Beyond market shares and cost plus pricing: designing a horizontal market power mitigation framework for today's electricity markets

Beyond market shares and cost plus pricing:
designing a horizontal market power mitigation
framework for today’s electricity markets



With the proposed interim guidelines for identifying competitive suppliers and
granting market-based rate authorizations, analysis of generation market power
has come to the forefront of policy development at the FERC and at other US
regulatory and market institutions. The design and selection of successful
monitoring programs, analytical screening tools, and mitigation procedures is
not straightforward, and this is reflected in experience to date. Most market
power regulatory initiatives have abandoned the three cornerstones of market
power analysis, skipping over the market definition stage, making short order of
the market power testing (sometimes using out-of-date approaches for the
deregulated power industry) and plunging heedlessly into cost-based market
mitigation. Thankfully, FERC’s two-part indicative screen is only an interim
solution – a number of important adjustments are necessary before we can say
that we have successfully addressed generation market power considerations in
today’s evolving competitive electricity markets. After outlining our
recommended three-stage framework for ex ante market power detection and
mitigation in wholesale electricity markets, the second section of this White
Paper analyzes the meaning of market power. The next sections then cover each
of the stages of analysis: defining the market, market power testing, and
designing mitigation rules.









August 2004

---

Table of Contents
AN APPROACH OF FIRST PRINCIPLES .............................................................................................................1
WHAT IS MARKET POWER IN A ‘WORKABLY COMPETITIVE’ MARKET? ...........................................4
STAGE 1: DEFINING THE MARKET....................................................................................................................6
GEOGRAPHICAL DIMENSION.......................................................................................................................................8
PRODUCT DIMENSION...............................................................................................................................................10
TIME DIMENSION......................................................................................................................................................13
FERC’S MARKET DEFINITION...................................................................................................................................13
EMPIRICAL APPROACHES TO MARKET DEFINITION....................................................................................................14
STAGE 2: CONSTRUCTING MARKET POWER TESTS .................................................................................25
PIVOTAL SUPPLIER TEST...........................................................................................................................................26
RESIDUAL DEMAND ANALYSIS .................................................................................................................................29
ANALYZING MARKET POWER IN THE CONTEXT OF NEW ENTRY.................................................................................33
CRITERIA FOR SELECTION OF MARKET POWER TESTS................................................................................................34
RECOMMENDED APPROACH FOR MARKET POWER TESTING.......................................................................................34
STAGE 3: MITIGATION MEASURES .................................................................................................................35
CRITERIA FOR WEIGHING MITIGATION OPTIONS........................................................................................................35
IN WHAT MARKET/MARKETS SHOULD MITIGATION MEASURES APPLY?.....................................................................36
THE COSTS AND BENEFITS OF INTERVENTION ...........................................................................................................38
RECOMMENDED MITIGATION MEASURES FOR WHOLESALE POWER MARKETS ...........................................................39









London Economics International LLC provides consulting services and expert testimony on economic, financial,
regulatory, and strategic issues to corporations, law firms, and public agencies world-wide from its Boston
headquarters, with affiliated offices in Mexico City, Melbourne, London and Paris. For further information on
London Economics’ electricity sector practice area or questions relating to market power advisory, please contact
Julia Frayer at (617) 494-8200 or julia@londoneconomics.com.

---

White Paper

An approach of first principles
We believe that it is important to start from first Key Building Blocks for a
principles by rigorously defining the market in which rigorous Market Power Analysis
there is a market power concern. That definition should
framework
cover three important dimensions: product, geography,
and time. Anti-trust and regulatory precedent (both 1. Define the market and
inside and outside the electricity industry, domestically
identify key attributes of the
and internationally) suggest a set of quantitative and
market (supply and demand)
qualitative approaches for doing this. Suitable 2. Test for potential market
quantitative techniques include price correlation
power or prima facie evidence
analysis, econometric and model-based studies for
of market power abuses
market boundaries including the widely used Granger
causality price test, as well as the FTC-practiced SSNIP
3. Propose mitigation
test. This could be further supported by industry
arrangements against those
experience — the commercial realities of how these
entities found to be capable
wholesale power markets operate. The selected
of market power (potential)
definition needs to be robust: application of mitigation
measures to an incorrectly defined market will, at best,
be ineffective, and at worst, damaging. Relying on “one size fits all” defaults is not sufficient in
today’s complex power markets.
The next step involves measurement of market power in the relevant market. It is of
fundamental importance to note that all market power measures will give misleading and
potentially harmful results if they are applied to improper market definitions - hence the
importance of the market definition stage. Market power measures can be applied ex ante,
designed to identify when there is potential market power, or ex post, to determine when
market power may have been misused in a relevant market. We focus here on ex ante measures
of market power. We do so mainly because of the adverse consequences of regulatory
intervention under an ex post framework. Moreover, ex ante measures are better suited and
more applicable for the analysis required by FERC in its market based rate authorizations, and
by other institutions in their analysis of market conditions and the competitiveness of spot
pricing.
There are a number of different measurement approaches that can be used. The most widely
used screen for market power is, perhaps, the Herfindahl-Hirschman Index (HHI) which is well
accepted and heavily relied upon in Department of Justice (DOJ) merger cases. Other
approaches include alternative concentration measures, pivotal supplier analysis, residual
demand analysis, tests based on the Lerner Index, and empirical model-based tests similar in
character to the SSNIP tests used to help define the market. While none of the measures is
definitive, most of them can be adapted to the specific circumstances of electricity market to
provide reliable results, so long as the tests account for the specific factors that may significantly

- 1 -
London Economics International LLC
350 Massachusetts Ave. Suite 330
Cambridge, MA 02139
www.londoneconomics.com

---

White Paper

influence firms’ behavior,1 such as the absolute and relative size of each supplier, the flexibility
of the supplier’s production capacity, and the supplier’s fixed commitments. Accordingly, we
recommend developing the pivotal supplier measure and its close analogue, residual demand
analysis, such that these indicators reflect and perform consistent with electricity market
fundamentals.
The third and last stage of a market power evaluation framework needs to address mitigation
measures. Effective means of intervention are needed in the event that there is evidence of
market power, and there are significant adverse consequences from its prospective misuse. In
the US, mitigation invariably is almost universally based on cost-based mitigation and price
caps, perhaps reflecting past regulatory practice in the US — let us do what we are used to!
There is much less reliance on these measures outside of North America. Indeed, cost based
mitigation is rare, and where used has been of doubtful benefit. Rather, the focus has been on
structural measures to mitigate market power, and transitional protections while the proper
market structures evolve. However, the choice of mitigation mechanism must reflect the reality
of market circumstance: the structural and transitional mechanisms adopted in, for example,
Australia, may not be realistic options in the US.
Neither of the two measures widely used in the US – cost plus and price caps - is ideal. Cost
based mitigation — in the sense of limiting the revenue of particular suppliers to reflect an
underlying measure of their costs — and price caps both present a high risk of delivering
inefficient outcomes in the long term. For example, cost-based mitigation of suppliers in
constrained areas of a transmission network sharply reduces incentives for other generators to
locate in those areas or for transmission investment, even though this is likely to be desirable in
the long term. Just as problematic, in a market that presents geographically differentiated
prices, cost-based mitigation can distort price differentials that are so important in determining
the need for transmission investment and the best locations for suppliers (and customers) to
locate. It is conceivable that cost-based measures applied in a selective and parsimonious
manner can promote a competitive fringe2 (though this is rarely the outcome).
Price caps also offer little prospect of efficient outcomes. The principal concern is that price caps
as applied in the US are invariably much lower than customers’ true willingness to pay for
power. As a result, market prices at times of shortage are destined to be inefficiently low (i.e.
they do not include an appropriate scarcity rent). Subsequently, a litany of other inefficient
practices emerges, perhaps the most damaging of which is that there is a reduction in the
incentives for power customers to hedge (contract) to provide price stability and protect against
price extremes. This, in turn, dramatically changes the conditions of market entry.
We therefore recommend designing a mitigation approach that, at the minimum, encapsulates
local scarcity rent. It is neither necessary nor efficient to apply a single mitigation approach to

1 This would include a discussion of the impact of market immaturity and market regulation itself on the nature of
and requirement for mitigation approaches.
2 Competitive fringe refers to all suppliers other than the mitigated supplier(s). They are denoted as competitive
based on the economic theory of the dominant firm and the competitive fringe. Please see discussion under
Residual Demand Analysis section for more detail.

- 2 -
London Economics International LLC
350 Massachusetts Ave. Suite 330
Cambridge, MA 02139
www.londoneconomics.com

---

White Paper

all suppliers in all markets. We recommend that mitigation measures are selective in
application, applying only to participants that demonstrably possess market power, and do not
necessarily extend to all suppliers in the relevant market. We further recommend defining
thresholds for each mitigation alternative based broadly on the costs of intervention versus its
likely effectiveness.

- 3 -
London Economics International LLC
350 Massachusetts Ave. Suite 330
Cambridge, MA 02139
www.londoneconomics.com

---

White Paper

What is market power in a ‘workably competitive’ market?
The classical definition of market power can be found in the US Department of Justice’s Merger
Guidelines – market power is the “the ability to profitably maintain prices above competitive levels for
a significant period of time”.3 In the wholesale electricity market4 as in other markets, a firm with
market power can achieve a sustained increase in profits by raising its own offer prices or by
withdrawing capacity from the market.5
There are a number of factors that predispose and mitigate market power in general, and
specifically in deregulated electricity markets in the US. Short-run inelastic demand, the
instantaneous nature of production, delivery, and consumption of electricity,6 potentially long
lead times for entry and the novelty of deregulated markets tend to predispose the possibility of
the existence of market power. Factors that tend to mitigate market power include the existence
of centralized market operators and exchanges, common carrier transmission and distribution
networks, and relatively simple access to fuel.
Markets, in general, do not always work efficiently in the sense participants sometime possess
market power and are able to sustain prices above efficient levels. The principal reason for
market failure is barriers to entry, such that new entrants are unable to enter the market to
discipline the behavior of participants with market power. 7
Due to the characteristics of the suppliers in this industry and the cost structure of supply,
certain conditions are not necessarily evidence of market power. For example, economic theory

3 U.S. Department of Justice and Federal Trade Commission, Horizontal Merger Guidelines § 0.1 (April 2, 1992)
(“Horizontal Merger Guidelines”).
4 When investigating market power, it is common to address the functional aspect of the market definition (e.g.
wholesale versus retail). While there may be market power issues in power retailing and in vertical relations
between wholesalers and retailers, the discussion in this policy paper focuses on the wholesale market —
the market into which generators sell their power.
5 The term ‘economic withholding’ is widely used by power market regulators and system operators. The term is not
widely found in the anti-trust literature, as it really connotes firms seeking to raise prices.
6 Our assumption here is that electricity cannot be stored in a commercially viable manner. The most common
electricity storage device is a battery; though commercial batteries neither last long enough nor provide
sufficient source of energy to be comparable to a power plant. Fuel cells, which are essentially large scale
batteries, have yet to be commercially competitive with traditional power generation. Another example of
storable electricity is pumped storage hydro plants. Aside from being highly costly to build, it is –for natural
reasons- also not possible to build pumped storage facilities in all markets or regions. Note that none of
these technologies store “electricity” per se. Rather, they store energy, which, once converted into electricity,
must be consumed.
7 There are other factors that may predispose a market to non-competitive actions such as high transaction costs. It is
beyond the scope of this document to characterize these, or to exhaustively review the question of barriers
to entry into electricity markets, except to note that there are substantial short-run barriers to entry, but
relatively modest barriers over the time period consistent with the permitting and construction time for new
capacity (as evidenced by the large amount of entry that has taken place in markets as diverse as ERCOT,
England and Wales and Southeast Australia — even California has seen substantial new entry).

- 4 -
London Economics International LLC
350 Massachusetts Ave. Suite 330
Cambridge, MA 02139
www.londoneconomics.com

---

White Paper

states that under perfect competition, all suppliers offer their output at marginal cost. For a
generator, typical marginal costs would include fuel costs, variable operating and maintenance
costs, emissions and water expenses (if those change in relation to output levels). However, a
significant proportion of costs faced by generators are fixed, and the ratio of fixed and variable
costs differ significantly across generators. Clearly, then, perfect competition cannot be the
yardstick for assessing behavior, because some generators will not be viable in such a market —
prices will not remunerate their fixed costs. If, in fact, were we to adhere to this strict definition
of perfect competition, it could be difficult to present a story of a sustainable industry, as we
would have successive amounts of capacity exiting the market. Of course, under a set of highly
restrictive conditions competitive power markets may deliver efficient outcomes when all
participants make offers equal to their short-run marginal costs. Perhaps the most important
condition is that real time prices clear the market such that the market price is sufficient to
curtail demand to meet available supply. In practice, this rarely applies. Rather, most markets8
have a series of mechanisms that preclude such outcomes, including, for example:

• offer and price caps that substantially understate customer willingness to pay;
• operating rules that fix aspects of supply such as reserve, supported by powers to
direct; and
• constraints on customers bidding into the market.
Proponents of such mechanisms have their reasons in supporting some of these. For example,
given inelastic supply and demand in the short-term, fixed procurement targets for reserve may
be required.9 However, justifying such constraints on one basis and ignoring their impact on
price formation and market efficiency is incomplete for the purposes of determining a
benchmark for assessing market power.

8 The notable exception is, perhaps, Australia, where offers and real time market clearing prices are allowed to rise to
levels close to the ‘willingness to pay’ at times when there is a high risk of a shortage.
9 Again, the Australian market, which works well in the absence of such fixed rules, might refute this presumption.

- 5 -
London Economics International LLC
350 Massachusetts Ave. Suite 330
Cambridge, MA 02139
www.londoneconomics.com

---

White Paper

Stage 1: Defining the market
Market definition
All questions of market power must be
preceded by an identification of the relevant Market definition seeks to identify an area of
market. In this section of the paper, we will activity over which market power can be
review existing approaches to defining exercised. For example, it would seem wrong to
speak of the supply of bottled non-carbonated
markets as they apply: in the US electricity water as a market, if consumers view bottled
markets, as per horizontal merger guidelines non-carbonated water as readily substituted for
and as part of existing market power by sparkling water. This is because no firm
measurement and mitigation protocols; in could, under these circumstances, exercise
other industries in the US; and in deregulated
market power over bottled non-carbonated water
electricity markets outside the US (i.e., the UK
alone: if it merely increased the price of bottled
and Australia, among others). We will also non-carbonated water, and every other price
describe the approaches and techniques remained unchanged, consumers would defeat
utilized by economists in the US and abroad in
the price rise by shifting their purchases to
developing an appropriate market definition. sparkling water. As a result, no proper inquiry as
In the last sub-section of this part of the policy
to market power, or as to its use, could focus on
paper, we will illustrate — through indicative
conditions in the supply of bottled non-
carbonated water and of that alone.
application of several empirical tests to US
markets — how quantitative analysis can be The same result could be achieved by supply-
utilized to aid in the development of a robust side substitution. Thus, it may be the case that
market definition.
there is an entrenched demand for bottled non-
carbonated water, and that consumers of bottled
A separate market constitutes a sphere of non-carbonated water do not regard sparkling
rivalry that is distinct from the others (please water as a close substitute. However, if
refer to the text box on the right). Under the producers of sparkling water can shift into the
Department of Justice guidelines, the production of bottled non-carbonated water
geographical boundaries of a market can be without substantial delays or sunk costs being
empirically estimated by examining whether a
required for the shift to occur, then an increase in
hypothetical monopolist would be prevented the margins being charged on bottled non-
from profitably imposing a monopoly price by
carbonated water alone will be defeated as
existing suppliers of sparkling water shift their
demand or supply substitution or by new production over. As a result, no proper inquiry
entry. In other words, a distinct market can as to market power, or as to its use, could focus
be said to exist if a hypothetical monopolist on conditions in the supply of bottled non-
over the whole supply in that market would carbonated water alone.
profitably be able to sustain a Small-but-
Significant Non-transitory Increase in Price,
* This discussion adapted based on considerations presented
which is the basis of the SSNIP test.
in the Nestlè /Perrier merger (see Case No. Comp/M. 190 –
Nestlè/Perrier; Article 8(2)(b). Decision of 22/07/1992.)
We recommend adopting a classical approach
to market definition, consistent with that used in most merger and anti-trust cases by the
Department of Justice and Federal Trade Commission (FTC) in the US, and competition
regulators abroad. The classical approach seeks to describe the market across four dimensions:
time, geography, function and product (see the figure below for their application to power

- 6 -
London Economics International LLC
350 Massachusetts Ave. Suite 330
Cambridge, MA 02139
www.londoneconomics.com

---

White Paper

markets).10 There is generally no debate regarding functional dimension: most agree that we are
analyzing the wholesale electricity market. However, the other three aspects are more difficult
to delineate.
Figure 1. Defining the market in electricity
Time
Geographical
Dimension
timeframe for
Dimension
market span limited
investment
on the basis of
response
SSNIP test*
(approx. 2-3 years)
Product
Dimension
spot and forward
Functional
markets for
Dimension
wholesale
energy and ancillary
energy market
services, such as
reserves and regulation


From an economics perspective, one needs to ascertain (across all dimensions) whether
demand-side or supply side (or market entry) substitution would defeat the attempt by a
hypothetical monopolist to profit by raising prices. In economic terms, this could be addressed
if one were to ask the following question: how would one product’s customers respond if prices
were to change for a different product? In the context of electricity markets, this could be
phrased as follows: how would consumers of electricity respond in the spot market if forward
market prices declined? The answer to such a question is captured in cross-price elasticities of
demand. In addition, one needs to consider how other suppliers’ strategy would change if one
supplier raised his prices? The answer to this question is captured in a measure of cross-price
elasticity of supply.

10 We mention all four dimensions here for reference purposes. However, there is broad agreement that the relevant
functional market in the case of horizontal generation market power is the wholesale market (rather than
the retail market). Thus, we will limit our discussion of this fourth dimension. However, it is instructive to
highlight that some of the same techniques that one applies to define geographical and product scope can
also be used to delineate functional market space, given the availability of necessary data.

- 7 -
London Economics International LLC
350 Massachusetts Ave. Suite 330
Cambridge, MA 02139
www.londoneconomics.com

---

White Paper

Figure 2. Stylized example on estimating cross-price elasticities
(PX) Price in Spot Markets
Cross-Price Elasticity of Demand:
(Product X)
∆QY * PX
Products X and Y are
∆PX QY
close substitutes, due
to high elasticity
all substitutes will have a
(slope of line)
positive cross-price elasticity.
Products X and Y are
weaker substitutes,
due to low elasticity
(slope of line)
(QY) Quantity of
Forward Contracts
(Product Y) Demanded


US Courts first began to use the concept of cross-price elasticities of demand for defining a
relevant market in the Times-Picayune anti-trust case in 1952.11 The concept of substitutes is
acknowledged in the opinion overturning the district court decision that Times-Picayune
Publishing Co. violated the Sherman Act through certain tying arrangements it had with
advertising: “For every product, substitutes exist. But a relevant market cannot meaningfully
encompass that infinite range. The circle must be drawn narrowly to exclude any other product to which,
within reasonable variations in price, only a limited number of buyers will turn; in technical terms,
products whose "cross-elasticities of demand" are small. Useful to that determination is, among other
things, the trade's own characterization of the products involved. The advertising industry and its
customers, for example, markedly differentiate between advertising in newspapers and in other mass
media.” Due to the perfect substitutability of the products in question, it was determined that
there was only a one market. With a single market, there could be no anti-trust case regarding
tying agreements.
Geographical dimension
The geographical dimension should be defined by the physical boundaries of the market. In the
case of electricity, we can rely on the extent of the transmission infrastructure as an indicator of
the ‘natural’ physical boundaries of the market. While this provides us with a starting point for
determining geographical market boundaries, the existence of transmission interconnections
between groups of generators in of itself is not sufficient to show that those generators operate
in the same market. Rather, this depends upon whether the transmission links between
generators have sufficient capacity such that suppliers in one region can discipline the pricing
behavior of suppliers in interconnected regions. Geographical market definition could be

11 Times-Picayune Publishing Co. v. United States, 354 US 594 (1952).

- 8 -
London Economics International LLC
350 Massachusetts Ave. Suite 330
Cambridge, MA 02139
www.londoneconomics.com

---