Incentive regulation in Australia - a hybrid approach?
Stephen P. King
The University of Melbourne
“Hybrid - an animal or plant resulting from a cross between genetically unlike
individuals. Hybrids between different species are usually sterile” (Collins
English Dictionary).
Introduction
In the last decade, Australia has undergone a revolution in the management,
ownership and regulation of key public utilities. Companies in the
telecommunications, energy and transport sectors have been privatised. Those firms
that remain in government ownership have been corporatised so that the public sector
managers are judged by quasi-profit performance. Regulatory powers have been taken
from the relevant firm and given to arms-length regulators. These regulators have
been granted new powers to control utility firms through both specific and general
legislation.
The aim of these reforms is to improve performance in a wide range of industries that
have previously been dominated by government firms. But simple deregulation will
be unlikely to achieve this aim in some utility industries. These industries often were
originally brought under direct government control because market-based competition
appeared inadequate to ensure appropriate performance. Some of these industries
involve natural monopoly technologies, so that competition might be socially wasteful
or unsustainable. In such situations, regulation is a necessary adjunct to improvements
in performance.
Unlike, say, the United States, Australia has relatively little experience in regulating
public utilities. When government, through a relevant minister, directly controlled
these utilities, explicit arms-length regulation was not relevant. As a result, reform has
forced Australia onto a steep learning curve to adopt workable regulatory structures.
1
In some situations, these structures have been adopted from overseas experience. In
others, such as the access regime under Part IIIA of the Trade Practices Act, the
regulations are largely ‘home grown’.
The Australian regulatory reform process has provided a wonderful opportunity to
adapt and adopt from overseas experience and academic learning about regulation.
But in many cases we have failed at this task. Rather than choosing innovative
modern regulatory techniques, much of Australia’s regulatory reform has involved re-
inventing the worst aspects of US rate-of-return regulation. This regulation, which
was first developed over a century ago in the United States, now forms the basis of
much of the regulation being implemented in Australia. Rate-of-return methods have
been altered to fit particular Australian industries. But it is the uniformity of
regulation in Australia, not its innovation and variation that are its outstanding
features. Where regulatory approaches differ this tends to reflect the relative
inexperience in this country. The result is a poorly designed ‘hybrid’ regulatory
structure that may undermine many of the potential benefits of the reform process.
The economics of regulation
Good regulation is difficult. At its simplest, the regulation of a private firm involves
trying to force the managers of the firm to act in a way that is not in the best interest
of the firm’s owners.1 Effective regulation means that the firm will make lower profits
than it could achieve in the absence of regulation. The government, either directly or
through its delegated regulator, tries to regulate firms from a position of weakness.
The regulator has only limited knowledge of the operations, costs and capabilities of
private firms. The regulator relies on the firm managers to provide it with much of the
information that it requires to create and enforce sensible regulation. But this means
that the regulator must rely on the people who have the most to lose from effective
regulation when the regulator tries to develop and implement such regulation. Good
regulation recognises both the relative ignorance of the regulator and the perverse
incentives that face the managers of a regulated firm.
1
For the purpose of this talk I will limit my attention to the regulation of private firms.
Corporatisation is a form of regulation of public firms, but is beyond the scope of this discussion.
2
In economic terms, those designing and implementing regulation face two incentive
problems. First there is adverse selection. The regulator does not know the exact
capabilities of the regulated firm and, as a result, cannot directly tell the firm how to
operate. The managers, who have access to this information, have no incentive to
reveal it to the regulator. Rather, they have an incentive to misrepresent their firm's
capabilities so that the resultant regulation will not impact too adversely on their
profits.
Second, there is moral hazard. The regulator does not know how the firm should be
operated over time. Because regulation changes the profits that a firm can achieve,
regulating a firm will alter the way that the firm is operated. Even if the regulator
could, ex post, receive perfect information about a firm’s actual capabilities, the
regulator does not know whether these are set at an optimal level or have been
reduced because of the regulation.
For example, suppose that a regulator decides to use a form of cost-plus pricing to
regulate a firm. Also, suppose that the relevant firm has not previously been regulated
and is currently operating efficiently. The firm has adopted minimum cost technology,
but the regulator does not know these costs. The regulator also has a limited ability to
determine the costs. If they ask the firm’s managers, then these managers will have an
incentive to distort their cost reports. If the manager’s can convince the regulator that
costs are higher then they truly are, then the firm will make more profit. The
regulators may be able to implement a cost-plus regulation – but it will probably be
based on a distorted measure of firm costs.
To the degree that the cost plus regulation limits firm profit, it will also alter firm
incentives to maintain minimum costs over time. If the managers can artificially raise
costs, for example, by incurring unnecessary but verifiable expenditures, then these
costs can be passed on to customers through regulation. Even if these costs do not
raise profits, they might benefit management. And, due to the regulatory regime, the
costs are bourn by customers not shareholders. Over time, regulation will lead to a
bloated firm that does not produce efficiently.
Good regulation recognises these incentive problems. In particular, it recognises that
regulation is imperfect and that it is always a second-best solution to problems of
3
market failure. Good regulation does not try to mimic a perfectly competitive market.
If such a market was feasible for the relevant industry then there would be no need for
the regulation. It is because competition is inadequate that there is a need for
regulation and regulations that are based on the idea of replicating competitive
outcomes, such as rate-of-return regulation, may worsen the problem.
Good regulation will have a number of characteristics.
1. The goals of the regime are clearly specified. For example, with access
regulations, what types of behaviour is access meant to prevent and how will the
regulation achieve this.
2. The alternatives (including no regulation) are carefully weighed and compared. In
particular, the effects on firm incentives and future firm performance that will
flow from current regulations are considered. Short-term gains in allocative
efficiency from more efficient pricing today can be quickly swamped by the long-
term cost of decreased firm performance and reduced investment incentives.
3. The regulatory rules are not considered in a vacuum but are considered within the
framework of industry structure and technological change. For example, the
original price cap regulations on British Telecom were designed to become
redundant after about five years.2 While this time frame has proved overly
optimistic, the rules were deliberately crafted as a short-term control during the
movement to competition. In contrast, these types of rules have been adopted in
Australia for long-term regulation. The price-cap rules were never intended for
this purpose.
4. Regulatory rules are carefully drafted to try and best achieve the relevant aims.
The rules do not allow for misinterpretation, regulatory uncertainty or
opportunism by either the regulator or the regulated firms. The 1996 regulatory
reforms to the US telecommunications sector highlight the consequences of
uncertain reforms. Rather than quickly opening up local telephone competition in
the US, these rules have led to numerous court battles and have stifled investment.
2
See Vickers and Yarrow (1988) and Armstrong, Cowan and Vickers (1994).
4
5. As rules adapt over time the relevant changes are carefully considered and
grounded in solid economic analysis. In contrast, there is a perception in Australia
that some state based regulators are acting on an ad hoc basis, adjusting rules as
problems arise without considering the full consequences of the changes.
6. Regulatory rules are based on 'state-of-the-art' understanding of incentives. As
noted above, regulation is all about incentives. There is no sensible distinction
between ‘incentive’ and ‘other’ regulation. All regulation effects incentives. The
real distinction is between good and bad regulation.
Much of the current regulation that has been drafted and implemented in Australia
fails to satisfy some (if not all) of these characteristics.
The state of regulation in Australia
Australian utility regulation involves numerous names and acronyms. We have price
caps, revenue regulation and incentive regulation. Asset values can be based on ODV
or DORC (called ODRC in New Zealand). There are glide paths, WACCs and
allocation rules. Costs may be based on LRIC, TSLRIC or TELRIC. Occasionally
even avoidable cost or LRAC is relevant. But like the conjurer who always makes you
pick the same card, regardless of what card you really want to choose, so too
regulation in Australia is really of one type in practice. Almost all Australian
regulation is rate-of-return regulation.
Rate-of-return regulation is a static tool to limit the profits that a firm with market
power might extract from the market. The idea is simple. In a static perfectly
competitive market equilibrium, each firm earns zero economic profits or,
equivalently, earns exactly the risk-adjusted market rate-of-return on its capital stock.
A firm can be made to mimic the behaviour of a perfectly competitive firm if it is
forced to make no more than this rate-of-return.
Rate-of-return regulation has three basic parts. First, the regulator must determine the
firm’s capital stock or asset base. This relies on information that is potentially
manipulable and is provided by the company. Secondly, the regulator must determine
the relevant return to be applied to the capital base. This weighted average cost of
capital (WACC) is generally based on comparisons with similar industries and uses a
5
capital asset pricing model. The allowed return and the capital base are multiplied
together to determine an allowed level of ‘variable profits’. These are sometimes
misleadingly referred to as allowed revenues in Australia, so that rate-of-return
regulation is sometimes called revenue regulation.
The key point is that any regulation, regardless of its name, that uses these basic
principles to determine a firm’s allowed profit, is rate-of-return regulation. For
example cost-based regulation, where the regulator allows a firm to just set a price to
cover average cost, is rate-of-return regulation if the average cost is calculated by
using a reported capital stock multiplied by an allowed ‘interest rate’. On-going price
cap regulation sets a maximum price that the firm can charge, where this price is
determined to allow the firm a ‘reasonable’ rate of return on capital. It is rate-of-
return regulation. Profit regulation involves the regulator setting a profit cap for the
regulated firm. The cap is checked by examining the firm’s revenue then subtracting
variable costs and an allowance for return on capital. If the return on capital is
established by setting a capital base and an allowed rate-of-return, then profit
regulation is rate-of-return regulation.
It is easy to see that all these forms of regulation are identical by looking at the
mathematics that underlies each type of regulation. If we denote the regulated asset
base by K , the allowed rate of return by r , and the firm’s reported price, sales and
variable costs by P , Q , and V respectively, then the profits ‘as measured by the
regulator’ are PQ − V − rK .
Profit regulation simply says that the firm cannot have measured profits that exceed
an allowed profit level; π a ≥ PQ − V − rK .
Rate-of-return regulation simply states that the firm cannot have a measured rate-of-
PQ − V
return that exceeds an allowed level; r a ≥ .
K
Revenue regulation simply restricts the firm’s variable profits to be less than an
allowed level. This level is usually calculated by multiplying the capital base by an
allowed rate-of-return. Even if this is not explicitly done, it is implicit to revenue
regulation. r a K ≥ PQ − V .
6
Cost regulation requires that a firm price at no more than an x percent mark up on
V + raK
average cost as calculated using an allowed rate of return; P ≤ (1 + x ) .
Q
On-going price cap regulation requires the firm to price below a fixed level, P . But
V + raK
this allowed price is often based on average costs so that P ≤ P = .
Q
There are well-known problems and limitations of rate-of-return regulation. First, it
tends to distort firm production choices. The regulator is almost certain to incorrectly
estimate the true cost of capital for a firm. In particular, because of fears about firm
bankruptcy or the firm running down its capital stock if it is given an inadequate
return, the regulator tends to overestimate the cost of capital. It sets the allowed rate-
of-return or WACC too high. This results in the Averch-Johnson effect.3 The
regulated firm will choose a sub-optimal mix of inputs, relying too heavily on capital
inputs. This raises the cost of production above an efficient level.
Secondly, rate-of-return regulation is a form of cost-plus regulation. These regulations
give firms little incentive to minimise cost. In particular, to the degree that the firm,
say, raises its variable costs, these are simply passed on to consumers. The firm’s
managers have incentives to cost pad and seize perquisites through the operations of
the company. The shareholders do not mind this inefficient cost padding as it is the
customer, not the firm’s owners, who bear the inflated costs.
Thirdly, because rate-of-return regulation rewards a firm on the basis of its measured
capital stock, firms have strong incentives to increase this measured capital stock. In
the short-term this is likely to be achieved through transitional asset valuations. It
should not be surprising that regulated firms in industries where infrastructure costs
have been rising over time, have been strong supporters of replacement cost
valuations for their assets. These valuations provide the firms with a windfall gain at
the customers’ expense. Of course, in industries such as telecommunications where
3
See Averch and Johnson (1962)
7
replacement costs have been falling, regulated firms have championed actual or
historic cost asset valuations.
In the longer term, rate-of-return regulation will lead to reluctance by firms to remove
redundant assets from their regulatory asset base. It can also lead to firms seeking to
add dubious assets to this base, particularly where these assets are also useful in
providing a non-regulated service or product. Over time, rate-of-return regulation
leads to an increasing involvement of the regulator in the firm’s capital investment
decisions. For example, in the US, state regulators require assets to pass a ‘used-and-
useful’ test before allowing these assets to continue to be part of the regulated firm’s
asset base. There is no reason to expect that the regulated firm will invest efficiently
over time under rate-of-return regulation.
In brief, in Australia we have introduced a regulatory scheme that is unlikely to lead
to efficient production in either the short-term or the long-term.
Basic rate-of-return regulation is often augmented, particularly in the case of multi-
product firms. For such firms, allocation rules are used to further distribute allowed
profits over types of products or groups of customers. For example, with price cap
regulation there can be baskets of products covered by a cap and individual products
covered by sub-caps. This is the situation in Australian telecommunications. For
revenue regulation, the allowed revenues can be allocated to different customer
groups. In electricity transmission, this is done on the basis of the amount of
infrastructure required to service that group. In some situations, the allocation can be
based on political considerations, beliefs of fairness or ability-to-pay. For example,
the NSW rail access regime treats coal lines differently to other railway lines.
Allocation rules are a commonly used to augment rate-of-return type regulation in
Australia. These rules are likely to exacerbate the problems with rate-of-return
regulation by further limiting a firm’s ability to operate efficiently. To paraphrase
William Baumol, allocation rules are like snatching defeat from the jaws of victory.4
4
This phrase was used by William Baumol at the 1997 Economics Society conference in
Hobart. He was responding to a presentation by the ACCC on its use of cost-based regulation for
telecommunications in Australia.
8
If rate-of-return regulation has these well-known problems, why has Australia
fervently embraced it? The answer lies in Australia’s lack of regulatory experience.
Rate-of-return regulation, under its many guises, appears fair and equitable. It can
easily be justified as avoiding excessive profits, and in the short-term, while the
phones still work and the lights still come on, it appears to be a simple answer to a
complex problem. But in the longer term it is likely to undermine most of the
potential gains from the last decade of microeconomic reform.
Examples and alternatives
Australian regulation can be improved by using a systematic approach based on the
economic principles of good regulation. The first step is to ask if regulation is needed.
In particular, is the relevant utility firm already subject to market constraints that limit
any significant abuse of market power? If this is the case, then regulation is a costly
intrusion that is likely to harm both the firm and consumers.
The Victorian gas industry provides a useful example. Victorian gas consumption is
dominated by household use. Experience from the introduction of natural gas into
Victoria in the early 1970s shows that households will change from one fuel to a
cheaper alternative if there is a significant price advantage. This shift is not
instantaneous, but is likely to occur over five to seven years. From a household
perspective, gas and other energy sources, such as electricity, are often reasonable
substitutes. Further, there is no significant household use of gas that is not subject to
competition from an alternative fuel.
This suggests that a private integrated gas producer might have some market power in
Victoria, but that substitute fuels will limit any abuse of this market power. A gas
utility that used its market power to raise prices and profit in the short term, would
quickly face a consumer backlash. Over time, a short-term grab for profit would
probably lead to a loss in market share and firm value.
Inter-fuel competition would not be a perfect regulator of the gas industry in Victoria,
but it might offer significant benefits relative to intrusive hands-on regulation by the
Office of the Regulator General.
9
Secondly, it is necessary to ask whether the firm needs regulation or reform. Many
government owned utilities required (and still require) reform. But this is different to
regulation.
Urban rail systems in Melbourne and Sydney provide a useful example. These
systems have traditionally been owned and operated by public sector bureaucrats.
Performance by these systems has been poor. But this does not mean that they require
regulation to either improve competition or replace absent competition. Quite the
opposite – urban rail systems face enormous competition from private transport and
other forms of public transport. The problem in urban rail is one of management not
competition. The use of regulatory tools, such as vertical separation and rate-of-return
regulation for track access, is unlikely to be appropriate and, as experience in the UK
shows, may possibly lead to deterioration in performance. Rather, the urban rail
system needs managerial reform. The use of franchise contracts together with the sale
of the rolling stock, as has occurred in Victoria, represents a more appropriate
solution to a problem of management rather then competition.
Thirdly, if regulation is needed, should it be transitional or long-term? In some
industries, the movement from monopoly government provision to private provision
involves transitory issues of market power. Competition takes time to develop. Short-
term regulation might be desirable to both protect consumers and promote
competition. These regulations may differ considerably from long-term regulation.
Local voice telecommunications provides a useful example. In the short-term, most
voice traffic is still carried by the fixed network. But an increasing amount of voice
telephony is carried by wireless methods. Further, voice telephony might, in the
future, be offered as a low cost addition to systems designed for video and data traffic.
If the bottle-neck represented by Telstra’s customer access network is only a short-
term problem for voice communications, then regulation should be geared towards
promoting competition through alternative investments. This would be closer to the
UK regulatory approach than that adopted in Australia. For example, should access
only be for data communications not basic telephone services? Is the long term
bottleneck really access by internet service providers, not standard telephone
companies? Should voice telecommunications regulation be focussed on encouraging
wireless local loop rather than sharing Telstra’s local network?
10
Note that changing the focus of the regulatory regime requires the regulator to address
significantly different questions. Under the current telecommunications regime,
competition is promoted through long-term access, and regulation is focused on the
price and quality of access. If, as an alternative, regulation is geared towards the
development of competing networks, then other regulatory issues, such as
interconnection pricing, will emerge. Short-term access might still be an important
concern to aid competitive entry, but it would only be a small and diminishing part of
the regulatory regime.
Suppose that an industry does require long-term regulation. The best solution might
be simple regulatory rules or restructuring rather than intrusive on-going regulation.
Even if intrusive regulation is to be used, simple structural or procedural changes may
significantly ease the regulatory burden.
Telecommunications again provides a useful example. In the long-term there might be
concern about interconnection fees between competing telecommunications carriers.
For example, mobile carriers might be able to anti-competitively set termination
charges for fixed-to-mobile calls. Market power, that is derived from being the carrier
connected to the relevant receiving party, can be reduced by informing calling parties
of the identity of the mobile carrier that they are calling. Requiring mobile carriers to
‘identify themselves’ before billing commences can help overcome customer
ignorance about which specific mobile carrier is associated with a particular number.
This can promote competition but does not require on-going intrusive regulation.
Structural processes can also be used to overcome the incentive problems associated
with regulation. For example, auctions and contracting can often be used to elicit
information and create a low-cost regulatory structure. Airports provide a simple
example. Rather than privatising the airports with price controls based on rate-of-
return procedures, the government could have used auctions to improve airport
management. ‘Slots’ at the major airports could have been auctioned on, say, a five
year basis. This would allow the market to determine the correct price of key airport
services rather than arbitrary rate-of-return procedures. Further, the slots could be
tradeable so that the allocation of airport services would be improved and could be
regulated by the market over time. A similar use of auctions to allocate capacity in gas
11
transmission pipelines could be used to avoid rate-of-return access regulation in New
South Wales and Western Australia.
If intrusive rate-of-return regulation is required (usually disguised under a more
acceptable name) then the regulator should be encouraged to use some simple
procedures to aid this regulation. The regulator should use appropriate non-linear
prices rather than average prices. While non-linear prices, such as two-part tariffs, can
be difficult to design, they offer considerable potential benefits. Urban water
regulation and reform provides a good example. Traditionally, in Melbourne, water
was charged at a low marginal cost but water rates were levied on households. These
rates were related to property value not water use. This pricing probably represented a
useful approximation of optimal non-linear pricing for a water system that is not
subject to capacity constraints. The marginal cost per litre of water is very low, and
the fixed costs of the water system were covered by fixed charges. These charges
were sensibly levied to avoid forcing people off the system – those more able to pay
were required to pay more, where property value was used as a proxy for ability to
pay. Of course, the non-linear pricing was not sensibly applied when capacity
constraints became relevant. In times of shortage, arbitrary restrictions were imposed
rather than increasing the marginal price of water.
Interestingly, recent reforms have moved substantially away from this economically
optimal pricing regime. A desire to see ‘user pays’ water charges has led to increased
marginal water charges. To the extent that these charges now exceed marginal cost,
‘user pays’ pricing is less efficient than the old non-linear water rates.
The regulator should also be encouraged to recognise their ignorance under standard
regulation. If possible, the regulator should use alternative sources of information to
reduce their reliance on the regulated firm. Careful use of benchmarks and yard-stick
comparisons can be useful. These comparisons must be carefully designed to avoid
firms colluding. But they can be useful in reducing the incentive that each firm has to
artificially raise its own reported costs.
To see the usefulness of benchmark comparisons, suppose that there are a number of
companies that operate in different regions, are all regulated, and have similar but not
identical operating conditions. The UK water companies provide a useful example. IN
12
this situation, the regulator could use cost information from all companies as an input
to determining each company's allowed price. Because each company's price is now
only partially dependent on its own cost report, each company has a reduced incentive
to exert effort to artificially raise its reported cost. The system could be enhanced by
explicitly rewarding a company reporting a low cost allowing it to set a higher price
than otherwise. While this might appear perverse, it simply recognises the regulators
lack of information and rewards a firm for providing information that is of higher
quality. Such systems need to be carefully designed to avoid the potential for 'gaming'
by the regulated firms. But in some cases, these systems can be used as a useful
adjunct to other regulatory tools.
Conclusion
As noted above, all regulation is incentive regulation. Because regulation, by its
nature, is trying to force firm managers to behave in a way that is not in the
shareholder's interest, all regulation deals with and alters incentives. But there is good
and bad regulation. Good regulation recognises that the regulator is always at an
informational disadvantage compared to the regulated firm. It also recognises that
regulation is an imperfect tool and is a second-best solution to market failure. As a
result, good regulation often involves structural or process-based changes that (a)
elicit information and (b) operate automatically with little day-to-day interference.
Rate-of-return regulation, under its many names, is often bad regulation. It is simple
to introduce, gives the appearance of solving the market failure, but has long-run costs
that often outweigh any benefits. But this is the regulation we have adopted in
Australia. The Australian 'hybrid' approach has brought together the worst aspects of
overseas experience to create a sterile framework that threatens to undermine the
benefits of microeconomic reform.
References.
Armstrong, M., Cowan, S. and Vickers, J (1994) Regualtory reform: economic
analysis and the British experience, MIT Press, Cambridge, MA.
Averch, H. and Johnson L (1962) "Behavior of the firm under regulatory constraint",
American Economic Review, 52, 1052-69.
13
Vickers, J. and Yarrow, G (1988) Privatization: an economic analysis, MIT Press,
Cambridge, MA.
14
