BEFORE THE PUBLIC UTILITIES COMMISSION OF THE STATE OF CALIFORNIA __________________________________ ) Order Instituting Rulemaking on ) the Commission's Proposed Policies) R.94-04-031 Governing Restructuring ) California's Electric Services ) Industry and Reforming Regulation ) __________________________________) ) Order Instituting Investigation on) the Commission's Proposed Policies) I.94-04-032 Governing Restructuring ) California's Electric Services ) Industry and Reforming Regulation ) __________________________________) REPLY COMMENTS OF THE COALITION OF CALIFORNIA UTILITY EMPLOYEES ON COMPETITIVE MARKETS AND MARKET INSTITUTIONS IN A RESTRUCTURED ELECTRIC INDUSTRY REPLY TO THIRD ROUND COMMENTS August 18, 1994 PREPARED BY: Marc D. Joseph David Marcus ADAMS & BROADWELL P.O. Box 358 651 Gateway Blvd. Berkeley, CA 94701 Suite 900 S. San Francisco, CA 94080 C1011.030 I. INTRODUCTION The Coalition of California Utility Employees joins virtually every party in this proceeding in endorsing the principle of competitive wholesale markets. The issues before the Commission are not about whether such markets are desirable, but how to achieve them and what they should consist of. The Commission already has before it the record of a full day of en banc hearings regarding competitive market structure. In addition SDG&E has filed a description of another full day's discussion of competitive markets held July 29th, and numerous parties filed extensive comments on July 26th. The Commission does not need another lengthy treatise on the relative merits of bilateral markets and pooled markets.1 Both types of markets allow full customer choice and access to any sort of contract between customer and generator. This filing focusses only on one critical difference between the two models: The extent to which each model allows least cost or merit order dispatch of plants. Professor Hogan characterizes this difference as "Operator Dispatch of Flexible Plants." (Hogan, 8/4/94, p. ii.) We focus on this issue because the Commission should recognize that the choice of market structures will affect the actual dispatch of generation and the overall costs paid by ratepayers. Under a bilateral scheme such as the natural gas industry model proposed by Enron or the "Opco" approach described in the testimony of Professor Hogan, the system operator's ability to control dispatch will be limited to, at most, the level required to 1 The terms "bilateral" and "Poolco" are used here as shorthand for the two main approaches which have been extensively debated before the Commission. (See the August 4 written testimony of Professor Hogan for additional characterization of these two approaches.) C1011.030 maintain system stability. Least cost dispatch will become a thing of the past. Bilateral contracts, like the current QF contracts, would require relatively expensive plants to be run even when cheaper energy is available, resulting in even more overpayments by ratepayers to generators. A pool approach on the other hand, whether with features proposed by SCE, by SDG&E, by Hogan, or by the Sierra Club or EDF, and regardless of the size of the pool, allows both customer choice and cost-minimizing dispatch. Only with least cost dispatch would the overall costs of electricity borne by ratepayers be minimized. II. WHO CONTROLS THE DISPATCH ORDER AND WHAT ARE THE RESULTS? The Commission has heard repeated references to "economic dispatch" or "merit order" dispatch. These are both terms for the straightforward concept that, in any given time interval, only the least cost units are operated which are necessary to meet load in that interval. The reality is that no system can produce pure economic dispatch. However, the closer to least cost dispatch the system operator can come, the lower the total cost to ratepayers. Thus, the model which best achieves the objectives of the Blue Book is one which allows both full customer choice and physical dispatch of plants based, as much as possible, on the least cost energy each hour. A. Operators have minute-by-minute control, but are constrained Under any approach, whether bilateral or Poolco, there will be a system operator. Currently, the system operator function is supplied by the joint efforts of several dozen control area operators throughout the Western Grid, each of whom is responsible for monitoring control area loads and resources, and balancing them at least every ten minutes. For the PG&E C1011.030 2 control area, that involves adjusting generation levels as frequently as every four seconds. Clearly, the system operator function cannot be executed by negotiated contracts, but must be done in real time under predetermined rules, with associated dollar payments made after the fact. The system operator does not have complete flexibility. Operators are always constrained by generator physics (e.g., if a unit is slow-start, it cannot go from off to full power from one hour to the next, even if such a situation might be otherwise desirable for either economic or reliability reasons). Operators are also constrained by transmission physics: if a transmission link is already fully used, then additional generation, no matter how desirable, cannot be moved over it. On the demand side, operators have very little flexibility: customers can add or subtract load unilaterally, and do, with very limited operator control via interruptible contracts or direct load management controls.2 Operators are also constrained by contracts and other legal limits and rules. At the system level, for example, the decision to maintain a given level of San Francisco-based generation on line at all times in the PG&E control area, for example, is a reliability-based dispatch rule which constrains operators. In general, "administrative" constraints which are not due to the laws of physics play a substantial role in determining dispatch even when dispatch is nominally in "merit order." The sections below discuss various constraints imposed on operators under the current system, and how those constraints would change under bilateral or Poolco competitive systems. 2 PG&E estimates total customer load subject to dispatcher control in its control area at about 500 Mw (PG&E, 1994 ECAC, Forecast volume, 4/1/94, p. 6- 31). C1011.030 3 B. Operators face significant, expensive constraints under the existing rules Under the current system, in which each control area nominally operates like a pool, with cost-minimizing dispatch as the general goal, operators are already significantly constrained, and routinely cannot achieve least cost dispatch. "System minimum conditions" or "minimum load conditions," as PG&E calls them, occur when lack of load causes non-oil/gas units to become the incremental source of supply while more expensive oil/gas units are still operating.3 Under such conditions, operators are not free to reduce costs by turning off the more expensive oil/gas units, and least cost dispatch does not occur. In 1993, PG&E reported minimum load conditions in over 1,400 hours of the 8,760 hours in the year, hardly an infrequent occurrence.4 The primary cause of minimum load conditions in California is "must- run" generation. Must-run generators are those which, for one reason or another, are not subject to operator control and thus can end up dispatched out of merit order. QFs are the most conspicuous example of must-run generation. Almost all existing QFs have contract provisions which make them nondispatchable in all or almost all hours of the year. QFs thus produce large amounts of generation which cannot be taken or rejected based on price considerations. For PG&E, for example, QFs are forecasted to provide 20 percent of all 1995 generation.5 3 See PG&E, 1994 ECAC, Forecast volume, 4/1/94, pp. 6-23, -24. 4 PG&E, monthly "12(b)" filings with the CPUC, 1-11/93, where PG&E reports that "Boxed hours indicate Minimum Load Conditions..." 5 PG&E, 1994 ECAC, Forecast report, 4/1/94, Table 6B. C1011.030 4 Nuclear units are another source of must-run generation. Because of their physical nature, nuclear units are not well-adapted to having their output dispatched up and down. Thus, nuclear units are typically run at the maximum level they are capable of running. This dispatch strategy is normally consistent with least cost dispatch, but not necessarily so.6 Hydro and geothermal are, to a lesser extent, nondispatchable as well. Geothermal generation is partially dispatchable under PG&E's renegotiated Unocal contract, but there are minimum requirements and annual requirements which limit the dispatcher's control of it. Some hydro generation is run-of- the-river and completely uncontrollable by dispatchers. Other hydro generation is driven by irrigation and other non-power releases, and is only partially controllable by dispatchers. Even "fully dispatchable" hydro is under dispatch control only as to the timing of generation, not the quantity. Operators can decide when to let water out of certain reservoirs, but they cannot control the total annual amount of water to be let out, which is controlled by rainfall and snow melt. Other resources are out of dispatcher control because they are scheduled by others. These include purchases and generation from municipal utilities which do not have integrated operating agreements with the larger IOUs. Finally, even oil/gas units owned by the IOUs themselves are not dispatched purely on merit order. PG&E, for example, dispatches its own oil/gas units out of merit order in order to meet its own reliability rules in 6 Bob Kinosian, DRA, 3/94, "Report on Cost-Effectiveness and Alternative Pricing Proposals for Nuclear Power Plants for SCE Company General Rate Case," A.93-12-035. C1011.030 5 San Francisco, the East Bay, and the Humboldt Bay area.7 It also dispatches oil/gas units out of merit order to comply with fish protection rules in the Delta in spring months, to provide regulation and spinning reserve, and to accommodate the riskiness or impossibility of shutting down large thermal units overnight and restarting them the following day.8 The net result is that in any given hour a substantial fraction of generation is actually committed based on non-price considerations. Even now, merit order dispatch is really just merit order dispatch for some units, not for all. C. With a bilateral trading approach, operators will be far more constrained The Commission has heard from advocates of pure bilateral trading that natural gas markets are the example to follow. It has been suggested that in natural gas markets, system operators only need to control about one percent of the resources in order to maintain system reliability. The implication is that a similar level of control will be sufficient to provide system stability and reliability in electrical markets. If the Commission chose to rely on bilateral markets, economic dispatch would be far more severely constrained than at present. Every generator owner would be free to enter into contracts for its generator which 7 PG&E, 1994 ECAC, Reasonableness volume, Chapter II. 8 PG&E (and the other IOUs as well) will operate oil/gas units out of merit order overnight in order to have them available to meet the following day's peak load. The economic appropriateness of this strategy depends on whether the cost of overnight out-of-order generation is exceeded by the savings from having the units on-line the following day, which in turn depends on what alternative resources would have cost both overnight and the following day. C1011.030 6 would then act as constraints on the dispatchers' flexibility, and could be expected to do so. That is the very point of bilateral contracting.9 As more and more generator output came under bilateral contracts, system operator flexibility would steadily decrease. If system operator flexibility were ever reduced to the one percent level asserted to exist in the natural gas industry, system reliability would undoubtedly be imperilled.10 But long before that, there would be major economic consequences for ratepayers. Decreased system operator control would result in increased frequency of minimum load conditions. Inexpensive resources without market access via a bilateral contract (e.g., nonfirm hydro sold by Pacific Northwest utilities) would be denied access to dispatch to make room for more expensive resources with contractual security (e.g., future combined cycle projects built with the security of a bilateral contract with a customer or group of customers). 9 CLECA's suggestion in its July 26 filing that a bilateral world would later adopt a pool is disingenuous at best. What is the point of a pool in which the pool operators have no flexibility because each generator can say, as QFs already do, "my unit is must-take; you can't control when to turn it on or turn it off; I do that"? Claims that bilateral contracts could be modified later to allow merit order dispatch are foolhardy. Would QFs today, with bilateral contracts with the IOUs, voluntarily amend them to accept merit order dispatchability? Some wouldn't do it at all, and none would without substantial financial reward. CLECA makes no showing as to how (assuming sanctity of contracts) any transition from bilateral contracting to a pool with merit order dispatch would work. 10 Consider, for example, PG&E's "San Francisco constraint," under which, for reliability reasons, up to 50 percent of on-peak San Francisco load is served by San Francisco generation. Suppose bilateral direct access were available to San Francisco customers, and suppose non-San Francisco generation was available at a lower cost than San Francisco generation (as it generally is). Then for any given customer, it would make sense to sign up for non-San Francisco generation, and let some other customer pay for the more expensive San Francisco generation. The end result would be that everyone would contract for non-San Francisco generation, and San Francisco reliability would decline for all. C1011.030 7 Municipal utilities would have less access to transmission, thereby increasing costs to municipal utility customers. Although some would pay less, costs would increase for others and the total costs paid for electricity would increase. This would be contrary to the principles expressed in the Blue Book. D. With a Poolco, constraints should be reduced Under a Poolco, customer choice in financial arrangements is arranged separately from physical dispatch. Customers have the full panoply of choices offered by generators,11 but generators must physically deliver to the Pool, which dispatches in merit order based on operating price bids from the various generators. If a generator owner wants its unit to be operated in certain hours and not others, or in all hours for that matter, it can achieve that result by bidding a low price in the hours it wants to operate. Poolco thus 11 Some have disputed this. NYMEX, for example, justifies its refusal to get involved in electrical futures market-making in England by blaming the English Pool for volatile prices which it says cause market-making to be too risky (Levin, 8/4 oral testimony). However, volatile hourly prices are a fact of electricity markets caused by the time-varying costs of electricity production, not by the presence of a pool. PG&E, for example, reports its historical incremental costs to the CPUC. For 1993, those reports show incremental cost changes from one hour to the next of as much as 390 percent (PG&E, 12(b) report, 5/28/90, 6-7 a.m.). In every month, the maximum hourly jump in incremental cost is at least 25 percent (PG&E, 12(b) reports, 1-11/93). For SDG&E, reported 1993 economy energy costs have at least one hour every month whose cost is more than double that of the preceding hour. (SDG&E, 1993 CPP 100 Mw Incremental Economy Energy Costs.) These are just changes in energy costs. As NYMEX acknowledges, reliability costs can be even more volatile. The CPUC already knows this, from filings such as SCE's real-time pricing tariff RTP-2-I in which reliability-sensitive hourly prices can vary by over 400 percent from one hour to the next. (SCE, advice letter 1057-E, 6/24/94). Thus, the mix of generating resources and costs is the cause of volatile hour to hour energy prices, not the presence or absence of a pool. C1011.030 8 does not constrain generator owners from operating any way they want to and meeting any physical or financial constraints they may face, but it requires them (rather than ratepayers) to accept the economic costs of doing so. A Poolco will end the historic practice of forcing customers to bear the operating costs of new generation whether or not it is economic to dispatch. New IPPs, like old QFs, will undoubtedly be largely financed based on contracts which provide for a steady revenue stream. But that steady revenue stream will have to come from a financial contract, not from the assurance of guaranteed access to physical dispatch without regard to merit order. A Poolco may also act to reduce some of the current sources of uneconomic dispatch. For example, spinning reserve and other reliability services could be bid to the Pool along with generation, as "ancillary services" are in England. That would create opportunities for lower-cost providers of those services to gain access to dispatchers who currently rely on IOU-owned resources to provide these services.12 More generally, Poolco would enable owners of currently must-run resources to provide society with the benefits of dispatchability for their units and be properly compensated for doing so. The Commission presumably 12 See, for example, the June 27 submission to the Commission of the comments of Energy Storage Partners, the would-be developer of the Lorella Pumped Storage Facility in southern Oregon. Such a project would be a source of spinning reserve and regulation which could reduce the need for out-of- merit order dispatch of oil/gas units. Under the current system, the IOUs have no incentive to buy reserves and regulation from this project because they get them at no marginal cost from their own rate-based units, and the costs of out-of-order operation of oil/gas units are recovered through the ECAC mechanism. C1011.030 9 does not intend (assuming it could) to break existing contractual commitments. Thus, under Poolco, existing QF contracts would be converted into contracts for differences, with the contract making up the difference between the QF's revenue entitlement under its PURPA contract and its revenue from Poolco.13 The resulting contracts would ensure QFs the exact same revenues as at present, but would give QFs a price signal allowing them to shut down when economically appropriate for both the QF and ratepayers. For example, suppose a QF such as a biomass producer has high variable operating costs, which exceed the utility marginal costs but are still lower than its contractual payments under PURPA. Then it is in the QF's interest to operate, but doing so costs results in out-of-merit order dispatch. Under the present rules, the QF does operate, and the IOU dispatchers can't do anything about it.14 Under Poolco, if the Poolco price was less than the QFs marginal cost of production, the QF could buy from the Pool rather than operating its own facility, deliver the purchased energy to the Pool to qualify for its contract for differences entitlement, and end up with the same revenue but lower costs than before. Ratepayers as a whole would be better off, because the high-operating-cost QF would have been turned of and something cheaper operated instead, yet the QF would not be penalized. 13 We expect the same treatment would be afforded Diablo Canyon under the Settlement Agreement, so that if the Settlement Agreement price was, say 12 cents per kwh and the Poolco price was 3 cents, the other 9 cents would be paid to Diablo Canyon as a contract for differences. The Commission should consider whether PG&E's stated opposition to Poolco is an unwarranted fear that Poolco would reduce Diablo Canyon revenues. 14 Except try and negotiate a voluntary bilateral contract amendment, at considerable transaction cost, with no assurance of success. C1011.030 10 III. CONCLUSION The existing rules, most notably QF contracts, already allow a substantial fraction of generation to escape merit order dispatch. The result is minimum load conditions and increased fuel costs which, via ECACs, are borne by ratepayers. Bilateral contracts which are allowed to affect dispatch will exacerbate the out-of-order dispatch problem. If unchecked, they could threaten reliability, and will certainly increase costs to utility service customers and overall ratepayer costs. With full customer choice and physical dispatch through the existing IOUs or a Poolco, all generators face the right marginal pricing signals. New generation and new bilateral financial contracts will not exacerbate existing minimum load problems, and some of the existing problems should themselves be ameliorated. Therefore, while allowing full customer choice, the Commission should ensure that physical dispatch is based on economic merit, not bilateral contract terms. Respectfully submitted, ADAMS & BROADWELL _______________________ Marc D. Joseph C1011.030 11 PROOF OF SERVICE I, the undersigned, declare as follows: I am now, and at all times herein mentioned, have been a citizen of the United States, over the age of 18 years, a resident of San Mateo County, California, and not a party to the within action or cause; that my business address is 651 Gateway Blvd., South San Francisco, California. At said address on August 18, 1994 I served a copy of the attached REPLY COMMENTS OF THE COALITION OF CALIFORNIA UTILITY EMPLOYEES ON COMPETITIVE MARKETS AND MARKET INSTITUTIONS IN A RESTRUCTURED ELECTRIC INDUSTRY (Reply to Third Round Comments) by placing said copy in an envelope addressed to the parties named on the attached service list. The envelope was then sealed and placed for collection, mailing and deposit on the above date, in the United States Postal Service, following ordinary business practices. I am readily familiar with the practice of this office for collection and processing of correspondence for mailing with the United States Postal Service; this correspondence would be deposited with the United States Postal Service on the above date in the ordinary course of business. Also on August 18 1994, service was made upon those addressees on the attached service list marked with "HAND DELIVERED" by hand delivery made by an employee of Professional Messenger Service. I declare under penalty of perjury that the foregoing is true and correct. Executed on August 18, 1994 at South San Francisco, California. C.W. Trent C1011.030