useful definition of market power is the ability of a generation company, using
one or more of its plants, to increase market price profitably over a
significant period of time. The
company ceases to be a price-taker. The game it is playing is static, a
one-shot event, and its strategy is rational only within a single point in
time. The company makes a unilateral decision to withhold capacity, to raise
price bids, or to do both. Its aim is to increase the slope of the market
supply curve and thus to raise the market price. It then is able to profit
handsomely from the supply it has not withheld. The clues for detecting market
power in generation are well established.  Low
demand elasticity produces a weak consumer response to a market price increase.
Large demand exhausts fringe capacity, and the dominant firm is a monopoly over
the large residual demand it faces. A firm can raise the price bids of its
marginal, price-setting plants in order to benefit its infra-marginal ones. Low
supply elasticity allows a firm to induce a market price rise without the
concern that, in response, its rivals might raise their output. Binding
transmission constraints divide the grid into isolated pockets and thus allows
favorably located generating plants to wield local market power. The submission
of bid curves that vary significantly across similar hours and market
conditions is another indicator. Finally, emission allowances also affect
market power. If and when a generator reaches its emission limit, assuming it
is operating below capacity, then its production is constrained.
mentioned above, market share and HHI are unable to capture the dynamics of
space-time competition in restructured power markets. Transmission constraints
alter the scope of the geographic market and render a market share calculation
meaningless. A generator could strategically induce congestion in order to
blockade the entry of imports and thus to capture the market solely for itself.
Indeed the FERC criteria for market-based rates, the market share “safe harbor”
of 20% and related HHI measures, were deeply flawed and never proven in power
Market shares and HHIs, therefore, are useful only as an initial screening
device and definitely not conclusive. To acquire meaning and depth, they have
to be combined with information on transmission congestion.
A typical approach to market power analysis is to
utilize a computer simulation model.
However, a model is only as good as the theory underlying it, and what is
needed is a scientific body of knowledge guiding its creation and
implementation. The field of economics fills this need. But a simplistic
application of economics models, such as single firm behavior or Cournot with a
competitive fringe, tends to have a bias in favor of finding an anticompetitive
effect. A forecast of market behavior has to include, among many other factors,
demand uncertainty, the cost of withholding capacity, entry, information
uncertainty, contracts, and market rules.
market price is determined by a confluence of several diverse events and
factors (see Figure 1). A decision by one plant could affect and be affected by
commercial and physical conditions both near and far. Supply and demand, with
all their nuances, is just another set of factors. All possible market design
loopholes and legal inconsistencies are exploited for profit. The status-quo
pattern of transmission constraints is usually beneficial to some generation
and transmission owners but detrimental to others. In California, the potential
for earning capacity payments in the ancillary service markets is a powerful
incentive to withdraw capacity from the energy market, in which payments are
purely on energy. Expectations of drought and unfavorable changes in weather
patterns increase the scarcity value of water and worsen any strategic behavior
exercised by a hydro unit. In short, many interacting factors are at work, 
and any proper analysis of generation market power quickly becomes intractable.
Figure 1. Price Discovery in Restructured
Specialist models, such as UPLAN, our proprietary
engineering economy representation of the Northern American electric power
system (see Appendix A), can do a proper job of capturing the key commercial,
physical, regulatory, and climactic factors driving market outcomes.
Thus, a methodology that appears to present itself points beyond market shares,
HHI, and the usual misuse of economics models. After defining the relevant
product and geographic market, market shares and HHIs could be calculated. A
dynamic analysis is then performed in order to account for
- The time-varying nature of the
- The relevant economic capacity
defined as the actual generation during the pricing period;
- The frequency of market dominance
- The duration of market
- The structure of bids.
Figure 5. Simulated Monthly Differences Between
the Entergy Area and SPP Average
Peak and Off-peak Prices
Appendix A. The UPLAN Modeling System