California Electricity Restructuring, the Crisis, and Its Aftermath

Transcription

1 California Electricity Restructuring, the Crisis, and Its Aftermath DRAFT James L Sweeney October 3, 2005

2 Table of Contents Abstract 1 California s Restructuring 1 California Electricity System Before Restructuring 1 Changing Federal Regulations 6 The Context for California Electricity Restructuring 8 Generation/Wholesale Markets 8 Stranded Costs 9 Retail Markets 10 The Restructuring Process 11 CPUC Regulatory Proceedings 12 State Legislation: Assembly Bill After Passage of AB Wholesale Markets under the Restructured System 18 The California Power Exchange (PX) 18 The California Independent System Operator 20 ISO Real Time Markets 20 CAISO Ancillary Services Markets 21 Divestiture of IOU Generating Assets 21 Forward Wholesale Contracts for Electricity (or Lack Thereof) 22 Wholesale Markets: In Summary 23 Retail Markets 23 Municipal Utilities 25 Market Issues in the Restructured System 25 PX and CAISO Market Clearing Price Relationships 25 PX Bidding Strategies for Electricity Generators 27 Market Risk 28 Wholesale Market Risk 28 Retail Market Risk 32 Western Electricity Crisis and California Financial Crisis 33 The Nature of the Crisis 33 The Electricity Crisis 34 Wholesale Price Increases 34 Energy Emergencies and Rolling Blackouts 37 Economic Forces During the Electricity Crisis 39 Limitations on Generation Capacity 39 Generation Plants Offline 41 Environmental limitations 41 Pacific Northwest Drought 42 Natural Gas Limitations 45 Limitations on Demand Adjustment 47 California s Dysfunctional Energy Market Structure 48 Risk Created by the Financial Crisis 48 Exercise of market power and market gaming 48 Energy Conservation 51 Draft. Do Not Quote. ii

3 The Financial Crisis 52 State and Federal Regulation Changes During the Crises 54 Federal Responses: FERC Rule Making 55 November 1 and December 15, 2000 Orders 55 April 26, 2001 FERC Price Mitigation Order 56 FERC June 19 Order Broadening Scope of Price Mitigation 57 California After the Crises 58 Current Conditions 58 Electricity Generation Capacity 58 IOU Status 58 Retail Prices of Electricity 59 Litigation 59 Changing Market Structure, Market Forces 59 Fundamental Market Flaws 60 Retail Competition on Hold 60 Electricity Demand Measures 61 IOU Wholesale Electricity Procurement 63 ISO Market Redesign 64 Capacity Incentives 64 Other Electricity Policies 65 Environment 65 Fuel, Transmission 65 Organizational Issues 66 Public Perceptions 67 Proposals to Re-regulate 67 Some Reflections 68 References 70 Draft. Do Not Quote. iii

4 Abstract In chapter 10, James Sweeney provides an overview of California s electricity restructuring in mid 1990s. California s goal was to deregulate both the wholesale and the retail electricity market, after a transition period during which utilities would be able to recover stranded costs. Three regulatory rules were in place during the transition period: a) regulators forced investorowned utilities to divest at least 50% of their generating assets, b) regulators precluded the utilities from entering forward contracts to acquire electricity, and c) regulators imposed retail price caps on the utilities. In the midst of the transition period, the Western electricity crisis, driven primarily by a sharp reduction in available hydropower and limitations on natural gas supplies, drove wholesale electricity prices throughout the western United States to triple-digit or quadruple digit levels. The thee regulatory rules together proved disastrous, driving one utility into bankruptcy and another near bankruptcy, and left the state with billions of dollars of losses. Flaws in the California market design encouraged market gaming by generators, traders, utilities, and the State itself. With the crisis over, California has now established resource adequacy rules for utility electricity acquisition, encourages long-term contracts, and has eliminated retail price caps. An ISO-led market redesign project is underway to restructure transmission congestion management. Retail deregulation has been temporarily put on hold. But the future regulatory direction remains uncertain, with some advocates trying to force a return to a vertically integrated monopoly structure and others striving to improve regulatory rules within a partially regulated, partially deregulated system. California s Restructuring California Electricity System Before Restructuring Prior to 1998, California's electricity system traditionally operated similarly to electricity systems throughout the United States. It included three large investor-owned utilities (IOUs), collectively selling most of the electricity in California. Each IOU had a franchise in one of three separate parts of the State Pacific Gas and Electric Company (PG&E) in Northern and Central California, Southern California Edison (SCE) in Coastal, Central, and Southern California, and San Diego Gas and Electric (SDG&E) in San Diego. In addition were several 1 much smaller IOUs, several electric co-ops, and numerous municipal utility systems, the largest of which were the Los Angeles Department of Water and Power (LADWP), and the Sacramento Municipal Utility District (SMUD). The IOUs served 78% of the California customers, the municipal utilities served 22%. Electric co-ops and the federal agencies collectively served less than 0.1% of the customers. In terms of total MWh of electricity, the IOUs sell 72 %, the municipal utilities, 24 %, and the federal agencies 3%. 1 Smaller IOUs were PacifiCorp, Sierra Pacific Power Co., and Southern California Water Co. Draft. Do Not Quote. 1

5 Each utility operated as a local monopoly, selling electricity in its own exclusive franchise area, with no direct retail competition from other electricity sellers. The three large IOUs, as well as some of the municipal utilities, were vertically integrated to include generation, transmission, and local distribution. Some municipal utilities operated as only local distribution companies; some participated in the other two functions. Figure 1 shows the service areas of the various utilities as of 1996, while the restructuring was being debated. Figure 1. California Utility Service Areas, Source. CEC For IOUs, almost all significant financial decisions involving any of the three functions were subject to the jurisdiction and control of the statewide regulatory body, the California Public Utilities Commission (CPUC). Customers paid retail prices for electricity based on operating costs plus a regulated rate of return on the prudently incurred used and useful invested capital. The CPUC would review whether costs were prudent and would determine the fair rate of return on invested capital. Pricing was based primarily on cost of service and only secondarily on market conditions. Draft. Do Not Quote. 2

6 The significant decisions made by the publicly owned municipal utilities were subject to the jurisdiction and control of their appointed or elected governing bodies. Thus, their strategies could be based on local decision-making, rather than on statewide regulations. However, they typically were operated so that over a span of several years their revenues roughly equaled their total costs of operation. Thus, for municipal utilities as well as for IOUs, pricing was based primarily on cost of service and only secondarily on market conditions. Figure 2 shows the mix of resource types used to generate electricity in California 2. The three largest sources of electricity are natural gas, nuclear, and hydroelectric. In 1998, at the time of the restructuring, 36% of electricity was generated using natural gas, 21% from hydropower, and 18% from nuclear. The 15% generated using coal is dominantly generated outside of California, but owned by California utilities All Other Wind Organic Waste Geothermal Coal Hydroelectric Nuclear Gas Average MW (Thousands) Figure 2. Electricity Generation in California 2 Source: CEC. The Commission provides this important explanation of a data change. Prior to 2001, utilityowned shares of coal, nuclear plants and some firm contract generation outside California were considered part of utility-owned generation category. Since 2001 most of this data is included in the energy imports category. However, two coal-fired plants, Intermountain and Mohave, which are located outside of California, but inside LADWP and CAISO control areas, respectively, are still treated as utility-owned generation. Draft. Do Not Quote. 3

7 California has been a relatively low user of electricity, when measured on a per capita basis. The industrial structure is not a heavy electricity consumer, relative to industry in other states. The State is blessed with a temperate climate and thus air conditioning loads are relatively low. High prices of electricity, coupled with aggressive energy demand management programs have kept consumption relatively low. California has had a long tradition of efficiency incentives and services that have helped to reduce California s electricity intensity in recent decades. In 1999, right after the restructuring, the residential electricity consumption per customer 3 was 37% below the US average during that year. Figure 3 shows the electricity consumption over time for the various major consuming sectors. From 1990 to 2000, use of electricity increased from 26,000 MW average consumption rate to just above 30,000 MW, a growth of 16% over ten years, or 1.4% per year. Growth in energy consumption, however, was somewhat faster from the 1997 through 2000 period, increasing by almost 2,000 MW during the three years, or an average growth rate of 2.3% per year average. Peak loads were growing at roughly the same rates. Average MW (Thousands) Residential Commercial Industrial Agricultural & Water Pumping Other Figure 3. Electricity Consumption in California 3 Data from U.S. Department of Energy. Data are based on number of customers in the residential sector. This figure corresponds closely to the number of households. Draft. Do Not Quote. 4

8 Although the IOUs in California operated as vertically integrated monopolies, they did purchase some electricity from external sources. California utilities had long-term contracts to purchase hydroelectric power from the Bonneville Power Administration (BPA), a federal powermarketing agency, which sells power generated primarily from hydroelectric projects in the Columbia River. Both Munis and IOUs also had other contracts to purchase electricity from Federal projects. California traditionally sold electricity to Pacific Northwest entities in the winter, when Pacific Northwest demands peak and purchased electricity from the Pacific Northwest during the summer, when California demands peak. This trade was part of a larger pattern of interconnections throughout the west. A grid connects the western states, plus the Canadian provinces of British Columbia and Alberta. Figure 4, created by the Western Governors Association 4 gives an indication of the transmission capacity between the various states and portions of states. The widths of the blue lines connecting the circles are proportional to the relative transmission capacities. There is much transmission capacity between California and the Pacific Northwest and between California, Nevada, and Arizona. These parts of the western region are linked particularly closely and the greatest flows of electricity occur among these areas. California represents 42% of the generation and 51% of the consumption in this tightly linked portion of WECC. Figure 4 Installed Capacity by Generation Type and Transmission in the Western Grid, as of January Source: Western Governors' Association Draft. Do Not Quote. 5

9 Changing Federal Regulations Several U.S. regulatory changes had important impacts on California. One of the most important was the Public Utility Regulatory Policies Act of 1978 (PURPA), enacted primarily to promote development of small-scale renewable sources of energy for electricity generation. Cogeneration was included as well, as a means of more efficiently converting primary energy into electricity and usable heat. PURPA mandated state regulatory commissions to establish procedures to require electric utilities to interconnect with and buy capacity and energy offered by any nonutility facility that qualified under PURPA. These so-called "qualifying facilities" or "QFs were typically small generating facilities based on renewable energy, waste products, or natural-gasfired cogeneration units. PURPA required utilities to pay a price for electricity from QFs equal to the avoided cost of electricity generation. This avoided cost was meant to be the totality of costs that a utility would avoid by purchasing electricity from these small alternative sources. PURPA left it to the state regulatory commissions to interpret what dollar price corresponded to "avoided cost" and the precise conditions under which the electricity and capacity must be purchased. In California, the CPUC took a very aggressive stance implementing PURPA. It set high prices for electricity purchased by the IOUs 5 under PURPA, requiring them to sign contracts based on standard offers with guaranteed prices that rose sharply over time 6. The financial incentives and guaranteed market for QF electricity, coupled with tax incentives established by the Federal government, created a significant industry of renewable electricity generation in California, including wind farms and electricity generation from wood wastes. These policy changes also led to large increases in cogeneration capacity 7, which was largely natural gas-fired. In the time between 1978 and the beginning of 2000, although the largest single source of new nameplate capacity was conventional generation, including both baseload facilities and peakers, QFs based on cogeneration or QFs based on renewables, each have a comparable nameplate capacity 8. The majority of the QF capacity was from small units (nameplate capacity smaller than 50 MW). Figure 5 shows the composition of the new generation capacity in California coming on line between 1978 and the beginning of January The black component of the bars shows the 4 Conceptual Plans for Electricity Transmission in the West - Western Governors Association - August Since the CPUC did not regulate the municipal utilities, these high prices were not relevant to these entities. 6 Under Interim Standard Offer No. 4 (ISO4), a QF based on renewable energy could sign a contract based on a fixed forecast of future electricity price. Such a QF entering a contract would be guaranteed $57/MWh in 1985, $81/MWh in 1990, and $109/MWh in After 10 years the contract price reverted to the short run avoided cost, which typically would be far lower than the fixed price guarantee. Gas-fired cogeneration units were not treated nearly as generously but were generally paid an annual average of about $25/MWh for capacity and about $25 to $30/MWh for energy. 7 Cogeneration now is the single biggest source of PURPA electric generation capacity in California. Of the roughly 10,200 MW of total QF nameplate capacity in California in 2001, about 5,700 MW is cogeneration, 4,500 is generation from renewables such as wind or organic wastes. (Data from CEC database of electric generating plants on line in California.) 8 Delivered electricity from many renewables, particularly wind and solar, are significantly smaller than suggested by the nameplate capacity. Draft. Do Not Quote. 6

10 plants with nameplate capacity 50 MW or greater; the grey component shows plants with nameplate capacity smaller than 50 MW. MW Nameplate Capacity Greater Than 50 MW Smaller Than 50 MW 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0 QF Cogen QF Renewables Cogen, not QF Conventional Nuclear Figure 5 New Plant Construction 1978 through January 2000 However, with long term contractual obligations to purchase electricity from QFs at a high cost, by the early 1990s the utilities were facing a high average cost of electricity generation. In addition, California utilities had invested in nuclear power plants, units whose construction costs turned out to be greater than initially predicted, further increasing the average cost of electricity generation. These factors together helped make California s retail prices among the highest in the nation. Only in the states of California, Alaska, Hawaii, and the Northeastern states did average retail prices to residential customers exceed 8 cents/kwh ($80/MWh). California's average revenue in 1998 was 9 cents/kwh ($90/MWh). Another important effect of PURPA was the separation of some ownership of generation from ownership of local distribution. However, utilities still controlled all electricity transmission lines. A utility that wished to stifle competition in electricity generation could do so by refusing to allow its competitors to transmit electricity along its transmission lines. Thus creating a competitive market for electricity generation required federal officials to deal with issues of utility control of transmission lines. Draft. Do Not Quote. 7

11 The Energy Policy Act of 1992 (EPACT) was an attempt to address this problem. Among its many provisions, EPACT opened access by non-utilities to the transmission networks. And in 1996, the Federal Energy Regulatory Commission (FERC) issued Order 888, which much more generally opened transmission access to non- utilities. These regulatory changes together started to transform the electricity transmission system into a common carrier system 9. With EPACT and Order 888, it became much more difficult to control electricity generation markets by restricting access to transmission lines. Utilities still made the investment decisions for transmission facilities and thus could still exercise some control over generation markets, but this form of control was less effective than direct control over access to transmission line. These developments, made while California was designing its electricity market changes, were important for establishing the opportunity for wholesale electricity competition in California. The Context for California Electricity Restructuring Generation/Wholesale Markets The high retail price of electricity in California was one basis for arguments that California s electric system should be deregulated to create a more competitive and presumably lower cost, electricity system. The major contributors to the high retail price were high average cost of generating electricity and high prices embedded in contracts for purchasing electricity under PURPA contracts. Some advocates of electricity generation deregulation expected deregulation to reduce retail prices of electricity quickly. But as described in this chapter, this expectation was based on a fallacy: it implicitly assumed that deregulation could correct the historical problems that had led to the high generation costs and high costs of purchasing bulk power under contracts. But the costly investments in nuclear power plants and the long-term contracts for QFs could not be reversed. Sunk costs cannot be changed by restructuring. Regulators can reallocate who will bear those costs, but sunk costs will continue to be borne by some combination of companies and individuals. And, as became apparent over time in California, laws and legal precedents, basic notions of fairness, and political forces can all sharply limit the ability of regulators to reallocate many of the historical sunk costs. Changing regulations, however, can halt practices that historically led to high costs. But in California, many of the practices that had increased generation costs had already been halted. At the time of the restructuring debate, the state was no longer investing in new nuclear power plants. New cogeneration plants and renewable energy investments were being made on a market basis when such investments were expected to be economically attractive. The high price standard offers under PURPA were no longer required for new contracts. There was a more subtle argument, however, that deregulation would change some practices that had elevated costs and that could be expected, absent deregulation, to continue elevating costs. Halting such practices would reduce costs, although the cost reductions would be gradual. There was a widespread belief that the regulatory system did not provide strong enough incentives for 9 These issues are discussed more fully in Chapter 14. Draft. Do Not Quote. 8

12 utility-owned electricity generators to minimize costs. Thus it was doubtful that the regulated system led to the lowest cost mix of energy generation technologies. Utilities seemed to be favoring their own generation over generation by independent power producers (IPPs) and thus not cost minimizing. There remained incentives and opportunities for utilities to block distributed generation and to rely instead on central-station power, even if distributed generation had lower overall costs. Economists and other industry analysts were arguing that creating competition could change economic incentives facing the utilities. The change in incentives would gradually reduce costs of electricity generation, which would gradually reduce retail prices. This argument, although not proven 10, was probably valid, even though the hope of fast cost savings was never realistic. In addition, it was expected that expansion of wholesale markets would encourage more California investment in new generating capacity by IPPs. Most analysts anticipated that the healthy California economy would continue to need more electricity. Many doubted whether the old regulated system would be responsive enough to those economic needs. And many argued that the old regulated system would lead utilities to discourage new investment by IPPs. Thus, there was the opportunity in California to deregulate electricity generation to expand the scope of the existing competitive portion of the electricity generation industry. Expanding competition in electricity generation was expected to create incentives for cost cutting, to encourage investments in new generating capacity by IPPs, and to provide a flexible system for a dynamic California economy. Stranded Costs A problem of wholesale market deregulation was the issue of stranded costs. If future costs would be low for new generation, then future wholesale prices would be low, as well. This implied that the existing high cost generating units would no longer be economically viable in a competitive environment. Costs incurred by the utilities in constructing these plants would be stranded ; utilities would incur losses, absent policy intervention. The issue of stranded costs was not a problem of going forward costs future total costs of electricity generation ignoring sunk costs -- even if those costs might be very high for some of the units. If future wholesale electricity prices turned out to be lower than the operating costs of these plants, the plants would shut down in a competitive environment and their entire remaining book value would be a loss to its owner. However, such plants should be shut down for economic efficiency reasons. On the other hand, if the future wholesale electricity prices turned out to be higher than the operating costs of these plants, these plants could sell electricity at the wholesale price and would find generation to be more profitable than shutting down. These plants could compete in a market environment, as would be desirable for economic efficiency. However, there would still be a loss: the owner would not be able to recover all of the remaining book value. 10 Those arguments were discussed more fully in the Yellow Book and the Blue Book. These documents were developed by the CPUC in the process of the restructuring. Draft. Do Not Quote. 9

13 These fixed losses were sunk costs, and therefore not expected to influence the market-clearing price. But someone would have to bear the losses. Who should bear these losses the utilities or their customers was a politically important issue, a legal issue, and an issue of basic fairness. Thus, the issue of stranded costs was the issue of who should bear the burden of these fixed costs, costs incurred by the utilities in good faith under the previous system. There were several possible alternatives consistent with deregulation of generation. One alternative would be to allow the utility to include those costs in retail prices, thereby keeping retail prices high. That solution would provide motivation for customers to bypass utilities with large stranded costs, by purchasing directly from generators or generating electricity themselves, say in cogeneration units. Customers most able to do so would be the large users of electricity. If enough large customers bypassed utilities, utilities would primarily sell to residential and small commercial consumers; small users would thus pay most of the stranded costs. Another option would be to require the utility to write off the assets as losses, requiring the stockholders to bear the consequences of stranded costs. The utilities argued persuasively that it would have been inequitable for their investors to bear the stranded costs of long-term contracts and generating investments that had already been approved by the CPUC as prudent under the old "regulatory compact. It had generally been understood that under the regulatory compact utilities would make investments to serve the needs of ratepayers and ratepayers would pay back the costs of those investments plus a fair rate of return on the investments, over the life of the equipment. The utilities argued that eliminating the obligation of the ratepayers to pay back the costs, after the utilities had made the investment, would be unfair. And because some additional costs, such as some of the QF contracts, had been forced on them by the CPUC implementation of PURPA, forcing the utilities to bear the costs would have been particularly unfair and may well have been illegal. The debate over who should be liable for the stranded costs became central to the regulatory hearings and legislation. And the size of stranded costs was difficult to predict, since they depended on the future sales prices of wholesale electricity and future natural gas prices. This debate was not limited to California. While California debated its restructuring, how to deal with stranded costs was a national debate. The debate was particularly intense because the stakes were large. Some analysts estimated the nationwide stranded costs as exceeding $100 Billion. Retail Markets There was also the possibility of creating competitive markets for retail electricity, since many recognized that the commodity, electricity, could be unbundled from its distribution services. Retail competition offered the possibility that competing retailers would provide differentiated energy services that would be attractive to consumers. Some retailers could provide green power to environmentally conscious consumers. Others could bundle energy efficiency measures with electricity to help consumers reduce the overall cost of obtaining energy services. Some retailers could provide higher reliability of electricity for customers for whom reliability was essential or interruptible service for those customers willing to accept service interruptions in exchange for a lower overall bill. Some could sell electricity at real-time prices for those Draft. Do Not Quote. 10

14 customers that wished lowest average cost but did not mind price variability and others could sell electricity at guaranteed prices, essentially selling risk management services bundled with electricity. A competitive retail market could enhance consumer options and create a more flexible system. The Restructuring Process California s regulatory agency, the California Public Utilities Commission (CPUC), took leadership to reduce economic regulation in electricity markets, spearheading the move toward electricity deregulation in California. The deregulation that they envisioned would rely on competitive market forces in both wholesale and retail electricity markets rather than governmental control over electricity production and use. The process was initiated in 1992, culminated with a final CPUC restructuring order in December 1995, and legislation Assembly Bill 1890 signed in September Table 1 provides a timeline of the process and the subsequent events. Each element of the table is discussed at a later point in this chapter. PURPA Passed 1978 EPACT opened access to transmission networks 1992 CPUC initiated review of trends in electricity markets 1992 Yellow Book published February 1993 Blue Book Published April 1994 CPUC Decision D May 1995 Memo of Understanding Presented September 1995 Final CPUC Restructuring Order December 1995 Request filed with FERC to create PX and ISO April 1996 Assembly Bill 1890 Passed and Signed into Law September 1996 FERC Order FERC Authorized ISO and PX 1997 First generating plant divestiture June 1998 Proposition 9 defeated November 1998 Last generating plant divestiture April 1999 Wholesale prices spike above $400/MWhr June 2000 Wholesale prices spike above $2,000/MWhr December 2000 California DWR begins buying electricity January 2001 FERC Price Mitigation Order April 2001 Wholesale Prices Fall Sharply May-June 2001 FERC Final Price Mitigation Order June 19, 2001 Scheduled end of transition period March 2002 Governor Davis recalled in special election November 2003 Arnold Schwarzenegger become Governor January 2004 Nunyez Bill 2004 Proposition Table 1. Timeline of California's Restructuring, Electricity Crisis, And Subsequent Events Draft. Do Not Quote. 11

15 CPUC Regulatory Proceedings In April 1992, the CPUC initiated a review of trends in the electric industry, which initially resulted in a staff report 11 published in February This report, commonly referred to as the Yellow Book, set the intellectual framework for the subsequent debate on deregulation of California s utilities, laying out the fundamental reasons why California should reform its regulatory structure: First, California's current regulatory framework, significant portions of which were developed under circumstances which no longer persist, is ill suited to govern today's electric services industry. Second, the state's current regulatory approach is incompatible with the industry structure likely to emerge in the ensuing decades. It expanded on these conclusions, describing problems of the then current regulatory system: The regulatory program blunts incentives for efficient utility operations. The current regulatory program increases the potential for inefficient investment due to unbalanced incentives governing utility investment options. The current regulatory approach requires many complex proceedings, which increase administrative costs and threaten the quality of public participation and Commission decisions. The current regulatory approach offers utility management limited incentives and flexibility to respond to competitive pressures. The current regulatory approach conflicts with the Commission's policy of encouraging competition in the electric services industry. The Yellow Book laid out four broad options, ranging from limited structural change to restructuring the utility industry through divestiture of utility generation assets and creation of a competitive retail electricity industry. One option included creation of a core/non-core retail market in which a limited segment of consumers could enter direct access contracts with IPPs, bypassing the utilities. Building on the conclusions in the Yellow Book, in April 1994, the CPUC issued an Order proposing a process of restructuring California's electricity industry. Often referred to as the Blue Book, this order envisioned competitive retail and wholesale electricity markets. This and all subsequent restructuring Orders maintained the utilities as monopoly providers of delivery services. 11 California s Electric Services Industry: Perspectives on the Past, Strategies for the Future, February 3, 1993, California Public Utilities Commission. Draft. Do Not Quote. 12

16 The Blue Book laid the foundation for California's subsequent electricity restructuring, proposing several fundamental changes, including replacing cost-of-service regulation with performance-based regulation, wherever regulation was needed. The Blue Book proposed to grant all purchasers of electricity voluntary and direct access to electricity suppliers, in a timephased manner. It envisioned that both the regulated utility and unregulated retail purchases would coexist. Electricity generation would be deregulated. Wholesale prices would be disciplined by competitive wholesale markets. The Blue Book addressed the issue of stranded costs by proposing a financial transfer from the utility customers to the generation side of the utilities, in the form of a limited-time-duration nonbypassable "competition transition charge" (CTC), imposed on each retail customer. The utility would receive all money collected through the competition transition charge, allowing it to recover stranded costs. The total of stranded costs itself depended on the market-determined wholesale price of electricity. The higher the wholesale price, the smaller would be the stranded costs, and the smaller would be the CTC. Thus the CTC would be determined only over time, as cost and price data was revealed in the marketplace. Although the Blue Book laid the foundation for the restructuring, many steps were required to complete the process. Each step seemed to add more complexity to the restructured system. And many of the restructuring design elements changed greatly as the process continued. In May 1995, the CPUC issued a Decision 12 that laid out two broad policy alternatives for organization of restructured wholesale markets and transmission management: a preferred (majority) and alternative proposed policy. The preferred structure was a wholesale power pool, managed by a newly created California Independent System Operator (CAISO). The ISO would dispatch generation based on a dayahead auction and would arrange transmission access for generators that successfully sold in that auction. Under this proposal, management of the grid, dispatch of generators, and wholesale trading would be integrated functions. Wholesale prices for electricity could vary sharply with supply and demand conditions, with risk for both generators and consumers. Risk management would be available only through financial instruments to hedge prices. The CPUC was to take no responsibility for establishing such hedge instruments, but the Decision envisioned private firms creating such markets. Physical, bilateral contracts would be allowed after two years. The alternative policy envisioned consumer choice through physical, bilateral contracts, separate from any pool bidding, to be available immediately. In September 1995, four major participants, including a utility, a group of generators, and two electricity user groups, presented a Memorandum of Understanding (MOU) with their joint recommendations. Although it addressed virtually all elements of the proposed restructuring, the MOU focused on market structure and stranded cost issues. The proposed market structure 12 CPUC Decision D Draft. Do Not Quote. 13

17 combined features of the preferred and of the alternative proposals from May. It proposed a system more complex and less coordinated than envisioned in the May Order. The MOU proposed creation of a Power Exchange (PX), creation of an ISO, and early phase-in of direct bilateral contracts between generators and individual customers or distribution companies. The ISO and PX would be separate entities, operating independently of one another. The PX would develop a visible electricity spot market with transparent electricity prices, open to all suppliers, both within and outside of California. The ISO would manage the grid. Normally these functions, integral parts of a smoothly functioning system, would be tightly integrated into one organization. This structure, ultimately adopted by California, almost assured that the functions would not be well coordinated. It would provide opportunities for energy traders to operate profitably; market inefficiencies could create profit opportunities through arbitrage and through selling financial instruments for managing the increased risks 13. A continued set of hearings and public submissions led to a final CPUC restructuring Order 14, issued in December 1995, often referred to as The Preferred Policy Decision. The CPUC restructuring Order followed the MOU recommendation to separate the ISO (to manage the grid) and PX (to create wholesale markets.) The organizational separation of the two closely connected functions promised to create an extremely complex and untested system. As shall be explained later in this chapter, the resulting inefficiencies and unusual incentives associated with this structure would help make the subsequent energy crisis worse than it would have otherwise been. Like the Blue Book, the restructuring Order proposed to deal with stranded costs through a CTC. The CTC would be designed to allow utilities to recoup virtually all stranded costs 15 by the year This CTC will be discussed more fully at a later point. At the retail level, the CPUC envisioned a system in which consumers would have many options for electricity purchases. Consumers could continue complete reliance on a local distribution company to purchase and deliver electricity or could opt for direct access through bilateral contracts to deliver electricity. Those relying on a local distribution company could agree to either pay the average cost of electricity throughout the year or pay a real time price, a price that varied on an hour-by-hour basis with changing wholesale market conditions. Those paying a real time price could choose hedging contracts with third parties to reduce the risk. The CPUC restructuring Order included one provision that might have led to a price cap on retail electricity prices. The provision would have limited the size of the CTC, assuring that if the CTC otherwise would have increased electricity prices above the January 1966 levels, then the CTC would be reduced or eliminated to limit retail price increases. 13 This seems to be a remarkable public policy concept: create market inefficiencies to make the system profitable for arbitrageurs, traders, and financial markets. 14 Order Instituting Rulemaking on the Commission s Proposed Policies Governing Restructuring California s Electric Services Industry and Reforming Regulation. Decision (December 20, 1995) as modified by D (January 10, 1996). 15 In addition, long term contractual obligations entered into before January 1, 1996 would be recovered over a longer time period. Draft. Do Not Quote. 14

18 Electricity distribution functions would remain with the utilities and would be regulated by the CPUC. The regulated distribution costs would include a separate non-bypassable component of retail rates that would have provided funds for other social goals: a Public Interest Energy Research Program and demand side management programs to promote energy efficiency. This restructuring Order left open many implementation details. Implementation rules were developed through a sequence of CPUC decisions 16. State Legislation: Assembly Bill 1890 Although the CPUC had issued the restructuring Order, such a fundamental reform would be politically more viable if it were the product of legislation, not simply regulatory rulemaking. Soon after the CPUC restructuring Order, the State legislature embraced this role. The legislative process culminated in Assembly Bill 1890 (AB 1890), passed by the California Legislature and signed into law in September It was originally to take effect on Janurary , but was delayed to March 1998 due to implementation delays. The bipartisan nature of the restructuring legislation was striking. Primary leadership for the entire legislative package came from a Democratic member of the State Senate and a Republican member of the State Assembly. The bill passed with no dissenting votes from legislators of either party, a rarity in contentious California politics. A Republican Governor, Pete Wilson, signed the bill. Moreover, the bipartisan legislation built upon an open public process led by the CPUC. The legislative process started from the CPUC restructuring Order of 1995, but modified several central provisions and added its own features. AB 1890, like the CPUC restructuring Order, allowed a utility to include generation, transmission, and local distribution, but would not allow coordinated decision making among these functions. Decision-making and control of the transmission function would reside with the ISO, not the utility owning the transmission lines. The market structure separated local distribution decisions from fossil-fuel-fired 17 electricity generation decisions, so that a utility that both generated electricity from fossil-fuel-fired plants and sold electricity at retail would operate as if two separate companies owned these two functions. This separation of control was accomplished by requiring the utility to sell all electricity it generated using fossil-fuel-fired plants through the PX or the ISO and to acquire all electricity for retail sale through these markets. It could not sell any electricity through bilateral contracts nor keep its own-generated electricity for its retail sales. Market clearing conditions operated independently of the identity of buyers or sellers. Thus the utility would be unable to treat itself preferentially either as a buyer or seller of electricity. In that way, market designers hoped to eliminate, or at least sharply reduce, the market power of the utilities operating as producers of wholesale electricity and simultaneously as dominant users of wholesale electricity, typically transferring the electricity it generated to itself. 16 Particularly important was the 1996 Roadmap Decision CPUC Decision D Hydroelectric generation could still be coordinated with retail sales and would thus provide the utility some opportunity of changing production with changes in load. Draft. Do Not Quote. 15

19 Unfortunately, this critical provision created a fundamental problem of risk management. Initially the PX operated markets that cleared only one day ahead of the electricity use. The ISO included only day-of markets. Thus this rule precluded acquisition through forward contracts, dangerously exposing the IOUs to all short-term volatilities in wholesale prices. In addition, as discussed later, the reliance on spot markets, rather than long-term contract markets increased opportunities for exercise of market power by generators. Only in 1999 did the PX created any markets that operated more than a day before the time of electricity use. AB 1890 partially separated distribution services from retail sales of electricity and confirmed that distribution services would continue to be subject to CPUC regulatory authority. The Act promised to create competition for retail electricity sales by authorizing direct transactions between electricity suppliers and end use customers and by allowing electricity aggregators, while the IOUs would be the default seller of electricity. This system was only a partial separation of distribution services and retail electricity sales, since the incumbent utility would supply both distribution services and electricity, while competitors would supply only electricity. Any customer who wished to pay only one monthly bill or wished to have only one company with which to deal, had no option but to remain with the incumbent utility. This gave the utility a great competitive advantage as seller of electricity. One option would have been to require a complete separation of the two functions, by allowing companies to either provide distribution services or sell retail electricity. Such a system, which has been adopted in Texas, is discussed more fully in Chapter 11. Under such a system, the incumbent utility might start with a competitive advantage if it chose to sell retail electricity rather than distribution services. But since the utility would have no lasting structural advantage, over time any competitive advantage could be expected to dissipate. But such a system was never seriously considered in the restructuring, perhaps because the utilities themselves played such a large role in the system design. Direct access was to start simultaneous with the initiation of the PX and ISO. Aggregation would be allowed by private sector marketers, as long as individual customers could freely choose remaining with the local utility or purchasing electricity from the aggregator. Aggregation by cities or other public agencies would be generally allowed. AB 1890 transformed the cap on the CTC into a retail price cap, a critical change. The law required that bundled electricity plus distribution services prices for residential and small commercial customers be reduced immediately by at least 10% below their June 10, 1996 levels. The transition period, during which the stranded costs would be recovered, was made much shorter than proposed under the CPUC restructuring Order. This period would end no later than March 31, 2002, or whenever the stranded costs had been fully recovered, whichever came sooner. The IOUs were not guaranteed that they would recover their entire stranded costs; they would be at risk for costs not recovered by March 31, Draft. Do Not Quote. 16

20 To recover its stranded costs, each utility would propose to the CPUC a cost recovery plan that included the capped retail prices. It was anticipated that wholesale costs would decline. Under this assumption, the differential between capped retail prices and costs 18 would increases, leading to financial accumulations each year and stranded costs would be recovered over a limited transition period. But wholesale prices could not be predicted well. AB 1890 included no statement of what might happen if the wholesale price exceeded the capped retail price. Perhaps if that issue had been clearly resolved, various parties would have more fully understood the financial risks and would have been able to take appropriate steps to manage risk once the law had been passed. A safeguard, at least in theory, was that AB 1890 imposed no restrictions to stop the CPUC from modifying or abandoning the stranded cost recovery plan, if so requested by the utility. The CPUC could agree to reduce the amount of stranded costs to be recovered to just the amount already recovered, thereby terminating the requirement that retail rates for that utility remain at their price-capped level. AB 1890, like the CPUC restructuring Order, kept organizationally separate the PX and the ISO. It directed the CPUC to work with the utilities to develop a PX, but it included no further guidance about the competitive auction, the bidding structure, or length of the advance period during electricity could be purchased. AB 1890 described the technical functions of the ISO but gave no guidance to its market functions. Implementation issues were left to the CPUC, the ISO Board, and to FERC. AB 1890 directed CPUC to work with the utilities to obtain authorization of FERC for the creation of ISO and the PX. Even before its passage, in April 1996, the three IOUs submitted requests to FERC 19 asking approval of those restructuring elements subject to FERC jurisdiction. FERC largely approved these proposals and in 1997, FERC authorized the first limited operation of ISO and the PX. The set of regulatory changes, culminating in AB 1890, promised to fundamentally change the electricity system from one strictly regulated from cradle to grave, into one in which market forces would play the primary role, at least once each utility passed its transition period. Wholesale markets were intended to allow competition to determine supply, demand, and prices of electricity in wholesale transactions. Although analysts envisioned that most small retail customers would continue obtaining their electricity bundled with distribution services, sold by regulated utilities, the seeds for a competitive retail market were planted. Given the anticipation that wholesale prices would fall, the transition period was expected to be just that, a smooth transition to a competitive market, a transition in which the problem of stranded costs would be solved. 18 The CTC would be equal to this difference. The CTC was a non-bypassable charge that would be imposed on all retail sellers in addition to the utilities. 19 FERC Dockets No. ER and EC The applications were filed after the CPUC restructuring Order but before passage of AB These applications were approved only after AB 1890 was signed into law. FERC took due note of the passage of AB 1890 during its proceedings. Draft. Do Not Quote. 17

21 After Passage of AB 1890 But the process was not yet completed with passage of AB The next year a group of consumer advocates gathered signatures to qualify what became Proposition 9, voted in the 1998 California ballot. Proposition 9 would have turned back the restructuring and would have reinforced a system of tightly regulated electricity transactions. Until its defeat, Proposition 9 created much uncertainty and reduced the incentives for private companies to invest in new electric generating facilities. By creating significant delays in the capital investment associated with investing in new generating plants, Proposition 9 increased the risks that were inherent in this newly restructured system. It was not until the defeat of Proposition 9 in November 1998 that assured AB 1890 would be allowed to operate. Like the restructuring order, AB 1890 was simply a framework, not detailed designs for system implementation. The structure of the PX and ISO, as well as of the markets they were to operate, was left to stakeholder committees and the CUPC in California to design and left to FERC to approve or disapprove the designs. Careful delineation of the jurisdictional split between the Federal regulators and the State regulators was left to the various parties to work out. Working out jurisdictional conflicts within the short time frame was close to hopeless. That jurisdictional split has never been fully resolved. AB 1890 came into effect (still under the cloud of Proposition 9) only eighteen months after it became law. With such a tight implementation schedule, the original applications to FERC were filed while the California legislature was still considering AB Once FERC approved the applications, the large size and diversity of the part-time stakeholder boards made it difficult, if not impossible, to seriously rethink or revise the original structure. The following examines the various system components, in more depth. Wholesale Markets under the Restructured System The California Power Exchange (PX) The IOUs and the CPUC developed, and FERC approved, plans for the PX and for the wholesale markets that the PX would manage. The PX organized a set of competitive auctions, open on a nondiscriminatory basis to all suppliers 20. The PX initially established one-day-ahead and day-of wholesale markets for electricity. Only much later did it establish markets that allowed contractual agreements extending longer than one day in advance. For both the one-day-ahead and day-of wholesale markets, the PX accepted bids to sell electricity hour-by-hour and bids to purchase electricity hour-by-hour. Hourly prices were determined on a market-clearing basis with all buyers paying the same market-clearing price (MCP) and all sellers receiving the same MCP. All bids to sell with offer prices lower than or equal to the MCP and all bids to purchase with offer prices greater than or equal to the MCP 20 The three large IOUs were required to sell all their remaining generation through the PX. For all other entities, use of the PX was optional. Draft. Do Not Quote. 18

22 would be accepted; all others would be rejected. This one-price auction system was designed to simulate a perfectly competitive commodity market. There were apparent alternatives to such a one-price auction system. One alternative, in principle, would have been to set up a single commodities futures market. Agents would enter bids to buy and to sell, prices would adjust, and MCP would be reached. Prices for a given delivery time would be free to adjust as the delivery time grew near. But adjustment process would take time; electricity markets would have to adjust on a much faster time scale than normal commodity markets. All adjustments would have to be completed, starting at most one day before the day of delivery. And there were 24 separate hourly markets for each day. This can be compared with the future market for natural gas, which sets a price for delivery for an entire month, not 720 different prices (24 x 30) for each month. Although development of such a structure would have fully integrated the various markets and would have reduced the opportunities for market gaming, it appeared (and still appears) too complex to be successfully implemented. A second alternative would be to design the system to pay bidders what they bid, rather than to pay them the MCP. The total cost of all purchases would be averaged, and the buyers would each pay the average bid price. Some have argued that a system of paying on an as-bid basis rather than on a market-clearing basis would result in smaller total payments by the buyers of electricity. After all, those bidding to sell at prices below the cut-off price would not receive the cut-off price but would receive only their bid prices. The fallacy of that reasoning is that it implicitly assumes that the sellers of electricity would offer the same bids under an as-bid system as they would under a market clearing system. But under an as-bid system, each firm makes the most profit by guessing the cut-off price and bidding at or just below that price, as long as the cut-off price is at least as high as its marginal cost. Thus, even in a competitive market, suppliers would not bid at their marginal costs. There was a difficulty with an hourly auction system. Many generating plants, typically operating as base-load plants, have long and costly periods for ramping up from no production to full capacity. These plants might be profitable to operate if they received at least a particular price, say $30/MWh, for a large fraction of the day or for all of the peak period during a day. However, if they were to operate for only a few hours, even at a higher price, say $40/MWh, they might not be profitable to operate, since the fixed costs of ramping up could be greater than the profit earned during those limited hours. For such plants, their offer price at any hour must depend on whether they would be generating electricity at the other hours during the day. For such plants, bidding based on unit commitments, commitments of the unit to operate for long blocks of time, would be more appropriate. But that solution was not chosen. Other plants, for example nuclear, had no option to cycle or reduce production, but rather had to be run as baseload. But for these plants the problem could be easily solved by bidding in their production at zero price, thereby assuring that they would sell their output at the MCP. Draft. Do Not Quote. 19

23 The California Independent System Operator The CAISO was given the responsibility for managing the transmission grid and for assuring that resources were available to assure safe operations of the grid, including assuring that there would be sufficient quantities of electricity available at all times. ISO Real Time Markets To perform these functions, the CAISO collected the schedules of electricity to be generated and of the electric loads to be served by the IOUs, by the Munis, and by a host of other entities 21, on an hour-by-hour basis. Forty-some scheduling coordinators or SCs reported their schedules to CAISO. CAISO integrated these schedules and assured that these schedules collectively did not overload any parts of the transmission grid or were in any other ways not feasible. If they were not feasible, then the CAISO could order adjustments to dispatch generation feasibly. Each scheduling coordinator was required to submit a balanced schedule in which the total loads and resources were equal to one another 22. That is, the total projected use of electricity at each hour and the total generating resources to provide that electricity were required to be equal to one another for each submission by every SC. Thus, in theory, the sum total of loads and resources would be balanced for the system, absent CAISO-ordered adjustments. However the participants in the market could not perfectly project the electricity needs and the loads and resources could become unbalanced. To correct such imbalances, the CAISO would run a real time energy imbalance market, buying and selling electricity after the PX day-of market had closed. Deviations between predicted and actual supplies and demand would be corrected on this imbalance market. System designers expected these differences to be small. The CAISO tariffs did not include penalties for imbalances, even for extreme imbalances. As discussed later, there was an incentive to mis-schedule. Imbalances turned out to be large fractions of scheduled power. CAISO was responsible for purchasing and selling electricity on behalf of the SCs. Because it was engaging in these balancing transactions, its decision rules were central to the real-time market for electricity. One decision rule was that the CAISO would acquire sufficient electricity to meet the loads it predicted. These predicted loads did not depend at all on the market price. CAISO would not reduce electricity acquisitions, even if the acquisition prices became very high. Possible demand-side adjustments were excluded from the market price determination. The mentality of 21 AB 1890 applied only to the IOUs and not the munis. Under the new system, CAISO does not control all load, only the load of the IOUs and others who wish to join. Due to high costs of the CAISO, many of the Munis have declined to use the services of the CAISO. 22 More precisely, the sum of all loads and the sum of all generation submitted by a scheduling coordinator must be within 2 MW of each other. Draft. Do Not Quote. 20

24 reliability at any cost at a later time led to exceeding high costs for all electricity users in California, costs often higher than the reliability was worth to them. A second decision rule was that CAISO would reject bids above some wholesale price cap level. This wholesale price cap (discussed later ) was during different times $250/MWh, $500/MWh, or $750/MWh. Together, these two rules implied that the demand function in the CAISO real time market was completely price independent or inelastic for prices up to the cap but perfectly elastic at the cap level. These rules together increased the volatility of prices on the real-time market, up to the capped price, but kept MCPs from exceeding the cap. CAISO Ancillary Services Markets The CAISO initially paid sellers of ancillary services to make their generating units available, should they be needed. If these units were needed, the sellers were paid for the electricity generated 23. Every load-serving entity or LSE, typically an electric utility, was responsible for its proportional share of ancillary services. Each scheduling coordinator could choose whether to provide its share of ancillary services or to have these services purchased on its behalf by CAISO. For those ancillary services not self-supplied by the scheduling coordinator, CAISO managed a singleprice bidding system that operated in day-ahead and hour-ahead periods. The CAISO could obtain additional ancillary services through supplemental bids offered during the hour the reserves were needed. Divestiture of IOU Generating Assets AB 1890 and subsequent CPUC rulings sharply reduced the degree of vertical integration in the industry. However, there was still a concern that common ownership of generation and retail functions would give the utilities wholesale market power 24, would make it difficult to operate a competitive wholesale market, and would lead to high wholesale prices. To address that concern, several options were considered. One option was to allow utilities to continue acquiring electricity directly from their own generators, as well as buying it from non-utility generators, either through organized markets or through bilateral contracts, keeping some vertical integration. 23 Initially this system was set up so that these units would be paid for the electricity plus paid for the ancillary services. However FERC later required ISO to change the rules so that generators would be paid for either generating electricity or providing ancillary services but not both. 24 In general, vertical integration itself does not create market power. However, as long as the incumbent utility remains dominant in the retail market, vertical integration allows it to extend its retail market power to the wholesale market. Even if the retail activities of the utility are regulated, if the utility s acquisition of bulk power is not perfectly regulated, then it can extend its market power that is derived from the retail market into the otherwise competitive wholesale market. Draft. Do Not Quote. 21

25 A second option was to require the utilities to divest themselves of their generating assets, at least their thermal assets. The CPUC ultimately implemented a two-fold solution. First, the CPUC required the utilities to divest themselves of 50% of their generation assets and provided financial incentives to divest the remainder. Second, all remaining thermal generation owned by the IOU could be sold only through the PX or the CAISO 25. Together, those rules would assure that the PX and the CAISO markets would include large volumes of transactions and that IOUs would be precluded from any meaningful self-dealing between their wholesale and retail operations. The incentives for divestiture were successful. As of 2000, only 29% of the electricity sold in the State was generated by the IOUs vs. 44% generated through plants that had been divested and owned by non-utility generators. Forward Wholesale Contracts for Electricity (or Lack Thereof) Divestiture of plants, however successful, created new problems. As the IOUs divested their generation assets, there were linked agreement both to sell the generator and to purchase electricity under a long-term contract from that generator, the CPUC feared that there would be financial incentives to distort the selling price and the long-term sales price to increase profits. Guarding against this potential would require more regulatory oversight. In addition, there was a fear that long-term contracts could simply substitute for a utility ownership of the generators, hence preventing the emergence of a competitive market. Moreover, potential new entrants into the wholesale market might be discouraged in the same way as would be the case absent divestiture of the assets. In order to assure that the PX markets would not be too thin, there was a desire to limit the long-term contracts at the wholesale level. As indicted above, the CPUC required utilities to acquire all electricity, not already under longterm contract, through the PX or CAISO. Since the PX and the CAISO markets operated no more than one-day ahead of the electricity delivery, this requirement effectively prohibited the utilities from entering any forward contracts. Perhaps intentionally, perhaps unintentionally, these rules together made the California IOUs overly dependant on a spot market, a market that turned out to be very volatile and subject to market gaming. The IOUs, realizing their potential vulnerability unsuccessfully tried as early as 1999 to gain the right to procure electricity on a longer term basis. In March 1999, SCE filed an application for a pilot program under which it could enter traditional power purchase agreements for electricity and for capacity. But the CPUC denied the application. In mid-1999 the PX applied to organize a block forward market, FERC approved the application, and the CPUC approved the request by SCE and PG&E to participate in that market. But the block forward market allowed contracts for no more than one year. However, until August 2000 the utilities had no right to enter bilateral contracts. This, along with the retail price controls, during the energy crisis proved to be fatal flaws in the system. 25 More precisely, no costs could be recovered for this generation unless the electricity were sold through the PX or the ISO. Draft. Do Not Quote. 22

26 Forward contracts could have substantially reduced the risk of large changes up or down in the acquisition cost of electricity. Utilities could have guarded against or limited the high risk of large fluctuations in the wholesale electricity price. But, inexplicably, that road was not taken by the CPUC. Wholesale Markets: In Summary AB1890 and CPUC rules created a complicated set of wholesale markets imperfectly coordinated with one another. These markets were given monopoly or near monopoly status and thus utilities could not escape problems associated directly with these markets. Risk management options were taken from the IOUs through divestiture and through reliance on PX s wholesale spot markets. Therefore, volatility in the wholesale markets was nearly assured. The interplay of these various markets, the resulting bidding strategies of utility-buyers, private generators, marketers, and municipal utilities, and the responses of the CAISO and the PX personnel were all untested at the time of the restructuring. In addition, there remained opportunities for exercise of market power by even those generators with small market share. Retail Markets Creating competitive retail markets was seen to be even more of a challenge, even though there had been extensive experience in other nations. An ultimate goal was to have a competitive retail market for electricity. However, at least two factors stood in the way: retail market power of the incumbent utilities and risk management. Local distribution companies had a natural monopoly for the delivery services, the wires, transformers and control systems 26. In addition, in the short term they could be expected to have a significant degree of market power for the supply of electricity itself, since electricity had always been bundled with the delivery services. Unless retail sales of electricity were fully unbundled from monopoly delivery services, the issue of retail electricity market power would remain. The new system partially decoupled delivery services from retail sales of electricity. Delivery services would still be provided by utilities, as monopoly franchises, earning a regulated fee. Unbundled electricity could be sold by aggregators or generators and delivered by the local utility. Under the new system, in principle, any customers could enter bilateral contracts with electricity suppliers and therefore to bypass the incumbent IOUs for electricity, even though the CTC could not be bypassed. But the decoupling was only a partial decoupling. The local utility could sell electricity bundled with distribution services. IOUs would operate as regulated retail sellers of electricity, subject to review and control by the CPUC. As discussed above, during the transition period, AB 1890 imposed price caps for retail electricity sales by the incumbent utilities during a transition period. This price cap created a 26 The CPUC had jurisdiction over onlyious, and thus they were treated differently from the Munis. This discuss is based on the changes for the IOUs. Draft. Do Not Quote. 23

27 dilemma. On the one hand, the IOUs were required by AB 1890 to reduce electricity prices for residential and small commercial customers by 10%, as discussed above. On the other hand, the CTC magnitude was to be chosen so that all stranded costs could be recovered over a small number of years, although recovery was not guaranteed. Thus with the CTC, the retail price of electricity alone would be approximately equal to the recent historical electricity component of the bundled retail price, not below that level. The dilemma was resolved through a financial instrument. The utilities were authorized during the transition period to issue rate reduction bonds to finance the difference between their cost for electricity (wholesale price plus CTC) and the price capped retail price as well as to refinance some of their existing capital equipment. These bonds would be repaid once all stranded costs had been recovered and the CTC was no longer in operation. This plan implied that the price reduction the customers thought they were enjoying would be repaid in later years. The IOUs would be default sellers of electricity, available for everyone who wished to purchase retail electricity from these utilities. Their price-capped rates would be available for all customers, even those that switched to other retail suppliers but subsequently chose to return. This implied an asymmetrical relationship between the utility and the new competitors: the new competitors could choose whether to take new customers, but the utility had no choice whether to take new or returning customers. New entrants could create distinctions among electricity delivered from different sources. For example, they could sell green electricity, advertised to be generated entirely or primarily by renewable sources. But this component of the market would necessarily be small, because most of the renewable electricity was being sold under contract to the large electric utilities. For residential customers, entrants could bundle energy efficiency measures with electricity to help consumers reduce the overall cost of obtaining energy services (e.g., warmth, lighting, cooking, clothes drying, refrigeration). But the IOUs themselves were offering these services using public benefit charges included in non-bypassable surcharges within the delivery fees. This limited the ability of new entrants to attract customers through provision of energy efficiency services. New entrants would find it very difficult, if not impossible, to compete on the basis of price, given that their customers would be required to pay the CTC and given that the utilities were selling electricity at retail rates 10% smaller than the pre-restructuring retail rates, with the reduction financed using the rate reduction bonds. However, they could compete on the basis of price if they marketed electricity primarily to those customers whose loads were less time variant than typical loads, for example some industrial customers. For such customers, a new entrant could save money on the wholesale purchases of electricity and might be able to sell electricity at a lower retail price than did the incumbent utilities. But, new entrants would have to pick customers carefully. In principle, some retailers could provide higher reliability or interruptible service for particular customers with differing needs. However, because the utility would deliver the electricity, it was not clear that an individual electricity retailer could offer such services without full cooperation from the utility providing the delivery services. Draft. Do Not Quote. 24

28 Residential retail customers typically would need a reason to switch their purchases of electricity from the well-know incumbent utility, particularly if they would remain dependent on that utility. But the retail restructuring gave very little room for new entrants to offer their potential customers any good reasons to switch. Thus, although the ability to bundle some services with the electricity or to offer an enhanced quality of electricity (e.g., more green, more dependable) provided some opportunities for new entrants, at least during the transition period, these opportunities were quite minimal. Finally, the IOUs during the transition period were not simply passive default sellers of electricity, indifferent to how much they sold. The more electricity they sold, the more rapidly they would collect funds paying their stranded costs. And because they were at risk for any stranded costs not recovered by March 31, 2001, the more electricity they sold, the higher would be the probability that they would recover all of their stranded costs. Thus the IOUs had a financial incentive during the transition to keep as many customers as they could themselves and to limit the role of new retail competitors. Thus, unsurprisingly, the success of retail competition was relatively limited, particularly for sales to residential and small commercial customers. The only significant success of retail competition was for large industrial users of electricity. Direct access contracts between these large users and independent generators of electricity represented the only significant amount of non-utility-sold electricity. Municipal Utilities California's many municipal utilities, serving 22% of California's customers, continued operating as they had prior to the restructuring. Each municipal utility had a governing board, either appointed or elected within the municipality, responsible for managing the utility to benefit residents. Typically, municipal utilities were expected to cover their costs through sales of electricity. The governing boards retained the ability to increase retail prices at which the municipal utility sold electricity, if the need arose. These utilities typically purchased electricity using a mix of short, medium, and long-term contracts so they were hedged from rapidly changing wholesale prices. Municipal utilities, therefore, differed sharply from the IOUs in that they retained all capabilities to manage their risks. Market Issues in the Restructured System PX and CAISO Market Clearing Price Relationships The electricity that could be bought and sold on the PX markets was exactly the same electricity that could be bought and sold on the CAISO markets. Thus, the PX markets and the CAISO markets were alternative venues for market participants. Buyers could choose in which venue to make their purchase commitments just as sellers could choose in which to sell their electricity. The tight linkage created important market relationships. Draft. Do Not Quote. 25

29 If firms could substitute between sales/purchases on the PX and the CAISO, the possibility of arbitrage implied that prices on one venue closely tracked prices on the other. If generators expected the price on the real time market to be higher than the price on the PX, they would choose to sell on the higher priced real time market, driving up prices on the PX and down on the real time market, until the expected price difference became small enough to not encourage additional shifts 27. Incentives were equivalent for wholesale purchasers of electricity. The CAISO tariffs, however, if followed in letter and in spirit, limited the ability to substitute between the markets and tended to increase market inefficiencies. In particular, as discussed above, each SC was required to submit a balanced schedule in which the total loads and resources were equal to one another. The balanced-schedule requirement implied that a utility that needed to acquire electricity would be required to purchase in the PX even if it expected prices to be significantly lower in the CAISO real time market. But utility managers understood that it would be most profitable to acquire electricity from the market with the lower expected price 28. There could be a strong financial incentive to violate the balanced-schedule requirement and to submit an unbalanced schedule if significant price differences were expected among the various markets. This conflict between the incentives for unbalanced schedules and the requirements for balanced schedules could be resolved through systematically biasing load projections. By systematically underestimating expected load, a utility could submit a schedule that would in fact be unbalanced but that appeared on paper to be balanced. Thus it could meet the letter, but not the spirit, of the requirement for balanced loads. This became a routine practice for utilities, thus implying that prices on the PX and the CAISO would tend to equilibrate, at least statistically. The price equilibration between the PX and the CAISO implied that wholesale price caps in the CAISO real time markets (described above) would limit market prices in the PX, even though there were no formal price caps in the PX markets. But price caps can create shortages. When price caps were binding, utilities might not be able to satisfy all remaining electricity demands through real-time CAISO purchases. Normally the likelihood of a shortage in the real time market would provide motivation for the utility to purchase in the PX, assuring acquisition of adequate electricity. However, under the CAISO rules, all utilities would equally share any shortage, independently of whether they had purchased enough electricity on the PX. Thus, bidding in the PX above the CAISO price cap would cost the utility more but would provide no additional protection. Therefore, there was no incentive for any utilities to bid above the price cap on the PX and the CAISO real time price caps effectively controlled the maximum prices on the PX. 27 During the electricity crisis the ISO real time price tended to be somewhat higher than the PX price. 28 For clarity of explanation, this statement ignores the possible exercise of market power by utilities. A utility might choose to purchase some electricity in the ISO real time market even if prices were higher in the real time market if by doing so it could exercise its market power as a purchaser of electricity and drive down prices in the PX sufficiently. Draft. Do Not Quote. 26

30 PX Bidding Strategies for Electricity Generators The PX market, as discussed above, was one-price auction system. The theory of bidding in such a market, for a supplier of electricity operating purely competitively, is that the optimal bid must equal the incremental cost of the electricity 29. The incremental cost of an electricity sale on the PX includes the opportunity cost of the PX sale. Since electricity sold in the PX could not be sold in the subsequent CAISO real-time or ancillary services markets, the opportunity cost of PX sales would be the expected price for selling in one of the CAISO markets. Thus the correctly-calculated increment cost for selling electricity in the PX is the higher of three amounts: a) the incremental cost of generating the electricity, b) b) the expected value of market-clearing price in the subsequent CAISO real-time market, or c) c) the expected value of returns 30 from selling in the subsequent ancillary services market. Thus, because the subsequent CAISO real-time market, for a competitive supplier of electricity, optimal bidding into the PX implied that the optimal bid would be equal to the higher of the three amounts 31. For a particular firm, the optimal bid would not be the marginal cost of generating electricity unless it expected the CAISO markets to clear at or below its marginal cost of generating electricity. Whenever, the CAISO s MCP was expected to significantly exceed the generation cost for an individual supplier, competitive optimal bidding into the PX would significantly exceed the generation cost 32. A similar analysis holds for bidding in the ancillary services market and all CAISO markets that clear before the CAISO real-time market clears. Only in the last of the markets to clear will optimal competitive bids not include an opportunity cost associated with the existence of other markets 33, since at that time, the opportunity to bid in the previous market would have passed. 29 If the bid is above the marginal cost and the market clearing price is below the bid, but above the incremental cost, then the supplier would lose a profit opportunity. See Carlton and Perloff (1994). 30 These returns include the payment for supply the services plus the expected value of revenue from supplying the electricity, if the electricity were needed. 31 The optimal bidding for a competitive firm is even more complex than this suggests. In addition to the PX dayahead market, there is a day-of market, five different ancillary services markets that clear sequentially, and the ISO real-time market. After the day-ahead market closes and the real time market closes there is a post-close quantity match that allows additional sales. The opportunity cost would include the price expected from the highest of these markets. But in addition, if each of these price are uncertain, but not perfectly correlated, then the opportunity cost will be greater than the maximum of the expected values. 32 A simple numerical example will illustrate this point. Assume that the cost of natural gas and operating expenses for producing electricity were $50/MWhr, but a firm expected the price to clear at $200/MWhr in the ISO real time market. If the firm bid $50 in the PX, and the PX MCP turned out to be $180, the firm would be paid $180. This amount would be $20 less than the $200 it could have expected from the ISO real time market. Thus the optimal bid in the PX of a firm operating perfectly competitively (as a price taker) would not be $50, but would be $200. However, if that firm were not able to sell in the PX, then the optimal bid in the ISO real time market would be $ This comment is not meant to preclude the possibility that there could be an opportunity cost associated with environmental constraints on generation from the plant. Draft. Do Not Quote. 27

31 However, the system did not preclude the exercise of market power. A firm with market power could submit bids to the PX and the CAISO above the correctly-calculated incremental cost, under the expectation it could increase the MCP by so doing. A high price may imply that not all of the electricity from the firm s generators could be sold. The sure-to-be-rejected bid on a portion of the available power could sacrifice profits for that portion but might increase prices on all the electricity it did sell. If the gain on the rest of the electricity were great enough, then bidding a high price only some of the available generation could be a profitable strategy. Such bidding strategies could be profitable as long as the firm had a large enough market share. Exercise of market power might be possible for some firms, but whether it would be generally be possible was not known at the time of the restructuring. What was known, however, was that those firms that did have market power were likely to exercise that power, since actions to maximize profits are normally to be expected for private-sector firms. It was less likely that firms would have market power when the system were well below capacity or when additional electricity could easily be imported into California; it was more likely when capacity was constrained, as proved to be the case in Low elasticity of demand for electricity at the wholesale level makes the existence of and the exercise of market power more likely. During the time period, California s particular regulatory rules for the IOUs 34 assured a near-zero electricity demand elasticity. Both opportunity costs and market power lead to bid prices in the PX that can be above the marginal generation costs. Thus the simple observation that PX bid prices significantly exceed marginal generation costs does not tell whether the firm is operating competitively and including opportunity costs or is exercising market power 35. Because both opportunity costs and market power will lead to PX bid prices higher than marginal generation costs, perhaps very much larger, it has been particularly difficult to use PX bidding patterns to infer whether firms were exercising market power in any particular instances. Market Risk Wholesale Market Risk The restructuring of markets created significant economic risks for the IOUs because the wholesale market for electricity could be very volatile, long term electricity supply contracts were precluded by regulation, and the utilities faced retail price caps. Wholesale price volatility could result from either energy limitations (limitations on the availability or cost of the energy used to generate electricity) or generation capacity limitations (limitations on the capital equipment to generate electricity.) Energy limitations could increase the price of electricity at all times, peak and off-peak. Generation capacity limitations could 34 As discussed at another point in this chapter, three factors led to almost zero demand elasticity: the ISO mentality of reliability at all cost, the normal regulatory lag between increases in wholesale price and retail prices, and the retail price control imposed in California. 35 In the example of footnote 32, one might observe the firm bidding $200 in the PX and conclude that a bid well above $50 would be evidence that the firm was attempting to exercise market power. Such a conclusion would be incorrect. Unfortunately, that logic mistakenly underlaid public statements and litigations after the electricity crisis, asserting that such bids were clear evidence of market power or of market gaming. Draft. Do Not Quote. 28

32 increase the price of electricity when the system was operating at or near peak, but would have little impact on prices when the system was operating in off-peak periods. Either limitation, if it became binding, could drive the incremental cost of generating electricity up sharply and thus increase the MCPs. The short-run elasticity of demand for electricity is typically low for end-users. In addition, utilities do not charge customers real-time prices, but rather average-cost based prices, thus reducing consumer response to wholesale price changes. The wholesale price elasticity is further reduced because of retail price control; wholesale price increases would not be passed on to retail customers, even over an extended time. Measured at the wholesale level, the elasticity of demand for electricity would be near zero. Thus, small reductions in supply or increases in demand could sharply increase prices. The role of demand elasticity is illustrated in Figure 6. The supply function is reduced, from the dashed gray line to the dashed black line. If the demand elasticity is very low, as represented by the dashed black line, then the new price will be much higher than the initial price. If, on the other hand, the demand elasticity is higher, the new price will increase by only a much smaller amount. Under the restructured system, all spot sales of electricity would sell at a price equal to the MCP, a price that would be at least as costly as the marginal cost of generating electricity from the most expensive unit being used. Moreover, the utilities were buying most of their electricity on these spot markets, because they had divested at least half of their generating assets and had not entered into long-term electricity supply contracts. Thus, total expenditure for acquiring electricity was very uncertain; risks were large. Draft. Do Not Quote. 29

33 Electricity Price New Price, Very Low Demand Elasticity New Price, Higher Demand Elasticity Original Price Quantity of Electricity Figure 6. Price Increases in Response to Supply Reductions, Two Demand Elasticities The mix of energy sources for generating electricity increased the risk that the system could approach energy limitations. During peak demand periods, much of the primary energy used for electricity generation in California was natural gas. (See Figure 2.) And the infrastructure of pipelines to move natural gas in California was extremely limited, as was the capacity of pipelines to bring natural gas into California. Therefore, the risk stemming from limitations on natural gas availability or from volatile natural gas prices was great. The second largest source of electricity supply was through hydropower, which is dependent on unpredictable rainfall during the previous year. And much of the electricity imports into California were derived from hydroelectric power in the Pacific Northwest, also subject to wide variations from year to year. The risks were highly correlated: low rainfall in California could be expected during those years of low rainfall in the Pacific Northwest. Divestiture, absent long-term power purchase agreements, increased the allocation of the risk to IOUs. If the utilities had not divested generating capacity, they would have faced cost variations that changed with the average generation cost. But, because they had divested the assets and were purchasing electricity on the spot market, they would face cost variations that changed with the marginal cost of electricity, because the MCP changed with the marginal cost of electricity generation. With divestiture, then, if the marginal cost of generation increased while the average cost of generation increased by only a smaller amount, the IPPs that owned the divested units would enjoy increased profits. At the same time the utilities would see reductions in their profits or would face losses. Absent divestiture, those profits would be earned by the generation side of the utilities while the losses would be incurred by the distribution side. The overall variation in profit would be reduced because losses by one side of the utility would be partially balanced by Draft. Do Not Quote. 30

34 gains to the other side 36. In general, the marginal cost is much more volatile than the average cost. Thus divestiture exposed the IOUs to increased profit risk whenever there was risk of wholesale market price variations. The absence of long-term power purchase agreements (beyond the pre-existing contracts) increased the risk that firms would possess and exercise market power. Generators could exercise market power by bidding above their (correctly calculated) costs in the PX and CAISO markets or, equivalently, not bidding at all into these markets. By reducing the capacity that was made available 37, the MCP would be increased for all other electricity being sold on the spot market. If the MCP increased enough, it would be more profitable to reduce supply of electricity. However, if a large share of the electricity were sold under either long-term or medium-term contracts, then this incentive would be sharply reduced. Consider two situations for a given generator, assuming that all other generators are similarly situated. In the first situation, the generator sells 100% of its electricity on the PX and/or CAISO spot markets; in the second situation, it sells 90% of its electricity on term markets and only 10% on the PX and/or CAISO spot markets. Now for the two situations, examine the economics for a generator deciding whether to reduce the supply it offered to the market at the competitive price, either by bidding well above its costs or by physically withholding supply. In the first situation, the firm could influence the price for all of its power, since it would have no term contracts. For illustration, assume it reduced its production by 50% of its total capacity. Assume that this reduction in supply to the spot markets would increase price by X%. If X% were large enough, then it would be profitable to reduce supply. The firm would have market power. In the second situation, as the firm estimated its changes in profits, it would recognize that the price of the contract power would not be influenced by the MCP in the PX or CAISO markets. It could influence the price for only the 10% it sold on the PX/CAISO markets. If it reduced its supply by that entire 10% available for these markets, it would make no profit; thus, its maximum profit could be achieved by a reduction in supply of less than 10% of capacity. For illustration, assume it reduced its production by 5% of its total capacity. Assume that this reduction in supply to the spot markets would increase price by Y%. If Y% were large enough, then it would be profitable to reduce supply. In the first situation, the firm would reduce supply by 10 times as much as would in the second situation. If all generators were similarly situated, the spot markets would be 10 times as large in the first situation. Thus, it might appear that the firm s reduction in supply would be the same percentage of the market in either situation and therefore that the price change would be the same 36 This discussion assumes that the retail price control remained in effect. Without retail price control, cost variations could be passed on to the consumers through changing rates and the utility would face very little risk in either case. 37 Note that such bidding strategies for natural-gas fired generation made the capacity constraint tighter, but would have only a small impact on the energy constraint. The natural gas that would otherwise be used by the generator would still remain in the system, either to be used by the particular generator at another time or sold to another generator. The capacity not used at that time would not be available any other way. Draft. Do Not Quote. 31

35 in both situations, that is that X% and Y% would be identical. Thus, it might seem that there would be the same amount of market power in either situation. But such a conclusion would be incorrect. In the second situation, the 90% of the electricity, although under contract, could be offered on the spot market for resale by the contractual purchaser. Thus, the elasticity of supply of electricity on the spot markets would be increased by the existence of the 90% of electricity under contract, even though the overall elasticity of supply and demand for electricity would be unchanged. Thus the Y% would be substantially smaller than the X% and it is substantially less likely that firms would have market power in the second situation. The difference between X% and Y% can be estimated more completely, by adding an assumption that all contractual buyers act as profit maximizing entities willing to resell on the spot markets whenever that would be profitable. In that case, the amount of electricity available on spot markets in the second situation (although not necessarily actually bid), would be the same as in the first situation. But the supply reduction from a firm deliberately withholding supply would be 10 times as much in the first situation. Thus, a rough approximation is that the price increase in the first situation (X%) would be 10 times as large as the price increase in the second situation (Y%). Thus with or without term contracts, a firm could bid above its cost, but the incentive to do so would be much smaller with term contracts. In fact, even if it were very likely that a firm without term contracts would have market power, that same firm with term contracts is very unlikely to have market power. The great reliance on the spot market sharply increased the risk that generators would possess market power and thus the risk that generators would exercise market power. The reliance on spot markets thus led to an important risk that wholesale prices would be sharply increased through exercise of market power. Retail Market Risk The deregulated retail market created incentives for consumers to shift back and forth between regulated utilities and the unregulated retailers if market conditions changed, at least once the transition period were over. Retail customers could choose to buy from competitors of the IOUs whenever those competitors offered electricity at a more attractive price and buy from the utilities when the opposite was the case. The IOUs were obligated to serve all customers, including those that switched back from an unregulated competitor. The retail price that they would charge would be based on the average cost of their acquisition of electricity, once the transition period were over so that there no longer was a retail price cap. If spot prices were lower than the average prices of electricity acquired by utilities, unregulated firms could sell electricity at the lower price and customers would shift purchases away from regulated utilities. With fewer customers, regulated utilities would purchase less electricity on spot markets, thereby increasing their average cost and the prices charged to remaining customers. The price increase would cause more customers to leave the regulated utility, thereby Draft. Do Not Quote. 32

36 further increasing the price for the remaining ones the so-called death spiral. If spot prices turned out to be much higher than the average acquisition cost, customers would abandon the unregulated competitors and purchase electricity from utilities, leading to a death spiral for the unregulated competitors. The increased retail sales would lead to increased spot market purchases, thereby increasing average cost and regulated price for the utility. Thus both the IOUs and the unregulated competitors faced high risk as a result of the asymmetrical role of the IOUs as the default providers of electricity. Western Electricity Crisis and California Financial Crisis Until early 2000, in terms of publicly visible goals, the system seemed to be operating as intended. Wholesale prices remained below historical average costs and there was sufficient headroom for the IOUs to collect the CTC, which they could apply to cover stranded costs. By 1999 SDG&E had collected sufficient CTC to cover all of their stranded costs and would no longer be subject to a retail price cap. As stipulated under AB 1890, henceforth, the CPUC would normally allow SDG&E to pass on increases in the wholesale price of power to its retail customers. However, that change turned out to be politically unacceptable beginning in 2000 when wholesale spot prices surged In California, applications to construct many new electricity generating units had been filed and approved; construction had begun for many thousands of megawatts of new capacity. But very little competitive retail market for electricity had developed. Although several companies entered the market to sell consumers green electricity, generated entirely by renewable sources, this component of the market was small. Less publicly visible, were operations of the PX and the CAISO. The market structure continued to prove unwieldy. The congestion management approach, relying on zonal pricing, a complex system of adjustment bidding to relieve anticipated transmission congestion, and the opportunity for gaming of the congestion markets, seemed to be fundamentally flawed. The CAISO submitted to FERC a sequence of proposed amendments 24 by the year 2000 to its tariffs. A congestion market redesign process was initiated but never completed. In mid-2000, however, came signs that all was not well. Soon after the electricity crisis was upon the entire western part of the U.S. The Nature of the Crisis The California electricity crisis was, in reality, two crises: 1) An electricity crisis associated with very tight wholesale markets for electricity in the West and California s poorly functioning electricity markets and; 2) A financial crisis affecting California s IOUs, resulting from rigidities of the State s regulatory control of these utilities. The California electricity crisis was part of the electricity crisis facing the Western states. Demands for electricity throughout the West had grown over the years but supply had not. New Draft. Do Not Quote. 33

37 generation capacity in California was under construction but construction had not been completed. Short-term supply reductions led to very tight markets and wholesale spot prices increased sharply in California and throughout the West. In this chapter I will focus attention on the California supply and demand issue. For more complete discussion of the western supply and demand balance, the reader is referred to the paper by Fisher and Duane 38. The financial crisis started primarily as a crisis for the investor owner electric utilities and turned into a crisis for the State budget. That the financial challenge turned into a financial crisis was the direct result of State regulatory, administrative, and legislative action or, more precisely, inaction. These two crises will be discussed in turn. The Electricity Crisis The first indications of the oncoming electricity crisis appeared in May 2000, with sharp wholesale price increases. By December 2000, the difficulties had grown into a crisis, with massive increases in the wholesale electricity prices and with frequent energy emergencies as the operating reserves were at dangerously low levels. The crisis remained severe all winter. Not until late spring 2001, did the electricity crisis start to disappear, with wholesale prices falling, electricity consumption declining relative to 2000 levels, and decline in the frequency of energy emergencies. By fall 2001, wholesale prices had declined to historical levels and energy emergencies had disappeared entirely. The crisis had passed. Wholesale Price Increases In May 2000 and in June of 2000 wholesale electricity prices in all of the regional markets reached 39 peaks above $400/MWh. In July, peak prices soared even higher on all of the Western markets, with all markets showing highest prices exceeding $500/MWh during one week. Then prices fell, only to peak again, but not as high, in August. After the August peak, prices started falling again and reached a low during the week of November 6. But the crisis had only begun. Spot prices began rising in November In December prices skyrocketed, exceeding $1000/MWh for the average of the high and low price peaks in some weeks. Prices were even higher in Washington and Oregon than in California 40. The following three figures illustrate these extraordinarily high prices. These figures show prices, estimated in The Western Price Survey, on a week-by-week basis, from Jan 3, 2000, through September 30, Data are present for three market centers at the California- Oregon Border 41 (COB), receipt points along the Columbia River (Mid-Columbia), and at 38 A complete discussion of the long term trends in the West appear in Trends in Electricity Consumption, Peak Demand, and Generating Capacity in California and the Western Grid, Jolanka V. Fisher and Timothy P. Duane, September, 2001, working paper from the University of California Energy Institute Program on Workable Energy Regulation (POWER). 39 The Energy NewsData publication, Clearing Up reported on June 16, 2000: Markets throughout the region were drawn into price spikes. Power prices at Palo Verde/Four Corners and Mid-Columbia/COB moved upward in huge increments all week, with almost no overlap from day to day. At the peak of trading on Wednesday Palo Verde peak power was fetching 450 to 500 mills/kwh. COB and Mid-C had been over 400 mills/kwh. 40 Although California demand is typically lower in the winter than in the summer, the reverse is true in the Pacific Northwest. And the hydropower reduction was greatest in the Pacific northwest. 41 The interconnection point at the California/Oregon border of the Pacific Northwest/Pacific Southwest AC Intertie Draft. Do Not Quote. 34

38 switchyard of the Palo Verde nuclear power plant, Arizona and for the exchanges in California. California prices shown are the PX MCP until January 29, After that date, as will be discussed later, the PX became defunct so that PX prices were not available. After January 29 prices are given for Northern California (NP 15) and Southern California (SP 15). The first graph, Figure 7, shows the average of the low and the high peak prices for Western markets, but the vertical scale is truncated to show detail before and after the December spikes. The second graph, Figure 8, includes the entire range in order to show how high the average of low and the high peak prices were in December 2000 in the Pacific Northwest. The third graph, Figure 9 shows the average of the low and the high off-peak prices for Western markets, with the vertical scale truncated as in Figure 7. Particularly striking is the observation that off-peak prices, not simply peak prices, were extraordinarily high during the crisis. The $450 $400 $350 $300 Mid-Columbia COB California PX California NP 15 Palo Verde California SP 15 $250 $200 $150 $100 $50 $0 3-Jan Jan Feb Mar Apr May Jun Jul Aug Sep-00 9-Oct-00 6-Nov-00 4-Dec-00 1-Jan-01 Week Beginning Date 29-Jan Feb Mar Apr May Jun Jul Aug-01 average of the low and the high off-peak prices in several weeks exceeding $250/MWhr. 10-Sep-01 Figure 7: Spot Power Prices: Average of High and Low Peak Prices. ($/MWhr) Source: Western Price Survey, Draft. Do Not Quote. 35

39 $2,800 $2,600 $2,400 Mid-Columbia COB California PX $2,200 $2,000 $1,800 California NP 15 Palo Verde California SP 15 $1,600 $1,400 $1,200 $1,000 $800 $600 $400 $200 $0 3-Jan Jan Feb Mar Apr May Jun Jul Aug Sep-00 9-Oct-00 6-Nov-00 4-Dec-00 1-Jan-01 Week Beginning Date 29-Jan Feb Mar Apr May Jun Jul Aug Sep-01 Figure 8. Spot Power Prices: Average of High and Low Peak Prices. ($/MWhr) Source: Western Price Survey, $450 $400 $350 $300 Mid-Columbia COB California PX California NP 15 Palo Verde California SP 15 $583, $678 $250 $200 $150 $100 $50 $0 3-Jan Jan Feb Mar Apr May Jun Jul Aug Sep-00 9-Oct-00 6-Nov-00 4-Dec-00 1-Jan-01 Week Beginning Date 29-Jan Feb Mar Apr May Jun Jul Aug Sep-01 Figure 9: Spot Power Prices: Average of High and Low Off-Peak Prices. ($/MWhr) Source: Western Price Survey, These figures suggest that different market forces were operational at different times. During peak periods, high prices could result from generation/transmission capacity limitations, Draft. Do Not Quote. 36

40 particularly during times of especially high peak demands. And, as will be discussed at a later point, these high peak prices may be partially the result of exercise of market power by generators who recognized that most generators are already at or near full capacity. However, at off peak times, generation was not near fully capacity and so generation capacity limitations could not explain the off-peak prices. Nor is it likely that the exercise of market power explains these prices, either in California or in the other western markets, given the large amount of unused capacity during off-peak periods. However, limitations in energy inputs to generation natural gas and stored water for hydropower were relevant in both peak and off peak periods. Since both natural gas and water were storable inputs, their market prices or opportunity costs did not change greatly, if at all, during daily cycles. As will be discussed more fully later, natural gas and water were in short supply throughout the electricity crisis. In addition, within some air quality districts, particularly in southern California, environmental constraints limited cumulative generation and thus also were relevant in off-peak periods, as will also be discussed at a later point. To the extent data are available, prices were similar among the Western markets for most of the time, with the exception of two weeks in December 2000 (the week of December 11 and the week of December 18). During those weeks, the maximum prices 42 during the peak period were about $5000/MWh at Mid-Columbia and $4000/MWh at COB, while the maximum prices recorded on the PX were $1400 and $950. On the other hand, the maximum prices during the off-peak period were about $700 and $1,100, respectively, on the PX, but did not exceed $400 during the week 43 of December 18 at Mid-Columbia and COB. Throughout January 2001 offpeak prices at the PX substantially exceeded the other western prices. Transmission capacity constraints between the Pacific Northwest and California allowed these price differentials. Note also that prices in SP 15 tracked Palo Verde price most closely; prices in NP 15 tracked COB and Mid-Columbia most closely. Transmission constraints between northern and southern California, along Path 15 allowed the difference between these regions to persist. Figure 7 through Figure 9 make it clear that the electricity crisis was not simply a California phenomenon. The high prices both peak and off-peak were felt similarly, although not identically, throughout the western electricity markets. Energy Emergencies and Rolling Blackouts The peak of the electricity crisis was marked by energy emergencies in California and the profound fear of rolling blackouts that occurred when some customers were blacked out to avoid instability in the entire grid. Although blackouts came to symbolize the electricity crisis, energy emergencies were the norm at the height of the crisis and blackouts were the exception. Figure 10 plots the energy emergencies declared by CAISO. Data are on a daily basis; each major division on the horizontal axis represents two weeks. Stage 1, 2, and 3 emergencies are indicated by the 1, 2, and 3 on the vertical axis. Rolling blackouts are shown as one level higher. 42 The average of the high and low peak prices during these two weeks were $2,700 and $2,687 at Mid-Columbia and $2,200 and $2,188 at COB. 43 I do not have data for off-peak prices in the week beginning December 11 for COB or Mid-Columbia. Draft. Do Not Quote. 37

41 Blackouts came to symbolize the electricity crisis in California. Nevertheless, although the threat of blackouts was frequent, actual blackouts were very rare. The CAISO ordered blackouts in fact on only six separate days, as shown in Figure 10. Moreover, blackouts were called for only a small fraction of the load at any time. The most severe was on January 18 in which 1000 MW of load was curtailed, accounting for 3.2% of the peak demand that day. Other rolling blackouts ranged from 300 MW to 500 MW, or 0.9% of the peak load to 1.7% of the peak load. Data on blackouts are shown in Table 2. Emergency Stage 4 Rolling Blackouts Jun Jun-00 3-Jul Jul Jul Aug Aug Sep Sep-00 9-Oct Oct-00 6-Nov Nov-00 4-Dec Dec-00 1-Jan Jan Jan Feb Feb Mar Mar-01 9-Apr Apr-01 7-May May-01 4-Jun Jun-01 2-Jul-01 Figure 10 Energy Emergencies and Blackouts in California. Source: CAISO Date Curtailment (all 2001) Ordered (MW) % of Peak January % January % March % March % May % May % Table 2 Rolling Blackouts in California. Source CAISO Draft. Do Not Quote. 38

42 Economic Forces During the Electricity Crisis Supply and demand conditions throughout the Western United States are tightly linked, as shown by the degree that the wholesale spot prices moved together. Transmission lines, across which electricity can be transmitted, connect adjacent states and adjacent regions within a state, as shown in Figure 4. Since electricity generation in any part of the West could serve the electrical loads in any other part of the West, electricity supply and demand changes in any part of the West could have price impacts on the entire region. A tight market in the Pacific Northwest could easily translate to a tight market and high wholesale prices in any western state, including California. The following nine economic forces together led the prices to skyrocket and/or placed limitations on market-induced downward pressures on prices. The tenth item was the only one pushing prices downward. First, limitations on generation capacity created an initial tight market situation. Second, generation capacity was reduced as a historically large number of generation plants were offline and not generating power. Third, the cost of marginal generation was increased because of the sharp rise in the price of NOx emissions permits. Fourth, extremely low rainfall in the Pacific Northwest and California reduced energy available to generate hydroelectricity. Fifth, limitations on natural gas supply and increases in natural gas demand led to sharp increases in price in California. Sixth, limitations on demand response implied that small changes in supply or demand could lead to large changes in price. Seventh, California s energy market, often described as dysfunctional, facilitated market gaming, particular for ancillary services and transmission congestion. Eighth, increased financial risk reduced electricity supply and led to risk premia for selling to California utilities. Ninth, exercise of market power and market gaming exploited the tight market situation and the California energy market, further increasing prices throughout the West. Each of these factors will be briefly discussed below, with a primary focus on California markets and the impacts of the regional markets on California, and only a secondary focus on the entire regional supply and demand balances. Limitations on Generation Capacity Each generator is constrained by an upper limit on the physical capacity to produce electricity; the operator cannot generate electricity from a plant at rates beyond physical capacity. This physical capacity constraint leads to a hockey stick shape of the electricity supply function: for operation well below capacity, small changes in price can lead to large changes in electricity supply; for operation near capacity, the same changes in prices motivate very small changes in electricity supply. Looked at the other way, for operation well below capacity, small changes in Draft. Do Not Quote. 39

43 available capacity or demand will have only very small impacts on price; for operation near capacity, small changes in available capacity or demand can have large impacts on price. Throughout the West there had been very little growth in generating capacity in the decade prior to the crisis, although there had been continuing growth in electricity use. Thus the peak period electricity demand had gradually become closer and closer to available supply 44. After California s restructuring, permits were granted for many new generation plants and construction had started on many. In the years 1997 through 2001, applications averaged about 4,500 megawatts per year, in comparison to an average of well less than 1000 MW per year between1980 and But construction of new generation capacity is always slow, taking several years from application for a permit to generation of electricity. Figure 11 shows the California new generation capacity that was initiated in the various years from 1980 through Data are grouped by the year in which the approval application was first submitted 45. None of the plants that had been proposed in 1997 or later were on line at the beginning of the crisis. From May through October of 2000, only 1250 MW of new generation capacity came on line in the west (including western Canada). Only during the last part of the crisis did new generation capacity place any significant downward pressure on prices. Total MW Capacity (Thousands) On Hold Under Review Under Construction Operational < 50 MW Operational > 50 MW Application Year Figure 11. New Generation Capacity in California 44 Since there were only few electrical interconnections between the West and the rest of the United States, electricity supply was not available from the eastern United States. 45 Plants under 50 MW capacity did not need CEC approval and thus data is available when they first went on line but not when they were started. For these plants it was assumed that the process was started two years before the plant went on line. Draft. Do Not Quote. 40

44 Generation Plants Offline In California, large numbers of generators went offline in during the late fall 2000 and winter Historically, between 1,000 and 6,000 MW average daily generating capacity would normally be off-line in a month. However in the period between October 2000 and May 2001, a monthly average of 12,000 MW generating capacity was off-line, reaching a peak out of 14,000 MW in April The generating capacity off-line resulted from a combination of causes. However, the greatest number was off-line reportedly for repairs or maintenance. But whether the maintenance and repairs were forced upon the generators, were part of a competitive cost minimizing solution, or were designed to increase wholesale price has not been fully resolved. Outages of the older natural-gas-fired units would have increased prices at the peak period, but not significantly in the off-peak periods. During off-peak times there was sufficient electric generation capacity available so that the binding constraint on production was the availability of natural gas, or for an individual plant, the price of natural gas. A gas-fired plant off-line did not use any natural gas; natural gas would be released for other generation. Thus there would be little or no impact on the off-peak electricity price. A new plant with a low heat-rate would have had some impact on the off-peak prices, as well as on the peak prices. The plant next in the merit order was likely to have a higher heat-rate. Thus a newer plan going off-line would result in increases in natural gas use for a given amount of generation, would increase natural gas price, and would thus increase the off-peak electricity price. At peak or near-peak times, generators would earn rents on the capital equipment. The less capacity available, the higher the market-clearing price of electricity. Thus, in peak times plants going off-line would increase electricity prices. A smaller, but significant number of the offline plants perhaps up to 3000 MW were not old gas fired units but rather were the QFs whose operators were not being paid for the electricity they were selling the investor-owned utilities. When these plants went off-line they would increase the demand for natural gas, thus increasing both peak and off-peak prices. Thus, some of the capacity reductions, but far from a majority, was the result of uncertainty imposed on the QFs and the reduction in their cash flow. Another outage was an accident at SCE s San Onofre nuclear power plant in February 2001, taking over 1000 MW offline for several months. Reductions in nuclear power output increased both peak and off-peak prices, since its shut-down would not release natural gas for other users. Environmental limitations In 1993 the California South Coast Air Quality Management District instituted the REgional Clean Air Incentives Market (RECLAIM) to force reductions in emissions of two criteria pollutants, NOx and SOx. The RECLAIM program required emitters of NOx and SOx to acquire enough RECLAIM Trading Credit permits to match their actual emissions each year. Through 1998, there was an excess supply of permits but during 1999, the number of allocated permits Draft. Do Not Quote. 41

45 diminished to equal total emissions. Through compliance year 1999, the price of NOx trading credits ranged from $1,500 to $3,000 per ton. In 2000, prices for the first ten months of the year 2000 increased to an average of about $45,000 per ton 46. This price increase stemmed from both the further reductions in the available permits, thereby reducing their supply, and increases in the amount of electricity being generated using gas-fired units, thereby increasing the demand for RECLAIM credits 47. A new baseload gas fired generating unit, releasing about 0.1 pounds of NOx per MWh of electricity, would pay only $2.25/ MWh at a price of $45,000 per ton. But old gas-fired turbines, emitting up to 4 pounds of NOx per MWh, would face a cost of RECLAIM credits of $90/MWh. Because the market clearing prices in both the PX and the CAISO were based on the highest accepted bids, this cost of RECLAIM credits added to the price for all electricity whenever bids based on generation from these old units were accepted. Since the old units were used primarily in peak periods, the cost of RECLAIM credits increased electricity price by the greatest amount during peak periods. Pacific Northwest Drought In 2000 the Pacific Northwest and northern California experienced extraordinarily low rainfall, following years of particularly high hydroelectric supply. Low rainfall reduced the hydropower that could be generated in those locations 48. The drought can be measured by the runoff at the Dalles, the downstream dam of the Columbia River system. Figure 12 shows this runoff 49 by year, from 1981 through 2004, as a percentage of the runoff averaged from 1961 through The average runoff for a water year (January through July) is 104 Million Acre Feet (MAF). The runoff for January through July 2001 was 58 MAF, just 58% of the average runoff. Not only was 2001 a year with particularly low water, but 2001 had been preceded by six years near or above average. 46 White Paper on Stabilization of NOx RTC Prices, South Coast Air Quality Management District, web site: 47 This increase in the increase in electricity generated using natural gas will be discussed more fully at a later point. 48 Similar issues are discussed in Chapter Data from the National Oceanographic and Atmospheric Administration., available on the web site: Draft. Do Not Quote. 42

46 160% 150% 140% 130% 120% 110% 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Percentage of Average Runoff Water Year (January - July) Figure 12: Water Runoff for the Columbia River, at the Dalles Low stream runoff in the Pacific Northwest lead to a 14% decline in hydroelectric generation from 1999 to 2000 and a further 31% percent decrease from 2000 to These data are shown in Figure 13. The reduced generation in the Pacific Northwest implied that there was less electricity available for importation into California during the crisis. Figure 14 shows the total imports of electricity into California 50. This figure shows how sharply energy imports into California were reduced in 2000 and 2001, putting great pressure on production within California, and driving up California prices. It should be noted that the 2000 and 2001 data in Figure 13and Figure 14 are for the entire year, while the energy crisis including only portions of each of these years. Thus generation and import reductions during those times were probably greater than indicated here. 50 In 2001 California changed its categorization of production, leading to changes in the official data series for imports. Prior to that year utility-owned shares of coal, nuclear, and some firm contract generation from outside California were include as internal to California. Since 2001 most of this data is included in the energy imports category. The data presented here for 2001 and beyond are energy imports estimated using the same methods as in Draft. Do Not Quote. 43

47 25 20 Total California Average MW (thousands) Note: Total includes California, Oregon, and Washington. Figure 13. Hydroelectric Generation in Electric Power Industry CA, OR, WA. Source: EIA Average MW (Thousands) Figure 14. Imports of Electricity into California Draft. Do Not Quote. 44

48 Natural Gas Limitations Natural gas used in California comes primarily from Canada and secondarily from California sources, the Rocky Mountains, the U.S. Southwest, and storage withdrawals. These supplies were all limited. Natural gas in storage in California and throughout the West at the beginning of the heating season was well below normal levels. Beginning-heating-season storage was 22% below the average of the previous five years in California and 30%-40% below the average in most of the West. 51 An explosion in the El Paso pipeline plus subsequent reductions in capacity of that pipeline reduced the available gas carried into California from the Southwest. Capacity limits on the pipelines from Canada limited the increases of Canadian gas. At the same time natural gas demand increased. With the decrease in available hydropower, in California, total electric generation increased 10.5% from 1999 to 2000, decreasing somewhat from 2000 to This increase in the total electricity generation, combined with the reduced California hydropower, increased gas-fired electricity generation. In addition, many older and less-efficient plants were kept online. Because these older units have higher heat rates than the newer units, using older units led to an increase in natural gas demand per MWh generated. The two effects an increase in megawatt hours of electricity generated using natural gas and an increase in the natural gas used per megawatt hour of electricity generation amplified each other in increasing the demand for natural gas in California. As a result, natural gas demand increased sharply. Figure 15 graphs deliveries of natural gas to all users in California, by month, for various years. Each line begins in June of one year and extends through May of the next year, showing the difference between the time of the electricity crisis in California and other years, before and after the crisis. Natural gas deliveries to all California users in the period December 2000 through April 2001 was 19% higher than in the corresponding period, December 1999 through April Source: Energy Information Administration, Electricity Shortage in California: Issues for Petroleum and Natural Gas Supply. 52 EIA Natural Gas Monthly. A similar pattern can be seen from data on deliveries Washington, Oregon, plus California. These delivery figures, it should be remembered, represent the equilibrium increase in demand for natural gas; the increase in the demand functions for natural gas were greater than the increases shown here, since natural gas prices increased sharply, reducing actual consumption. Draft. Do Not Quote. 45

49 Billions of Cubic Feet Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Figure 15. Delivery of Natural Gas to California, All Users. Source EIA Natural Gas Monthly The resulting increase in demand for natural gas tightened gas markets throughout the western U.S. and increased natural gas prices. The natural gas price increased sharply. Part of the reason for the large increase in the natural gas prices was the low elasticity of demand for natural gas. A large fraction of the demand for natural gas was derived from the demand for electricity. But since electricity wholesale demand was virtually inelastic in California, that component of demand was also virtually inelastic. Natural gas demand for residential uses and commercial/industrial uses in California could be expected to have a small, but non-zero, elasticity of demand. Therefore, the overall short-run elasticity of demand for natural gas was very small. For that reason demand increases or supply decreases led to disproportionately high increases in the natural gas price. Figure 16, based on data from Gas Daily Annual Price Guide, shows increases in the California natural gas spot prices from the beginning of January 2000 through August, Prices are shown for two delivery points to utilities in California. Figure 16 shows that natural gas price at both California locations more than tripled from January 2000 through mid-november Then in late November and December, the California natural gas price increased even more sharply, to about $50 per million Btu in December. Although this price peak lasted for only two weeks, the spot natural gas prices in California remained for most days above $10 per million Btu until May Draft. Do Not Quote. 46

50 $60 $55 $50 $45 SoCal gas, large pkgs Price PG&E Citygate Price $40 $35 $30 $25 $20 $15 $10 $5 $0 1/1/00 2/1/00 3/1/00 4/1/00 5/1/00 6/1/00 7/1/00 $/MMBtu 8/1/00 9/1/00 10/1/00 11/1/00 12/1/00 1/1/01 2/1/01 3/1/01 4/1/01 5/1/01 6/1/01 7/1/01 8/1/01 Figure 16 Natural Gas Spot Market Prices. Source: Gas Daily Annual Price Guide 2001 This increase in the natural gas price greatly increased the cost of electric generation. The lowest cost impacts were on new generation plants. With heat rates for a new combined cycle power plant of 6.8 million Btu per MWh, every $1/MMBtu increase in the natural gas price increased generation cost by $6.8/MWh. For older, less efficient plants with twice the heat rate, every $1/MMBtu increase in the natural gas price increased generation cost by $13.6/MWh. Thus a $40/MMBtu increase in the price of natural gas would increase the cost of operating such older plants by $544/MWh. Increases in the price of natural gas led to increases in the price of electricity in both peak and off-peak periods, since natural gas generating units generally set the market-clearing prices. However, since the higher heat-rate units were likely to be the marginal units at peak times and somewhat lower heat-rate units were likely to be the marginal units at the off-peak times, the increase in natural gas price had a greater impact on electricity prices during the peak hours. Limitations on Demand Adjustment The nature of the supply and demand adjustments in this market implied that small variations in supply or demand for electricity could lead to disproportionately large variations in price. In no states did retail prices change simultaneously with wholesale prices, thus demand adjustment did not serve to reduce the very short run (that is, hourly) price increases. In California, however, retail price controls assured that retail prices did not change at all until 2001, and then by very little, relative to the wholesale price increases. These price controls thus greatly reduced the longer-run demand adjustments and stopped longer-run demand adjustments from limiting longer-run price increases. Thus limitation on retail price adjustment blunted the incentives on users of electricity to reduce their consumption; thus demand reductions, which would otherwise Draft. Do Not Quote. 47

51 have placed downward pressure on prices in either the very short-term or longer-term were very limited. California s Dysfunctional Energy Market Structure As discussed above, California s energy market structure, typically described as dysfunctional, included a complicated set of wholesale markets, not well coordinated with one another, and congestion markets in severe need of redesign. Because these markets were given monopoly or near monopoly status, California utilities were forced to purchase through them. The CAISO was charged with avoiding blackouts, almost at any cost. As such, when shortages seemed likely, the CAISO would search for electricity outside the California electricity markets, which it purchased at very high prices. The anticipation of selling at high prices to the CAISO provided an incentive for electricity traders and brokers to hold back their supplies or demand greatly elevated prices, confident that they would ultimately be able to sell at high prices to the CAISO. The structure forced exceptionally large dependence on the spot market. The interplay of these various markets, the resulting bidding strategies of generators of electricity, electricity marketers, and traders, and the responses of the CAISO personnel together created a very complex market, rife for market power and market gaming. Risk Created by the Financial Crisis Because of the financial crisis, discussed subsequently, utilities began delaying payments to electricity generators, promising to pay later. However, it was becoming clear that absent State policy measures, utilities were unlikely to be capable of paying for electricity in a timely manner, if at all. This created significant uncertainty among suppliers. Uncertainty of future payments would lead a rational competitive supplier to increase the price at which it was offering to sell supplies into the California market, further increasing the wholesale prices. This phenomenon was particularly important in the November 2000 through January 2001 period when utilities were not paying for electricity and their credit ratings were declining sharply. In January, the state of California stepped in as the creditworthy buyer, seemingly guaranteeing the payment for all electricity purchased on behalf of the utilities. But California refused to pay its own spot market wholesale power bills. During 2001, continuing nonpayment caused some suppliers to drop offline; continuing risk of nonpayment caused others to include a credit premium in their bids. FERC subsequently approved a 10% credit premium to compensate for continuing financial risks for sales to California. Exercise of market power and market gaming A final element is the incentives for exercise of market power by electricity generators selling into California markets, for gaming of the system by electricity marketers, such as Enron, and for exercise of market power by the highly concentrated set of wholesale buyers of electricity. Exercise of market power and gaming was facilitated by the complex rules of the California electricity markets. As discussed above, a supplier with market power would have an incentive to bid into organized California markets at prices higher than its correctly-calculated incremental cost or could refrain Draft. Do Not Quote. 48

52 from bidding some units. This incentive was strongest when the entire system was near capacity, as was the case during peak periods of the crisis. The degree to which exercise of market power was responsible for increased prices is currently the subject of many investigations and litigations. Serious research work, such as work of Borenstein, Bushnell, and Wolak (2000) and Joskow ( ) attempt to quantify market power and conclude that prices were elevated above marginal cost during the crisis as well as before the crisis, that is, that market power led to significant price increases. Work by Hogan and Harvey questions these conclusions. In what follows I discuss reasons why market power may have been less important than is generally concluded in the political discussions, even though it was some part of price increases. The IOUs were dominant buyers of electricity. However, they had virtually no control over their loads, other than through their advertising, in conjunction with California State advertising, urging consumers to conserve electricity and to shift their use of electricity to off-peak periods. But they could attempt to exercise market power by under-scheduling the load submitted to the CAISO, thus modifying their mix of demands between the PX and the CAISO markets and reducing PX MCP, at least initially 53. In mid January 2001, California Governor Davis ordered the State Department of Water Resources (DWR) to start purchasing wholesale electricity on behalf of the electric utilities, who by that time were no longer creditworthy buyers. Other than electricity self-generated or purchased through pre-existing contracts, investor-owned utilities acquired all their electricity through the DWR. This change fundamentally altered the market structure in California, shifting market power toward the State. Instead of competing buyers of electricity, buying on the PX or CAISO, the DWR become the dominant buyer of electricity in California. The DWR could choose how much electricity to purchase in forward contracts and how much to acquire on the real time CAISO market. If DWR chose, it could acquire all electricity through bilateral contracts outside the CAISO. As dominant buyer, the DWR could not change total electricity consumption. But it could negotiate forward contracts of varying time duration, with various sellers. And it was not required to disclose prices it paid for particular transactions 54. DWR could price discriminate and exercise its market power as a dominant buyer. Figure 17, based on data released by DWR, illustrates the large number of entities from which DWR was buying electricity and the wide range of prices it paid. Each bar represents an entity selling electricity to DWR during the first six months of Black bars represent public entities and gray bars represent private entities. Entities selling the most electricity are furthest to the left. The bar heights show the cumulative fraction of electricity purchased, including that 53 It is likely that this strategy allowed them to exercise market power only temporarily, since generators would take into account the likely CAISO MCP when submitting their PX bids. Increases in the likely CAISO MCP would lead generators to increase their bid prices in the PX and this would counteract the IOU bidding strategy, once generators understood the changed price patterns resulting from the IOU strategies. 54 Data on purchase prices ultimately had to be released, but the releases were enough after the time of transactions that the information would have little or no value to the generators trying to defend against discriminatory pricing. Draft. Do Not Quote. 49

53 seller. Mirant supplied 26% of the DWR purchases. Dynegy supplied 8%; the cumulative supply for Dynegy and Mirant is 34%. Williams sold another 8% and thus the cumulative percentage is shown as 42%. The graph shows that four sellers accounted for almost 50% of the purchases by DWR, eleven sellers accounted for 75%, and twenty-one sellers together accounted for 90%. Thus the market structure was one in which a single dominant buyer, DWR, was purchasing electricity from many competing sellers. Cumulative Percentage 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% Mirant $230 Dynegy $187 Williams $252. Powerex Corp. $425 AES $158 Duke Energy $128 LADWP $276 Sempra Energy $366 DWR $129 BPA $229 TransAlta $298 El Paso $210 BP Energy $244 MIECO, Inc. $200 Public Service of New Mexico $223 Constellation Power Source $118 Reliant Energy $236 Merrill Lynch Capital Services $202. Calpine Energy Services $165. SMUD $268 City of Burbank $306 Pinnacle West $195 Coral Power, L.L.C. $336 Eugene Water & Electric Bd $381 Enron $181 Morgan Stanley Capital Group $90. Commonwealth Energy Corp. $181 Salt River Project $118 PG&E Corp. $168 Grant County PUD $335 Trans Canada Power Co $307 Nevada Power Company $182 No. California Power Agency $161 Tuscon Electric Power $156 0% Figure 17 DWR Electricity Purchases, January through June 2001, Source DWR Average prices of electricity sold to DWR are shown above the names of sellers. The prices negotiated by DWR varied widely across sellers. Of the four largest sellers, DWR paid Mirant an average price of $230/MWh, Dynegy an average price of $187/MWh, Williams, $252/MWh, and Powerex $425/MWh. The highest price of these four was more than twice the lowest. Among the sellers accounting for 90% of the transactions, DWR paid prices ranging from $128/MWh (Duke Energy) to $425/MWh (Powerex Corporation). Monthly data show the same pattern of price discrimination for each month. DWR was willing and able to price discriminate among the sellers. A skillful price-discriminating dominant buyer, purchasing from a number of competing sellers should be able to reduce the total acquisition cost, even without reducing the total amount of electricity purchased, although it could not reduce the acquisition price below the prices at which Draft. Do Not Quote. 50

54 the generators could sell electricity out of state, where there were competing buyers. And DWR had experienced, capable staff members. Thus, it appears as if DWR was able to exercise its market power as a price-discriminating dominant buyer. Once DWR was purchasing electricity, there could be no exercise of market power on the PX, since purchases were no longer going through the PX auction. Market power could still be significant for the CAISO purchases. But the DWR could assure that few purchases were needed through the CAISO. In its acquisition of electricity, DWR s ability to price discriminate reduced any residual market power. If it suspected that a seller was trying to manipulate market prices, the DWR could discipline that seller by purchases from other sellers, even those offering electricity at higher prices. Such a strategic purchase option implied that as of late January 2001, a generator could exercise market power only 1) for sales through the CAISO real time market and ancillary services markets or 2) through actual reductions of generation. In addition, for a period of time the CAISO was providing DWR with confidential information on generator bids, giving DWR and additional information advantage, until CAISO was ordered to stop that practice. Undisputed, however, is the conclusion that in addition to market power to increase the price of electricity, there were other possibilities of market gaming, particularly gaming of the ancillary services markets and of congestion prices. The famous Enron smoking gun memo detailed some methods by which Enron gamed markets, in particular the ancillary services 55 markets and the markets for transmission services. Although some of the strategies were complex arbitrage schemes, some created congestion in transmission lines and thus increased electricity prices. Energy Conservation One force operated to reduce wholesale energy prices during this perfect storm : the incentives and information campaigns encouraging energy conservation and time shifting of peak load. Reductions in the demand function reduced prices in both the peak and off-peak periods. Electricity demand, both measured in terms of peak demands and total megawatt hours of electricity, declined from 2000 to During the spring and summer of 2001, reductions in electricity consumption and peak demands took pressure off the tight electricity market. Figure 18 shows the reduction in average electricity consumption and in peak demand, based on California Energy Commission (CEC) data 56. The bars show the peak demand reductions on a month-by-month basis from 2000 to 2001; the lines show the reductions in average electricity use. Monthly peak electricity demand was reduced, on average, by 1,900 MW (4.4%) and the monthly average use was reduced by 1,200 MW (4.3%). These demand reductions reduce price in both the peak and the off-peak periods, reducing the use of generation capacity and of energy needed for generation. 55 In order to manage and electric supply system, given the special characteristics of electricity, the system operator needs agreements with generators to provide generation reserves, reserves that could be called upon at short notice to increase or to decrease the total generation of electricity. These reserves plus additional resources or loads that can be controlled to keep the system stable are referred to as ancillary services. 56 Data are published by the CEC on its www page. Draft. Do Not Quote. 51

55 MW or Average MW (Thousands) (0.5) (1.0) (1.5) (2.0) (2.5) (3.0) (3.5) JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Reduction in Peak MW, 2001 vs 2000 Reduction in Average MWh, 2001 vs 2000 Figure 18. Reductions in Electricity Use, 2001 vs. 2000, Source: CEC In addition, there were reductions in electricity use throughout the west. In the Pacific Northwest, in particular, there were significant numbers of industrial shutdowns, including in the aluminum industry, which together reduced the demand for electricity in the west. These demand reductions stem from a combination of factors -- expectations of increased electricity prices, high retail natural gas prices 57, the energy demand management programs, energy efficiency and conservation programs, publicity about electricity problems, and the decline in the California economy. The Financial Crisis The high wholesale prices, the lack of long-term contracts, the divestiture of utility generation assets, and the retail price control together created a tremendous financial stress on the utilities. By spring of 2000, they had divested about 60% of their generating assets. Divestiture itself would have not been particularly harmful if the utilities had replaced these generating assets with long-term fixed-price contracts to acquire the electricity, but they had not been allowed to do so. Both PG&E and SCE made requests of the CPUC and appealed to Governor Davis to allow them to enter such contracts. The initial response was negative. 57 Many customers, particularly in northern California, received one single bill for natural gas and electricity purchases. When the natural gas prices increased, newspapers carried storied energy bills increasing. Although a reduction in electricity use in response to an increase in natural gas price is not what economists normally predict, it seemed to be occurring in California. Perhaps subsequent empirical work will be able to examine whether this phenomenon in fact occurred in significant amounts. Draft. Do Not Quote. 52

56 On June 22, 2000 the CPUC voted to allow the utilities to buy power outside the CAISO and the PX markets. However, the California Legislature subsequently overrode the decision. Thus the utilities were forced to continue purchasing and selling all electricity on these two markets. In August 2000, once prices had jumped sharply, the CPUC did file an order allowing the utilities to enter limited amounts of bilateral contracts. Even then, any purchases other than from the PX would either need pre-approval or face an after-the-fact reasonableness review, while all purchases from the PX were considered per se reasonable. This failure to encourage long-term contracts and the resultant over-reliance on spot markets help create the financial crisis. The over-reliance on spot markets assured that when wholesale prices began to soar, the utilities would pay these high prices for almost half of the electricity they would need to serve their loads. This lack of risk management options was central to the financial crisis. The utilities could have avoided a financial crisis if they had been allowed to raise their retail rates to cover the increased wholesale costs. Consumers and businesses would have faced higher prices of electricity but the price increases would not have been of crisis proportions. But the State of California continued to enforce 58 the retail price control established through AB This continued enforcement assure the financial crisis. By fall 2000, the utilities were selling electricity to their customers at an average price of about $65/MWh for the electricity (plus another $60/MWh for transmission, distribution, and other delivery services) yet the wholesale price of electricity ranged from $150/MWh to $1000/MWh for half of the electricity they were selling. Almost one-half of the remainder, from the QFs, was already being purchased at high prices. Generation from their own hydroelectric assets and nuclear power plants accounted for only 30% of their electricity. The accumulations into the transition accounts to pay for stranded costs quickly reversed, with large negative competition transition charges. And as a utility, they could not refuse to sell electricity. Each of the utilities requested increases in retail rates to cover the average cost of the electricity they were selling. Until January 2001, no rate relief was made available to the utilities. In fact, SDG&E, the only one of the three major investor-owned utilities that under AB 1890 would be no longer was subject to retail price controls, remained free of the controls only for a limited time. The California Legislature voted to reimpose controls. As of December 2000, the financial crisis had come to a head. PG&E ultimately filed for reorganization under Chapter 11 of the bankruptcy code in April 2001; SCE teetered on the brink of bankruptcy for month until that utility negotiated a rate agreement with the CPUC. Once the utilities were no longer creditworthy, the State, through the California Department of Water Resources, became the buyer of electricity, as noted above. Under Governor Davis order, 58 This provision of AB 1890 could have been overridden during the crisis either by legislation or by emergency decree of the California governor. Draft. Do Not Quote. 53

57 the DWR would begin purchasing electricity on behalf of the utilities, the retail prices would remain low, and the high cost of wholesale power purchases would be borne initially by the State treasury. The State would subsequently issue long-term revenue bonds to reimburse the State treasury. Repayment of these bonds would be a surcharge on retail electricity prices and thus ultimately the ratepayers would pay all of these costs. In essence, the State temporarily put itself in the place of the investor-owned utilities, except that ratepayers would at a later time repay all State losses. After the State became the primary buyer of electricity, there were no longer any transactions on the PX and that institution had no way of raising money to pay its costs. The PX ultimately declared bankruptcy in March. One of the two market institutions established by AB 1890 was thus eliminated 59. The significance of California regulatory controls to the financial crisis can be seen by comparing the financial impacts on the IOUs and Munis within California and those throughout the Western states. The electricity crisis was a phenomenon across the West; increases in spot wholesale prices were similar throughout the western states. Yet only the IOUs of California faced a fundamental financial crisis. Some IOUs in other states and some municipal utilities in California faced short-term difficult financial problems but others were not brought to the brink of bankruptcy. Only the California IOUs faced the double whammy: they purchased over 50% of their wholesale electricity on the highly volatile spot market and they were constrained by retail price control. These two issues were the product of California s unique regulation of the utilities; in no way were they necessary components deregulation. State and Federal Regulation Changes During the Crises State policy and Federal policy sharply conflicted during the crisis, making any solution even more difficult. California political leaders emphasized direct government intervention in the market place, reliance on retail and wholesale price controls, plus strong regulatory intervention. The CPUC, by then with changed leadership, stopped trying to improve markets, strongly enforced retail price controls (which the State could control), and lobbied for wholesale price controls (which the State could not control.) The California Governor relied on a public relations campaign of blaming California s electricity problems on the Federal government, including Federal regulators, on electricity generators, on deregulation, all but on his own policy inaction. When the Governor and Legislature did respond to the financial crisis, their response relied on direct governmental wholesale purchases of electricity and negotiations to acquire assets of the utilities, increasing State participation in electricity markets. Federal action, through FERC, though slow, sometimes misdirected, and inconsistent over time, seemed designed to address the underlying flaws in the market design and implementation and to strengthen the role of markets. FERC operated to reduce reliance on direct control of market transactions and market pricing. 59 The bankruptcy of the PX had no impact on the market because all purchases, other than through the CAISO were through bilateral trades. The PX was no longer being used. Draft. Do Not Quote. 54

58 The State and FERC each had jurisdiction over important parts of the restructured system so neither could fully impose its views on the other and each needed actions of the other to be fully effective. The fundamental ideological differences were never fully resolved. However, ultimately the FERC did develop a set of market-mitigation measures, including limits on wholesale prices. Federal Responses: FERC Rule Making November 1 and December 15, 2000 Orders In the first major rulemaking, on November 1, 2000, FERC issued a market Order proposing remedies for California s wholesale electricity markets and on December 15, with an Order directing the remedies. FERC reiterated its earlier conclusion that the wholesale prices were not just and reasonable and identified problems of market design as fundamental to the problem 60. A first goal of the Orders was to reduce the reliance in California on spot markets for wholesale electricity. FERC eliminated the requirement that the investor-owned utilities buy and sell all power on the PX or CAISO, thereby in theory permitting utilities to enter bilateral forward contracts. Utilities could use their self-generated electricity directly. Although an important long-term change, given the financial crisis, its short-term impact was limited to allowing the utilities to use directly self-generated electricity. And that change probably had little or no impact on either the energy crisis or the financial crisis. FERC did impose strong incentives on utilities and generators to complete almost all scheduling of load and generation with CAISO ahead of time, rather than in real time transactions. FERC announced a large penalty on all real time energy transactions greater than 5 percent of the prescheduled amount. CAISO ancillary service market rules were modified so that a supplier bidding into this market would receive only the energy payment if it were called upon to deliver energy, not the capacity payment in addition. These modifications were designed to make market scheduling more rational, the acquisition process more deliberate, and energy emergencies less frequent. The FERC Orders rejected California s proposal to give the PX authority to impose wholesale price caps. It denied the CAISO request to continue managing price caps, but ordered the CAISO to impose a $250 price cap on its purchases, without modifying the $100 price cap it had imposed for replacement reserves, until the end of December. FERC ordered that all single-price auctions run by the CAISO or the PX be temporarily modified to a hybrid system, often described as a soft price cap. A single market clearing price would be used if it were no greater than $150/MWh. If the market did not clear at $150/MWh or below, suppliers bidding less than $150/MWh would be paid $150/MWh and those bidding above would be paid their actual bid. This rule changed optimal bidding strategies for those times participants expected that the market would not clear at a price at or below $150/MWh. In those cases optimal bidding strategies 60 Language from the November 1 Order. Draft. Do Not Quote. 55

59 would be the same as in any as-bid auction system, not like those in the previous single-price system. The PX and CAISO were required to report confidentially to FERC all bids in excess of $150 and each seller was required to file bid price, electricity quantity, and marginal generation cost of each such transaction. Bidders above $150/WHh would be required to justify their bids. Lack of valid justification could result in the supplier refunding the higher price. Such a requirement for bid justification could provide strong motivation if generators believed that price bids above costs would be detected and that refunds would include a penalty in addition to the difference between price and marginal cost. However, if generators believed that bids well above costs were unlikely to be detected and that, if detected, the only penalty would be a requirement to refund the difference, then this requirement would provide little or no motivation for limiting bids. Wholesale prices increased after these Orders were issued. April 26, 2001 FERC Price Mitigation Order By Spring of 2001, FERC agreed to impose more rigid wholesale controls in California, but not in the other Western States. The April 26 Order established a mitigation and monitoring plan for the California wholesale electric markets, to become effective May 29, The plan required that all planned outages by units with Participating Generator Agreements 61 (PGAs) with CAISO be coordinated with, and approved by, CAISO. Such a requirement would presumably make it easier to judge whether outages of generating units were for nonmanipulative reasons. The price mitigation plan retreated from the hybrid auctions, moving back toward single-price auctions for the CAISO (the PX was no longer in operation.) All price caps would be eliminated and replaced with bid caps, limitations on maximum seller bid prices. Bids could be no higher than an administratively determined estimate of the marginal cost of operating the plant, based on the unit s heat rate and emissions rate, prices of natural gas and emissions credits, and a $2/MWh fee for operation and maintenance costs. Bids would determine market price through a single-price auction. Bid caps would be imposed only when an energy emergency had been declared. No caps would be imposed in the absence of an energy emergency, in the theory that competitive forces would be sufficient in those times. Exceptions to bid caps were allowed if a firm could establish that its cost was greater than the cap. Such bids would not help determine market clearing prices. Firms would be required to justify their costs or pay refunds. Similarly, resources located outside California could bid, but their bids would not determine market-clearing price during mitigated periods. 61 A Participating Generator Agreement is a legal agreement between a generator and the ISO that establishes the terms and conditions on which the ISO and the Participating Generator will discharge their respective duties and responsibilities under the ISO Tariff. The agreement allows the Participating Generator to schedule energy and submit bids to ISO through a Scheduling Coordinator. Draft. Do Not Quote. 56

60 The plan imposed a must offer requirement on fossil-fuel plants 62, a requirement that all sellers with PGAs offer all their available power to the CAISO in real time if not already scheduled. This rule was designed to stop generators from exercising market power by not bidding at all. The Order prohibited other bidding practices if they were seen as potentially allowing anticompetitive market gaming, for example, bids that vary with unit output in a manner not related to its cost characteristics. Finally, the Order made it clear that demand response should be part of the market response. Load- serving entities were required to establish demand response mechanisms in which they identified the price at which load should be curtailed. FERC June 19 Order Broadening Scope of Price Mitigation The April 26 Order applied only in California, ignoring the fundamental linkages between all of the Western markets. This problem was corrected in the June 19 Order. That Order extended the price mitigation plan to all western states and to all times and expanded its scope. The new Order was scheduled to remain in effect until September 30, The June 19 Order set rules to be implemented in a common manner 63 across all of the markets in the Western region. The bid caps and the must-offer requirement would remain during energy emergencies. 64 But the bid cap would no longer allow costs of emissions credits or other emissions costs 65. These costs and start-up fuel costs could be directly billed to CAISO and thus would have no impact on CAISO market-clearing prices. It also imposed price caps on bilateral sales. The market-clearing price in the single price auction would set an upper limit on the allowable prices for these transactions. Generators and marketers would be treated differently. Generators could justify costs higher than the administratively determined bid caps, but these higher costs would not be used in setting the market-clearing price. Marketers would not be allowed to a price higher than the marketclearing price. That is, they would be required to act as price takers. For non-energy-emergency times the price for spot market sales could be no greater than 85% of the highest CAISO hourly market clearing price established while the last Stage 1 emergency (not Stage 2 or 3) was in effect. 62 The requirement would not be imposed on hydroelectric plants. 63 In price control times a 10% price premium was allowed for sales to California, to compensate for market risk. 64 FERC: Order On Rehearing Of Monitoring And Mitigation Plan For The California Wholesale Electric Markets, Establishing West-Wide Mitigation, And Establishing Settlement Conference (Issued June 19, 2001) 65 Such a system would no longer assure that the lowest cost units would have their bids accepted. In particular, emissions and start-up costs would not count in determining which plants would be dispatched. Unfortunately, this would create a bias toward selecting the plants with less good environmental performance if their costs otherwise were low. Draft. Do Not Quote. 57

61 California After the Crises Although the electricity crisis itself was a short-term event California still faces risk of inadequate supply. The financial crisis has left a continuing legacy, the financial overhang, as the historical costs are allocated, partly by regulation, partly by litigation. Another legacy is the motivation for further reforms of the reforms. Some of those reforms have been implemented; some issues are still being debated; some are not yet fully understood. And one group, generally opposed to deregulated markets, has qualified an initiative that would limit competition. And a final part of the legacy is the continued distortions in public perceptions. In what follows, I discuss the current California system. Current Conditions Since the crisis, demand for electricity is California has continued to grow, albeit slowly, while supply growth has been limited. As a result, the overall supply/demand balance has become tighter, particularly in Southern California. The risk of supply shortages in within the next few years is real, absent reductions in growth of electricity use, reductions in peak demands, or increases in supply to California. The financial crisis of the IOUs has been resolved. But consumers currently face high retail prices of electricity, through surcharges design to recover the historical costs. Increasing nationwide natural gas prices can be expected to further increase wholesale costs and retail prices. Electricity Generation Capacity The restructuring led to increased investments in new generation capacity right after the restructuring and until the end of the electricity crisis. Approximately 23,000 MW of new capacity was approved for construction since AB Of this total, about 10,000 MW is now on line, and about 5,000 MW of new capacity is currently under construction 66. Another 7,700 MW has been cancelled or is currently on hold. (See Figure 11 for year, by year data.) However, since the crisis, the rate of new capacity applications has declined sharply. Figure 11 shows that new applications to construct generating plants declined to 1,000 MW in 2002, 500 MW in 2003 and to 250 MW in With the low rate of new capacity additions, the retirement of some old thermal plants, and growth in electricity usage, markets promise to become tighter and tighter over the next few years. The need for additional capacity and/or additional transmission capacity is a particular problem in rapidly growing southern California. At the same time, however, independent power producers, such as Calpine Corporation, face severe financial difficulties and are unlikely to invest in new generation capacity absent longterm contracts with retail utilities or well-functioning capacity markets. Utilities, uncertain about their regulatory future, have been hesitant to enter new long-term power purchase agreements. IOU Status As of this writing, the immediate consequences of the financial crisis for the IOUs have been resolved. Pacific Gas and Electric has emerged from bankruptcy. All the IOUs are credit- 66 These figures represent about 20% and 10% of California peak demands, respectively. Draft. Do Not Quote. 58

62 worthy. The State is no longer entering new electricity purchase contracts on behalf of the utilities; utilities are now buying their own electricity. Utilities are managing the long-term electricity purchase contracts negotiated by the State of California. Retail Prices of Electricity The financial losses during the crisis must be borne by some entities, either generators, utilities, the State, or electricity consumers. The great majority of these costs are being borne by electricity consumers, in the form of increased retail rates or surcharges on the retail rates. The high costs of the DWR-negotiated long-term electricity supply contracts are being paid entirely by consumers. A large portion of the financial losses of the utilities during the crisis is now being recovered in the retail prices, under either court order or CPUC negotiation. The financial losses incurred by the State in buying electricity during the crisis have been financed through over $12 billion in revenue bonds. Debt service of these bonds is being paid by the electricity consumers. The retail price increases were heavily weighted against industrial/commercial users of electricity and against residential customers who for whatever reason were large users of electricity. As currently implemented, the rates (including surcharges) are highly tiered, with households that use low amounts of energy paying prices about equal the cost of delivery services plus acquisition cost of electricity and those using larger amount of electricity paying about three times that marginal price (depending on the particular geographic zone, about $0.24 per KWhr.) During the last year, the well-head price of natural gas has risen sharply, partly as a result of hurricanes Katrina and Rita. At the time of this writing, the price of natural gas at Henry Hub has increased to over $14/mcf, in comparison to $6.50 as of September Although natural gas prices can be expected to decline after the heating season, absent additional damages to the natural gas production facilities, these natural gas prices are expected to remain considerably higher than they have been historically. Such high natural gas prices can be expected to further increase the wholesale prices of electricity and therefore the retail prices. Litigation Some of the historical costs of the crisis may not be borne by consumers, but may be recovered through litigations of various types. Prominently, FERC continues its proceedings to order refunds to California for times when generators and marketers California for electricity prices that were not justified under FERC rules. Many such refunds have already been ordered. Private litigations attempt to reallocate costs from electricity users, back to generators and marketers. Some firms that signed long-term power purchase contracts during the time of the crisis have sued or negotiated for modifications to the price terms in these contracts. In addition, the U.S. Department of Justice has brought criminal indictments against one company and its employees. The legal processes continue. Changing Market Structure, Market Forces Draft. Do Not Quote. 59

63 The dual crises the electricity crisis and the financial crisis motivated many reforms of the reforms. Some changes occurred during the crisis, other have been being implemented in the years since the crisis was over. Fundamental Market Flaws Happily, the most fundamental market flaws the reliance on spot markets and the retail price controls have been repaired. The inappropriate over-reliance on spot markets has been eliminated. Utilities are no longer relying on spot market purchases for other than small fractions of electricity purchases. Almost all of their electricity is acquired through a combination of self-generation (dominantly, but not entirely, from generating units in existence before the crisis), from the long-term contracts negotiated by the DWR, and from QF contracts from before the crisis. A few new shorter-term contracts have been entered by the utilities. Very little other electricity is being transacted on the organized spot markets. The PX is defunct; no electricity has been sold on that market since Real time purchases on the CAISO are sharply reduced from during the crises, since IOUs are no longer systematically underscheduling on those markets. The FERC has instituted penalties for under-scheduling into the CAISO and the IOUs, with most of their electricity purchased on term markets have no incentive to try exercising market power through under-scheduling. The CPUC has ordered resource adequacy rules for IOUs, including minimum planning reserve margins of 15%. At least 90% of the necessary resources must be acquired at least one year in advance 67. This represents a fundamental shift in the regulations, a shift motivated by the experiences of the crisis. Retail price controls no longer are operational; normal ratemaking procedures now govern retail pricing. Of course, normal procedures do not allow wholesale market price fluctuations to be quickly incorporated into retail prices. Thus, short-run demand functions for electricity still are quite unresponsive to wholesale price changes. However, longer-term trends in cost changes can be incorporated into retail rates. This possibility allows demand functions to adjust over the course of a few months to changes in wholesale prices and it helps assure that IOUs will not face a similar financial crisis if wholesale prices do systematically increase 68. Retail Competition on Hold During the crises, the California legislature recognized that the long-term power purchase contracts, then being signed by the DWR, would obligate the State to acquire those contractual quantities of electricity. They further recognized that after the crises were over, the contracts 67 However some of the contracts seem to not be based on identifiable generating assets, but rather liquidated damage contracts. Liquidated damages will protect the utility but will not necessarily assure adequate supplies of electricity. Regulatory rule making in this area is continuing. 68 At the time of this writing, natural gas prices at Henry Hub have increased to $14/mcf, up from $6.50 one year ago. This will increase the electricity generation costs and therefore can be expected to increase retail rates. Draft. Do Not Quote. 60

64 would become obligations of the utilities to purchase the bulk power. In order to protect the State and the utilities from risks that the contractual power would either be a greater quantity than needed by the utilities or at an uncompetitively high price, the legislature passed a law suspending the right for electricity users to enter new direct access contracts. This suspension of all retail competition would last as long as the state contracts were still operational, until roughly As originally envisioned, this law would preclude companies and individuals from entering contracts to purchase electricity, except from their local utilities or from municipalities. Subsequently, the CPUC established some limited opportunity for firms to enter direct-access contracts, bypassing the utilities, if those firms paid an exit fee, designed to cover that firm s share of the historical costs. This law and the CPUC implementing rules assure that the costs of the DWR long-term contracts would be recovered from ratepayers. Governor Schwarzenegger has supported the reopening of direct access to acquire electricity, through a core/non-core structure in which large users of electricity would be able to enter direct access contracts, but smaller core customers would rely on the utility to supply electricity 69. Such a regulatory change is still under debate. In the meanwhile, there now is no effective retail competition. Electricity Demand Measures The CPUC, the CEC, and utilities, are developing and implementing systems to improve electricity consumption response to changing conditions. Demand management contracts have been entered with large commercial and industrial users of electricity. Some contractually require those users to curtail use of electricity when conditions are tight; others give the utility the ability to remotely curtail that use. Additional demand response mechanisms are under debate or experimentation. However, much more research and policy analysis is needed to increase the short-run and the longer-run demand response to wholesale prices. This is particularly important because the current market structure so severely blunts the connection between hourly wholesale prices and the retail prices that most consumers face. A second approach is tighten the connection between hourly wholesale prices and the retail prices, possibly through real-time pricing, or some variant of real time pricing, such as critical peak pricing. Experiments are underway to test real-time pricing. Interval meters have been installed in facilities of the largest industrial users, allowing the possibility of real time pricing. Tariffs that include variants of real time pricing for electricity have been developed and tested, but have not been broadly implemented. PG&E and SDG&E have proposed to deployed advanced meters to all customers over the coming years. The significance of demand response can be seen in Figure 19, developed by the CEC, which estimates peak demands for electricity in California for the various usage categories. Almost 30% of the peak demand is for residential or commercial air conditioning. This use, in principle, can be shifted significantly over time within the course of a day. However, absent incentives or controls to do so, there is no reason to expect that a large fraction of this load will shift away from the times of the greatest demands on the system. With appropriate incentives, and with appropriate information technology, such uses can be shifted in time so as to optimize the overall 69 Such a core/non-core structure was one of the options suggested in the Yellow Book. Draft. Do Not Quote. 61

65 system, reducing the costs at the highest peak loads, and reducing the need for new peaker power plants. California is continuing and expanding its emphasis on longer-term energy efficiency measures. Energy efficiency programs organized through the IOUs have been expanded through a CPUC rulemaking process. The State of California continues its Flex Your Power publicity campaign, providing information about and encouragement for energy efficiency. The Governor has ordered all State-owned building to be managed so as to meet minimum energy efficiency standards. New appliance efficiency standards have been established; others are under consideration. Energy efficiency in incorporated into California building standards and additional rulemaking can be expected. These actions collectively can be expected to continue the trend of per-capita electricity use in California growing significantly less rapidly than in the rest of the United States. Draft. Do Not Quote. 62

66 Figure 19. California Peak Demands by Usage Category IOU Wholesale Electricity Procurement CPUC has recently established rules that now require IOUs to acquire new electricity supplies through a competitive process, intended to level the playing field across different potential suppliers of electricity, in particular between IOU self-supply of electricity and IPP generation. Such a competitive procurement process increases the probability that the utility portfolio will be based on minimum cost, for the given risk, reliability, and environmental performance. But such rules do not eliminate the incentive 70 for utilities to purchase from themselves in preference to IPPs; the rules simply make it more difficult for the utilities to act in accordance with that incentive. Moreover, the rules can be applied to allow the utilities to exercise market power over existing generators that do not yet have long-term contracts to sell their power 71. However, the 70 IOUs are not compensated for the credit risk they take on when they enter long-term contracts, although they are compensated for risks when they build their own plants. This provides an incentive to avoid long-term contracts in favor of self-generation. Most of the IPPs are in difficult financial situation, so that there is risk of long-term contracts with these entities. Regulatory uncertainty makes long-term commitments risky, but whether long-term contracts with IPPs are more risky than construction of utility-owned generation is unclear. Finally, though, utilities earn a regulated rate of return on own-generation; they do not earn a return on contracts. 71 For example, the ongoing PG&E procurement process excludes already completed IPP plants, such as the Calpine Metcalf plant, from bidding to supply electricity. This restriction is being negotiated at the time of the Draft. Do Not Quote. 63

67 alternatives simply ban the utilities from any self-generation or any new self-generation would create important dysfunctional incentives and would preclude utility self-generation when such an option was truly the best for consumers. In addition, competition among wholesale buyers of electricity is currently very limited in California. Because California no longer has any meaningful retail competition in electricity markets, the IOUs are virtually the only wholesale electricity buyers in California. This market domination allows the IOUs to exercise market power over the IPPs. IPPs, understanding the dominance of the IOUs as buyers, are unlikely to start construction of new power plants, absent long-term contracts with the IOUs. Moreover, IOUs seem unwilling to enter long-term contracts with IPPs to purchase electricity. Thus while in principle wholesale procurement rules promote a level playing field, whether the field will turn out in practice to be truly level is not clear. ISO Market Redesign The CAISO is developing a new market design in order to correct flaws in California wholesale electricity markets. The current design includes nodal pricing for electricity, rather than the original zonal pricing, and firm marketable transmission rights. Market software to support the new market design is under development. But the process is very slow and costly. Originally entitled MD 02, for Market Design 2002, the process has been renamed, reflecting the reality of the process speed. Capacity Incentives It is generally realized that absent payment for providing unused generation capacity, electricity suppliers are unlikely to have economic incentives to invest in enough new capacity to assure that there will always be an adequate reserve margin of unused generation capacity. In order for generators to provide sufficient capacity to assure that they can cover very uncertain peak loads, they need incentives to do so. One option is for generators to enter long-term contracts to sell a bundled combination of electricity and generation capacity, with the generation capacity some agreed-upon fraction greater than the expected electricity generation. But such a bundled combination would be more costly than an electricity-only contract. Thus, absent regulatory rules that require utilities to enter such bundled combinations, it is unlikely that such combinations could compete with electricity-only contracts. even when it is not used. In periods of adequate electricity supply, the short-term prices for bulk power tend to be too low to repay the capital costs. In periods of short supply, in theory, prices might increase enough that these limited time periods could allow enough profits to cover the long-term capital costs. In practice, however, the electricity crisis has made it clear that wholesale price mitigation can be expected. Such price mitigation measures virtually assure that generators could not repay the long-term capital costs through such short-time periods of short supply. In addition, generators can only wait a limited time to recover capital costs. The need to writing. But it does illustrate the ability of the IOU to adopt procurement rules that do not truly level the playing field, since PG&E does not exclude itself from acquiring electricity from its existing generation assets. Draft. Do Not Quote. 64

68 obtain financing through normal financial markets implies that waiting too long to recover capital costs may be a route to bankruptcy. In response to this recognized problem, the State is examining how to create an electricity generation capacity market that would allow utilities and generators flexibility in meeting resource adequacy requirements. The market could include tradable capacity rights or obligations. However, design is far from complete. Thus it is not clear when and whether such a market will be created. Other Electricity-Related Policies Environment Environmental protection has been an important component of electricity policy, both before and since the crises. However, since the crises, issues of environment have become more explicit. Broad classes of electric generation/use technologies viewed as environmentally attractive are given priority over those that have greater adverse impacts on the local environment or on the global environment, through a loading order among classes of technologies. The highest priority is energy efficiency energy use technologies and practices that reduce the overall cost of providing energy services, without reducing the available services. Second priority is demand response and conservation measures, as discussed above. Third highest is renewable energy for electricity generation. The renewable portfolio standard in current law requires each IOU to increase its use of renewable sources of electricity to 20% by Governor Schwarzenegger is supporting acceleration of this standard to 20% by 2010; the CPUC has adopted a policy supporting that acceleration. The California administration is examining whether a 35% standard would be viable at some future year. The CEC is funding research and development to improve use of renewable technologies in California. In addition, policies initiatives, such as the Million Solar Roofs initiative, a subsidy program for installation of distributed photovoltaics, are under debate 72. Finally, Governor Schwarzenegger has established greenhouse gas emissions goals for the state and has directed members of his administration to develop strategies to meet those goals. Although many options are being examined, it is likely that some of the strategies will significantly influence electricity generation and use technologies. Transmission During the electricity crisis, transmission capacity significantly limited movement of electricity between northern and southern California. Over a longer term, although existing transmission lines allow electricity movement between California and neighboring states, there is very little transmission capabilities between California and the more distant states. 72 The Million Solar Roofs initiative was submitted to the legislature in It did not pass during that legislative session; further legislative development can be expected during the next session. Draft. Do Not Quote. 65

69 Since the crises, the most important transmission bottleneck in California, Path 15 linking northern and southern California has been upgraded to increase its capacity. Many minor upgrades have been completed or are underway within California. Four Western governors 73 have proposed a transmission line, the Frontier Line, with a 6000 MW capacity, from Wyoming to California. This would enable large increases of electricity from the Intermountain west to be developed and transmitted across the region, including to California. Advocates of the program stress the possibility of importing more electricity generated using renewables and clean coal technologies. Yet there is the possibility of importing electricity generated from high-carbon-dioxide-emitting coal plants. Work continues to identify alternative routes, and to examine regulatory, financial, environmental, economic, and engineering issues. Prominent among the regulatory rules will be possible electricity acquisition policies to assure that the electricity imported into California is dominantly based on renewables and clean coal technologies. If the Frontier Line is completed, it would reduce the current risky dependence on natural gas for electricity generation. Natural Gas Natural gas limitations were fundamental to the electricity crisis, as discussed above. Since the crisis, natural gas storage facilities have been expanded 74. Many proposals have been made and are under State consideration for siting liquefied natural gas (LNG) facilities in California. Although no particular site has been approved, approval of some site is highly likely. Improvements of such natural gas infrastructure could reduce the risk of electricity problems stemming from natural gas restrictions or high natural gas prices. However, it is likely to be many years until the gas can be delivered from the LNG facilities currently being considered. Organizational Issues During the crisis, the state legislation created a State power authority 75, with broad authority to construct new electric generating facilities and to acquire existing facilities by use of eminent domain procedures, making the state as a major electricity generator. It initially signed letters of intent to finance over 2000 MW of renewable energy and over 3000 MW of natural gas-fired gas fired peaker units, but all these plans have been cancelled. The authority has been dismantled. Governor Schwarzenegger has proposed a consolidation of the State energy agencies in order to improve the capability of the state to deal with its energy problems. The reorganization plan, currently being revised, would create a California Department of Energy and would integrate policy, planning, conservation/efficiency programs, and R&D functions into that agency. Such a plan would make energy more central to decision making within the Governor s office and could be expected to improve governmental reactions to further energy problems. However, whether such a reorganization will be approved by the legislature is still uncertain. 73 Wyoming Gov. Dave Freudenthal, Utah Gov. Jon Huntsman Jr., Nevada Gov. Kenny Guinn and California Gov. Arnold Schwarzenegger. The cost is estimated to be $3.3 billion. 74 The Wild Goose Storage expansion was complete in Formally, the California Consumer Power and Conservation Financing Authority Draft. Do Not Quote. 66

70 Proposals to Re-regulate The discussion above is based primarily on the policy and regulatory actions being taken by the California administration and legislature, actions generally consistent with reforming the reforms, by fixing the problems that exist in the California markets, generally moving in a direction of competitive markets where and when possible. However, some California groups do not share the vision underlying these actions but would return to the old system of vertically integrated rate-of-return-regulated utilities. In 2004, the State legislature passed a bill, AB 2006, authored by speaker of the Assembly, Fabian Nunez, that would have moved toward reestablishing vertically integrated utilities. It would have assured that the utility could recover any costs it incurred in building its own generation units and would encourage utilities to build and operate their own generators rather than entering term contracts with IPPs. It would have made rules for direct access so onerous as to preclude any new direct access contracts, thus permanently eliminating retail competition. The Nunez bill enjoyed strong support, and considerable lobbying, by Southern California Edison, one of the two largest utilities. The Governor vetoed that bill. At the time of this writing, an initiative, Proposition 80, entitled by its sponsors the Repeal of Deregulation and Blackout Protection Act, is scheduled for California vote in November Sponsored by The Utility Reform Network (TURN), that initiative would repeal many of the provisions of AB It would permanently prohibit new direct access contracts. It would subject all electricity suppliers to CPUC regulation. It would make real-time pricing or any electricity pricing that varies with time of day, day of week, or month of the year very difficult or impossible to implement for residential consumers. It would mandate a renewable portfolio standard of 20% by 2010, but would not allow additional increases of that standard. It would require a 2/3 vote of both houses of the legislature to modify any provisions of the initiative 76, and would thereby sharply reduce the flexibility to pursue reforms of the reforms. By the time this book is published, California voters will have accepted or rejected Proposition 80, and will have thus greatly influenced the future directions of change in California electricity markets. Whatever the vote, the qualification of Proposition 80 as an initiative provides strong evidence that there is a strongly felt opposition to market-based reforms within California s restructuring. Public Perceptions The crises have left a legacy of public perceptions, many of them quite distorted or, at best, incomplete. The electricity system, its operations, and the economics of electricity are hard for even experts to understand fully. For non-specialists these issues are even more difficult and thus non-specialists tend to rely on information from news media. During the energy crisis, political leaders released and the press reported a tremendous amount of politically inspired miscommunication about the nature and causes of the price increases, the financial problems, 76 In fact it would allow the legislature to amend this act only to achieve its purposes and intent. This language assures that virtually no further reforms would be possible without prolonged litigation. Draft. Do Not Quote. 67

71 and the blackouts 77. Official press releases villainized electricity generators and traders, particularly out-of-state generators. The communication distortions have now been muted, but they continue. Now that the crises are over, the impressions of many people are based on what they learned during the crisis. Unfortunately, the more distorted are the public perceptions, the more difficult to implement good public policy. An example of politically motivated distortions appears in the 2005 California voters pamphlet arguments in favor Proposition 80. Proposition 80 advocates assert: California s failed experiment in electric deregulation cost our people and businesses billions of dollars, implying that these costs occurred because Enron and other traders held California hostage. 78 As described above, these advocates offer a ballot initiative which, if it passes, will permanently eliminate retail competition, restrict time varying pricing, and mandate particular levels of renewable resources, policy changes quite unrelated to Enron, energy traders, of costs of the electricity and financial crises. Some Reflections Unfortunately, one message repeatedly communicated is that the California experience proves that electricity deregulation has been an utter failure in California, and by extension, is likely to be a failure elsewhere. Yet a fair assessment of the California experience cannot reach such a conclusion. In California, restructuring has set the stage for widespread wholesale market competition and adequate electric generation capacity. California s restructuring has resulted in increases in new electric generating plants proposed, approved, and under construction. An oft-repeated contention is that deregulation caused the electricity crisis. However, the Western electricity crisis appears to be primarily the product of a perfect storm, a combination of several adverse conditions, of flawed market rules, and only secondarily of exercise of market 77 For example, during the crisis, Governor Davis public communications, carried on television, radio, and newspapers, assign all virtually blame for California s electricity problems on electricity generators and marketers. Words and phrases such as profiteering, plunder, unconscionable, price gouging, exorbitant profits, market marauders, pirate generators, privateers, obscene profits, and outrageous wholesale prices were essential parts of his formal statements. 78 The initiative justification in the Voter s pamphlet begins as follows: Five years ago, California was devastated by an electricity crisis. Enron and other energy traders held California hostage, extorting tens of billions of dollars from us. They manipulated the electricity market, driving up wholesale prices 1000%. California faced rolling blackout and untold economic damage. Audiotapes released by the U.S. Justice Department revealed Enron energy traders boasting of making buckets of money by creating power shortages. One trader laughed about all the money you guys stole from those poor grandmothers in California, while another ordered a power plant worker to just go ahead and shut her down. California s failed experiment in electric deregulation cost our people and businesses billions of dollars. We learned many lessons from that disaster. The state has taken some positive steps to clean up the mess but not nearly enough. Amazingly, legislation to require sufficient supplies of electricity [AB 2006, the Nunez bill] was vetoed by the Governor last year. That s why Proposition 80 the Repeal of Deregulation and Blackout Prevention Act is on the ballot. The statement was signed by Robert Finkelstein, Executive Director of TURN; Richard Holober, Executive Director, Consumer Federation of California; Nan Brasmer, President, California Alliance of Retired Americans. Draft. Do Not Quote. 68

72 power and market gaming. The financial crisis was the direct result of California regulatory actions. However, the financial crisis was not the result of deregulation, but rather of inappropriate regulation. Even though all municipal utilities and IOUs throughout the entire West faced the electricity crisis, only the IOUs located in California, facing the CPUC regulatory rules, experienced the financial crisis. The California experience illustrate that actions economically isolating the supply side of markets from the demand-side create problems. Particularly retail price control isolated consumer from the changing economics of the supply side of electricity markets. Retail price control discouraged energy demand reductions energy conservation even though energy conservation was California s best hope for forcing decreases in wholesale prices. The California experience also brought into sharp focus the importance of managing risks associated with implementation of any policies, especially policies that radically change the system. California electricity restructuring radically changed regulations, altering an electricity system that itself has never been free of economic risk. Risks exist and they should be managed, distributing risks appropriately. The California experience highlights something well known to everyone who has managed, participated in, or observed large-scale changes in complicated organizations or economic structures. Any major restructuring of such important systems will continue to require modifications well after the initial changes. System operation requires monitoring and may require leadership to identify and implement changes that are needed after unintended adverse consequences of the system change become apparent. California's electricity system includes both complicated organizations and complex economic structures. The restructuring was fundamental and sweeping. No one was able to predict with any certainty how all the changes would work. Nor could anyone rule out unintended adverse consequences of the restructuring. System flaws requiring changes can be expected. Thus, it was important that State carefully monitored system operation and it was important that leadership was available to respond to the problems once identified. The California experience shows the real lasting harm that can befall a state when its leaders fail to take the appropriate leadership roles. Draft. Do Not Quote. 69

73 References Severin Borenstein, James Bushnell and Frank Wolak, 2002 Measuring Market Inefficiencies in California s Restructured Wholesale Electricity Market, American Economic Review, December, 92(5), Severin Borenstein. The Trouble with Electricity Markets: Understanding California s Restructuring Disaster, Journal of Economic Perspectives, 16(1): California Public Utilities Commission. California s Electric Services Industry: Perspectives on the Past, Strategies for the Future, (The Yellow Book) February 3, 1993, California Public Utilities Commission. Decision D California Public Utilities Commission. Order Instituting Rulemaking on the Commission s Proposed Policies Governing Restructuring California s Electric Services Industry and Reforming Regulation. (The Blue Book) Decision (December 20, 1995) as modified by D (January 10, 1996). California Public Utilities Commission. Roadmap Decision CPUC Decision D Dennis Carlton and Jeffrey Perloff. Modern Industrial Organization. Harper Collins College Publishers (1994), p 139 and p 922. Ralph Cavanagh. Revisiting the Genius of the Marketplace : Cures for the Western Electricity and Natural Gas Crises, Electricity Journal, June Robert Thomas Crow, Not Invented Here: What California Can Learn from Elsewhere about Restructuring Electricity Supply. Stanford Institute for Economic Policy Research Working Paper, December Scott Harvey and William Hogan, On the Exercise of Market Power Through Strategic Withholding in California, April 2001 William W. Hogan. Electricity Market Restructuring: Reforms of Reforms, May 25, 2001, Harvard University. Jolanka V. Fisher and Timothy P. Duane. Trends in Electricity Consumption, Peak Demand, and Generating Capacity in California and the Western Grid, September, 2001, working paper from the University of California Energy Institute Program on Workable Energy Regulation (POWER). Energy Information Administration. Electricity Shortage in California: Issues for Petroleum and Natural Gas Supply. Draft. Do Not Quote. 70

74 Energy Information Administration. Natural Gas Monthly. Federal Energy Regulatory Commission: Order On Rehearing Of Monitoring And Mitigation Plan For The California Wholesale Electric Markets, Establishing West-Wide Mitigation, And Establishing Settlement Conference (Issued June 19, 2001) Federal Energy Regulatory Commission: Report On Plant Outages In The State Of California Prepared by: Office of the General Counsel, Market Oversight & Enforcement, Office of Markets, Tariffs and Rates Division of Energy Markets, February 1, 2001 Federal Energy Regulatory Commission: Order Directing Remedies For California Wholesale Electric Markets Issued: December 15, 2000 P. Joskow and E. Kahn, A Quantitative Analysis of Pricing Behavior in California s Wholesale Electricity Market During Summer 2000, NBER Working Paper 8157, March 2001 Paul Joskow and Edward Kahn, A Quantitative Analysis of Pricing Behavior in California s Wholesale Electricity Market During Summer 2000, The Energy Journal, 23(4),1-35. South Coast Air Quality Management District. White Paper on Stabilization of NOx RTC Prices, web site: James L. Sweeney. The California Electricity Crisis Hoover Institution Press; Stanford Institute for Economic Policy Research. Stanford, California. U.S. General Accounting Office. Energy Markets: Results of Studies Assessing High Electricity Prices in California, June Western Governors Association, Conceptual Plans for Electricity Transmission in the West August Draft. Do Not Quote. 71