This report was prepared for EnerNOC.

Executive Summary

The Western Australia Public Utilities Office (PUO) outlined its proposal for reforming the Reserve Capacity Mechanism (RCM) into a competitive auction-based mechanism that meets reliability objectives at lower cost. However, the PUO was concerned that if it instituted the auction immediately, a sudden large drop in capacity payments could impair suppliers and threaten the sustainability of the electricity market. The PUO therefore proposed a transitional mechanism intended to balance the competing objectives of reducing excess capacity and customer costs, while mitigating price impacts on suppliers.

A key element of the PUO's transition proposal is that it would discriminate against demand response (DR), paying a much lower price for it than for generation, to incentivize some DR to exit. Generation would be paid according to an RCM payment curve that declines somewhat more steeply than today's curve. The curve would remain static until the reserve margin tightens to 5–6%. If and when that occurs, the final auction design would be triggered, with a much steeper demand curve for capacity, and with DR and generation competing on a level playing field.

The PUO's proposed transition mechanism will meet or exceed the reliability objectives, and it will partly meet the objective of reducing customer costs at least compared to the status quo. However, we believe the proposed transition mechanism will largely fail to meet its goals of reducing costs and eliminating economically inefficient resources, and that these inefficiencies will persist for well over a decade before shifting to the more efficient competitive auction design. This is due to the proposed resource discrimination and other related problems:

  • Discriminating between generation and demand response will distort investment and exit decisions away from least-cost, economically-efficient outcomes. Generators will have incentives to reinvest in their capacity even when it is unneeded and less costeffective than demand response.
  • The proposed static pricing schedule will maintain uneconomically high prices for a protracted period that may last more than a decade before the auction is triggered.
  • The proposed outcomes-based trigger for moving to an auction (with lower prices) further distorts generators' incentives to uneconomically maintain and invest in their capacity in order to delay triggering an auction.

In this paper, we propose an alternative transition mechanism designed to balance the PUO's competing objectives of reducing customer costs and mitigating supplier impacts, while correcting the concerns we have identified. This proposal would treat all generation and demand response resources on a level playing field, move to an auction format as quickly as possible, and mitigate price shocks with a phased introduction of a demand curve. The initial demand curve would be relatively flat but would become steeper each year according to a pre-determined schedule. The curve would match the final proposed PUO curve after a known period, possibly five or ten years.

The result of our proposal would be to introduce an initial price shock that is similar to that under the PUO proposal, then gradually decrease prices as the demand curve becomes steeper, and move to the final auction design over the course of a pre-specified number of years. This alternative approach would limit the time of the transition period and provide a more concrete timeline for achieving the final design, while mitigating price shocks along the way. And it would eliminate the inefficiencies associated with resource discrimination.

I. Background on the PUO Transition Mechanism Proposal

The Western Australia Public Utilities Office (PUO) outlined its proposal for reforming the Reserve Capacity Mechanism (RCM) into a competitive auction-based mechanism that meets reliability objectives at lower cost. The proposed auction design is generally consistent with best practices from successful capacity markets in the U.S. The auction would determine a clearing price and quantity based on the intersection of an administratively-determined demand curve and a supply curve made up of offers from all types of qualified suppliers competing against each other on cost. The demand curve would be constructed to procure just enough capacity to meet reliability objectives, without being so steep as to produce excessively volatile prices or invite the exercise of market power. Because the proposed curve for the final auction will be much steeper than the current RCM payment function, prices would decline more rapidly with excess capacity. At a high reserve margin, the price would be much lower than it is under the current RCM, so capacity would exit and avoid the present situation with high prices under excess supply conditions.

However, the PUO was concerned that if it instituted the auction immediately, the large and sudden price drop would be "financially disruptive for participants and create risks for the sustainability of the market as a whole."1 The PUO noted a risk of "flow on effects to the energy market...from a widespread impairment of generation assets."2 Hence they proposed a transition mechanism to ease that shock, while reducing customer cost, discouraging new entrants, and encouraging the exit of inefficient resources.3

The PUO's proposed transition mechanism administratively sets the capacity price using a curve proposed by the Independent Market Operator (IMO) with a slope of -5 instead of the more gradual curve in effect today.4 The PUO suggested using a single slope throughout the transition in lieu of a progressively steepening slope year to year.5

A key element of the PUO's proposed transition mechanism is that it would discriminate against demand response, paying a much lower price for it than for generation to incentivize some DR to exit. The PUO's rationale for this discrimination appears to be that it wants to reduce the excess capacity of both generation and demand response, but that a lower price is needed to cause demand response to exit due to its fundamentally different cost structure:

The reforms to the capacity price formula during the transition period will result in the price paid for capacity being discounted more heavily when there is excess capacity than is currently the case. This adjustment will result in a lower capacity price and reduce incentives for generation capacity to be maintained in the market or new generation capacity to enter the market.

A lower capacity price will not have the same effect on incentives for demand side management capacity, as it has fundamentally different cost drivers to other forms of capacity.6

The PUO explains that demand response's cost structure reflects minimal capital costs and high dispatch costs that exceed the energy price cap. DR would incur high opportunity costs when dispatched, but the likelihood of being dispatched is low under excess supply conditions, so demand response will stay in the market until the capacity revenue is at levels far below what the proposed transition mechanism would produce.

II. Assessment of the Proposed Transition Mechanism

The PUO's proposed transition mechanism will meet or exceed the reliability objectives, and it will partly meet the objective of reducing customer costs at least compared to the status quo. However, we believe the proposed transition mechanism will largely fail to meet its goals of reducing costs and eliminating economically inefficient resources, and that these inefficiencies will persist for well over a decade before shifting to the more efficient competitive auction design. This is due to the proposed resource discrimination and other related problems:

  • Discriminating between generation and demand response will distort investment and exit decisions away from least-cost, economically-efficient outcomes. Generators will have incentives to reinvest in their capacity even when it is unneeded and less costeffective than demand response.
  • The proposed static pricing schedule will maintain uneconomically high prices for a protracted period that may last more than a decade before the auction is triggered.
  • The proposed outcomes-based trigger for moving to an auction (with lower prices) further distorts generators' incentives to uneconomically maintain and invest in their capacity in order to delay triggering an auction.

These effects are likely to extend economic inefficiencies and high prices for many years, without providing price stability or predictability to the market.

A. RESOURCE DISCRIMINATION WILL REDUCE ECONOMIC EFFICIENCY

The PUO's proposal recognizes that the current excess of capacity is costly and provides little marginal value, so some resources must be let go. Ideally, the resources that stay should be the most competitive, and those that go should be the costliest. Indeed, one of the PUO's goals is "encouraging mothballing/retirement of inefficient capacity" in the transition.7 But this is not what the proposed transition mechanism would do. The proposal instead targets demand response, but without having demonstrated that all demand response is less economic than all generation, other than referencing its higher dispatch costs. But dispatch costs are only one component of a resource's total costs. It would similarly miss the whole story if one were to focus on aging generators and note that they would likely have the highest capital reinvestment costs and so determine that the oldest generators should be targeted for price discrimination.

The missing pieces of this story are:

  • First, capacity is a different product from energy. It is a megawatt of supply that has committed to be available so the system operator can avoid involuntary load shedding due to total supply being inadequate to meet demand. The ability of any resource to provide such a product should be established on technical grounds that are indifferent to the underlying asset's dispatch costs (those differences are fully recognized in the energy market). Resource qualification should however account for the true availability of each resource to help avoid load shedding. In the case of demand response, qualification requirements should include accurate measurement and verification standards as well as a measure of performance during reliability events (or surprise test events if there are no realized reliability events).
  • And second, the most economically efficient resources to keep online to meet the capacity requirements are those with the lowest net costs, and that are willing to accept the lowest capacity payment.8 The resources needing the highest capacity payment are the costliest and so should be the first to exit.

In many cases, existing generators will have lower net going forward costs than demand response and would be willing to accept very low capacity prices before retiring or mothballing. This is because existing generators' capital costs are largely sunk, and they would earn sufficient margins out of the energy market to stay online even without earning any capacity revenues. However, other existing generators would have very high net going forward costs, for example if they faced a major retrofit or plant reinvestment to continue operating. Overall, there will be a diversity of high-cost and low-cost resources amongst both demand response and generation. We have seen this in PJM, where demand response resources have out-competed many existing generators.9

The PUO's proposed mechanism is not set up to discover the relative economics of generation and demand response resources because it discriminates between them. With generation receiving a much higher price, high-cost generators may stay online while low-cost demand response exits. The high price paid to generation will shield it from the pressure to retire even when uncompetitive and uneconomic. This will create incentives for generators to undertake high-cost retrofits to uneconomically prolong the life of the unit even under current overbuilt conditions. Generators may even find it profitable to invest in uneconomic capacity uprates to their existing plants. Such capital expenditures would be economically wasteful and run counter to the goals of the energy market review.

The best way to identify the lowest-cost resources is through a competitive auction with a level playing field for all qualified resources, rather than administratively selecting winners and losers.10 Every resource will offer to sell capacity at the minimum price they would need to recover their costs. Those with the highest offers would not clear. The result is to minimize total costs across the resource base, and minimize the customer costs associated with meeting a particular capacity requirement. The PUO recognizes the value of using a competitive auction: "When the auction arrangements commence, demand side management capacity will be subjected to the correct signals to compete on a level playing field with other capacity providers."11 But a level playing field would be just as valuable during a transition, when it will be equally important to allow competitive forces to push the most uneconomic resources out of the market.

It is true that under the PUO's proposed transition mechanism, demand response would be compensated at uneconomically high prices if it were included in the payment scheme. But the same is true for generation. The medium-high prices contemplated in the transition mechanism are lower than under the current RCM, but still far above the economic value of capacity, far above the long-term proposed demand curve level, and far above the price levels that are likely needed to encourage substantial exit (of both generation and demand response) from the market. These concerns suggest that an even steeper demand curve should be implemented more quickly to reduce the magnitude of overcompensation across the board.

To continue reading this update, please click here

Footnotes

1 See PUO (2015), p.13.

2 Ibid.

3 See PUO (2015), p. 45.

4 The proposed curve is defined as Price = Price Cap * ((1-(Surplus % + 3%))*slope))-1. A more negative slope parameter would correspond to steeper payment curves. The current formula sets Price = Price Cap * (Surplus %)-1. An illustration of both curves is provided in Figure 3 in a later section. See IMO (2013), p. 32.

5 See PUO (2015), p. 46.

6 See PUO (2015), p. 47.

7 See PUO (2015), p. 45.

8 A capacity supplier's net costs are given by the going-forward fixed operating and maintenance (FOM) and capital costs minus net energy revenues. "Net energy revenues" are the revenues earned in the energy market minus any fuel and other variable costs, thus fully reflecting each resource's value related to its dispatch costs. Net revenues from the energy market help cover a portion of a resource's fixed costs. This is particularly true for baseload generation that accrues substantial net revenues from the energy market. This is why the more capital-intensive baseload units need not earn a higher capacity payment than less capital-intensive combustion turbine units. Both baseload and peaking capacity will recover their capital costs in an efficient market, but baseload will earn proportionally more out of the energy market and less from capacity. Similarly, demand response will earn nothing from the energy market if it is not dispatched, or even face net costs from dispatch given the low energy price cap in Western Australia. If all resources are paid the same energy price for dispatch and the same capacity price for availability, then competitive forces will incentivize the lowest-cost combination of these resources to meet both energy and capacity needs.

9 See Pfeifenberger et al. (2011), p. 17.

10 By "level playing field" we mean equal competitive treatment across all resources that provide the same or nearly the same capacity value, being available to balance supply and demand and thus prevent involuntary load-shedding during peak conditions when supply can become inadequate. Differences in resources' availability and reliability can be accounted for through appropriate adjustments to their capacity ratings.

11 See PUO (2015), p. 14.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.