Kay, J., “The Future of UK Utility Regulation,” The Institute of Economic
Affairs lecture on regulation, delivered at the RSA in
London on 31 October
1995.

London Economics 32
October 1996
Notes

13
Starkey, M. and J Van Pelt (1995), “Productivity measurement and price cap
regulation: Issues for local exchange carriers in the USA,”
Telecommunications Policy, Vol. 19, No. 2 pp. 151-160.

14
To see this consider the steady state where depreciation equals new capital
expenditure - the regulatory asset value on which shareholders receive a full
return therefore remains constant over time and the replacement cost
valuation will also be constant but at a higher level.

15
Although the Modigliani-Miller theorem holds that the company’s cost of
capital is invariant to the structure of the balance sheet, it is well
known that
taxes and market imperfections allow the firm to manipulate its cost of
capital. In a regulatory context, the firm may adjust its balance sheet in
order to appear to have a higher cost of capital,
only to reverse the financial
engineering once the price review is over.

16
In our example in Annex 2, if capital expenditure was doubled, resulting in
an increase in the present value ofcapital expenditure from 77 to 154 units,
allowable revenue would only increase by 12 units.

17
An alternative, where the Regulator becomes more involved in the minutia
of corporate information and seeks to establish how much ofthe underspend
is due to efficiency savings and how much isdue to information asymmetry
may be a preferable solution but raises further information questions

18
To the extent that marginal cost pricing wouldlead to overrecovery of costs,
this may present an argument for an ‘inverse’ Ramsey price structure where
inelastic customers are charged a lower price than elastic customers, given
the same cost of serving each. Thus, if a regulatory decision is made
to hold
overall prices below marginal cost, then a charging structure which
favoured households over businesses could in fact be optimal.

19
The value of negotiations to users depends on reference tariffs being
determined independent of the outcomes of these. Otherwise, AGL could
negotiate discounts to larger customers with an effective ability to rely on
rises in reference tariffs to other users.

20
Zoltan Biro and John Kay, “Utility Cost Allocation in Principle and Practice,”
in P. Burns (ed.), EffectiveUtility Regulation- The Accounting Requirements,
CRI: London, 1994, Chapter 3.

21
However, this definition highlights the importance of the time horizon that
is implicit behind this discussion. Costs which may be fixed in the short
term, such as ongoing network maintenance and upgrading, may be
avoidable over a medium- to longer-term time horizon.

22
The imprecision in the IPART proposal is also reflected in the following
statement when they argue for lower connectioncharges for new customers:
“These new customers will make a contribution greater thantheir avoidable
costs, further reducing the cross subsidy.” (page 17)

London Economics 33
October 1996
Notes

23
There is another issue here which reflects the above discussion regarding
the tension betweena pricing rule (such as the Baumol-Willig definition of
cross-subsidies referred to in IPART’s document) which is based on a
replacement cost concept, and an asset valuation approach for determining
aggregate revenues linked to historical costs (or, at any rate, a valuation less
than full replacement costs). In this case, if aggregate revenues are capped
below replacement costs, it will be impossible to put in place a charging
structure that is consistent with the Baumol-Willig conditions (i.e. that prices
are at least equal to incremental cost). IPART’s proposed framework for
limiting AGL’s overall revenues is therefore appears to be inconsistent with
the definition of cross-subsidies.

24
These concerns for gaming have arisen in the literature on incentives for
workers within firms (see George Baker, Robert Gibbons and Kevin J.
Murphy, “Subjective Performance Measures in Optimal Incentive
Contracts,” Quarterly Journalof Economics,Vol. 109, No. 4, 1994, pp.1125-
1156.

25
Kay, J., op.cit.

26
For a discussion of this issue see Stephen King and Rohan Pitchford, “A
Theory of Privatisation and Corporatisation,” mimeo., Australian National
University, 1996.

27
Janos Kornai, “The Soft Budget Constraint,” Kyklos, Vol. 39, No. 1, 1986,
pp.3-30.

London Economics 34
October 1996