Cal-ISO files a new market design, but has it traded efficiency for software?
Bruce W. Radford is editor-in-chief for Public Utilities Fortnightly.
Eyeing a launch date of November 2007, the California Independent System Operator (Cal-ISO) at last has come forward with plans for revamping its widely disparaged wholesale market design. The formal proposal, known as the MRTU (Market Redesign and Technology Upgrade), was filed this past February at the Federal Energy Regulatory Commission (FERC). It embraces many, if not most, of the ideas that already have won wide acceptance among utilities, regulators, traders, and power producers that deal with markets operated by regional transmission organizations (RTOs) in the Northeast and Midwest.
All told, the package includes some 5,000 pages of tariffs, testimony, and explanatory language, governing the management of some 4,000 geographic market nodes. The proposal appears so complex that Western Area Power Administration (WAPA), a skeptic of RTO-style trading regimes for the Western Interconnection, has estimated that it will need some 12.5 GB of computing power to provide portfolio management and scheduling services to wholesale power customers under the MRTU.
This daily data requirement stems from a complex market design that for the first time will introduce Californians to such features as: (1) a day-ahead spot energy market; (2) a centralized, bid-based, and security-constrained dispatch of generation; (3) locational marginal pricing (LMP) for both supply and load; and (4) tradable financial rights to manage grid congestion.
The tradable grid rights will be known as CRRs (congestion revenue rights), to distinguish them from the Cal-ISO’s previous physical construct of “firm transmission rights” (not to be confused with Eastern-style “FTRs”). Thus, the Cal-ISO’s MRTU would build on a fully nodal network grid model, thus avoiding the inefficient zonal approximations that in the past have led to trouble.(See, FERC Docket No. ER06-615, filed Feb. 9, 2006, and industry comments filed through May 12, 2006.)
In a key departure from Eastern RTO practice, however, the Cal-ISO’s MRTU would not immediately include an ISO-administered capacity market, in the manner of the ICAP, UCAP, and LICAP models, which have proved so controversial. Instead, to assure fixed-cost recovery for power producers, the MRTU would look to the state public utilities commission (Cal-PUC), and its resource adequacy requirement (RAR).
All the same, this lack of a proposed capacity market has not necessarily simplified the new California market structure. To the contrary, the ISO now must coordinate its tariff with state-imposed rules, which may not prove easy.
The Cal-PUC’s RAR rule, enabled by state legislation (Assembly Bill 380, enacting new sec. 380 to the state Public Utilities Code) obliges load-serving entities (LSEs) under PUC jurisdiction to sign contracts that will ensure access to power supply resources. (See Cal-PUC Decision 05-10-042, Oct. 27, 2005, 244 PUR4th 341.) Importantly, Assembly Bill 380 also enacts new Public Utilities Code sec. 9620, which assigns a prudent resource planning obligation to publicly owned utilities outside PUC jurisdiction, with a reporting obligation to the California Energy Commission.
To echo the locational aspects of capacity markets proposed back East (e.g., New England’s LICAP plan, and PJM’s Reliability Pricing Model, or RPM), the Cal-PUC promises by June 15 to issue amended RAR rules setting resource requirements for “local capacity areas,” known as LCAs. (See, Order Instituting Rulemaking, Cal-PUC No. R.05-12-013, Dec. 20, 2005, www.cpuc. ca.gov/Published/Final_Decision/ 52265.htm.)
That step will entail some controversy, since Cal-ISO in its MRTU has undertaken to conduct its own independent analysis of LCA capacity adequacy, thus serving as a federal “backstop” for the PUC, and perhaps even mandating extra resources, which could add uplift to ISO markets, and could create jurisdictional conflicts. Moreover, the PUC plans later this year to address the idea of a formal capacity market, but the jury is still out on that notion. (See, Assigned Commissioner’s Ruling and Scoping Memo, Cal-PUC No. R.05-12-013, Dec. 15, 2005, www.cpuc.ca.gov/published/rulings/ 54059.htm.)
For all this complexity, however, and despite the project’s huge cost (some $170 million sunk so far, by best estimate), the Cal-ISO’s actual proposal remains surprisingly incomplete. In fact, many technical details remain uncertain. One missing detail is the exact algorithm for calculating nodal prices. Another is how the ISO will incorporate gas price indices into the cost adders it uses to calculate the “default energy bids” that Cal-ISO will use to mitigate market power for transmission paths deemed noncompetitive on account of failing the ISO’s proposed “three pivotal supplier” test. Other examples abound as well.
Instead of pinning everything down, the ISO has left these unfinished details to be fleshed out later in practical operating guides to be known as “business practice manuals.” And while the ISO has advised FERC in its most recent MRTU status report that the BPM effort is now “kicked off and on schedule” (see, Status Report of Cal-ISO, FERC Docket ER02-1656-009, filed May 1, 2006), the overall process still appears problematic.
On one hand, the FERC itself has made it clear on more than one occasion that any protocols that go beyond the typical internal utility operating manual—instructions that contain the kind of detail that ordinarily would appear in a power purchase agreement—must be filed with the commission in much the same manner as a tariff. As Chairman Joseph T. Kelliher noted in March, in his separate statement announcing a key decision on markets for the Southwest Power Pool, “We are applying the hard-learned lessons of the California crisis, by assuring that clear and complete market rules are established before markets open. It is better to get it done right than to get it done fast.”
At the same time, however, the Cal-ISO has argued that strict adherence to schedules—not only for project implementation, but also for developing and testing all the necessary software—dictates that the proposal must be “frozen” at an early date so that software developers and IT vendors can begin and complete their work in parallel, even as the proposal awaits regulatory approval. Listen to how the ISO described the problem as recently as March 15:
“Although the MRTU market design incorporates many features from the markets of Eastern ISOs, the MRTU software is based on a wholly different architecture …
“Features employed in Eastern ISOs cannot simply be incorporated into the MRTU markets without substantial effort.” (Status Report on Convergence Bidding, FERC Docket ER02-1656-030, filed March 15, 2006.)
As a result, and in order not to saddle its software developers with moving targets, the ISO has chosen to implement the MRTU in stages, and to postpone until an unspecified day-two release any implementation of virtual trading (also known as “convergence” bidding), by which market participants may bid to buy or sell power without actually possessing any power-producing capacity or load-serving responsibilities. It has done this despite a clearly stated preference from FERC for virtual bidding in RTO markets, and a show-cause order directing the Cal-ISO to explain its recalcitrance.
In similar fashion, owing in part to software complexity, the ISO has decided to decouple the calculation of transmission-line losses from congestion costs (though they play a key role in LMP determination), necessitating certain arbitrary assumptions in the reimbursement of losses and the allocation of CRRs. This decoupling thus will create new and different sets of winners and losers, and in fact has led some critics to complain that the burden of paying for marginal transmission line losses imposed on electric consumers under Cal-ISO’s new market design actually will outweigh objections from having to pay locational prices.
The ISO’s catering to computers, in fact, has become a fault in its own right, irking some opponents. The California Municipal Utilities Association (CMUA), in comments filed April 10, takes notes of the “considerable time” that Cal-ISO has spent it explaining its “dependence” on software development and warns FERC not to defer to the needs of IT vendors.
“The commission,” urges CMUA, “should not allow itself to be held at gunpoint in this fashion.”
Vetting the Issues
In reality, a fair number of MRTU elements proposed by Cal-ISO already have passed muster, as the ISO since January 2002 has been conducting a series of stakeholder conferences and has vetted numerous ideas before federal regulators, winning commitments from FERC in many areas, including:
• Locational Marginal Pricing. Cal-ISO proposal approved Oct. 23, 2003 (105 FERC ¶61,140);
• Market Power Mitigation. PJM-style model approved, July 1, 2005, targeting potential for gaming, instead of the Cal-ISO’s current AMP procedure, that being a form of “conduct/impact” test, as is used in New York and New England (112 FERC ¶61,013); and
• Nodal Averaging for Load. Cal-ISO proposal approved July 1, 2005, to assign LMPs to load based on geographic averaging at three LAPs (“load aggregation points”) corresponding to traditional service territories of major investor-owned electric utilities (112 FERC ¶61,012). Later, on Nov. 14, 2005, FERC allowed the ISO to forgo creating specific LAPs for certain large wholesale power customers to enable special case nodal pricing for load (113 FERC ¶61,151).
This combined process of step-by-step vetting and FERC authorization has narrowed the range of legitimate disagreement, making it more difficult for opponents to use Cal-ISO’s February tariff filing as a forum for rehearing the electric restructuring agenda from top to bottom.
As writes Don Garber, senior regulatory counsel for San Diego Gas & Electric Co., “The Great War over locational marginal pricing has been brought to armistice.” As he explains, SDG&E is reasonably satisfied and will not force the ISO to prove that every new tariff element will go off without a hitch:
“SDG&E rejects this review standard,” writes Garber, “as too friendly to the devil we know and too hostile to the friend we can reasonably anticipate.”
In fact, an objective look at the record would indicate that all three of the state’s large investor-owned electric utilities now appear to stand behind most features of the ISO’s new market design, with some caveats. Even the state PUC now “generally supports” the new tariff, including LMP.
Nevertheless, as with any 5,000-page proposal, enough questions remain unresolved to fill a dozen or more magazine columns. Here is a short list of some of the more serious criticisms and disagreements concerning Cal-ISO’s proposed new market design:
1. Resource Adequacy. ISO insistence on providing reliability services as a backstop to state-mandated rules on resource adequacy requirements and for allocating such costs to ISO market participants, even for publicly owned utilities not under PUC jurisdiction, could create unconstitutional interference with state’s RAR standard.
2. Resource Exports. MRTU gives too much discretion to ISO to curtail exports of gen resources that qualify as satisfying resource adequacy requirements imposed by state PUC.
3. Transmission Line-Losses. Recognition of losses in LMP pricing algorithm on a marginal, rather than average-cost basis, will lead to over-collection of losses at the ISO ($200 million per year) and impose a crushing cost burden, especially on PG&E and other LSEs using PG&E’s LAP pricing node for load. PG&E historically has shown a lower customer density and higher reliance on distant generation, yielding higher loss factors, and it will receive inadequate reimbursement under the ISO’s scheme to repay excess loss collections through average load-share ratios.
4. Transmission Outages. Stretching the mandatory notice period for scheduled transmission outages, from 72 hours to 45 days, appears unworkable to many.
5. Long-Term FTRs. The ISO in some minds appears hostile to the issuance of long-term transmission congestion rights (LTTRs), as mandated by Congress in the Energy Policy Act of 2005. Need for LTTRs remains clear, they say, since the California PUC bars use of financial rights payable on liquidated damages clauses to satisfy state-enforced requirement for resource adequacy.
6. Hour-Ahead Scheduling Process. Rules for HASP, which bar self-scheduling of exports (they must schedule day ahead), but which allow self-scheduled imports (but cleared not against demand bids, instead against Cal-ISO load forecast), reveal alleged tendency of ISO to favor load and over-regulate supply. Critics say HASP rule could backfire and prevent ISO from ridding itself of excess generation—a real threat because market design makes it difficult to de-commit a resource. Timing of HASP closing also fails to mesh comfortably with gas markets, raising fuel costs and minimum-run and ready-to-serve costs for gen resources seeking to provide real-time ancillary services.
7. CRR Terms. Rights should be monthly, some say rather than seasonal, as ISO proposes, since demand and congestion data show significant monthly volatilities that seasonal rights will leave unhedged.
8. CRR Allocations. Initial eligibility rules will penalize competitive retailers (still 12 percent of big-three IOU load), who cannot easily substantiate future load requirements. Grandfathering rules will discriminate against LSEs that gain load from year to year. Non-member LSEs in California serving load external to ISO will suffer especially, as rule for proving “legitimate need” is too strict, and since ISO-member LSEs enjoy residual CRR allocation rights once needs are met for existing contracts (ETCs) and transmission owner rights (TORs). ISO explanation that non-members can use imports or counter-party deals to schedule alternative gen sources to avoid use of ISO grid and exposure to unhedged congestion costs proves false, since renewable portfolio standards impose a certain physical inflexibility.
9. Market Power Mitigation. ISO reliance on PJM-style mitigation, testing mere potential for exercise of market power through 3-pivotal-supplier test, rather than more concrete conduct/ impact test, could fail to distinguish price increases driven by scarcity rather than gaming, and thus fail the judicial standard set in D.C. Circuit’s 2005 ruling, Edison Mission Energy v. FERC, 394 F.3d 964.
10. Demand-Response Bids. Postponement of demand-response bidding by participating load until the second-phase implementation of MRTU will penalize competitive retail energy service providers and large customers, such as the state water project.
At this point, the lion’s share of opposition to Cal-ISO’s market-reform initiative tends to come not from state regulators or investor-owned utilities, but from the public power sector, including municipal utilities and irrigation districts, which remain loyal to a physical-rights model.
For example, the Control Area Coalition, including Bonneville Power, WAPA, and the Sacramento and Los Angeles municipal utilities, asks why California should now rush to embrace a new market design when the chief shortcoming of the current regime— the ISO’s inability to manage with intra-zonal congestion—has abated of late. In particular, the coalition cites a 52 percent drop in intra-zonal congestion costs, from $426 million in 2004 to $203 million in 2005, as reported by Cal-ISO. (See, 2005 Annual Market Performance Report, March 2, 2006, www.caiso.com/ 17b1/17b1887672ff0.pdf).
Virtual Bidding: A Moral Affront?
Many industry players fault Cal-ISO for proposing to delay implementation of convergence bidding in the day-ahead market, especially as Cal-ISO proposes also in the MRTU to eliminate the current requirement, introduced in November of last year, in Cal-ISO tariff Amendment 72, that otherwise has required that scheduling coordinators must submit day-ahead schedules that reflect at least 95 percent of their forecasted daily demand, and which provide data to Cal-ISO on a weekly basis regarding actual daily loads. These players fear that just as power producers can withhold supply to drive up the day-ahead wholesale energy price, so can utilities withhold demand to drive the energy price down, and thus take advantage of relationships between the day-ahead and real-time markets. The solution, they argue, is to allow market participants to bid day-ahead and real-time, even if they lack physical positions (those without generating capacity or load to serve), in order to ensure day-ahead and real-time prices “converge” (hence the term) and thus weed out any arbitrage opportunities. Coral Power puts the case succinctly:
“Allowing LSEs, alone of all market participants, to shift load from the day-ahead market to the real time market is an invitation to do what suppliers may not do: bid strategically and settle in real time.”
Coral explains further: “The under-scheduling of day-ahead schedules artificially suppresses the day-ahead energy price that the LSEs pay for a large volume of load, which more than counterbalances the effects of potentially driving up the real-time price, which is paid on a much smaller volume of load.”
So far, Cal-ISO has continued to insist that to incorporate virtual bidding would delay the MRTU by a full year: “There is no single conceptual design of convergence bidding that all the other ISOs have adopted and that the Cal-ISO could adopt … without any stakeholder engagement. …
“For example, the PJM virtual bidding feature is based on a nodal approach while … the New York ISO markets utilize a load zone [and] hub-based approach.”
Nevertheless, after perhaps feeling stung by all the criticism, Cal-ISO suggested in a status report update filed on March 15 that it might consider moving up implementation to a “Release 1A.” In other words, of all the possible day-two additions of market features, the ISO said it would put virtual bidding at the top of the queue.
History buffs should note that this issue first came up a dozen year ago, at a hearing held June 14, 1994, to consider the California PUC’s infamous “Blue Book” vision for electric competition.
On that occasion, Robert Levin (then and still a senior vice president at NYMEX, the New York Mercantile Exchange) had recommended a purely financial market for California, such as NYMEX had designed for natural gas. In the NYMEX gas market, said Levin, virtual bidding typically caused daily trading volumes to run about four times the volume of daily consumption.
Whereupon the PUC, led by its then president Dan Fessler, came to question the wisdom of such a practice, as if it were a moral affront:
“What is the social value that is being added by [this] excess number of transactions that are being engaged? What are they providing to the economy? They have provided to certain people now famous in the American body politic a means of instant fortune, but aside from that circumstance, one would have to wonder.” (See, “Real Time in California,” Public Utilities Fortnightly, Oct. 15, 1994, p. 8.)