ubject of capacity payment that motivated this work and from the presentations of Larry Ruff and Harry Singh at that workshop. The positions expressed in
this paper,however, are the author's own views.

2

Department of IEOR, University of California at Berkeley, Berkeley, CA 94720 USA Oren@IEOR.Berkeley.Edu

Abstract: This paper discusses alternative approaches that
have been adopted around the world for guaranteeing the
appropriate level of investment in electric generation
capacity. We argue that the use of "capacity payments" is
the least desirable approach that undermines the long-term
efficiency objectives of the electric industry restructuring.
We explain how in an energy only market, long term supply
contracts in the form of call options with premiums that
depend on the contracts' strike prices can meet the need for
ensuring supply adequacy and the financial health of the
generation sector.
Keywords: Capacity payments, Generation planning,
Reliability, Adequacy.
1. I
NTRODUCTION

The reliability of electricity supply has been one of the
overriding concerns guiding the restructuring of the electric
power industry. The slogan "keeping the lights on" has been
the principal motivation for many technical and economic
constraints imposed on market designs. The term supply
reliability, encompasses, however, a mix of system
attributes that have diverse economic and technical
implications under alternative market structures. NERC
(National Electric Reliability Council) defines reliability as:
"the degree to which the performance of the elements of the
technical system results in power being delivered to
consumers within accepted standards and in the amount
desired". Imbedded within this definition is the notion of the
"obligation to serve" which is arguably out of step with the
notion of a deregulated industry with competitive supply. In
fact, the concept of reliability as defined by NERC
encompasses two attributes of the electricity system:
Security, which describes the ability of the system to
withstand disturbances (contingencies) and Adequacy,
which represents the ability of the system to meet the
aggregate power and energy requirement of all consumers at
all times.
The notion of system security identifies short term
operational aspects of the system which are characterized
through contingency analysis and dynamic stability
assessments. Security is provided by means of protection
devices and operation standards and procedures that include
security constrained dispatch and the requirement for so
called ancillary services such as: voltage support, regulation
(AGC) capacity, spinning reserves, black start capability
etc.. The notion of adequacy on the other hand represents
the systems ability to meet demand on a longer time scale
basis in view of the inherent fluctuation and uncertainty in
demand and supply, the non-storability of power and the
long lead time for capacity expansion. Generation adequacy
has been traditionally measured in terms of the amounts of
planning and operable reserves in the system and the
corresponding loss of load probabilities (LOLP) that served
as criteria for planning and investment decisions.
From a technical perspective security and adequacy are
clearly closely related since a system with abundance of
reserve capacity provides more flexibility in handling
unforeseen disturbances. However, while a system with
limited planning reserves may experience shortages it can
still be operated in a secure manner while a system with
ample reserve can be operated insecurely. From an economic point of view security and adequacy are
quite distinct in the sense that the former is a public good
while the latter is a private good. Security is a systemwide
phenomenon with inherent externality and free ridership
problems. For instance, it is not possible to exclude
customers who refuse to pay for spinning reserves from
enjoying the benefits of a secure system. Hence, like in the
case of other public goods such as fire protection or military
defense, security must be centrally managed and funded
through some mandatory charges or self-provision rules.
The resources for such central provision, however, can be
procured competitively through ancillary service markets,
long term contracts or other procurement mechanisms.
Adequacy provision on the other hand, as will be explained
later, amounts to no more than insurance against shortages,
which in a competitive environment with no barriers to
entry translate into temporary price hikes. Such insurance is
clearly a private good of which the adequate quantity to be
provided can be decided through customer choice. In an
environment in which "obligation to serve" is replaced by
"obligation to serve at a price", the concept of loss of load
probability is not well defined unless a distinction is made
between probability of lost load due to system collapse vs.
lost load due to inadequate supply. It is a prerogative of
consumers and producers to decide what is the appropriate
level of price insurance they wish to procure and how much
they are willing to pay for it as long as they are able from a
technical point of view to bear the consequences of their
decisions without affecting others. In the remainder of this
discussion we will only focus on adequacy provision.
The traditional approach to ensuring generation adequacy in
vertically integrated utilities was to build planning reserves
based on load forecasts LOLP calculation and estimates of
the value of lost load (VOLL) and assign the cost of the
extra capacity as a rate uplift. More elaborate schemes,
which will be discussed below, attempted to allocate the
cost of capacity according to time of use so that peak
consumption bears a larger portion of that cost. In an ideal
competitive market were prices of energy vary continuously
to reflect the equilibrium between supply and demand at
each moment, payment to inframarginal generators (above
marginal cost) should cover their capacity cost. Economic
theory tells us that in a long-term equilibrium, the optimal
capacity stock is such that scarcity payments to the marginal
generators when demand exceed supply will exactly cover
the capacity cost of these generators. Furthermore, the
optimal generation mix (were generators are characterized
by their fixed and variable cost) will be such that the
operating profit of each generator type will exactly cover
their capacity costs. This optimal equilibrium mix is
achieved through exit of plants that do not cover their cost
and entry of plants whose cost structure will yield them
operating profits that exceed their capacity costs.
The critical role of electricity in the economy and the
political ramifications of widespread electricity shortages
have prompted many regulators around the world to take
steps above and beyond reliance on market forces in order
to ensure generation adequacy. While in theory, allowing
the prices of energy to reflect short run supply and demand
equilibrium will create market signals and provide adequate
financing for proper capacity expansion, many regulators
have been concerned that energy prices occurring in the
various restructured systems are not sufficiently high to
cover generators' capacity costs and to prompt adequate
investment. The prevalence of regulatory intervention to
suppress energy prices even when they reflect legitimate
scarcity rents justifies the concern that indeed generators
would not be able to cover their fixed costs through energy
sales alone. Ruff [1] argues that "suppression of energy
prices" is inherent in all current market designs that accept
only hourly or half hourly energy prices that could not
possibly reflect the second by second changes in the supply
and demand balance. He claims that such mechanisms tend
to suppress the price spikes that would arise in an idealized
continuous double auction reflecting true spot prices.
Furthermore, the absence of demand side bidding in most
energy markets (that would determine the spot price in case
of shortage) and the tendency to dispatch reserves to
mitigate shortages obscure the scarcity rents. On the other
hand one could argue to the contrary that in a continuous
double auction with full demand side participation prices
would be lower than when demand is treated as inelastic
over hourly intervals regardless of changes in supply
conditions. There is also the strong possibility that the
measures taken to ensure generation adequacy have the
effect of suppressing energy prices due to excess capacity or
perverse incentives so that the necessity of such measures
becomes self-perpetuating. This is clearly the case in
Argentina, for instance, were a large capacity payment paid
on the basis of generated energy induces generators to bid
below marginal cost so as to increase production and
capacity payment revenues.
2. A
PPROACHES TO ENSURING ADEQUACY

There are currently three basic approaches to dealing with
generation adequacy in restructured electricity markets.
i) Energy Only markets
This approach has been adopted in California, Nordpool
and the Australian Victoria pool. Generators in such
markets bid only energy prices and, in the absence of
constraints, all bids below the market-clearing price in each
hour get dispatched and paid the market-clearing price. The
primary income sou