Resource Adequacy in California Options for Improving Efficiency and Effectiveness October 2012 Johannes P. Pfeifenberger Kathleen Spees Samuel A. Newell Prepared for Copyright © 2012 The Brattle Group, Inc. This material may be referenced subject to citation. Reproduction or modification of materials is prohibited without written permission from the authors. About the Authors Johannes P. Pfeifenberger and Samuel A. Newell are Principals and Kathleen Spees is a Senior Associate of The Brattle Group, an economic consulting firm with offices in Cambridge, Massachusetts, Washington DC, San Francisco, London, Rome, and Madrid. They can be contacted at www.brattle.com. Acknowledgements and Disclaimer The authors would like to thank Matt Barmack, Avis Kowalewski, and Mark Smith of Calpine, Peter Fox-Penner of The Brattle Group, Gary Stern and Kevin Duggan of Southern California Edison, and others for providing valuable comments and insights for developing and refining this report. Opinions expressed in this report, as well as any errors or omissions, are the authors’ alone. The examples, facts, results, and requirements summarized in this report represent our interpretations. Nothing herein is intended to provide a legal opinion. TABLE OF CONTENTS Executive Summary .......................................................................................................................1  I.  Motivation ..............................................................................................................................5  II.  California’s Resource Adequacy Framework .....................................................................6  A.  Description of Resource Adequacy Provisions ..............................................................6  1.  Utility Long-Term Procurement Plans ...................................................................7  2.  System and Locational Resource Adequacy Requirements ...................................9  3.  CAISO Backstop Capacity Procurement Mechanism ..........................................10  4.  CAISO and CPUC Efforts toward Meeting Flexible Resource Needs .................11  B.  Inefficiencies Introduced under the Current Framework .............................................12  1.  Price Discrepancies among Different Types of Capacity Resources ...................12  2.  Lack of Competition between New and Existing Resources ................................15  3.  Uneconomic New Generation Investments Driven by Planning Uncertainties ....17  4.  Potential for Inefficient Once-through-Cooling Replacements and Retrofits ......18  5.  Forward Backstop Mechanisms that Could Preempt Market Alternatives ..........19  6.  Inefficient Cost-Effectiveness Tests for Demand Response ................................20  7.  Difficulty in Attracting Third-Party Demand Response .......................................21  8.  Lack of Liquidity and Transparency in Short-Term Bilateral Transactions .........22  III.  The Value of Market-Based Resource Adequacy Mechanisms ......................................22  A.  Non-Discriminatory Capacity Procurement ................................................................22  1.  Advantages of Non-Discriminatory Capacity Procurement .................................22  2.  Ability to Attract Low-Cost Alternatives to New Generation ..............................25  3.  Ability to Attract Cost-Effective New Generation ...............................................27  B.  Forward Resource Adequacy Requirements ................................................................29  1.  Advantages of Multi-year Forward Procurement .................................................29  2.  PJM Forward Market Performance with MATS Regulation ................................31  IV.  Improving California’s Resource Adequacy Framework ...............................................34  A.  Reforming LTPP for Non-Discriminatory Capacity Procurement ..............................35  B.  Increasing RAR Program Efficiency Through Centralized Auctions ..........................36  1.  Residual Capacity Auctions to Fulfill RAR Obligations ......................................37  2.  State-Administered, IOU-administered, or CAISO-Administered Auctions .......39  3.  Prompt vs. Forward Auctions ...............................................................................40  4.  Single-Year vs. Multi-Year Commitments ...........................................................40  5.  Other Capacity Auction Design Options ..............................................................42  C.  Reforming both LTPP and RAR ..................................................................................44  1.  RAR Capacity Auctions with Non-Discriminatory LTPP Procurement ..............44  2.  Replacing LTPP and RAR with an Integrated Forward Capacity Market ...........45  V.  Recommendations ................................................................................................................46  Bibliography .................................................................................................................................47  List of Acronyms ..........................................................................................................................53  i EXECUTIVE SUMMARY We have been asked by Calpine, a generation owner and developer in California and nationally, to evaluate the efficiency and effectiveness of California’s approach to attracting and sustaining investments for electric resource adequacy. Our analysis focuses on: (1) the Local and System Resource Adequacy Requirement (RAR) that the California Public Utilities Commission (CPUC) imposes on retail suppliers; and (2) the Long Term Procurement Plan (LTPP) process through which the CPUC oversees the procurement of new conventional generation by investor owned utilities (IOUs) on behalf of all CPUC-jurisdictional load serving entities (LSEs) as well as the IOUs’ procurement of new and existing resources on behalf of bundled customers. These interrelated but uncoordinated mechanisms were initially introduced in response to the Western power crisis and have evolved over the past decade with changing circumstances and policy goals. Looking out over the coming decade, California faces two pressing challenges to meeting the State’s environmental and reliability objectives that were not anticipated at the time the current framework was developed. The first challenge is the once-through cooling mandate that will require approximately 16,000 MW of existing generators, representing one-third of California’s fleet, to either retire or invest in costly environmental upgrades over the next decade. The second challenge is California’s renewable portfolio standard (RPS) to procure 33% of customers’ energy from renewable resources by 2020. Bringing these large quantities of intermittent wind and solar generation into the system will suppress energy market prices and require additional resources that can operate flexibly enough to balance the variable renewable generation output. These new challenges highlight the need to take a fresh look at California’s resource adequacy framework so that it is better aligned with policy objectives and better able to meet these objectives at lowest cost. In reviewing the current mechanisms that were independently developed for separate purposes, we identified a number of inefficiencies as explained in Section II of this report. Many of these inefficiencies stem from the fact that these mechanisms do not constitute an integrated approach that fosters competition among different types of capacity resources. These inefficiencies include: Price discrepancies among different types of capacity resources – Currently, different types of capacity resources are paid very different prices for providing the same product, with existing resources earning only approximately $18-38/kW-year (or less) under the RAR program while IOUs are paying the equivalent of $150-300/kW-year in capacity payments for contracts with new resources under LTPP. This large price discrepancy indicates that California is likely substantially overpaying for new generation and forgoing lower-cost opportunities to retain existing resources. Many existing resources may retire prematurely if they are compensated at these levels, even though they could be retained for less than the cost of building new generation. Lack of competition between new and existing resources – California does not currently have a bilateral or auction-based market mechanism for fostering efficient, direct competition between new and existing generation resources. Greater competition between new and existing resources would create opportunities to identify relatively low- cost uprate, retrofit, and repowering opportunities for existing generators, enable the 1 efficient retirement of units that are no longer economic, and postpone the need to invest in costlier new generation resources. Uneconomic new generation investments driven by planning uncertainties – There are large uncertainties in the outlook for load growth, retirements, imports, and demand resources. This means that IOU and CPUC projections under LTPP of whether and when new generation will be needed will necessarily prove to be imperfect. In some cases, these uncertainties led to over-estimating the quantity of new resources needed for resource adequacy, and therefore to prematurely building new generation resources at contract prices that far exceed the current cost of alternative supplies. Potential for inefficient once-through-cooling retirements and retrofits – It is not clear what portion of the 16,000 MW of resources subject to the once-through-cooling mandate over the coming decade will retire and how much (if any) may be retrofitted cost effectively. California does not have a market-based competitive mechanism for evaluating which of these resources could be upgraded cost-effectively and which should retire because lower-cost alternatives exist. Forward backstop mechanisms that could preempt market alternatives – While the CAISO’s two-year, forward-looking backstop procurement authority has not yet been used, this mechanism has the potential to distort prices in the energy markets and RAR program. Forward out-of-market contracting may pre-empt the market from identifying lower-cost alternatives for meeting the RAR objectives. The new proposal to extend forward backstop procurement authority to five years for flexible resources would also be implemented without any market-based mechanism for assuring that the selected resource is the lowest-cost option for meeting the system’s future flexibility needs. Inefficient cost-effectiveness tests for demand response – The CPUC cost-effectiveness tests for demand response (DR) resources assume that peak load reductions provide customers $136/kW-year of savings in capacity costs. This administratively-determined parameter values capacity as if California were in a long-run equilibrium in which the marginal cost of capacity is equal to the net cost of new entry (CONE) for procuring new generation. In reality, California currently has substantial excess capacity and the cost of alternative capacity supplies is only $18-38/kW-year (or less), meaning that customers may be paying for DR programs whose costs currently exceed their benefits. Difficulty in attracting third-party demand response – Despite a high DR cost-effectiveness threshold, California may actually be under-procuring low-cost DR by effectively precluding third-party DR suppliers from accessing capacity payments. By allowing third-party curtailment service providers (CSPs) to monetize the value of peak load reductions without going through LSEs, eastern U.S. power markets such as PJM observed rapid growth in low-cost DR, sufficient to cover 10% of the system’s peak load for 2015-16. These third-party DR resources have taken on resource adequacy commitments at capacity prices far below the cost of new generation and the “capacity value” assumed in California’s DR cost-effectiveness tests. Lack of liquidity and transparency in short-term bilateral transactions – The current RAR program relies solely on bilateral transactions, which are inherently less efficient and transparent than centralized auctions or over-the-counter trading platforms. The bilateral approach also increases transactions costs and makes it more difficult to estimate the value of incremental investments in short-term capacity resources such as DR, generation 2 uprates, and imports. Creating more transparency through a centralized auction or an over-the-counter capacity exchange would reduce these inefficiencies and better inform appropriate bilateral contract prices. Overall, these inefficiencies in California’s resource adequacy construct will lead to uneconomic investments in new capacity resources when lower-cost alternatives exist, ultimately increasing system and customer costs. California could gain substantial efficiency benefits by reforming the current LTPP and RAR programs in two important ways. First, the programs could be refined to incorporate non- discriminatory procurement practices that invite competition among all types of capacity resources including: (a) new generation; (b) existing generation, including resources with low going-forward costs, as well as those that need major reinvestments or retrofits to continue operating; (c) investments to uprate existing generation facilities; (d) imports; and (e) demand- side resources including DR and energy efficiency. Unless each of these types of resources has the opportunity to compete to supply capacity at the same price and under the same terms, it will not be possible to meet resource adequacy objectives using the lowest-cost mix of supply resources, as we explain in Section III.A. Second, California may gain additional efficiency benefits by procuring all of its needed capacity commitments on a 3-4 year forward basis, as explained in Section III.B. Meeting the local and system RAR objectives on a forward basis will increase the scope of competition by awarding capacity commitments at approximately the same time that suppliers need to make major irreversible investment decisions for retrofits and new construction. To improve procurement efficiency and achieve these benefits, we recommend either: (1) reforming RAR and LTPP to incorporate non-discriminatory procurement practices; or (2) replacing both mechanisms with a forward capacity market. Further, the CPUC and CAISO appear to require assurance of resource adequacy on a forward-looking basis, particularly in light of upcoming once-through-cooling and flexibility challenges. Considering these concerns, the efficiency of either approach could be improved by including 3-4 year forward obligations covering all or most system and local capacity requirements, including requirements for operationally flexible capacity. Ideally, these forward obligations would be met through non- discriminatory, transparent, single-price auctions conducted by CAISO, a state agency, or the IOUs. If administered by a state agency or IOUs, CAISO would need to develop supplemental mechanisms to also cover the resource adequacy requirements of non-CPUC-jurisdictional entities. Option 1: RAR Capacity Auctions with Non-Discriminatory LTPP Procurement – The current resource adequacy construct would be more efficient and better integrated if all long, intermediate, and short-term procurements under LTPP were conducted on a non- discriminatory basis, with all remaining capacity needs procured at the RAR compliance deadline through a non-discriminatory residual capacity auction with transparent clearing prices. The auction could be administered by a state agency, the IOUs, or the California Independent System Operator (CAISO). Both LTPP and RAR procurements would invite offers from all types of capacity suppliers, including new generation, existing generation, uprates, imports, and demand resource, thereby rationalizing prices and enabling cost-effective tradeoffs among these types of resources. Capacity procurements under LTPP would also be separated from energy procurements, to reduce the potential 3 for administrative error in evaluating tradeoffs among very different types of resources such as combined cycle (CC) plants and DR that may have the same resource adequacy value but very different energy value. Implementing non-discriminatory procurement under LTPP would remove the restriction that long-term system procurements are only open to new generation, and would likely require re-framing the “needs determination” assessment to determine what portion of total capacity obligations should be procured at each forward period (e.g., 30% seven years forward, 50% five years forward, and 100% three years forward at the time of the RAR capacity auction). LTPP capacity procurements would also invite offers of any term from one year to many years, creating an opportunity to identify low-cost, short- term resources that may postpone the need to contract for long-term contracts with new resources. Only a portion of system capacity needs would be procured through the RAR auction, because most capacity commitments would continue to be procured on a forward basis prior to the auction. The increased transparency of an RAR auction would also enable more efficient bilateral contracting and LTPP procurements by more clearly identifying system needs and signaling the price at which alternative supplies are available. Option 2: Replacing LTPP and RAR with a Forward Capacity Market – An alternative option is to completely replace LTPP and RAR with a forward capacity market similar to those in New England and PJM, to be administered by a state agency, the IOUs, or CAISO. Similar to the first option, this approach would include a centralized capacity auction, but because capacity procurements under LTPP would be substantially reduced or eliminated, the auction would produce greater residual clearing volumes and become a more important part of California’s resource adequacy framework. This option would provide the highest level of efficiency by fully leveling the playing field among all resource types. Because this capacity auction would be the primary mechanism for assuring resource adequacy and attracting incremental investments in new and existing generation, it would be important to conduct the auctions on a 3-4 year forward basis, i.e., far enough forward that many suppliers still have the flexibility to make or reverse major investment decisions. A forward capacity market would also provide an efficient platform for co-optimizing procurement for both flexibility requirements and resource adequacy needs. We provide a more detailed explanation of these options for improving the California resource adequacy construct in Section IV. 4 I. MOTIVATION We have been asked by Calpine, a generation owner and developer in California and nationally, to evaluate the efficiency and effectiveness of California’s approach to attracting and sustaining investments for electric resource adequacy.1 Our analysis focuses on: (1) the Local and System Resource Adequacy Requirement (RAR) that the California Public Utilities Commission (CPUC) imposes on retail suppliers; and (2) the Long Term Procurement Plan (LTPP) process through which the CPUC oversees the procurement of new conventional generation by investor owned utilities (IOUs) on behalf of all CPUC-jurisdictional load serving entities (LSEs) as well as the IOUs’ procurement of new and existing resources on behalf of bundled customers. Resource adequacy has been a focus of California energy policy since the state experienced severe high prices and supply shortages during the Western power crisis of 2000-01.2 Over the past decade, the CPUC has developed a resource adequacy framework that relies partly on market-based incentives, partly on regulated planning and long-term contracting, and partly on backstop reliability interventions. However, these mechanisms have yet to be harmonized into an integrated approach to meeting system and local resource adequacy needs at lowest cost. In particular, the current approach does not foster direct competition among all types of capacity resources, including: (1) new generating plants; (2) existing generating plants, including cost- effective reinvestments to uprate, life-extend, or retrofit power plants that might otherwise retire; and (3) utility-sponsored and independently-developed demand response (DR) resources. Rather, California’s various mechanisms use different criteria to evaluate each type of supply, making them unlikely to achieve the most cost-effective result overall. Looking out over the coming decade, California faces two pressing challenges to meeting the State’s environmental and reliability objectives that were not anticipated at the time the current framework was developed. One challenge is the State Water Resources Control Board’s once- through cooling mandate.3 This mandate applies to 19 California power plants using once- through cooling with water drawn from coastal or estuarine sources, and requires them to upgrade to closed-loop cooling systems or implement other measures for comparably reducing impacts on aquatic life. The mandate will affect approximately 16,000 MW of resources, or approximately one-third of California’s generating capacity, over the next decade.4 In some many cases, the lowest-cost retrofit option may be clearly uneconomic (e.g., a costly new cooling tower investment on an aging plant), but the economics may not be as clear-cut in other cases 1 As a generation owner and developer in California, Calpine has an interest in the state’s mechanisms to ensure resource adequacy. While Calpine has commissioned this whitepaper, its contents represent our independent view and assessment of California’s resource adequacy framework. The conclusions that we draw in this whitepaper are solely based on our review and analysis of the California market design for resource adequacy and similar analyses we have undertaken in other markets. For example, see Newell, et al. (2009, 2010, 2012); Pfeifenberger, et al. (2008, 2009, 2011a-b); LaPlante, et al. (2009). 2 See Wolak (2003), pp. 17-18. 3 See SWRCB (2012). 4 Based on an analysis of individual units’ compliance dates and summer capacity, excluding resources with compliance dates prior to 2012 (which have already retired or complied) and excluding resources with compliance dates after 2022. Data from SNL Energy (2012); SWRCB (2012); CAISO (2010). As of August 2010, the 57,345 MW of resources were committed to CAISO for resource adequacy purposes, see CPUC (2011a), p. 13. 5 (e.g., alternative upgrades that are somewhat lower-cost, or a cooling tower on a plant that may operate for many more years). Ideally, California’s resource adequacy framework would use bilateral or auction-based market mechanisms to weigh the costs of reinvesting in these facilities against the costs of alternative sources of replacement supply. Unfortunately, current resource adequacy processes do not facilitate tradeoffs of this sort. A second emerging resource adequacy challenge arises from California’s renewable portfolio standard (RPS) to procure 33% of customers’ energy from renewable resources by 2020.5 Most of the new supply will be from intermittent wind and solar generation. Such a large amount of intermittent generation will create a need for additional resources that can operate flexibly enough to balance the variable renewable generation output. The California Independent System Operator (CAISO) and CPUC have both initiated efforts to address these intermittent resource integration challenges. However, these efforts have not yet developed into a market-based approach to assuring sufficient flexible resource capability in a manner that is consistent with the resource adequacy framework and the CAISO-administered ancillary service (A/S) markets. This whitepaper discusses the inefficiencies in the interplay among California’s current resource adequacy mechanisms, summarizes relevant experience from other power markets, and presents options that could be pursued to resolve these concerns. The options we describe rely on market mechanisms that foster competition among all types of capacity resources, while maintaining consistency with what we believe are the CPUC’s policy goals, including: (1) meeting system and local reliability needs at lowest cost; (2) meeting environmental objectives cost-effectively; and (3) maintaining state regulators’ ability to pursue environmental and other state policy objectives. II. CALIFORNIA’S RESOURCE ADEQUACY FRAMEWORK California’s resource adequacy framework relies on a number of interrelated but uncoordinated mechanisms. In this Section, we describe these mechanisms and describe their various inefficiencies, many of which are related to a lack of competition among different types of capacity resources. A. DESCRIPTION OF RESOURCE ADEQUACY PROVISIONS California has three mechanisms for achieving resource adequacy within different timeframes: (1) the LTPP needs-determination process, under which the IOUs determine whether to solicit long-term contracts for new generation to meet projected system-wide needs, typically contracting 3-7 years prior to the facility’s online date;6 (2) the Local and System RAR 5 See DSIRE (2012). 6 The LTPP process is also the mechanism under which IOUs can procure a portfolio of short-term and intermediate-term contracts for energy, capacity, fuels, and hedges on behalf of their bundled retail customers. However, the most important aspect of the LTPP for the purpose of this report is its long-term procurement process under which new generation is contracted if the IOUs’ project a system-wide need in their service area (including both the IOUs’ bundled customers as well as other LSEs’ customers). 6 mechanism that requires all retail suppliers to procure sufficient capacity to meet their customers’ capacity needs just prior to delivery; and (3) the CAISO’s backstop reliability mechanism for curing any LSE capacity deficiencies under the RAR mechanism and preventing generation retirements that could introduce prospective resource adequacy concerns.7 The CPUC and CAISO are also currently considering enhancements to these mechanisms and to the CAISO’s A/S markets to attract and retain sufficient flexible capacity resources. 1. Utility Long-Term Procurement Plans In the aftermath of the Western power crisis, the CPUC assigned the IOUs the responsibility of procuring new generation supplies within their service territories through the LTPP process.8 The CPUC oversaw the IOUs’ first biennial LTPP processes in 2004, and the 2012 LTPP proceedings are currently underway.9 The LTPP process begins with a 10-year outlook on supply, demand, and other system conditions, with the CPUC recommending extending that outlook to 20 years starting with the 2012 LTPP.10 Plans use this information for two purposes. The first purpose of the LTPP is to develop a short-and intermediate-term procurement plan that specifies how each utility will conduct procurement to meet its own bundled customers’ needs at the lowest expected cost, including a portfolio of short-term and long-term energy, capacity, A/S, fuels, renewables, and hedging positions up to ten years forward.11 A second, and for our purposes more important, aspect of the LTPP process is the long-term system-wide needs determination analysis, which determines whether and when the IOUs will contract to build new conventional generation resources.12 This determination is based on system needs, including for the IOUs’ own bundled customers as well as customers buying from competitive Electric Service Providers (ESPs) and Community Choice Aggregators (CCAs). In analyzing whether new resources will be needed for system-wide resource adequacy, the IOUs incorporate substantial direction from the CPUC and input from stakeholders to develop detailed projections of future system and market conditions, including: (a) anticipated future peak load and average energy demand; (b) projected generation supplies after considering projected new builds, uprates, and renewable generation growth; (c) projected retirements, including those that may be driven by supplier economics and others driven by environmental policies such as the once-through-cooling mandate; (d) quantities of supply that may or may not be available through imports; (e) expectations about transmission infrastructure changes; (f) demand response and energy efficiency programs not already accounted for in the demand forecasts; (g) market price projections for fuels and carbon dioxide equivalent (CO e) emission allowances; (h) the need for 2 flexible resources; and (i) other factors.13 Because these projections are highly uncertain, the CPUC requires that LTPPs consider several possible resource scenarios and develop a supply 7 Typical forward period of LTPP contracts based on a current summary of active PPAs procured under this mechanism. See CPUC (2012a). 8 See p. 26, CPUC (2007a), (2004a-b). 9 See CPUC (2012b). 10 See CPUC (2012c), p. 7. 11 See the 2010 IOU LTPPs, CPUC (2012d); p. 8, CPUC (2012c). 12 See, for example, Section 2.1 of CPUC (2007a). 13 See CPUC (2007a), (2012c). 7