Toward a Capacity Demand Curve Market

The market for electric generation facilities has passed through several stages
in the transition to more competitive markets. Questions over the viability
of competitive markets gave way to fundamental market restructuring, investor
optimism, tremendous enthusiasm and investment, overbuilding and price depression
in seemingly rapid succession. Market designers are now looking toward the future,
trying to determine how the markets can be structured to attract investments
when and where they are needed. Problems are caused by both too much and too
little investment, and competitive markets will only succeed if they result
in maintaining a reasonable balance between these alternatives. In this environment,
one approach gathering momentum uses an administratively set demand curve to
set prices based on supply levels. Suppliers still face market risks, but the
risks are tied to the level of installed resources, under the hope that this
predictable market incentive will lead toward better stability and balance.

System reliability requires enough capacity to meet load in every hour during
the year and that requires maintaining resources that will only rarely be needed.
Maintaining reliability through the incentives available in a competitive market
is the challenge. Energy markets alone could provide this incentive if energy
prices were allowed to spike to very high levels. But, such pricing has not
been allowed due to concerns over market power, bidding behavior and consumer
reaction to extreme prices. A market with unrestricted energy prices is not
a practical alternative. The development of significant demand-response from
high prices could also help solve this problem, but such response has been slow
to develop. Nevertheless, enough, but not too much capacity, must be built.
Failing to achieve this delicate balance can either cost a lot of money associated
with overbuilding or threaten system reliability.

Too Much, Too Fast

One measure of the success of the competitive wholesale market is its ability
to attract new investment. The burst of new investment in the late 1990s and
early 2000s would seem to suggest that competitive wholesale markets have been
successful. It is now clear, however, that too much capacity was built and came
on line far too fast. Reserve margins that were once considered a problem if
they exceeded 20 percent now reach 30 percent in many areas. Prices have plummeted,
however, and past investor optimism cannot be relied upon to maintain high,
or even adequate levels of capacity.

The present excess capacity situation – along with heightened FERC oversight
of participant behavior – has dampened energy market price spikes. In markets
where energy prices are often correlated with the short-run marginal cost of
the last unit dispatched, the highest-cost units do not make much money. Energy
prices alone are not providing sufficient revenues to cover the total cost of
new investment (see Figure 1).

In the past, regulated utilities were required to maintain adequate capacity
levels, which were frequently tied to a specific reserve margin. This concept
has continued in competitive markets, generally by requiring the payment of
capacity deficiency charges for any load-serving entity that falls short of
its requirement. In a short-term capacity market, this creates the incentive
for a binary price: Prices for capacity equal the deficiency charge when the
market is short, but fall quickly to zero with higher levels. This price signal
is not proving to be adequate to provide the kind of stability investors or
regulators desire. It is compounded by the disincentives many load-serving entities
face in contracting for long-term power supplies in the current environment.

Demand Curve Concept

The solution, or at least the next approach to be tried, is the demand curve
concept. New York, New England and PJM have all introduced capacity markets
based on some sort of a downward sloping demand curve. New York has been operating
its demand curve-based capacity market since the summer of 2003, while New England
and PJM are currently developing capacity markets based on the demand curve
structure. The New England ISO has proposed such a market structure under a
FERC mandate, and the details of that design are currently being litigated at
FERC. A decision is expected in June 2005 and the market is to be operational
in calendar year 2006. In PJM, there is a stakeholder working group process
under way that is collectively determining the specific structure of the future
demand curve based capacity market. PJM anticipates holding auctions under the
new approach for the 2006–2007 planning year.

In a demand curve capacity market, the demand curve is set administratively,
based on the amount of capacity that is in the market. This price ranges from
substantially above the expected annual carrying and operating (fixed) cost
of a new peaking unit during times of relative shortage, down to zero in times
of substantial excess. Payments are made to all generators in the region, with
allowances for imports and exports. The cost of the program is then assessed
on load on a pro-rata basis. Properly structured, these payments will provide
the right incentive for entry, and units will have the opportunity to make a
reasonable return on investment, as capacity levels in the region generally
trend around the desired level.

Figure 2 presents the general form of the demand curves for each of these regions.
This comparison should not be taken too literally, because there are many differences
regarding such issues as adjustments made for expected energy revenues that
would otherwise be earned. But the graph gives a general idea of how these curves
look. While this graph has reflected a specific geographic sub-region for each
of the markets, all three systems have recommended different demand curves at
different geographic locations.


Source: ISO-NE, PJM and ISO-NY ICAP Demand Curves

The demand curve provides some predictability and stability to capacity revenues.
There is some debate over the conceptual framework upon which the parameters
of the demand curve should be based. Some look to the pricing dynamics of an
uncapped energy market (UEM) or a value of lost load framework (VOLL). In the
authors’ view, these theoretical constructs can be limiting, because they rely
on hypothetical market designs where prices are allowed to spike to unlimited
levels (UEM), or require controversial calculations of the value of supply interruptions
and the probability of such interruptions at varying levels of supply (VOLL).
Instead, a more straightforward method to maintain a reliable supply and lower
costs to consumers is desirable. Based on these objectives, the demand curve
mechanism provides a way to effectively put one’s “thumb on the scale” in the
supply-demand balance. The additional capacity payments should increase supplies
in support of reliability, and wherever possible the curve parameters should
be designed to lower the cost of that capacity in the market. Lowering investor
risks will ultimately translate to lower costs to consumers. The “thumb on the
scale” metaphor implies some deliberate intervention, yet – in theory – still
allows market participants on the supply side to respond to price signals to
achieve a balance between supply and demand. It also provides the flexibility
to lower costs to consumers, by structuring the payments in ways that make it
easier for new entrants to raise capital and by tying payments to performance
that can increase system reliability.

Unlike the capacity deficiency programs of the past, the downward sloping demand
curve provides a more continuous payment stream to generators across varying
capacity levels that are higher at critically low levels of reserves and lower
when the system has sufficient (but not too much) reserves. Investors will better
understand how prices can be expected to change, but are given no guarantees.
Investment decision will now focus on the potential for shortages and surpluses
– which matches squarely with the need to maintain reliability without excess
costs. New units are unlikely to enter unless they expect to earn at least equilibrium
returns in their first year of operation. This price level – associated with
the total cost of entry – is critical because the market should equilibrate
around this level. Estimates of the cost of new entry are used in designing
the demand curve, but market performance will ultimately be based on the actions
of participants. Prices will reflect suppliers’ actions, which are based on
their actual costs, not the initial estimates. Market exit decisions will also
be based on market prices, and retirements are likely to occur during times
of relative surplus, ameliorating excess supply conditions. Investors will bear
the risk of their decisions – whether profit or loss – based in part on their
own projections of the future supply/demand balance. This creates the incentive
to keep the system in balance.

Developing Demand Curves

While no demand curve will eliminate the risk of excess capacity eroding revenues,
the slope of the demand curve has significant implications. Keeping the demand
curve steep around the expected equilibrium level of the market – which is centered
on the cost of new entry in each of the proposed ICAP markets – will keep equity
investors focused on targeting new investment for periods when it is needed.
In this way, the signal to attract investment is highly focused on the level
of investment that is needed by the system to maintain reliability, and the
range of deviation around that point is minimized. A steep demand curve also
minimizes the cost of errors associated with the estimate of the cost of new
capacity. On the other hand, flattening the curve provides more price stability.
Within limits, providing some price stability will provide the revenue predictability
that makes it easier to attract capital – particularly low-cost debt.

The different perspectives of equity and debt investors should be considered
in evaluating demand curve parameters. Equity investors look at expected returns
over the long term. Large risks can be worthwhile, as long as the potential
returns are sufficiently high. Lenders are much more focused on the likelihood
of meeting debt payments on a yearly basis from project cash flows. A great
equity investment can be a very poor opportunity from the lender’s perspective.
Imagine, for example, an investment that hinges on the flip of a coin and pays
three dollars for “heads,” nothing for “tails.” While an excellent overall investment,
a lender sees a 50 percent chance of total failure; hardly the performance of
investment grade commercial paper. The lender’s return in a good year is capped
at the debt payment, and it generally does not have the luxury of offsetting
underperforming years with overperforming years over the course of the project’s
life. This difference in perspective between debt and equity comes in greatest
contrast for projects with the most volatile revenues.

A demand curve can be structured to consider the needs of both investors. The
equity investor’s money can be placed at risk, with the potential for appropriate
returns, while the need for greater certainty debt investors can be accommodated.
The risk to equity investors provides proper incentives to maintain system reliability,
while greater certainty provided to debt investors lowers the cost of capital
and ultimately lowers prices for consumers. While not explicit, these considerations
support the concept of a kinked demand curve, which is relatively steep in the
range of the carrying cost of new capacity, but with a flatter slope and therefore
greater payments at lower levels. This gives equity investors the incentive
to maintain just the level of capacity needed in the market, but makes it easier
for them to borrow money to build the new units.

Another key consideration pertains to how capacity payments are adjusted to
reflect energy rents to the hypothetical new generator. Investments are made
based on the expectation of total payments, but some of these will come from
short-term profits in the energy market. These are often called energy rents
and are the difference between energy revenues and the variable costs incurred
in producing the energy sold. The energy rent adjustment is made in all demand
curve markets, but there are differences regarding how the adjustments are made.
The closer the adjustments are tied to actual energy prices, the more accurate
the adjustment and therefore the more accurate the targeting of the desired
capacity level. Energy rent adjustments in New York (and proposed adjustments
in PJM) are based on generic projections, while New England proposes that the
adjustment be based on actual data for 12 months prior to the monthly auction.

As one digs into the details of these markets, other controversial issues arise.
Adjustments for unit availability and performance are important, and different
approaches can favor different participants. The goal for consumers is to incent
the behavior that provides the greatest reliability at the lowest cost. The
implications for trade between regions must be considered as well. These and
other details matter a great deal. A substantial portion of the total cost of
wholesale energy will be covered, either directly or indirectly, by these capacity
markets.

There are also questions about the effectiveness of demand curvebased capacity
markets in solving reliability problems in small zones. To some uncertain extent,
these designs may help reduce the need for reliability-must-run contracts. The
tradeoffs are likely to be based on the size of the payments needed to ensure
a competitive solution to local reliability problems. At some point the costs
needed to provide for a fully competitive solution to all circumstances can
be unreasonable, and nonmarket solutions are likely to be part of the market
for some time.

Will it Work?

While economic theory suggests that a properly designed demand curvebased ICAP
market should be superior to past attempts to address reliability issues, unfortunately
there is no guarantee of success and many issues have yet to be resolved. For
example, risk created by the potential for regulators to change the program
in the future will continue to exist. If the regulators do not let the market
respond to the price signals and instead intervene or force a regime change,
confidence could falter. In addition, the demand curve-based capacity markets
may not work as planned. The immense dollars at stake in these programs raises
concerns about unintended consequences and gaming. This concern will not dissipate
until the market has been in operation for enough time to ensure that such behavior
will not occur, or identify and address the behavior as it arises. In any event,
recent FERC actions clearly indicate that the demand curve concept will be tested
in the marketplace.