FLORENCE SCHOOL OF REGULATION

ANNUAL TRAINING COURSE

REGULATION OF ENERGY UTILITIES
2018-2019

Lesson 4
‘Regulation of natural monopolies’

Rafael Cossent*

*This reading material is largely based on a book chapter written by Professor Tomás Gómez from Comillas
University. Full citation information is: Chapter 4: Monopoly Regulation, in “Regulation of the Power
Sector”, Editor Ignacio Pérez-Arriaga. Springer-Verlag. ISBN 978-1-4471-5033-6. 2013.

-1/42-
Table of contents

1.  Introduction .......................................................................................................... 3 

2.  Fundamentals of monopoly regulation ................................................................. 4 
2.1  Drivers of monopoly regulation ................................................................... 4 
2.2  Costs and regulation ..................................................................................... 6 
2.3  Goals and role of regulation ......................................................................... 7 
2.4  Overview of regulatory approaches.............................................................. 8 
3  Cost of service regulation ..................................................................................... 9 
3.1  Determining the cost of service .................................................................. 10 
3.2  The regulatory asset base............................................................................ 13 
3.3  The rate of return ........................................................................................ 15 
3.4  Strengths and weaknesses of cost of service regulation ............................. 17 
4  Incentive-based regulation.................................................................................. 17 
4.1  Price cap ..................................................................................................... 18 
4.2  Revenue cap................................................................................................ 20 
4.3  Mechanisms for sharing earnings and losses.............................................. 21 
4.4  Design of an incentive-based regulation scheme ....................................... 23 
4.5  Strengths and weaknesses of incentive regulation ..................................... 25 
5  Implementation details for price or revenue caps ............................................... 25 
5.1  Regulatory asset base, investments and depreciation ................................. 26 
5.2  Operating costs ........................................................................................... 28 
5.3  Regulatory benchmarking .......................................................................... 29 
5.4  Present value of costs and revenues: the smoothing X ............................... 32 
5.5  Tax, cash flow and profitability.................................................................. 35 
5.6  Consumers or companies risks ................................................................... 37 
5.7  Quality of service and other issues ............................................................. 38 
6  Summary............................................................................................................. 39 
References .................................................................................................................. 41 

-2/42-
 1. Introduction
This document provides an introduction to the fundamentals of natural monopoly
regulation. The principles and concepts presented in this document have been frequently
applied to utilities that provide what have been traditionally considered to be public
services: electricity, water, telecommunications and gas. Notwithstanding, the focus
herein will be specifically placed on the regulation of energy networks, especially
electricity.
As it will be shown in section 2, economic regulation is necessary to prevent a
monopolistic provider from overcharging consumers or providing inadequate levels of
quality of service. When the monopolist is a State-owned company, the State may, in
principle, act as both owner and regulator without clearly separating the two functions.
However, energy network companies have been privatized in many countries nowadays.
In this context, independent regulatory commissions are necessary to implement and
supervise the regulatory principles defined by the State.
For example, in the US these tasks are carried out by the public utility commission
(PUC) of each state. Additionally, there are nation-wide regulatory agencies, such as the
Federal Energy Regulatory Commission (FERC) or the Federal Communications
Commission (FCC), whose responsibilities comprise interstate transactions or services
that fall outside the powers of individual states. Similarly, in the European context,
independent national regulatory agencies (NRAs) act as energy regulators1 in each
member state, whereas ACER (Agency for the Cooperation of Energy Regulators)
provides guidelines for NRAs at a European level.
Until not long ago, many sectors that have by now been successfully liberalized were
considered natural monopolies and, presumably because of this, the services were
provided by public companies. Among these, one may find air transportation, railroads
or phone services. Similarly, liberalization and competitive markets have been introduced
in sectors such as telecommunications, electricity and gas or water supply. Nonetheless,
competition has been limited to wholesale production and retailing of the products
themselves, e. g. the kWh of electricity or the cubic meter of water. The planning,
building, operation and maintenance of underlying network infrastructures are still
considered as natural monopolies (at least partly). The main common characteristic of
these activities is that they all require a physical network that is necessarily linked to a
particular geographical region.
As a consequence, these activities are characterized by the need of large investments
in fixed assets to set up the required network infrastructure. This feature constitutes one
of the major drivers for monopoly regulation. In fact, the clearest example illustrating the
inefficiencies of competition in network services is that of two firms supplying electricity
to consumers in the same area using duplicated grids. Being this the case, the required
network redundancy would end up with consumers paying much more than in the case of
a single company supplying electricity to all the consumers in that area.
The remainder of this document is organized as follows. Section 2 introduces the
principles to be followed when regulating monopolies and defines some basic concepts.
In section 3, the most conventional approach to monopoly regulation, i.e. cost of service

1
Note that a specific NRA may not only act as energy regulator but they may also have competences
over other sectors such as telecommunications or postal services.

-3/42-
or rate of return regulation, is presented and discussed. In order to address the major
shortcomings of cost of service regulation, some form of the so-called incentive
regulation has gained in popularity among regulators. The reasons for this transition and
the most common approaches to incentive regulation are presented in section 4. Section
5 will discuss in more detail how incentive regulation is applied in practice: price review
processes, efficiency assessments, quality of service, etc. Lastly, a summary of the main
ideas contained in this document is provided in section 6.

2. Fundamentals of monopoly regulation

A monopoly exists when a single firm becomes the only supplier of a specific product
or service. Normally, this situation is to be avoided as market competition is deemed to
yield more efficient outcomes for consumers as well as encourage innovation. However,
an efficient market competition is not possible under certain circumstances due to market
failures. A natural monopoly may be seen as the most extreme form of market failure,
and regulation is usually necessary to correct it. This section discusses the characteristics
of monopolistic activities and provides an overview of the main problems and concepts
to consider in their regulation.

2.1 Drivers of monopoly regulation

As mentioned above, a natural monopoly arises when it is less costly for a single
firm to provide a certain product or service. Natural monopolies are characterized by the
existence of strong economies of scale, i.e. marginal costs fall below average costs. An
unregulated monopoly could potentially charge consumers prices well above the costs of
production, creating economic inefficiencies (Joskow, 2007).
Additionally, several other conditions which denote the existence of a natural
monopoly can be found. Broadly speaking, natural monopolies may be characterised by
one or more of these features: i) economies of scale, ii) capital intensity, iii) non-
storability with fluctuating demand, iv) location-specific delivery generating location
rents, v) production of essential services for the community, vi) the product or service
supplied by two firms are close substitutes, vii) increased costs as a result of duplicating
facilities and viii) direct connections to customers (Joskow, 2007).
Notwithstanding, the existence of a market supplied by a single firm, or the fact that
this is the least-cost market outcome, does not necessarily imply that implementing
economic regulation is necessary. If it is relatively easy for new firms to start providing
a similar service (and stop doing it at a low cost), firms in the market would be deterred
from charging excessive prices due to the threat of new entrants. In the economic
literature, this is known as a contestable market. For example,
Box 2-1 describes the case of monopoly-like firm that is not subject to revenue
regulation as it is considered to perform a contestable activity.
The most relevant and common barriers of entry are caused by sunk costs, i.e. those
that once incurred cannot be recovered. These barriers are particularly noticeable under
the presence of (long-lived) fixed sunk costs and when these account for a large fraction
of total costs (Kahn, 1988). Consequently, sectors which require significant investments
in fixed assets may not be considered contestable, thus requiring the implementation of
economic regulation.

-4/42-
 Box 2-1: Example of a contestable activity in the energy sector (Cossent, 2013)
CLH (Compañía Logística de Hidrocarburos) is the main distributor of liquid fuels and
petroleum products in Spain. This company provides services related to storage,
distribution in tankers, pipelines, ships, etc., and logistics solutions. CLH also provides
some other value added services such as analysis and control of products, maintenance
services for petrol stations, control of fuel additives, etc. This is an activity that shows
cost subadditivity given that the integrated operation of storage and transport of fuels at
national level allows reducing operating costs and increases the geographical access of
all operators. However the regulatory intervention is limited to some restrictions on the
shareholders. Operators with petrol refining installations cannot own more than 45%
altogether and each individual shareholder cannot surpass 25%.
The organisation of the sector resembles a light-handed or self-regulation approach as
no further restrictions are placed on prices or revenues. CLH ensures transparent and non-
discriminatory access to all the operators by charging published prices which are equal to
all suppliers regardless of the volume contracted. Annual price updates linked to the RPI
and efficiency gains are made by CLH itself. Despite the fact that pipelines require indeed
significant investments, the existence of alternative transportation means such as trains
or tankers prevent CLH from charging abusive prices. In that case, alternative operators
could enter the market providing similar services. Therefore, the activities of CLH can be
considered as contestable, thus allowing for self-regulation.

In some sectors, competition for the market (franchise bidding) has been
implemented as a substitute for revenue regulation as proposed by (Demsetz, 1968). This
form of competition is relatively common, for instance, in the transportation or the
telecommunications sectors. However, its implementation in electricity networks
presents important limitations particularly related to the valuation of long-lived fixed
assets, as very long-term contracts would be needed under these conditions.
Notwithstanding, tendering processes have sometimes been implemented to award
licenses to electricity network companies. For example, electricity distribution licenses
in Panama are auctioned every 15 years. Nevertheless, distribution companies are still
subject to a revenue cap regulation. Therefore, the main purpose of franchise bidding in
these cases is not to use it as an alternative to revenue regulation but rather as a
complement to it. Thus, incumbents face the risk of losing the franchise and are
encouraged to adequately invest and deliver value to end consumers.
Coming back to the electricity sector, deregulation and liberalisation introduced
market competition in the generation and retail, whereas network activities, i.e.,
electricity transmission and distribution, were considered to be natural monopolies thus
remaining subject to regulation. It can be seen that transmission and distribution networks
present most of the characteristics of natural monopolies and important sunk costs, which
justify monopoly regulation. The remainder of this section will introduce the general
principles and concepts that should be considered when regulating network monopolies.

-5/42-
 2.2 Costs and regulation
Network companies need to incur in significant costs in order to perform their duties.
These costs can be broadly classified into two capital expenditures (CAPEX) and
operational expenditures (OPEX).
The former include mainly investments required to provide network access
(connections), reinforce the grid to accommodate any growth in demand and asset
replacement costs. Moreover, some additional investments are also necessary to ensure
adequate levels of quality of service, e.g. minimize interruptions, or to comply with
environmental requirements, e.g. building power lines underground in urban areas. Many
of these investments present long useful lives of several years that can go beyond 40
years. Additionally, CAPEX may also include other types of assets necessary to ensure
the smooth running of the firms’ operations such as protection systems or monitoring and
control devices. Lastly, some costs can be considered as CAPEX in spite of not being
related to the networks themselves such as office buildings, computer hardware, vehicles,
or other tools and equipment. A detailed list of all costs for an electricity distribution
company can be found in (OFGEM, 2007).
Similarly, OPEX comprise several categories being network assets maintenance
(preventive or corrective) and repair a prominent one. Furthermore, OPEX include
personnel costs, building rentals, expenditures in innovation, business support costs,
outsourcing, etc. Lastly, taxes are usually treated as OPEX.
In regulated monopolies, regulators must define the prices or revenues that network
firms can collect from their customers. These revenues must be sufficient to recoup their
(efficient) OPEX and make the necessary investments, including an adequate return on
capital. CAPEX remuneration is normally calculated as the sum of a term accounting for
depreciation and a term that represents the return on investments. Depreciation (D) is
computed according to gross assets and their useful lives, either linearly or through a more
complicated formula. On the other hand, capital remuneration is calculated as the product
of net assets (gross assets minus depreciation), also referred to as assets base (AB) and a
rate of return (r). Additionally, OPEX and taxes would be added to the previous
components. Equation (1-1) illustrates the main remuneration components.
R  AB  r  D  OPEX  Tax ( 2-1 )

The rate of return is frequently calculated as the weighted average cost of capital
(WACC). This means that the final rate of return is obtained as the weighted sum of the
cost of the different sources of financing used by DSOs, i.e. debt and equity.
Debt Equity
WACC   rdebt   requity ( 2-2 )
Debt  Equity Debt  Equity
For example, if 40 % of a company’s capital is debt and 60 % equity, and the interest
paid on debt (rdebt) is 5 % and the cost of equity is 8 %, the WACC is 6.8%.
Section 3 will provide further details about how to compute the WACC and its role in
the regulation of energy network companies.
It is important to remark that the values of several of the previous parameters can be,
and most likely will be, different in the companies’ accounting books and in the regulatory
calculations. For instance, it is frequent to use the terms regulatory asset base (RAB) in
order to refer to the amount of assets considered by the regulator to calculate the capital
remuneration or the term regulatory asset life, different from the actual useful lives of

-6/42-
assets, which is used to calculate the regulatory depreciation. As it will be shown
throughout this document, how the regulator determines the values of some of these
parameters can also be relevant to provide efficiency incentives.

2.3 Goals and role of regulation

The regulation of natural network monopolies comprises two complementary tasks: i)
determine the revenues that regulated firms are allowed to collect2, and ii) designing
network tariffs to allocate these costs among network users. Hereinafter, the focus will be
placed on revenue regulation. The issue of tariff design will be addressed in detail in
Module 13 of the Annual Training.
The economic regulation of natural monopolies faces a trade-off between two
conflicting objectives: 1) ensuring the financial viability of the regulated firms, and 2)
encourage productive efficiency so as to reduce the tariffs paid by end consumers.
On the one hand, the firms’ revenues must be sufficient to enable the utility to cover
its operational expenditures (OPEX) as well as allow them to carry out necessary
investments. In other words, revenues should ensure the network companies’ medium-
and long-term economic and financial viability. Note that if the regulator yields to
political or public pressure to lower tariffs in such a way that the return on the companies'
investment is insufficient, the regulated firms will likely decide to cut investments or, in
an extreme case, to revoke the utility license.
On the other hand, the rates paid by end consumers should be kept at an adequate level
by encouraging regulated companies to be efficient. This is particularly relevant in the
energy context as it is an essential product for the competitiveness of a country’s
industrial sector. Moreover, the well-being of people strongly depends on the access to
energy services, being energy poverty and vulnerable consumers’ protection a major
concern in many countries.
In the end, as it will be explained in subsequent chapters, the main differences between
the two main regulatory approaches that will be discussed in this document, i.e. cost of
service and incentive regulation, lie in whether the stronger emphasis is placed on cost
sufficiency or on cost efficiency.
Additionally, regulators should encourage firms to keep and appropriate balance
between OPEX and CAPEX and to exploit potential tradeoffs between OPEX and
CAPEX in order to gain in efficiency. For example, it may be more efficient to implement
a more advanced maintenance strategy (preventive maintenance) in spite of increasing
OPEX since this will allow lower CAPEX in the long-term as the useful life of assets can
be extended.
Furthermore, regulators should bear in mind that the incentives to increase efficiency
may jeopardize quality of service in an attempt to reduce costs at the expense of quality.
Therefore, quality standards are usually set, especially under incentive regulation
schemes, so as to prevent the deterioration of quality. Broadly speaking, two main
components of quality of service can be distinguished: commercial quality (consumer
satisfaction with the service provided, assistance through telephone, etc.) and technical

2Price regulation (e.g. limiting the average prices that can be charged), which is an alternative to
revenue regulation, usually involves an estimation of the sales of regulated firms. Hence both
approaches are very similar in essence, although their incentives properties can be very different.

-7/42-
quality. However, measuring quality of service is not always straightforward, as it may
have different meanings and implications across sectors, especially technical quality.
In the electricity sector, technical quality of service comprises the following: 1)
continuity of supply, i.e., number and duration of power supply interruptions; 2) voltage
quality, which measures disturbances that may affect the proper operation of apparatus
and equipment connected to the grid. Note that these indicators are directly related to
operation and maintenance costs as well as network investments.
Furthermore, regulators also control the investment in new infrastructure proposed by
the company in its transmission or distribution grid. Note that, the primary long-term
objective of regulators is to ensure that sufficient capacity is installed to meet the expected
demand at suitable levels of quality. Therefore, the regulator may attempt to solve this
difficult problem by establishing criteria to assess the suitability and necessity of the
investment plans proposed by the firms or the use of network planning models to
evaluate these plans. Nonetheless, this increases regulatory costs related to the
concomitant control and information gathering.
Lastly, a last function of regulation is to control the entry into and exit from the
regulated business. This is necessary to prevent the creation of inefficient parallel
networks, which may arise in case companies other than the incumbent monopolist tried
to provide the service. Thus, the incumbent is usually granted the exclusive rights to
provide network services in a certain area, in exchange for submission to regulatory
control. Moreover, the monopolist is normally obliged to supply to all users within a
certain distance of their grid, regardless of the associated cost, due to the fact that energy
is an essential service. Note that otherwise, companies may attempt to refrain from
servicing high-cost consumers or areas and focus on areas where costs are lower.
Nevertheless, this obligation may not hold for consumers located in secluded areas.
For instance, rural electrification in many countries requires specific regulatory
mechanisms, either by adapting the remuneration of the incumbent providing these
connections or establishing incentives for other potential suppliers to enter the market for
instance by granting territorial franchises to small local cooperatives.

2.4 Overview of regulatory approaches

In practice, several regulatory designs have been defined to meet the tradeoffs between
efficiency and cost sufficiency, whilst maintaining adequate quality of service. The
traditional type of economic regulation, used in the power sector for many years, is known
as cost of service regulation. A pure cost-of-service regulation consists of letting the
firms recover the costs they have incurred including a fair return on the investments they
have made. This is why this regulatory approach is sometimes referred to as rate-of-return
or cost-plus regulation.
In the previous paragraph, two adjectives have been written in italics. This is because
these are key issues in real-life cost-of-service regulation. Firstly, setting the allowed rate
of return in practice is a challenging task. This should be such that it attracts the required
level of investments but not as high as to exceed this level and make consumers pay more
than needed. Note that if this is too high, regulated firms would have an incentive to over
invest. Secondly, the definition of a pure cost-of-service regulation is based on two
assumptions: i) the firms’ costs are not scrutinised and ii) revenues are adjusted on a
continuous basis to match costs. However, as it will be shown in section 3, these
assumptions do not reflect actual regulatory practices.

-8/42-
 Incentive regulation, which can be seen as an extension of cost of service regulation,
was introduced in an attempt to overcome the scarce of efficiency incentives that cost-of
service regulation provided regulated firms with. This is achieved by decoupling the
revenues earned by regulated companies from their costs for a number of years. Thus,
their earnings will increase if they manage to reduce costs. Undoubtedly, the most popular
form of incentive regulation is RPI-X regulation. According to this remuneration
formula, the prices or revenues earned by network companies are updated annually
according to inflation and a factor X which represents the yearly targeted efficiency gains
defined by the regulator. This regulatory approach is nowadays widely spread,
particularly in Europe and Latin America. Thus, section 4 dealing with incentive
regulation will focus on RPI-X regulation, whereas section 5 will discuss how these
schemes are implemented in practice.
Some alternative regulatory approaches have been proposed, although these are rarely
applied. These are mentioned here briefly to complete this section. For instance,
yardstick competition consists in setting the tariffs that a company is allowed to charge
as a function (e.g. the average) of the costs declared by other regulated operators in the
same business (Shleifer, 1985). This encourages companies to lower their costs, since
their yardstick does not depend on their actual behaviour. Nonetheless, some practical
problems may arise. On the one hand, regulated companies may not be fully comparable
so as to assume that the differences among them are due exclusively to different degrees
of efficiency. On the other hand, collusive behaviour may be encouraged.
Notwithstanding, the concept of comparing firms among them is frequently used in RPI-
X incentive regulation through regulatory benchmarking.
Finally, a last regulatory approach that may be found is light-handed regulation or
self-regulation. Under this scheme, the company freely sets the tariffs to be charged to
consumers, whereas the regulator performs relatively lax ex-post controls of the prices
charged or the rates of return obtained. Thus, the regulator would only intervene in case
abnormal deviations occur. Light-handed regulation is usually accompanied by
information disclosure requirements to the utilities.

3 Cost of service regulation

Under cost-of-service, also known as rate-of-return regulation, the tariffs charged by
the utility, or the revenues earned by them, are set and authorised by the regulator. Thus,
periodical “re-negotiations” take place between the regulator and the companies in what
are generally known as rate cases or price control reviews. These processes essentially
entail two successive stages: revenue determination and tariff design3. As mentioned
above, Module 13 of this course will address the issue of tariff design in more detail.
Therefore, the focus will be placed on revenue setting.
Allowed revenues (rate level) are determined. This process involves: i) quantifying the
company’s costs and investments, and ii) setting a suitable rate of return. Rate cases rely
on the data furnished by the company for the preceding accounting period, usually in a
standardized way known as regulatory accounting system; and a forecast of the
investments needs for the next control period. The resulting tariffs will remain in place

3 The subsequent discussion mostly refers to the traditional cost of service regulation implemented
in the US, which was frequently applied to vertically integrated utilities. Hence, this may vary slightly
from the arrangements in other countries.

-9/42-
until the next price review. This can either occur annually, or at irregular intervals. For
instance, rate hearings in the US occur at irregular intervals, when regulated firms or
regulators ask for it. Rate cases or price reviews usually comprise the following steps
(Rothwell and Gómez, 2003):
 The tariff revision process is initiated either because the established period has
lapsed or, more frequently, because the utility requests it.
 The firms submit the information required by the regulator, including detailed
accounting information and investment plans.
 A negotiation process takes place, after which the regulator determines the rate of
return to calculate the capital remuneration and the allowed level of expenses.
 Finally, tariffs are adjusted to the allowed revenues calculated in the preceding
step, by taking into account the expected demand in the period. Since demand
may change with prices depending on its elasticity, information about this
parameter is likewise required.
The remainder of this section reviews the main parameters assessed by the regulator
during rate cases (allowed costs, asset base and rate of return) and summarizes the major
pros and cons of this kind of regulatory approach.

3.1 Determining the cost of service

The cost of service, i.e. the amount of costs that the regulator will allow regulated firms
to pass through to rate payers, is calculated as the sum of all different cost components.
As shown in Figure 3-1, these include operational costs, depreciation, return on capital
(rate of return times the rate base and others (e.g. taxes, minus additional revenues). At
every rate case, the investments approved by the regulator are added to the rate base
whereas depreciation is subtracted. Then, the capital remuneration is computed as the
regulatory asset base times the allowed rate of return.

Allowed costs
Approved capital
investment Operating costs

+ _
Depreciation
Rate base

x
Allowed returns

Rate of return

Others

Figure 3-1. Allowed revenues under cost-of-service regulation

Box 3-1 presents a typical general breakdown of the allowed revenues for a vertically
integrated electric utility. The revenues are intended to recover the total costs of providing
the service, which in this case include the generation, transmission, distribution and retail
activities. A similar structure would apply to any regulated monopoly.

-10/42-
 Box 3-1. Regulated revenues under cost of service

The following accounting formula represents the balance that the regulator must
strike when reviewing a rate case.
AR  CoS  O & M  D  RAB  r  Tax  ADR ( 3-1
)
where
AR is the allowed revenues,
CoS is the total cost of service,
O&M is the allowed operating and maintenance costs,
D is the depreciation expenses on the company’s gross assets,
r is the allowed rate of return,
RAB is the regulatory rate base, calculated as its net assets, defined to be its
gross assets less depreciation,
Tax is the taxes for which the company is liable, and
ADR is the additional revenue.

The precise definition of terms in the previous formula for vertically integrated
electric utilities used in cost-of-service regulation in the US is as follows (see
(DOE/FERC, 1973) for further details):
 Operating and maintenance costs: cost of fuel, material and replacement parts,
energy purchases, supervision, personnel and overhead.
 Depreciation: the straight line method is generally used. Fixed assets under
construction are not depreciated.
 Tax: all taxes for which the utility is liable, i.e., on profit, revenue and property,
as well as social security and construction tax (except as relating to fixed assets
in progress, since such tax is built into the value of the asset).
 Rate base: net fixed assets (power plants, transmission and distribution
facilities, other tangible and intangible fixed assets and nuclear fuel, less the
cumulative depreciation for all these items) and current assets (fuel and other
material and replacement part inventories, advance payments and deferred
revenue, research and development expenses and current asset requirements).
 Rate of return: average weighted interest rate on the company’s long-term
financial resources (bonds, debt certificates, shares and preferred shares).
 Additional revenue: expenses/revenues deriving from the sale of the company’s
property, revenues from wholesale energy sales and other revenues not directly
related to producing electric power.

-11/42-
 There are two issues that take up most of the effort in the negotiations between utility,
regulator and stakeholders to determine the allowed revenues for the next price control
period: the allowed rate of return, r, and the investments to be included in the rate base or
RAB. These will be addressed in subsequent subsections. Additionally, note that the
formula above accounts for the allowed OPEX, which may not necessarily be the same
as actual expenses. This provides an incentive for the company to enhance its efficiency;
an inefficient company management would be penalized.
During the rate case processes, regulators may also correct revenue allowances so as
to keep firms in a good financial health. For instance, the simple numerical example
displayed in Table 3-1, which presents the list of the accounting items used to calculate
the rate of return. An adjustment is made in the calculation of the allowed revenues to
improve the allowed rate of return, which would otherwise be lower than deemed
advisable by the regulator.

Anticipated Revenue Rate case after

rate case Adjustment adjustment
Revenues 30000 1600 31600
Expenses
Fuel 24000 24000
Operating 3000 3000
Depreciation 1000 1000
Total expenses 28000 28000
Net operating revenues 2000 3600
Rate base (RB)
Net assets 42000 42000
Working capital 350 350
Total rate base 42350 42350
Rate of return 4.72% 8.50%
Table 3-1: Example of utility accounts and calculation of rate case (in k€). Source:
Rothwell and Gómez (2003)

Generally, the resulting tariffs, calculated once the revenue allowances have been
determined, would remain unchanged until the next revision. Thus, companies are
encouraged to cut costs. If operating costs are lower than anticipated in the rate case, a
higher rate of return is attained. Conversely, if costs are higher than anticipated the rate
of return will decline. This intrinsic efficiency incentive is more important as the time
lapse between tariff revisions, a circumstance known as regulatory lag. Section 4 will
show incentive-based regulation is essentially a formalization of this regulatory lag. On
the contrary, if frequent mid-term revisions are made in such a way that the regulator
constantly adjusts the rate of return, the companies will see no incentive to reduce costs.

-12/42-
 3.2 The regulatory asset base
One of the most important aspects of rate case negotiations is the criterion used by the
regulator to calculate the regulatory rate base (RAB) The regulatory asset base is the
amount of net assets considered by the regulator as the basis for the calculation of the
CAPEX remuneration.
The methodology that should be followed to determine the RAB at the beginning of
each regulatory period essentially depends on the particular circumstances of each
country at the time the price review is carried out. Theoretically, two opposing approaches
can be found, namely book values or purchase costs and replacement or duplication costs.
An extreme implementation of the duplication cost concept is the new replacement value
(NRV) used in Latin American countries to regulate electricity distribution (Rudnick et
al., 2007).
In principle, book values are preferable over duplication costs as it provides higher
regulatory certainty to firms. However, implementing book values in practice can be hard
and burdensome considering that some investments may be over 30 years old, past
information may not be 100% reliable due to changes in the regulatory framework,
technology, firm ownership or heterogeneous accounting rules. Moreover, the regulator
may doubt that some of the past investments could be deemed efficient or decide that
consumers should not pay for certain investments that despite the fact that they could
have been considered efficient in the past, these are not “used and useful” nowadays
(Kahn, 1988). Under these circumstances, book values may be over-compensating some
firms at the expense of ratepayers.
Nevertheless, presumably the most relevant implication of the asset valuation
approach selected is related to how technology and costs evolve over time. If the asset
remuneration is higher than present costs, firms may earn too high returns or overinvest,
whereas the cash flows generated by CAPEX remuneration may be insufficient to drive
investments under the opposite circumstances (Kahn, 1988).
Hence, several intermediate RAB valuation methods have been developed to attain a
balance in the existing simplicity and cost-reflectivity trade-off. The most relevant
approaches are depicted in Figure 3-2. Since in many cases gross assets are computed, it
is also necessary to assume the remaining regulatory lives of assets to compute the RAB
(net assets), either by estimating the average life of assets or assuming new assets
(replacement).

Book Implicit Reproduction Reproduction New

value RAB costs costs replacement
(historic (modern value
average costs) costs) (from scratch)

Figure 3-2: Main approaches to calculate the RAB (Cossent, 2013)

A very simple alternative to book values is the calculation of the implicit RAB. The
only input data that are required are the initial distribution revenues, the share of CAPEX
over total distribution revenues, the average age of assets, the regulatory asset life and the
WACC. The mere multiplication of the two first parameters would provide the total initial

-13/42-
CAPEX remuneration for each DSO (from the previous rate case). The remaining input
data allow estimating the RAB as shown in equations 3-2 to 3-5. The major drawback of
this approach is that it must be assumed that the initial CAPEX remuneration is adapted
to the real RAB of the firms.
GA Life  Age
CAPEX  D  RAB  WACC   WACC   GA ( 3-2 )
Life Life
GA
CAPEX   1  WACC  Life  Age   D  1  WACC  Life  Age  ( 3-3 )
Life
CAPEX
D ( 3-4 )
1  WACC  Life  Age 
CAPEX  D
RAB  ( 3-5 )
WACC
Where:
CAPEX Annual CAPEX allowance
D Annual depreciation remuneration
GA Gross assets implicit in the CAPEX remuneration
Life Regulatory life of assets
Age Average age of assets

Closer to a NRV one may find the network reproduction cost, sometimes referred to
as the replacement cost of existing infrastructure. This method essentially consists in
computing the gross assets through the inventory information and some unit costs
determined by the regulator. These standard unit costs can correspond either to the
historical purchase costs averaged across DSOs and updated to present prices or the
current purchase costs. The former option is closer to a book value valuation whereas the
latter is closer to the NRV. In any case, reproduction costs introduce elements of yardstick
competition among DSOs in terms of the costs of inputs. Moreover, this approach
presents some of the advantages of a duplication cost approach with the mitigation of the
risk of creating stranded costs as in the NRV methodology.
Evaluating the asset base requires significant efforts from both the regulator and DSOs
and may lead to litigations. Therefore, instead of reassessing the RAB at the beginning of
every regulatory period, process known as reopening the RAB, the regulator could
decide to include in the RAB from the last rate case the non-depreciated investments
already allowed in previous rate cases. This is known as consolidating the RAB. This
approach mitigates regulatory instability and reduces the regulatory burden as only the
investments corresponding to the last regulatory period are subject to regulatory scrutiny.
Consolidating assets does not require such a detailed accounting as book values. It is
just needed to define a general RAB structure, i.e. the different asset categories to be
considered. For each one of these classes a distinct regulatory life of assets and WACC
can be defined. Then the regulator would have to keep track of the additions and
depreciation for every year and asset category. Table 3-2 shows an example of a possible
structure for the RAB of an electricity distribution company.

-14/42-
 Asset category Regulatory asset life WACC

Lines/transformers 40 8%
Control centres/communications 20 8%
Protections/measuring equipment 10 8%
Smart grid investments 15 9%

Table 3-2: Example of RAB structure definition for an electricity distribution company
(Cossent, 2013)

However, this method usually can only be applied up to a certain point in the past,
before which it is necessary to evaluate the whole RAB through one of the other methods
described. This can be caused either by lack of data or due to financial issues. For
instance, reopening the RAB has been frequently the case in power sector privatization
processes around the world. Conversely, in the US electricity industry, regulators have
traditionally used book value to assess the rate base and have focused their efforts on
adjusting the rate of return; hence the name rate-of-return regulation.

3.3 The rate of return

The most common method for defining the allowed rate of return is to calculate the
weighted cost of the different forms of financing used by the company, such as bonds or
shares traded on equity markets. This manner of calculating the rate of return is known
as the WACC (see section 2.2). An example is shown in Table 3-3. The WACC not only
determines the capital remuneration, but it may also be used as the discount rate applied
to calculate the present cost of projections when calculating allowed revenues throughout
several years. This application of WACC will be shown in more detail in section 4
(smoothing X factor).

Capitalisation ratio Rate (%)

Bonds 28 4.00
Quoted shares 12 12.00
Equity 60 6.50
Total 100 6.46
Table 3-3: Example of calculation of allowed rate of return

The interest rate on debt is usually lower than the rate of return on equity, since
shareholders are more exposed to the financial failure of the company than the lenders.
Then the WACC formula (Equation 1-2) would appear to infer that raising the percentage
of debt in the company’s overall capitalisation would lower the WACC. However,
increasing debt also raises the likelihood of bankruptcy, and therefore the interest rate on
debt, since the risk of default of the firm increases intolerably. Since risk rises with the
debt/equity ratio, lenders will demand a higher rate of return.
The most controversial item is usually the rate of return on the company’s own capital,
i.e., the remuneration of its equity, which the regulator typically establishes in accordance
with the rates set for regulated companies with a similar level of risk: electric power
distribution compared to gas distribution, water supply or communications, for instance,
or the actual rates of return of other firms, again with an estimated similar risk level. The
most widely used method is the capital asset pricing model (CAPM), which determines

-15/42-
the cost of capital as the sum of a risk-free rate plus a market risk premium according to
the formula shown in (3-6).
requity  R f    Rm  R f  ( 3-6 )

Where:
Rf Risk-free rate, generally drawn from State bonds interest rate
β Parameter representing the volatility of the value of the company’s shares
compared to average market volatility
Rm Expected return on the market of an efficient portfolio

CAPM is based on the assumption that the rate of return for any activity is equal to:
 the rate of return on risk-free assets in the economy in question, regarded to be the
amount received by investors placing their money in the safest financial assets,
typically State bonds (this was correct before the financial crisis), computed as the
average for the last few years for long-term rates, to establish a basis consistent
with the life of the company’s assets,
 plus a risk premium based on the degree to which the asset tracks the securities
market: in other words, the specific additional risk associated with the asset over
and above the average market risk (different methods are used to appraise debt and
equity).
The risk premium applied to company equity is assumed to be proportional to a
coefficient β which represents the volatility of the value of the company’s financial
assets (shares) compared to average market volatility (see Rothwell and Gómez (2003)
for further details).
In countries with no international securities exchange or which lack sufficient liquidity
for the type of industries regulated, or where the regulatory or financial risk is perceived
to be high, the cost of equity is adjusted upward to include a country risk premium.
Summing up, the utility’s average cost of capital is calculated as follows:

 R f    ( R m  R f   R c 
Equity Debt
WACC   R debt ( 3-7 )
Debt  Equity Debt  Equity

Where:
Rc Country risk premium
Rdebt Cost of debt
Since these rates may be expressed in nominal or real (net of inflation) terms, the first
step is to decide whether nominal or real WACC is to be used. The following expression
indicates the relationship between the two.
1  WACC nominal  (1  WACC real )  1  Inflation  ( 3-8 )

Finally, it should be considered that the interest paid on debt is corporation tax
deductible in earnings and cash flow calculations. Economically speaking, this means that
the company pays interest at only 1-t of the before tax interest rate.Consequently the
WACC value must be defined as a before tax or after tax rate. Assuming t to be the tax
rate per unit, the following relationships hold:
WACC after_tax  (1  t )  WACC before_tax ( 3-9 )

-16/42-
 Debt Equity
WACC after_tax   R debt  (1  t )   Requity _ after _ tax ( 3-10 )
Debt  Equity Debt  Equity

3.4 Strengths and weaknesses of cost of service regulation

Cost of service regulation has been progressively abandoned for incentive regulation
schemes. Its most commonly discussed shortcoming is the so-called Averch-Johnson
effect (Averch and Johnson, 1962). These authors showed that a regulated firm under a
constraint on the rate of return would not minimise its costs. If the allowed rate of return
exceeds the cost of capital, firms would tend to substitute labour inputs for capital inputs
in order to maximise their profits. In practice, this means that if the rate of return is too
generous, regulated companies would tend to over-invest. Conversely, if the rate of return
is lower than the cost of capital, the utility will invest very little and its operating costs
will rise, likewise generating economic inefficiencies.
However, the empirical evidences for the existence of the Averch-Johnson effect in
practice are not definite (Joskow, 2005). Notwithstanding, regardless of whether in the
real world the company behaves as predicted by the Averch-Johnson model, this type of
regulation requires the regulator to accurately calculate the utility’s real cost of capital.
Thus is not straightforward due to the existence of information asymmetries between
the regulator and the company. Being this the case, regulated firms may also justify over-
investment through technical arguments that may be difficulty assess ed by the regulator.
This asymmetry is mitigated by the regulator requesting as much information as
possible. This means that regulated firms must do the work involved in furnishing it. The
more involve regulators become in investment planning and operating cost management,
the greater is their understanding of the problems they must regulate. However, regulatory
burden and costs increase as well because more specialised personnel and more
sophisticated analytical tools are required.
Nevertheless, the most relevant drawback of cost of service regulation, and which has
been the main reason for its abandonment in many countries, is that regulated firms
receive very poor incentives to gain in efficiency. This is because when tariffs are
revised frequently, for instance every year or two, the company can recover all costs
accepted by the regulator. Furthermore, in some cases it can evolve into an intrusive and
legalistic regulation.
In spite of the aforementioned drawbacks, cost of service regulation, when judiciously
applied, presents important advantages too. Besides its apparently simplicity, the most
relevant pro is presumably that, in principle, regulated firms are certain that costs will be
recouped. This provides a strong financial stability and may reduce capital costs in the
long-term, thereby reducing tariffs, due to the lower risk faced by the utilities. Moreover,
a favourable environment for investments is created, thus ensuring adequate levels of
quality of service.

4 Incentive-based regulation
The basic principle behind incentive-based regulation is to decouple revenues from
actual costs so that regulated firms see an incentive to reduce costs in order to earn the
differential. However, if this separation is kept indefinitely, consumers would never
benefit from the cost reduction. Therefore, price reviews similar to the rate case
proceedings described in the previous section are carried out periodically every few years.
For each regulatory period, generally ranging from three to five years, a specific

-17/42-
revenue/price path is defined, thus creating an incentive to lower costs and thereby
increase profits. The price reviews performed at the end of each regulatory period result
in a new formula to set revenues or prices for the following years. Box 4-1 presents a
simple example to illustrate the differences between a pure cost of service regulation and
a forever incentive regulation.

Box 4-1. A scheme for illustrating the power of incentives

Laffont and Tirole (1993) introduced a simple model to establish the link between
traditional and incentive-based regulation:
R  (1  b)  Costsexante  b  Costsex post ( 4-1 )

Where:
R Final revenues earned
b Coefficient between 0 and 1, defined ex ante
Costsex-ante Estimated allowed costs (ex-ante)
Costsex-post Actual costs incurred (ex-post)
According to the previous formula, the incentive for the utility to reduce costs is
inversely proportional to the value of the parameter b. A pure cost-of-service regulation
can be likened to a low incentive type of regulation in which b = 1, where the company
is allowed to recover all the costs incurred. On the contrary, a forever incentive
regulation could be seen as setting the value of b to 0. Thus, no ex-post adjustment of
revenues is made. Nonetheless, real-life regulatory frameworks could be considered to
be in between. Actual implementations of cost of service would correspond to high
values for b, whereas other schemes, such as tariff freezing arrangements or incentive-
based regulation with long regulatory periods, would correspond to low values of b.

Incentive-based regulation normally involves setting an ex-ante revenue allowance

together with an ex-post correction depending on actual costs incurred. Thus, efficiency
gains achieved by the company are passed through to consumers in the next regulatory
period. Regulation involving incentive mechanisms poorly adjusted to ex-post costs may
lead to excessively high or low total revenues or may even jeopardise the feasibility of
the company’s long-term business plan. Hence the need for a price control review every
few years. Regulatory design should strike a balance between lower short-term costs,
global economic efficiency and long-term viability.
The most common incentive regulation approach is the so-called RPI-X regulation.
Two basic alternative schemes for incentive-based regulation: revenue cap and the price
cap. Moreover, these schemes may be combined with schemes for profit- or loss- sharing
with users to reduce risk. The remainder of this section will address these topics in further
detail.

4.1 Price cap

Price cap regulation consists in setting the maximum yearly price (or the maximum
average price) that the company can charge for each service provided, for a period of
several years. These prices are adjusted annually to account for inflation minus a
correction factor associated with expected increases in productivity:

-18/42-
Pm,t  Pm,t 1  (1  RPIt  X )  Z ( 4-2 )

Where
Pm,t Maximum price that the company can charge for service m in year t

RPIt Annual price variation per unit (retail price index, RPI, or inflation rate) in year t
X Productivity factor in per unit
Z Adjustments owing to unforeseen events beyond the control of the utility, such as
natural catastrophes, environmental regulation or tax hikes.
This form of regulation has been used in the United Kingdom, where it is known as
“RPI - X”, to regulate telecoms, gas and electricity network utilities; and in the United
States to regulate telecommunications companies, under the term “CPI – X”. In the power
sector, revenue cap regulation (presented in section 4.2) is nowadays much more
common. Note that cost-of-service regulation with tariffs frozen for several years can be
viewed as price cap regulation with no correction for enhanced productivity.
Figure 4-1 shows a possible price path during a regulatory period. In this case, a
positive value of X has been assumed, hence prices decline in real terms. Broadly
speaking, the larger the value of X is, the stronger the incentives to cut costs are. If the
company is able to lower its costs below the price trajectory set by the regulator, its
benefits will increase. Note that in some situations prices may tend upward rather than
downward, e.g. when the regulator recognises high investments in a given regulatory
period. At the end of the price control period the regulator will establish the price for the
following years, according to the updated estimates of costs, and will set a new value of
X for the next period. The benefit for consumers results from the combination of the new
price for the next year and the new value of X. Section 5 will go deeper into this
implementation details.

Figure 4-1: Price trajectories under price cap regulation. Source: (Ajodhia, 2005)

The general RPI or CPI is not the optimum indicator for this purpose because it reflects
variations in consumer prices but not in specific industry costs. In practice, mixed

-19/42-
indicators found as a weighted average of individual industry (or company) cost indices
and the general inflation rate are also used.
Moreover, the price limitation may be applied to: i) the average price for the company
as a whole, after duly weighting each of the services provided; ii) the average price to be
charged to each consumer class; or iii) the fixed price of each of the terms comprising the
end user tariff. Additionally, regulators may choose either to allow regulated firms to
design its end consumer tariffs themselves or alternatively design the end user tariff
herself to prevent cost shifting among consumer classes.
For example, electric power distribution companies in the UK used to be regulated
under a price cap regime known as “revenue yield” cap. According to this mechanism the
maximum revenue per kWh of electricity distributed is defined by the regulator.
Moreover, British distribution companies are free to design their end-user tariff structure
as long as the aforementioned limit is observed, albeit under regulator supervision.
Nowadays, the remuneration of distribution companies has evolved into a revenue cap,
but they are still allowed to set the end-user distribution tariffs. A revenue yield
remuneration formula, limiting the average revenue per unit of demand supplied, is shown
in equation (4-3).
Rt R
 t 1  (1  RPI t  X )  Z ( 4-3 )
Dt Dt 1
On the contrary, regulators in Chile, Peru or Argentina sets the maximum prices that
can be charged for the various cost items, such as investment, operation and maintenance,
customer management, and the adjustment coefficients applicable to those prices
throughout the regulatory period. Price caps are included in the resulting formulas for
calculating the distribution tariffs to be paid by end consumers. In this case, price caps
are translated directly to end-user tariff design. The component of the tariff corresponding
to electricity distribution in these countries is known as distribution aggregated value (or
Valor Agregado de la Distribución –VAD-, in Spanish).

4.2 Revenue cap

Revenue cap regulation is very similar to price cap regulation, although the cap is
placed on the maximum yearly revenues allowances. The simplest expression for a
revenue cap is given by:
Rt  Rt 1  (1  RPIt  X )  Z ( 4-4 )

Where:
Rt Allowed remuneration or revenues in year t
RPIt Annual price variation per unit (retail price index, RPI, or inflation rate) in year t
X Productivity factor per unit
Z Adjustment for unforeseen events beyond the control of the utility.

Nonetheless, many different variations can be found. Revenue cap formulas may
include the addition of revenue drivers (number of consumers, total energy supplied or,
in network companies, the size of the network), Z factors to account for unexpected events
causing cost deviations (natural catastrophes, environmental regulation or tax hikes), an
extra term representing non-controllable costs exempted from efficiency gain

-20/42-
requirements, terms accounting for incremental investments or even differentiation per
voltage levels.
For example, the formula in (4-4) includes a revenue driver term, which can account
for load growth, new customer connections, etc, or a combination of these:
Rt  Rt 1  (1  RPIt  X )  (1    Dt )  Z ( 4-5 )

Where:
 is the economies of scale factor, typically lower than 1, which provides an
indication of how much the regulated costs and therefore revenues increase in
proportion to a cost driver, represented by D.
Dt is the increment in year t per unit of the selected cost driver (the formula could
include just one or several cost drivers), such as units of energy supplied, number
of customers, network size or any combination of the three.
Revenue caps set at the beginning of the regulatory period and adjusted yearly in terms
of RPI-X only are known as “fixed revenue” caps. “Variable revenue” caps are adjusted
yearly by both the RPI-X factor and other cost drivers.
Under a revenue cap, tariffs must be designed so total revenues do not exceed the cap.
Nonetheless, deviations are bound to occur. Therefore, an adjustment mechanism should
be designed to correct for such deviations. This adjustment mechanism could also take
into consideration when actual revenues fell short of the allowed revenues by providing
the corresponding compensation to the company next year.
Both revenue and price cap provide similar incentives to lower costs. However,
variations in sales or energy distributed generate very different effects in the two schemes.
Whereas price cap cap encourages higher sales, revenue caps, depending on how
remuneration for sales growth is incorporated in the formula, is always more neutral to
this effect. Therefore, in sectors where energy savings or demand-side management
programmes are desirable, revenue caps are better suited.
Moreover, due to the characteristics of energy networks, the variation of costs with
demand is normally quite small in the short and medium term. This variation is more
apparent in the longer term as capacity upgrades are required. Consequently, revenue caps
including appropriate cost drivers duly adjusted for economies of scale are the most
popular scheme for price control in this type of regulated businesses. In the US, a good
number of state regulatory commissions have implemented “revenue decoupling” from
sales to regulate utilities. Note that this is complicated to achieve in vertically integrated
utilities which have a natural incentive to increase demand. Examples of how this partial
decoupling can be designed and work in practice can be found in (National Action Plan,
2007).

4.3 Mechanisms for sharing earnings and losses

In purely ex-ante incentive regulation systems, regulated firms are fully exposed to
any deviation between actual costs and allowed revenues/prices. Thus, they may obtain
excessively high earnings or suffer inadmissible losses as a result of deviations between
ex-ante allowances and actual costs. In order to mitigate this undesirable characteristic of
purely ex-ante regulation, several mechanisms can be found to share the risk between
companies and consumers by distributing cost deviations according to some predefined
rules. Sliding scale and profit sharing schemes are the most common approaches.

-21/42-
 (Viscusi et al., 2005) describes a possible application of sliding scale regulation based
on an ex-post correction of the allowed rate of return, r, earned by the regulated firms,
shown in (4-6). A value of unity for the sharing factor h would correspond to a cost-of-
service regulation, whereas a value of zero for this factor would represent a pure incentive
regulation system. Intermediate values for h correspond to a profit sharing system, being
the power of the incentive to reduce costs the closer to zero the value of the sharing factor
is. For example, a value of h = 0.5 would indicate that the profit or loss would be equally
shared by the company and its customers.

r  rt  h  r   rt  (4-6)

Where:
h represents the sharing factor, which is a constant ranging from 0 to 1,
rt is the rate of return obtained by the company in year t as a result of the tariffs set
in the preceding rate case
r* is the reference/target rate of return.

Sliding scale mechanisms may be progressive or regressive. In progressive

mechanisms, the part of the profit retained by the company rises with cost savings. For
example, a company might receive 20 % of the first 5 % saved, 40 % of the second 5 %
and so forth. Since cost savings are more difficult to achieve as total cost declines,
progressive mechanisms that provide for retaining a higher proportion of the savings
constitute more effective incentives for the company than regressive mechanisms.
Different examples of sliding scale mechanisms applied to the regulation of electricity
companies in the US can be found in (Comnes et al., 1995). These are summarized in
Figure 4-2.

Figure 4-2: Examples of earning sharing mechanisms in the US (Comnes et al., 1995)

A similar sharing mechanism can be applied to overall revenues or prices instead of to

rate of returns, as a complement to incentive or performance-based regulation (see section

-22/42-
4.4). Being this the case, regulated firms would not retain the whole difference between
ex-ante allowances and actual costs, but only a share of it. In these cases, the mechanism
is frequently called profit or earning sharing scheme. It is important to remark that
regulators should define the sharing factors in an ex-ante basis so as to avoid regulatory
uncertainty. Thus, ex-post corrections are made based on ex-ante defined rules.
Despite mitigating risks, sliding scale mechanisms do not fully overcome information
asymmetries. (Laffont and Tirole, 1993) argue that a regulator can perform better by
offering regulated firms a menu of contracts. Some of these contracts would be closer to
a pure cost of service regulation and others to a pure price/revenue cap regulation. In
Viscusi’s formulation, this could be implemented through different pairs of values for the
reference rate of return and sharing factors. Firms which are capable of achieving large
gains in efficiency would tend to opt for a high powered regulatory contract (closer to
price/revenue cap), whereas firms with less opportunities to do so would choose a low-
powered scheme (closer to cost of service). Hence, companies would tend to reveal their
true cost opportunities to the regulator, thus reducing asymmetries of information.
The so-called information quality incentive (IQI) implemented in the UK for
electricity distribution is an example of the application of these mechanisms. Therein,
firms are offered a menu of profit-sharing contracts or, better said, the regulatory contract
that applies to each company depends on the ratio of the firm’s estimation of revenue
needs and regulator’s revenue prognosis. This is a two step process. At the beginning of
each regulatory period, the companies’ cost estimations are compared against a baseline
determined by the regulator. This baseline would be similar to the revenue allowances in
a conventional revenue cap regulation. At the end of the period, the actual costs of each
company are compared against the ex-ante revenue allowances and final revenues are
computed following the rules defined in the corresponding contract. Further details on
the UK’s experience can be found in (Crouch, 2006).

4.4 Design of an incentive-based regulation scheme

Designing an incentive-based regulation scheme such as a price or a revenue cap
requires making certain key decisions, addressed in the discussion below (see Navarro
(1996) and RAP (2000) for further details).
Definition of the regulatory period
Four or five years is normally a good compromise, as it leaves sufficient time to create
incentives for the company to lower its costs (productive efficiency) without running the
risk of prices or revenues deviating too far from costs (seeking both financial viability for
the utility and cost-of-service efficiency for consumers). If particularly significant
changes take place that entail substantial deviations from the initial estimates for the
regulatory period, a rate case may be initiated to review the price or revenue formula
before the regulatory interval expires. Nowadays, there are discussions about the need to
lengthen regulatory periods for better ensuring the recovery of riskier investments in new
clean energy technologies and innovation in energy infrastructures. Being this the case,
intermediate, partial or less intrusive, price reviews would be presumably carried out.
Determination of baseline and adjustment parameters
Throughout the review and before the regulatory period begins, regulators must
proceed to analyse all the information furnished by the company on past costs and future
investment plans. Thereafter they must reach a decision on the costs that the company
will be allowed to recover in the following period. This may involve the use of detailed

-23/42-
analyses of each cost item and, wherever possible, engineering, operations research or
econometric models that can also process data on the other companies in the group for
benchmarking purposes.
In practice regulators classify costs in several categories:
 Operating expenses (OPEX): personnel, maintenance and operation.
 Capital expenditure (CAPEX): yearly depreciation, the return on the RAB and
projected investment during the price control period.
 Uncontrollable costs: taxes, upstream fees and other exceptional cost items.

Furthermore, regulators have to set several parameters that determine how revenues or
prices evolve during the regulatory period. The chief parameters involved are inflation,
productivity (X factor) and variation in market size, e.g. the variation in the number of
consumers or energy delivered.
Broadly speaking, there are two approaches to set the baseline for revenue or price
caps and adjustment parameters i) building blocks and ii) the total expenditure (TOTEX)
approach. For a fuller discussion see Ajodhia (2005) and Petrov and Nunes (2009).
Under the building blocks approach, the regulator assesses the OPEX, CAPEX, and
uncontrollable cost caps separately for each year of the regulatory period. OPEX and
investment efficiency can be determined by benchmarking. Under this approach,
baselines can either be set separately for OPEX and CAPEX or as a total maximum
allowed costs, which is the sum of each year’s OPEX, CAPEX and uncontrollable costs.
The former allows regulators to monitor more closely specific cost items, whereas the
latter reduces the possibility of firms’ capitalizing OPEX and vice versa. This is especially
relevant when cost gains can be achieved through balancing trade-offs between OPEX
and CAPEX, e.g. reducing investment needs by increasing maintenance.
Under the TOTEX approach, separate assessments are not explicitly performed for
regulated OPEX and CAPEX. Rather, a single TOTEX cap is calculated by the regulator.
In TOTEX, a company’s productivity factor X for efficiency improvements is usually
determined by the regulator by benchmarking. In theory, this approach is simpler and
give regulated companies greater freedom to make the optimal choice in the trade-offs
between OPEX and CAPEX to reduce costs. However, it makes it harder, for instance, to
evaluate investment plans.
In practice, regulators use a mix of procedures or strategies to define the revenue cap
formula. Thus, the resulting method can seldom be classified entirely under a single
approach. Regulators may, for instance, use the building blocks approach to calculate the
revenue cap formula and then set the cap on TOTEX, evaluating the firms’ efficiency on
the grounds of total savings.
Definition of secondary objectives/additional incentive schemes
In addition to the primary objective pursued with this type of regulation, i.e., lower
costs, other specific aims relating to the characteristics expected of the service provided
by the utility may also be sought. These objectives are related to company performance.
This type of mechanism is also known as performance-based ratemaking or regulation
(PBR), where the company’s remuneration depends on its actual performance in meeting
the specific targets defined by the regulator. Examples of such objectives include
improvements in quality of service and consumer satisfaction, universal service for all
consumers within the franchise area, reduction of the environmental impact caused by the

-24/42-
company’s activities and implementation of programmes in the public interest such as
research and development or energy efficiency schemes. The regulatory scheme must
include explicit financial mechanisms to reward or penalise the company for reaching or
failing to reach the targets set for each objective.

4.5 Strengths and weaknesses of incentive regulation

As noted above, the main advantages to incentive-based regulation are related to the
provision of clear and simple incentives for efficiency through cost reductions. In
addition, the amount of information required from the companies and the cost of
regulation itself are lower than under traditional cost-of-service regulation with frequent
price reviews (Joskow and Schmalensee, 1986). Incentive regulation has also proved to
be very useful in countries with scantly developed auditing systems, where State-owned
companies were divided and privatised. In such cases, incentive regulation had to strike
a balance between company and consumer interests under conditions in which the
information available was incomplete.
The implementation of incentive-based regulation schemes also has its weak points,
however. The first is the potential for a decline in quality of service. The incentives for
companies to lower costs may have an adverse effect on service quality. Facility
maintenance and investment costs are directly related to the quality of the service
delivered. Consequently, the regulator must necessarily set both quality standards and
financial penalties for the failure to meet them.
Excessive concern over the profits that companies may be making may lead
regulators to gradually revert to cost-of-service regulation. Such a concern may induce
regulators to increase the frequency of rate cases to review costs, ultimately slipping back
into traditional regulation and losing the potential benefits of cost reductions.
Aiming to achieve a fair profit-sharing between customers and the regulated company,
the key aspects that the regulator should resolve in each price control review are the initial
price or initial revenue and the choice of both the X factor and the inflation index, which
may include a mix of price indicators. All these decisions impact the delicate balance
between profit sharing and achieving efficiency and stability in the medium and long
term.
Since incentive regulation may lead to laxer cost supervision by the regulator,
companies may tend to shift the costs incurred in the non-regulated line of business to
their regulated activity. This could happen, for example, in an attempt to show that their
profits are falling and thereby obtain a larger tariff hike than would be strictly necessary.
Additionally, oftentimes the requirements for CAPEX are less strict than OPEX, due to
the difficulties in investment assessment and fears for underinvestment. This introduces
the risk of operating costs being capitalised through investment (NAO 2002).
Lastly, in schemes that allow companies to establish the end user tariffs, it may be
necessary to define mechanisms to prevent shifting costs to consumers with fewer options
or influence over the company, e.g. from industrial to residential consumers.

5 Implementation details for price or revenue caps

This section contains a practical description of some of the notions and issues
mentioned above that should be considered by regulators during price control reviews
when setting caps for the next regulatory period.

-25/42-
5.1 Regulatory asset base, investments and depreciation
The determination of the RAB is crucial in the tariff revision process. As mentioned
earlier, the RAB is updated every year by adding investment and subtracting depreciation.
In this regard, a key issue is how to determine of the RAB at the beginning of the
regulatory period and the criteria for inclusion of new assets in the RAB.
The RAB for existing facilities or opening RAB
The first step in RAB assessment is deciding whether the RAB can be consolidated
according to depreciation and new investment allowances, or to re-open it. The latter
would be necessary when a comprehensive method, such as the one described hereinafter,
is implemented for the first time. Additionally, this re-evaluation is typically done during
deregulation and restructuring processes entailing, for example, the separation of
regulated and competitive activities, the unbundling and privatisation of State-owned
companies or the unbundling of vertically integrated enterprises into separate business
units. The most common methodologies to re-assess the RAB have been described in
detail in section 3.2.
Inclusion of new assets in the RAB
Incremental assets are usually driven by three main reasons: 1) deliver electricity to
new customers or to meet growing demand, 2) replace aged, deteriorated or obsolete
facilities, and 3) improve quality of service or comply with new legal or environmental
requirements.
OPEX evaluation and regulation under incentive regulation is relatively
straightforward. However, determining ex-ante efficient investment is less obvious and
controversial due to their long-lives and information asymmetries. Utilities normally have
no lack of estimates, forecasts and projections on the amounts and characteristics of the
infrastructure that will be needed to meet the needs created by expected market growth
under different scenarios. Companies have an obvious incentive to over-estimate
requirements to raise their RAB and consequently their allowed revenues. Therefore,
company proposals must be critically re-assessed by the regulator, drawing from the
necessary expertise, including both asset needs and unitary costs.
Assessing the efficiency on investments actually carried out on an ex-post basis is not
easy either. Two situations may arise during the regulatory period: 1) the company invests
less than initially estimated, consequently reaping higher profits, or 2) it invests more
than planned, in which case it makes a loss.
The challenge is to be able to distinguish between the intentional deferral of necessary
investments and genuine cost reductions. This is particularly hard in sectors where
regulators who are not in a position to monitor each new facility individually, such as
distribution grids. Under these conditions, sliding-scale or profit-sharing mechanisms
(see section 4.3) are particularly useful. Conversely, when significant individualised
investment, such as in transmission assets, for instance, is not undertaken within the
expected time frame, some regulators have implemented an ex-post review known as the
trigger approach. The company is allowed extra revenues when the investment becomes
operational or is penalised when it is delayed with respect to when the asset in question
was expected to come on stream (Alexander and Harris, 2005).
Therefore, in practice it is frequent to adopt an ex-ante/ex-post approach. See
(Alexander and Harris, 2005) for other alternatives. Under this approach the assets
approved in the ex-post review are the ones that replace the assets included in the RAB

-26/42-
ex-ante. The same two situations as described above are possible here: 1) the actual
investment made by the company is smaller than initially included in the RAB, in which
case the most straightforward review is to include actual investment in the RAB; or 2) the
company invests more than initially forecast in the RAB, in which case the ex-post review
must include a detailed investigation into whether the investment decisions were adopted
prudently. If that investigation deems some of the investments to be unnecessary, the
respective assets may either not be included in the RAB or deemed to be eligible for
depreciation expenses but not rate of return. The investments found to be necessary and
efficient, by contrast, are included in the RAB for full recovery (Alexander and Harris,
2005).
Another issue that must be addressed is ex post review timing. Three options can be
considered:
1) At the end of the regulatory period: assuming regulatory periods of five years, the
incentive for efficient investment would be 5 years for facilities coming on stream
in the first year of the regulatory period and 1 year for facilities commissioned in
the final year.
2) At the end of the next regulatory period: the incentive would range from 10 to 6
years.
3) On a rolling basis: a 5-year incentive would be established for any investment
irrespective of the year when it is actually made
The longer the time lapsing between the inclusion of an asset in the RAB and the ex
post review, the greater is companies’ incentive to lower actual investment costs.
Consequently, the second of the three options listed generates the most powerful
incentive, followed by option 3. See Alexander and Harris (2005) for numerical examples
and a fuller discussion of other alternatives.
Depreciation
Assets may be depreciated by a number of methods4 (Green and Rodriguez-Pardina,
1999). The two most common ones are defined below:
 Annuity method: a flat annual allowance is set for the whole asset life. Annual
depreciation is computed so that the sum of amortisation plus return on the
remaining capital is constant.
 Straight line method: yearly depreciation expense remains constant throughout the
life of the facility, and therefore the total annual charge, including the return on
investment, declines over time. This is the most common approach in regulation.
These two methods are illustrated in Table 5-1.
RAB and DEPRECIATION (Annuity) Present value Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Year 7
Regulatory Asset Base 100,00 89,46 77,86 65,11 51,08 35,65 18,67
Depreciation 10,54 11,59 12,75 14,03 15,43 16,98 18,67
Return on assets (WACC*RAB) 10,00 8,95 7,79 6,51 5,11 3,56 1,87
Depreciation + return on assets 100,00 20,54 20,54 20,54 20,54 20,54 20,54 20,54
WACC 10%

RAB and DEPRECIATION (Linear) Present value Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Year 7
Regulatory Asset Base 100,00 85,71 71,43 57,14 42,86 28,57 14,29
Depreciation 14,29 14,29 14,29 14,29 14,29 14,29 14,29
Return on assets (WACC*RAB) 10,00 8,57 7,14 5,71 4,29 2,86 1,43
Depreciation + return on assets 100,00 24,29 22,86 21,43 20,00 18,57 17,14 15,71
WACC 10%

4 Note that regulatory depreciation and the regulatory life of assets considered do not necessarily

match the actual accounting rules followed by the firms or the actual useful life of assets.

-27/42-
 Table 5-1: Depreciation methods: 10 % rate of return over 7 years. Source: adapted
from (Green and Rodriguez-Pardina, 1999)

If correctly applied, the two methods yield the same present value of revenues to be
received by the regulated company (see Table 5-1), although the impact on present and
future customers differs. Straight line depreciation would require present customers to
pay more than future customers, while under constant annuity arrangements the two
groups of customers would be equally impacted.
Moreover, accelerated depreciation practices are also common. In accelerated
depreciation either the expense is higher in the early years or the depreciation period is
shortened. Like the other methods, accelerated depreciation would not affect the present
value of consumer repayments, although today’s customers would pay more than
tomorrow’s (see Table 5-2). This can be done in order to encourage more investments
through accelerated depreciation or solve cliff-edge situations where significant
investments may be necessary but their financing complicated.
RAB and DEPRECIATION (Accelerated) Present value Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Year 7
Regulatory Asset Base 100,00 77,14 57,14 40,00 25,71 14,29 5,71
Depreciation 22,86 20,00 17,14 14,29 11,43 8,57 5,71
Return on assets (WACC*RAB) 10,00 7,71 5,71 4,00 2,57 1,43 0,57
Depreciation + return on assets 100,00 32,86 27,71 22,86 18,29 14,00 10,00 6,29
WACC 10%

Table 5-2: Accelerated depreciation: 10 % rate of return over 7 years

Moreover, depreciation timing also affects the amount of corporation tax paid by the
company, for as a deductible expense, it lowers the net operating income on which the
tax liability is computed (see section 5.5). With accelerated depreciation the company
could pay less tax in the early years but more in the final years of the depreciation period.
In such cases the regulator should determine how to pass tax savings on to customers
(Green and Rodriguez-Pardina, 1999).

5.2 Operating costs

An estimate of the company’s efficient OPEX is needed to establish the price or
revenue path for the next regulatory period. The starting point is generally the audited
accounts of costs incurred in the preceding period and a business plan furnished by the
company with projections for all the years in the next regulatory period.
The company must also break down this information as much as possible to show the
different cost items by: activities (facility maintenance, delivery of supply to new users,
repair of equipment and facilities and new infrastructure); categories (labour, materials,
office material and expendables, energy consumption); type of consumer by service area
(residential, commercial, industrial, street lighting). Regulatory accounting systems
usually determine the cost breakdown deemed necessary by the regulator.
The problem faced by the regulator is how to define a feasible operating cost objective
for the period. This target must be able to serve as an incentive for efficiency and the
reduction of present costs, as well as to maintain company medium- and long-term
sustainability; i.e., the level of efficient costs the company should strive to attain. Once
the regulator somehow determines the operating cost objective for the regulatory period,
where growth in demand or in number of customers have been already taken into
consideration, the result can be expressed by applying a productivity factor, X, to these
costs throughout the regulatory period:

-28/42-
OPEX t  OPEX t 1  (1  RPIt  X OPEX ) (5-1)

Where:
OPEXt Allowed operating costs in year t
RPIt Inflation rate per unit in year t
XOPEX Productivity factor for the allowed operating costs
An additional term may be included in the equation (5-1) in order to take into
consideration variations in operating costs induced by market growth (number of
consumers or energy delivered).

5.3 Regulatory benchmarking

Regulatory benchmarking consists in developing certain measurements, typically
related to costs or other variables affected by the performance of the company, against
which the behaviour of actual firms can be compared. If the benchmark is appropriately
obtained, these techniques can be used to encourage regulated firms to become more
efficient over time. Moreover, if the standards used are common for all the firms carrying
out a regulated activity, benchmarking can be used to introduce some kind of competition
among these companies, which would not naturally appear in the market.
Over the years, regulatory benchmarking has gained greater importance in incentive
regulation schemes (Jamasb et al., 2003). Nowadays, many distinct benchmarking
methods can be found, ranging from very simple approaches based on comparisons of
cost ratios to sophisticated approaches relying on elaborate econometric or engineering
models. Benchmarking techniques may set comparisons among regulated companies in
the same sector or against a reference (sometimes hypothetical) company. A review of
the most common benchmarking techniques is presented in Box 5-1.
The most common applications of regulatory benchmarking could be summarized as
follows: assess OPEX efficiency under the building blocks approach, estimate efficient
total costs under the TOTEX approach or evaluating investment plans submitted by
regulated firms.

Box 5-1. Benchmarking techniques

Broadly speaking, a company is more productive or more efficient than others if it
requires fewer inputs to attain the same outputs, or if it produces more outputs with the
same inputs. When a company is producing at its highest ideal productivity it is said to
be at the productivity frontier (Coelli et al, 1998). Reviews on the benchmarking
techniques used by regulators to assess network monopoly efficiency can be found in
(Jamasb and Pollit, 2003), (Ajodhia, 2005) and (Cossent, 2013).

-29/42-
 Figure 5-1: Productive efficiency and productivity frontier. Source: (Ajodhia, 2005)

Figure 5-1 shows inputs and outputs for three companies. Companies B and C are
efficient because they have already reached their productivity frontier. Company A,
however, has room to enhance its efficiency, to “catch-up” with the others: its input is
the same as B’s but its output lower, while its output is the same as C’s but its input
higher. Companies that have already reached their productivity frontier can raise their
long-term productivity by adopting newer technologies or innovative processes. This
is what is known as the “frontier shift”. While productivity calculations should ideally
take both effects into consideration, regulators often focus on benchmarking current
best practice. The distance between a company’s current productivity and its projection
on the frontier is a measure of its inefficiency. The higher the inefficiency, the higher
is productivity factor X (or the lower the initial revenue allowance).
As shown in Figure 5-2, regulators have a number of different methods or
benchmarking techniques from which to choose. These methods involve applying
statistical techniques to compare the efficiency of different companies providing the
same service in the same or similar countries: electric power distribution or
transmission companies, for instance. Correlation analyses are run to compare
individual cost items in the various markets. The results of such analyses can be used
to define an average efficiency pattern or identify the most efficient companies (best
practice) as models that others should emulate. Benchmarking requires a substantial
amount of duly validated information.

-30/42-
 Figure 5-2: Classification of benchmarking techniques. Adapted from (Cossent,
2013)

According to the previous taxonomy, there are two main categories. On the one
hand, reference methods build a bottom-up benchmark based on an in-depth
knowledge of the functioning of the sector being subject to regulation. These rely on
either engineering simulation/optimization models or other non-formalised analyses of
the firms’ costs and its comparison with other firms operating in similar conditions On
the other hand, black-box methods see regulated firms as black boxes for which only
inputs and outputs can be observed and measured. The benchmark is usually
constructed through econometrics or operations research approaches.
One traditional, simple and practical method for comparing efficiency among
companies in the same industry is based on the calculation of certain partial cost ratios.
Some typical ratios could be the costs per unit of energy delivered or the number of
customers served per employee. Nonetheless, these simple ratios neglect the input-
output relationships process cannot be reflected. Thus, the total factor productivity
(TFP) generates a ratio based on the weighted sums of all outputs and inputs.
Frontier methods are the most popular ones. These are analytical techniques for
finding the optimal weighting with which to combine outputs and inputs.
Data envelopment analysis (DEA) is a non-parametric technique, i.e. a pre-defined
functional relationship between outputs and inputs is not required. Each company’s
efficiency factor is calculated by solving a linear optimisation problem. The efficiency
factor calculated lies between 0 and 1. Companies with an efficiency factor of 1 have
reached the productivity frontier which is defined as the linear combinations of the
most efficient companies in the group. Figure 5-3 displays a DEA analysis for four
companies considering two inputs (X1 and X2) and output (Y). Firms 1 and 3 determine
the productivity frontier, whereas the efficiency factor for company 2 would be
measured as the distance OA divided by the distance O2.

X2/Y

2
0.5

0.4

1
0.3
A
4
0.2
3

B
0.1

0 700 1000 1300 1600 1800 X1/Y

/
Figure 5-3: Data envelopment analysis: two inputs, one output, four companies

Parametric methods require an assumption about the form of production/cost

functions. Figure 5-4 depicts production costs (Y) vs outputs (X) for several companies.
The average cost pattern is obtained by fitting a regression line to the points using the
ordinary least squares (OLS) method. The corrected ordinary least squares method

-31/42-
 (COLS) defines the efficiency frontier in terms of the performance of the most efficient
company. For company B, the efficiency factor is calculated as the distance OB’ over
the distance OB.

Figure 5-4: Ordinary least squares (OLS) and corrected ordinary least squares
(COLS)

Stochastic frontier analysis (SFA) is similar to the COLS technique, except that it
takes stochastic measurement errors into consideration when estimating the
productivity frontier and efficiency factors. Each firm’s distance to the frontier is
explained here as the sum of a symmetrical error term (associated with relative
efficiency) and a random error term to account for noise in the observations. Known
probability functions for the distribution of those errors must be assumed (Jamasb and
Pollit, 2003). This provides more robust results against outliers.
Among reference methods, norm or reference model method (Mateo et al., 2010)
may be highlighted. These models resort to engineering optimization/simulation
models to build an ideal or reference network that complies with similar constraints as
actual grids. These techniques are particularly suited when the number of companies
is very low or exogenous factors hinder inter-company comparisons.
The different approaches present their own pros and cons. Nonetheless, it must be
borne in mind that benchmarking results should not be directly translated into revenue
allowances. Nonetheless, they can be used as a useful tool for decision making.
Moreover, how to apply benchmarking (e.g. selection of input/output variables) can be
as important as the selection of the benchmarking model.

5.4 Present value of costs and revenues: the smoothing X

This section relies on the concept of present value of costs and revenues, i.e., the
change in value of money with time. In rate cases, present value calculations are used to
determine the revenues required to cover the utility’s expected costs (including a suitable
rate of return on assets).

-32/42-
 Under cost-of-service regulation, each year’s revenues should equal the sum of the
annual operating costs, the rate of return on net assets during the year and the annual
depreciation on gross assets. Assuming that both costs and revenues for the year
materialise at year end, all these amounts should be discounted at the beginning of the
period to calculate the present net value. Net assets, in turn, should be re-calculated each
year by adding the new assets resulting from investment made during the year and
subtracting the annual depreciation on gross assets in service.
Under price or revenue cap regulation over a regulatory period of several years,
regulators may distribute the utility’s revenues in different ways in each year, as long as
its present value for the entire price control period is maintained. If the revenue
allowances have already been subject to an efficiency assessment, incentives are already
embedded within allowances. Therefore, the X factor can be used to smooth the revenues
along the regulatory period rather than as an efficiency requirement5. Thus, the X factor
can be computed in such a way the net present value of non-smoothed allowed costs equal
the net present value of the smoothed ex-ante revenues. This will result in balanced cash
flows for DSOs and tariff stability. The process is illustrated in Table 5-3.
REGULATORY ASSET BASE Year 0 Year 1 Year 2 Year 3 Year 4 Year 5
Opening RAB 450,00 467,50 473,00 481,25 489,15
Gross capital expenditure - Investment 50,00 40,00 45,00 47,00 50,00
Regulatory depreciation 32,50 34,50 36,75 39,10 41,60
Closing RAB 450,00 467,50 473,00 481,25 489,15 497,55

Average RAB 458,75 470,25 477,125 485,2 493,35

WACC (real pre-tax) 8% 8% 8% 8% 8%

REGULATORY DEPRECIATION Life Year 1 Year 2 Year 3 Year 4 Year 5

Existing assets (15 years) 15 30,00 30,00 30,00 30,00 30,00
New investment in year 1 (20 years) 20 2,50 2,50 2,50 2,50 2,50
New investment in year 2 (20 years) 20 2,00 2,00 2,00 2,00
New investment in year 3 (20 years) 20 2,25 2,25 2,25
New investment in year 4 (20 years) 20 2,35 2,35
New investment in year 5 (20 years) 20 2,50

Total regulatory depreciation 32,50 34,50 36,75 39,10 41,60

OPERATIONAL COSTS (OPEX) Year 1 Year 2 Year 3 Year 4 Year 5

OPEX 60,00 57,90 55,01 51,98 49,90

X factor for OPEX (%) 3,5% 5,0% 5,5% 4,0%

ALLOWED COSTS Year 1 Year 2 Year 3 Year 4 Year 5

Return on assets (WACC * RAB) 36,70 37,62 38,17 38,82 39,47
Depreciation 32,50 34,50 36,75 39,10 41,60
Operational costs (OPEX) 60,00 57,90 55,01 51,98 49,90

Total allowed costs (Revenue requirement) 129,20 130,02 129,93 129,90 130,97

SMOOTHED REVENUES Year 0 Year 1 Year 2 Year 3 Year 4 Year 5

Quantities 100 100 100 100 100
Average Prices (Revenue/Quantities) 1,320 1,313 1,306 1,298 1,291 1,284

Smoothed revenues 131,27 130,55 129,84 129,12 128,41

PRESENT VALUE ANALYSIS

PV (Total allowed costs) (WACC is discount factor) 518,85
PV (Smoothed revenues) (WACC is discount factor) 518,85
PV difference 0,00
Calculated X-factor (%) 0,55%

Table 5-3: Calculation of the X factor in a revenue yield cap scheme

5 In fact, the conventional interpretation of the X factor as an efficiency gap that ought to be reduced
is difficult to implement in practice, especially when technologies change or significant investments are
necessary.

-33/42-
 First the regulatory asset base (RAB) is updated yearly taking into account allowed
investments and depreciation. All the figures in red are input data. Next, the rate of return
or WACC is set by the regulator. The WACC would be calculated as explained in section
3.3 and applied as a discount rate. Note that the rate of return is expressed in real values,
i.e. no explicit provision for inflation is taken into account in this exercise, and the WACC
is a before tax figure, i.e., the revenues that the company would obtain before paying taxes
are taken into consideration. This subject is explained in greater detail in section 5.5. In
this simplified example the demand has been assumed to be constant during the entire
price control period.
The regulator then sets the allowed operational costs for each year of the regulatory
period. In our example these costs decline over time due to the regulator’s efficiency
requirements, i.e. an ad hoc productivity factor for OPEX that is estimated by the
regulator6, and are also expressed in real terms, i.e., net of inflation.
Each year’s total allowed costs are the sum of the return on assets, plus yearly
depreciation, plus the allowed OPEX. In this exercise, the smoothed revenue
requirements are calculated in such a way that the average price in each year, revenue Rt,
divided by the estimated quantities delivered, Dt, can be expressed as shown below, where
the global smoothing factor X is given in percentage and the initial value in year 0, R0/D0,
is known:
Rt Rt 1  X 
  1   (5-2)
Dt Dt 1  100 
This type of cap was referred to as the “revenue yield” cap in section 4.1. The inflation
rate was not included since this exercise was performed with real costs and prices.
Nonetheless, including inflation in all the cost items is straightforward.
The X factor is calculated with an Excel spreadsheet using the “goal seek” tool, by
equating the present value of costs to the present value of revenues and taking the WACC
set by the regulator as the discount factor. The computed X can be considered to have a
double meaning. On the one hand, it is the adjustment factor that makes possible that the
net present values of costs and revenues are equal; on the other hand, X can be seen as a
global productivity factor7.
Note that, while OPEX declines by around 4.5 % yearly, the CAPEX rises in keeping
with the investment pattern approved by the regulator, causing the RAB to climb
progressively. As a result, the total productivity factor, at 0.55 %, is much lower than
4.5 % but still positive, signifying a reduction in average prices in real terms.
One final observation that should not be overlooked is that sales or the quantities
delivered by the company remain flat throughout the regulatory period. In other words,
projected investment and operating costs are calculated on the assumption of constant
sales. In revenue yield schemes, revenues rise with sales. Variable revenue cap formulas
in which economies of scale are taken into consideration call for explicitly calculating the
rise in investment and operating costs with increasing sales volumes or selected cost

6 Note that this parameter is different from the global smoothing X factor in equation (5-2), which

is the outcome of table 5-3.

7 Nevertheless, note that under this approach, the X factor may even be negative without implying

that the firm is inefficient.

-34/42-
drivers. In other cases the regulator makes direct estimations of the diverse costs for each
year of the regulatory period without recurring to prescribed formulas.
A critical issue when designing price or revenue caps is how to connect prices or
revenues between two consecutive regulatory periods. Both the X factor and the initial
revenues can be used by the regulator to determine different revenue paths so as to control
cash flows or the variation of network tariffs between consecutive regulatory periods
(Green and Rodríguez-Pardina, 1999). When it is desired to mitigate the price changes
between regulatory periods, the initial revenues should be consistent with the
remuneration in the last year of the previous regulatory period and set the X factor as
required to meet the expected future costs (left part of Figure 5-5). On the contrary, if the
regulator prefers to reduce the gap between remuneration and actual costs much faster,
the alternative shown on the right of Figure 5-5 ought to be adopted. Note that this choice
also affects the period during grid regulated firms retain efficiency gains.

Figure 5-5: Smoothed revenue path vs. one-off price adjustment (Green and
Rodríguez-Pardina, 1999)

5.5 Tax, cash flow and profitability

Utilities must pay a series of duties or fees and taxes levied on their earnings that affect
their cash flow and business profitability. Regulators must take such expenses into
account in the tariff revision process when calculating the rate of return and costs that the
company is allowed to recover.
The charges to be paid by a utility are associated with the type of business conducted.
Distributors must pay a series of local duties, including: municipal tax, property tax,
chamber of commerce dues, business licence fees and other local charges. As tax rates
cannot be controlled by the company, they should be taken into account when calculating
its allowed revenues. Moreover, companies must pay the corporate tax, which for instance
can amount to 30 - 35 % of the earnings before-tax. The effect of taxes on the calculation
of the allowed WACC was explained in section 3.3.
The following remarks are relevant to determine how taxes affect utilities:
 Like operating expenses, duties (including local taxes and similar) must be paid out
of allowed revenues. They differ from operating expenses, however, in that the
company has no power to reduce or otherwise control them.
 Taxes have a direct effect on the company’s earnings. Depreciation policy affects
its tax liability. Depreciating more in the early stages, for instance, implies smaller
earnings in those initial years, lower taxes, and consequently higher cash flows.
 Allowed rate of return calculations must be consistent with allowed company
earnings calculations. For example, if the WACC defined by the regulator refers
explicitly to before-tax income, to be comparable, the company’s actual rate of
return must be calculated as before tax earnings divided by the rate base. On the

-35/42-
 contrary, if the rate of return established by the regulator refers to the after-tax
results, the actual rate of return must be calculated by dividing net earnings (i.e.,
net of taxes) by the rate base.

The example presented in Table 5-4 illustrates how a regulated company’s actual
profitability is calculated for a given period when allowed revenues are set purely ex-ante
by the regulator. The same example presented in Table 5-3 has been used. It has been
assumed that the performance throughout the 5-year regulatory period is exactly as
predicted by the regulator. In this example, as earlier, the values shown are net of inflation
and company sales are flat throughout the period.
ACTUAL ASSET BASE & INVESTMENT Year 0 Year 1 Year 2 Year 3 Year 4 Year 5
Opening AB 450,00 467,50 473,00 481,25 489,15
Gross capital expenditure - Investment 50,00 40,00 45,00 47,00 50,00
Depreciation 32,50 34,50 36,75 39,10 41,60
Closing AB 450,00 467,50 473,00 481,25 489,15 497,55

Average AB 458,75 470,25 477,125 485,2 493,35

ACTUAL DEPRECIATION Life Year 1 Year 2 Year 3 Year 4 Year 5

Existing assets (15 years) 15 30,00 30,00 30,00 30,00 30,00
New investment in year 1 (20 years) 20 2,50 2,50 2,50 2,50 2,50
New investment in year 2 (20 years) 20 2,00 2,00 2,00 2,00
New investment in year 3 (20 years) 20 2,25 2,25 2,25
New investment in year 4 (20 years) 20 2,35 2,35
New investment in year 5 (20 years) 20 2,50

Total depreciation 32,50 34,50 36,75 39,10 41,60

FINANCIAL STRUCTURE & COST OF CAPITAL Year 1 Year 2 Year 3 Year 4 Year 5
Equity (%) 60% 60% 60% 60% 60%
Debt (%) 40% 40% 40% 40% 40%
Rate of return applied to equity (post-tax) 7,0% 7,0% 7,0% 7,0% 7,0%
Rate of return applied to equity (pre-tax) 10,8% 10,8% 10,8% 10,8% 10,8%
Rate of interest on debt (Rdebt) 3,8% 3,8% 3,8% 3,8% 3,8%
Tax rate on benefits 35% 35% 35% 35% 35%
WACC (pre-tax) 8,0% 8,0% 8,0% 8,0% 8,0%

ACTUAL OPERATIONAL COSTS (OPEX) Year 1 Year 2 Year 3 Year 4 Year 5

Actual OPEX 60,00 57,90 55,01 51,98 49,90

ACTUAL INCOMES SET BY THE REVENUE YIELD CAP Year 1 Year 2 Year 3 Year 4 Year 5
Actual revenues 131,27 130,55 129,84 129,12 128,41

BALANCE SHEET Year 1 Year 2 Year 3 Year 4 Year 5

Incomes 131,27 130,55 129,84 129,12 128,41
Accounting costs 99,38 99,45 98,91 98,36 98,90
Operational costs (OPEX) 60,00 57,90 55,01 51,98 49,90
Depreciation 32,50 34,50 36,75 39,10 41,60
Debt payment (AB * Debt(%) * Rdebt) 6,88 7,05 7,16 7,28 7,40
Benefit before taxes 31,89 31,10 30,92 30,76 29,51
Taxes (Benefit * Tax rate) 11,16 10,88 10,82 10,77 10,33
Benefit after taxes 20,73 20,21 20,10 20,00 19,18

ACTUAL REMUNERATION OF EQUITY Present value Year 1 Year 2 Year 3 Year 4 Year 5
Actual equity (AB * Equity (%)) 1.169,33 275,25 282,15 286,28 291,12 296,01
Actual benefit after taxes 82,29 20,73 20,21 20,10 20,00 19,18
Actual rate of return on equity after taxes (%) 7,0%

Make difference PV(equity)*ror - PV(benefit) = 0 0,00

Table 5-4: Actual rate of return of a regulated company under revenue-yield cap

It can be seen that company investment and OPEX are as predicted (see Table 5-3),
and its financial structure is divided into 60 % equity and 40 % debt. Note that the 8 %
before tax WACC is computed based on that 60/40 structure, the 7 % rate of return on
equity, the 3.8 % interest rate on debt and 35 % corporate tax. The before tax WACC is
the one considered by the regulator in the ex-ante calculations. Net income is calculated
as revenues minus operating expenditure, depreciation, and interest payments. Lastly, the

-36/42-
actual rate of return on equity is checked to verify that it concurs with the value allowed
by the regulator, i.e. 7 %.
In the example, the actual rate of return, found as the present value of the after tax
profit divided by the shareholders’ equity, proves to be equal to 7 %, the value assumed
for calculating the 8 % before tax WACC. Note that the present value is calculated by
discounting at the unknown actual rate of return, using the Excel “goal seek” tool.

5.6 Consumers or companies risks

A key issue in implementing revenue or price caps is to analyse how inflation
variations, sales growth or recession would affect regulated revenues and company costs,
and who, consumers or companies, would bear the risk of such deviations.
Under traditional cost-of-service regulation, the prices or tariffs set during the rate case
process remain in effect until the next rate case is studied one or two years later. In this
approach, the company’s revenues obviously rise or fall in proportion to sales. If costs
rise faster than sales, because inflation is high, for instance, or sales are lower than
expected, the company bears all the risk (unless specific corrections are made), and
applies for a tariff revision. Conversely, when costs grow below expectations and/or sales
are growing faster than initially predicted, the company benefits.
Viewed from the perspective of incentive-based regulation, price cap regulation is
equivalent to cost-of-service regulation as far as the risks of variations in demand or sales
are concerned, but with a heavier economic impact because the regulatory period lasts for
several years. Under price cap arrangements, however, the general effect of inflation on
costs and revenues can be handled directly with the cap formula. In some cases a weighted
average of price indicators, i.e., a mix of general consumer price and specific industry
indexes, is adopted. Although in general inflation may not correspond to the actual
evolution of the cost of the utility.
In revenue cap regulation, revenues can be made to grow by raising only the value of
the market variables that have a direct effect on cost increases. One example would be a
revenue cap formula for a distributor that takes account of larger revenues collected as a
result of a rise in the number of consumers, but not as a result of higher demand for energy
by existing customers. From a practical standpoint this type of revenue cap for distributors
is consistent with the implementation of energy efficiency and savings programmes or
the introduction of distributed generation under net-metering.
In brief, with revenue caps (in the absence of ex-post adjustments) consumers may
bear all the risk associated with demand variations, while with price caps that risk is
transferred to the company. This is relevant because differences always arise between the
expected sales taken as the grounds for the revision process and the company’s real sales
during the period the revenue cap is in effect. Moreover, companies may have incentives
to manipulate the sales estimates submitted to regulators in order to influence them in
their favour, e.g. to increase investment allowances or to raise prices.
In revenue cap regulation revenues can be adjusted ex-post according to revenue
deviations resulting from differences between ex-ante predictions and the actual value of
variables beyond the company’s control, such as demand growth or inflation forecasts.
The ex-post approaches to acknowledge investments and their inclusion in the RAB raise
the risk assumed by the company, which is uncertain whether the investment cost will be
acknowledged by the regulator. The ex-ante approach, by contrast, affords companies
greater certainty, transferring risk to consumers, who would pay or benefit for any

-37/42-
deviations from actual infrastructure needs. Some mechanisms combining both
approaches, which may mitigate these problems, have been previously discussed.

5.7 Quality of service and other issues

As discussed in the foregoing, under cost-of-service regulation companies may have
an incentive to over-invest to achieve, for instance, self-imposed quality of service levels
primarily on the grounds of technical considerations or with any other excuse, simply
because the capital invested earns a generous rate of return. Under incentive-based
regulation the quality of service situation is exactly the opposite. Regulatory incentives
to lower costs and enhance efficiency can lead to deterioration of the quality of service
delivered. Quite obviously, both investment in infrastructure and operating costs for
maintenance and repair in the event of failures have a direct impact on quality.
In incentive-based regulation, performance parameters or indicators that the company
must meet should be defined, along with penalties that reduce its revenues when it fails
to do so. This is the alternative provided by incentive- or performance-based regulation
to solve the difficult problem discussed earlier in connection with cost-of-service
regulation. The solution consists not of monitoring each and every company investment
to determine its technical and economic justification, but rather of monitoring the results
delivered by the company in terms of both total cost and quality of supply.
Consequently, the regulator must establish a scheme to measure and monitor the
company’s performance indicators, i.e., the quality of service offered. Electricity
distribution monitoring, for instance, might involve verifying factors such as service
restoration time after an outage, response time in meeting requests for new service
connections or support for customers filing claims or complaints.
One light-handed regulatory tool to encourage companies to improve quality of service
levels is public disclosure of their results. The advantage is that the regulator need not
define quality targets or the economic implications associated with failure to meet them.
Regulatory costs are therefore low. This type of regulation may be insufficient, however,
if specific problems need to be solved or when public opinion fails to influence utility
behaviour.
Other stronger mechanisms for regulating quality of service are based on penalties and
incentives.
Penalties are established when the company fails to comply with the individual quality
standards set by the regulator. Here, the company must pay a penalty to consumers for
poor quality service, which may be measured in terms of the number and severity of
supply outages in a year or failure to respond in a pre-established time to a request for a
new connection to the grid, for instance. Such penalty schemes are based on individual
user quality monitoring systems.
Furthermore, integrated price quality regulation has been implemented in several
countries to regulate network infrastructures such as electricity distribution. A utility’s
yearly revenues may be increased or reduced by a certain percentage, for example,
depending on the degree of compliance or non-compliance with the quality standards
established for given areas or for the company as a whole. In practice this entails inclusion
of a Q-factor or a quality incentive term in the RPI formula. In this case the regulator sets
target values for specific system level quality indices, such as customer minutes lost or
percentage of customers with a certain number of outages. The utility’s measuring and
monitoring systems must, moreover, be open to regulator audit.

-38/42-
6 Summary
This document discusses the fundamentals of and most common methods for
regulating monopolies, with specific reference to their implementation in connection with
network industries.
The essential ideas set throughout this document are summarised below:
 Economic regulation is necessary to prevent a monopolist service provider from
overcharging consumers and to encourage economic efficiency.
 The most traditional regulatory method is known as cost-of-service or rate-of-
return regulation. Every year or two, the regulator analyses the company’s costs,
assets and investments and establishes the new revenue requirement for the
following period. Such revenue requirement enables the company to cover its
operating costs and asset depreciation expenses as well as to earn a rate of return
on capital set by the regulator.
 Two main drawbacks can be identified in cost-of-service regulation: 1) it provides
no incentive for reducing costs, since the regulator tends to acknowledge all costs
incurred; and 2) the rate of return may constitute an incentive for companies to
invest more than is economically optimal.
 To mitigate such problems, incentive-based regulation methods are becoming
more popular. The chief characteristic of this approach is that allowed
tariffs/revenues are decoupled from costs for longer intervals, typically 4 or 5 years.
This provides an incentive for the company to lower its costs and be more efficient
than under cost-of-service regulation. The two most commonly used methods are
price caps and revenue caps.
 The main incentive regulation methods set prices or revenues that the company may
charge or receive throughout the regulatory period with a formula with a yearly
adjustment factor known as (RPI-X).
 The major difference between price and revenue caps is that under price caps, any
increase in sales leads to higher revenues; i.e., costs are assumed to increase
proportionally with sales. Under revenue caps, by contrast, revenues do not increase
proportionally with sales. Thus, revenue cap regulation is better suited to
remunerate firms characterised by fixed investment and whose costs do not largely
depend on the intensity of use of the resulting assets.
 Incentive-based regulation induces companies to lower costs. However, this may
lead them to do so at the expense of quality of service. Thus, firms’ performance
regulation usually goes hand in hand with incentive regulation. The most important
of these objectives are arguably the quality standards that the company must meet.
Regulators may establish system level quality targets to verify average company
performance or individual standards to guarantee a minimum quality to each
consumer. In both cases, when the company fails to comply with the standards its
allowed revenues decrease.
 In each rate case or price control procedure, regulators must project future
efficient costs with which to align allowed revenues. Some of the key elements of
this process are: i) determine the opening RAB, ii) determine investment
allowances, iii) update the RAB with depreciation and asset additions, iv) setting
the yearly level of allowed OPEX, v) establish the rate of return (WACC), vi)

-39/42-
acknowledge other non-controllable costs such as charges, duties and taxes, and
vi) calculate the smoothing X factor and annual allowed revenues for the regulatory
period.

-40/42-
 References
Ajodhia V. S. 2005. Regulating beyond price: Integrated price-quality regulation for electricity
distribution networks, PhD Thesis. Delft University of Technology.
Alexander I.; and C. Harris. 2005. The Regulation of Investment in Utilities: Concepts and
Applications The World Bank Working Paper No. 52. Washington, D.C.
Averch, H. and L. Johnson. 1962. Behavior of the firm under regulatory constraint. American
Economic Review 52(5): 1052-1069.
Coelli, T., D.S. Prasada Rao, and G.E. Battese. 1998. An introduction to efficiency and
productivity analysis, Kluwer Academic Publishers, Boston.
Comnes, A., S. Stoft, N. Greene, and L. Hill. 1995. Performance-based ratemaking for electric
utilities: A review of plans and analysis of economic and resource-planning issues, Vol. 1.
Lawrence Berkeley National Laboratory, Berkeley, California (November): LBNL-37577.
Cossent, R. 2013. Economic Regulation of Distribution System Operators and its Adaptation
to the penetration of Distributed Energy Resources and Smart Grid Technologies. PhD. Thesis,
Universidad Pontificia Comillas, Madrid.
Crouch, M. (2006). "Investment under RPI-X: Practical experience with an incentive
compatible approach in the GB electricity distribution sector." Utilities Policy 14(4): 240.
Demsetz H., 1968. Why Regulate Utilities. Journal Of Law & Economics 11(APR): 55-65.
DOE/FERC. 1973. Uniform Systems of Accounts Prescribed for Public Utilities and Licenses
Classes A, B, C and D. Federal Energy Regulatory Commision. D.O.E./FERC-0028, Washington.
Gordon, C. R. 1981. Reforming the Regulation of Electric Utilities. Priorities for the 1980.
MIT Energy Laboratory working Paper MIT-EL-81-033WP. June
Green R., and M. Rodriguez-Pardina. 1999. Resetting Price Controls for Privatized Utilities.
A Manual for Regulators. The World Bank. Washington, D.C. February.
Jamasb, T., P. Nillesen and M. Pollitt (2003). "Gaming the Regulator: A Survey." The
Electricity Journal 16(10): 68.
Jamasb T. and M. Pollitt. 2003. International benchmarking and regulation: an application to
European electricity distribution utilities. Energy Policy 31:1609-1622.
Joskow, P. and R. Schmalensee. 1986. Incentive regulation for electric utilities. Yale Journal
on Regulation 4(1): 1-49.
Joskow P. L. 2006. Incentive Regulation in Theory and Practice: Electricity Distribution and
Transmission Networks. Cambridge Working Papers in Economics CWPE0607, Electricity
Policy Research Group Working Paper EPRG 0511.
Joskow P. L.. 2007. Regulation of Natural Monopoly, Handbook of Law and Economics.
Elsevier.
Kahn A. E. 1988. The Economics of Regulation: Principles and Institutions. Cambridge: MIT
Press.
Laffont, J.J. and J. Tirole. 1993. A theory of incentives in procurement and regulation. MIT
Press.
Mateo Domingo, C., Gómez San Roman, T., Sánchez-Miralles, A., Peco González, J.P., and
Candela Martínez, A. 2010. "A Reference Network Model for Large-Scale Distribution Planning
With Automatic Street Map Generation." IEEE Transactions on Power Systems,
National Action Plan for Energy Efficiency (2007). Aligning Utility Incentives with
Investment in Energy Efficiency. Prepared by Val R. Jensen, ICF International.
www.epa.gov/eeactionplan

-41/42-
 National Audit Office (NAO). 2002. Pipes and Wires. Available on-line at
Navarro, P. 1996. Seven basic rules for the PBR regulator. The Electricity Journal 9(3):24-30.
OFGEM. 2004. Electricity Distribution Price Control Review. Final proposals, Office of Gas
and Electricity Markets. Available: www.ofgem.gov.uk
OFGEM 2007. Electricity Distribution Cost Review 2006-2007. 10 December 2007.
http://www.ofgem.gov.uk/Networks/ElecDist/PriceCntrls/CostRep/Documents1/Elec%20Dist%
20Cost%20Review%20200607%20ref%2028907.pdf
Petrov K., and N. Nunes. 2009. Analysis of insufficient regulatory incentives for investments
into electric networks. An update. Final report. KEMA Consulting GmbH submitted to European
Copper Institute. Available on-line at http://www.leonardo-energy.org/
Posner R. A. 1999. Natural monopoly and its regulation. 30th Anniversary edition. CATO
Institute. Washington, D.C.
Rothwell G., and T. Gómez. 2003. Electricity Economics: Regulation and Deregulation. IEEE-
Wiley Press.
Rudnick, H., A. Arnau, S. Mocarquer and E. Voscoboinik (2007). "Stimulating efficient
distribution." Power and Energy Magazine, IEEE 5(4): 50
Shleifer, A. 1985. A Theory of Yardstick Competition. Rand Journal of Economics, 16:314-
327.
Suelflow J. E. 1973. Public Utility Accounting: Theory and Application, Michigan State
University, Public Utilities Studies.
The Regulatory Assistance Project (RAP). 2000. Performance-based Regulation for
Distribution Utilities. Available on-line at www.raponline.org/Pubs/General
Viscusi W. K., J. E. Harrington, and J. M. Vernon. 2005. Economics of Regulation and
Antitrust, 4th ed. Cambridge: MIT Press.

-42/42-