Must-run generation: Can we mix regulation and competition successfully?

Electricity Journal

Volumn 10, Issue 10, 1997, Pages 44-55

  Jurewitz, John L ^a   Walther, Robin J ^a  

^a NONE

Author keywords

[No Author keywords available]

Indexed keywords

COMPETITIVE MARKETS; CRITICAL TIME; MARKET POWER;

COMMERCE;

EID: 0031471798     PISSN: 10406190     EISSN: None     Source Type: Journal    
DOI: 10.1016/S1040-6190(97)80319-3     Document Type: Article

References (11)

1. "Must-run" generation must be distinguished from so-called "must-take" generation. In California, stakeholders have agreed that a utility's QF contracts and nuclear generation are "must-take" resources that will be dispatched by the ISO virtually whenever they are available. By contrast, "must-run" generators must be dispatched by the ISO during certain hours to assure local reliability.
2. Scale economies achieved by local generators and environmental constraints on siting new generation and transmission facilities may prevent the complete dissipation of local rents through entry, but one must question whether this possibility really warrants the continued regulation of this limited "local natural monopoly."
3. Assuming some scale economies persist and prevent the complete dissipation of locational rents through competitive entry, this remedy would essentially regulate this local natural monopoly using Harold Demsetz's classic suggestion to auction the right to the utility franchise.
4. Phillips Petroleum Co. v. Wisconsin, 347 US 672 (1954).
5. We assume the generator will be able to appeal this designation to FERC, but we also assume that FERC will be very deferential to the ISO regarding reliability issues.
6. While this title may be useful as shorthand, it is important not to misunderstand its meaning. More accurately, these generators should be called "otherwise uneconomic." In fact, their must-run services are economically valuable and, if they are designated as must-run resources, they should be a more economical source of such services than the next best alternative.
7. The call provisions of a must-run contract need not and should not interfere with the participation of the generator in the market. The only concern of the ISO should be that the generator operate when needed for local reliability. But the ISO does not "purchase" the generator's output in the usual sense. Rather, the ISO merely requires the must-run generator to operate, pays the call price, and substitutes the generator's output in the dispatch for that of another generator willing to decrease its output by an identical amount.
8. This criticism does not apply when the contract is imposed on a new owner at the time of asset purchase. The new owner would then capitalize these financial benefits in its bid price so they would accrue to the seller of the asset. If this seller were the utility, presumably the regulator could require that these capitalized benefits be credited to customers.
9. There are two additional criticisms of a RAP structure. The first is that it may not send the ISO the right long-term price signal for terminating the contract in favor of a substitute. The ISO will view the level of the reduced availability payment as the cost of continuing the contract. This amount may be higher or lower than the actual social cost of continuing the contract, depending on the level of net revenues being generated. For instance, if actual net revenues are high, the social cost of continuing the contract will be lower than indicated by the simple availability payment and the ISO will be inefficiently eager to cancel the contract in favor of an alternative. This potential source of inefficiency can be easily eliminated by permitting the generator to compete with offers from other must-run generators by reducing the availability payment it will accept.
10. One possible solution would be to set the call price at the higher of variable cost or the market-clearing price. The problem with this solution is that, by itself, it fails to control market power. The generator would be motivated to withhold its generation, drive up the market price, and then receive the higher market price when called by the ISO. This problem could be solved by requiring the generator to bid its generation into the market whenever available and be subject to certain minimum availability requirements. The generator could then be paid the higher of its variable cost or the market clearing price whenever called by the ISO. Designing and administering such participation requirements and protecting them against possible gaming would be an added regulatory burden of a Pure ZAP format. Moreover, the approach, at best, is a very limited solution because it can be applied successfully only to "economic" generators.
11. The problem arises to the extent the call price exceeds the market clearing price during any hour. When the generator expects this, it may be motivated to refrain from market sales if it is confident of being called by the ISO and receiving the call price. Remedying this perverse incentive is not easy. The difficulty is that, as a practical matter, the adder cannot be perfectly time-differentiated to assure that it always equals the amount of net revenues that would have been received in the competitive market if the generator had no locational advantage. As a practical matter, the adder will be set constant over all hours or groups of hours. Moreover, for "uneconomic" generators, an additional amount must also be collected, since market revenues alone would not be adequate to recover availability costs if the generator had no locational advantage. Thus, in both the economic and uneconomic cases, the owner will frequently have the ability and incentive to refrain from bidding in order to receive the adder embedded in the call price because the call price will be higher than the expected market price. One possible remedy is to structure the ZAP-with-Adder so that in hours in which (1) the generator is selling energy in the market, and (2) the ISO would have called the generator to operate, the ISO will pay the generator the difference between the call price and the market price. But for this to be an effective remedy, the generator must believe that the ISO will not game the situation by denying that it would have called the generator. This suspicion will be difficult, if not impossible, to overcome.