Back in January 2004, when AEP subsidiary Appalachian Power (AP) asked Virginia utility regulators for permission to transfer grid functions to the PJM regional transmission organization (RTO), AP claimed that RTO membership would not force it to carry a larger reserve of generating capacity than already was required under rules imposed by ECAR, the East Central Area Reliability Council.
In fact, AP had promised Virginia that PJM membership might well produce a modest amount of net benefits.
For that assumption, the utility had relied on testimony from PJM itself that integration of AEP into the RTO likely would reduce AP’s reserve requirement, owing to overall expansion of the PJM footprint (with Commonwealth Edison also joining). Also, it would improve regional load diversity and spread risk over a larger collection of generating units.
Of course, ECAR’s required 12 percent reserve fell short of PJM’s 15 percent parallel standard for installed reserve margin (IRM). However, in response to interrogatories, PJM’s manager of capacity adequacy planning, Thomas Falin, had provided data showing that the time of AEP’s peak demand diverged somewhat from PJM’s system peak, creating a 2.5 percent diversity factor between the two peaks. This adjustment for diversity ([1.0 – 0.025] x 115–100) would allow AEP (and AP) to hold a smaller margin of 12.125 percent.
The story was much the same for Dominion Virginia Power (DVP), where witnesses testified that diversity between system demand peaks would allow DVP to meet reliability rules with a 12.5 percent reserve capacity, again despite PJM’s nominal 15 percent benchmark. (See Testimony and Exhibits, Va.S.C.C. Case Nos. PUE-2000-00550 [AP] and PUE-2000-00551 [DVP]; and Comments of Div. of Consumer Counsel of Atty. Gen. of Va., FERC Dkt. No. ER05-1410, filed Oct. 19, 2005.)
Nevertheless, all this has now changed.
On Aug. 31, PJM pulled the plug on AEP, Dominion, and the Virginia state regulators. After a long and widely watched gestation period, plus a tumultuous stakeholder process that failed to produce anything resembling a vote of confidence, PJM applied to the Federal Energy Regulatory Commission (FERC) for authority to impose a new regime of requirements for reserves of electric generating capacity. This new construct, known as the reliability pricing model (RPM), would replace PJM’s current capacity market, known as UCAP (the same basic structure as an ICAP market for installed capacity, but with an adjustment for unforced outages, which can affect availability). PJM’s current UCAP market features what many call a “vertical” demand curve, set by administrative fiat, which stands as a proxy for market demand for reserve generating capacity. It does not involve auctions or bidding, but by matching the total available quantity of capacity with the artificial demand curve, it sets a deficiency charge to be assessed to utilities and load-serving entities (LSEs) that might lack the required reserves.
The RPM plan has rocked the electric utility industry. It has spurred new interest in reliability and resource adequacy, especially since FERC switched gears this fall regarding New England’s proposed locational capacity market, known as LICAP.
Moreover, by opening settlement talks in the LICAP case and inviting New England to consider all reasonable alternatives, FERC gave new hope to all sides in the PJM case as well. The result has thrown the ICAP/UCAP/LICAP debate wide open, encouraging theoretical discussions of all sorts.
For example, should a capacity construct include a market auction, or just a mandate for forward contracting? Should it include a locational aspect, if spot energy markets already provide for locational marginal pricing (LMP)? And what about energy markets: If price caps are removed, can the energy price alone, without any capacity market, create enough incentives to encourage adequate investment in generation?
Playing a key role in the debate is Harvard Professor William Hogan, often cited as the spiritual father of PJM’s energy spot market. Hogan had attracted considerable attention in September when he released an article suggesting an energy-only model, and attached it to comments filed by the California ISO in answer to an unrelated proposal to adopt a LICAP structure in California. (See “LICAP: A Mad Dash to the Finish,” November 2005)
To justify its RPM plan, PJM cited what it called “a dramatic spike” in power plant retirements—especially in New Jersey and the Mid-Atlantic, with many closures announced on no more than the required 90 days’ notice. These retirements had uncovered violations of reliability criteria for 2005 and each subsequent year through 2009, said PJM’s Steven Herling, vice president, planning, in a sworn affidavit. He added that trends in other areas, such as Baltimore-Washington and the Delmarva Peninsula, showed similar vulnerabilities for 2008.
The data also demonstrated, however, that PJM enjoyed a robust system-wide reserve margin across its entire footprint—as high as 33.5 percent in 2004, and 24.7 percent in 2005, with some forecasts showing a 20.4 percent margin in 2006-2007, declining to 16.8 percent in 2010-2011.
Moreover, according to Assistant Counsel Scott Perry, writing for Pennsylvania’s Department of Environmental Protection, information available on the Internet site of the U.S. Energy Information Agency this fall showed that in 2004 approximately 34 percent of the electricity generated in Pennsylvania was exported out of state.
So why the need for cpacity incentives? According to PJM’s Andrew Ott, vice president, market services, the region also had experienced a significant decline in recent years in power plants with flexible load-following characteristics and quick-start capability. Over the prior four years, load-following gen capability offered in PJM had declined, he said, by nearly one quarter. Also, its gen-fleet suppliers now offered fewer daily plant startups: from an average of 4.6 plant starts per day per seller in June 2000, to only 3.1 starts per day per seller in August 2004.
What was needed, claimed PJM, was a program for marshalling gen-plant capacity in specific locations with improved short-term operational characteristics, through the incentive of a substantial and dependable stream of capacity market revenue payments. Hence the RPM proposal, slated to take effect June 1, 2006, if FERC grants timely approval. (See FERC Dkt. No. ER05-1410, filed Aug. 31, 2005.)
Despite these apparent needs, PJM failed repeatedly to win stakeholder approval for its RPM construct, and was forced eventually to file its application unilaterally with FERC. Many, however, find no surprise in that.
As William McCoy, managing directory for Morgan Stanley Capital Group, wrote in comments filed at FERC in mid-October, RPM is an “entirely artificial, non-economic, centrally planned structure.”
Yet at the same time PJM obscures this point, he adds, by use of market nomenclature.
McCoy repeated Hogan’s warning that it could prove dangerous to define “market failure” as being “when the market doesn’t do what a central planner would.”
Virginia’s State Corporation Commission, echoing a criticism noted by many, said it was “mystified” as to why PJM would seek to solve delivery issues in a few localized areas by instituting a solution across the entire RTO footprint.
Writing for the American Public Power Association, Susan Kelly, vice president, policy analysis, and general counsel, saw it as “very questionable” to burden consumers with high-priced incentives to call forth the needed infrastructure.
“Pay and pray,” she called it.
“While the RPM filing,” Kelly added, “stresses the importance of price stability, what likely is more important to investors is rule stability that provides consistent outcomes over time.”
PJM’s proposed RPM plan borrows heavily from prior experiments by grid system operators in the Northeast, but at the same time breaks new ground in many key areas.
On one hand, it borrows from the New York ISO (the innovation of a sloping ICAP demand curve to dampen capacity price volatility in capacity prices). PJM labels its demand curve as the VRR—“Variable Resource Requirement”—reflecting the fact that demand for capacity could be supplied by conservation or new transmission facilities, as well as generating capacity. RPM constructs the VRR curve to achieve a target equilibrium where the capacity resources clearing RPM auction equal the standard 15-percent installed reserve margin (IRM) plus 1 percent (a 16-percent IRM).
RPM also borrows the idea of capacity pricing zones from New England’s LICAP construct. These zones (known in PJM as LDAs—“Local Deliverability Areas”—would reflect the reduced delivery capabilities and higher capacity prices anticipated in areas with import/ export constraints. RPM also copies New England’s concept of CTRs—“Capacity Transfer Rights”—LICAP’s zonal analog to financial transmission rights (FTRs) in regions with spot energy markets featuring LMP.
Beyond that, however, RPM also is notable for several reasons:
Now consider what these differences could mean in practice.
Under traditional ICAP and LICAP constructs, suppliers don’t submit actual bids. The “price,” as it were, is a benchmark, not an equilibrium point between willing buyers and sellers. It denotes a deficiency charge that LSEs must pay to the ISO (for reallocation to suppliers) if they fall short of the capacity reserve mandated by passive reliability standards, or by planning mandates imposed by state regulators.
In practice, the benchmark creates a default price floor for bilateral capacity trading, which goes on daily, weekly, and monthly, since capacity-short LSEs presumably would be willing always to buy capacity privately at the benchmark level or higher, at least to avoid assessment of the deficiency charge.
By contrast, under PJM’s proposed RPM plan, the RTO can force LSEs to buy more capacity than would be warranted by the nominal reliability standard. That occurs because the RPM plan sets the reserve requirement by inviting suppliers and LSEs to bid, and then committing them to the price and quantity that clears the auction. In short, it is bidding behavior—not a passive set of reliability standards—that dictates the required amount of installed reserve margin.
In a position paper submitted in PJM stakeholder discussions back in February, AEP said PJM’s RPM model would penalize vertically integrated utilities that own generation and provide their own supplies and reserves. According to AEP, LSEs might have to purchase a different mix of capacity in the auction process than has been approved by state regulators in the state’s process for integrated resource planning:
“If the supply curve from the auction intersects the demand curve at 20 percent, then the LSE will be required to purchase enough capacity to meet a 20 percent reserve margin, even though the IRM could be set at 15 percent.
“If AEP, through the IRP process, has long-term capacity acquisition plans designed to meet 13 percent, or even 16 percent … it must purchase at least 4 percent more capacity (if the auction clears at 20 percent) than has been found to be reasonable in ECAR.
“For a vertically integrated, self-supplying utility such as AEP, which has constructed its own generation to meet IRP plans, it means that the company must actually pay out additional funds to meet the PJM auction.”
(Of course, as a supplier bidding into the RPM capacity market, AEP eventually would be entitled to an allocation of revenues covering those same capacity costs, by the time of the delivery year at the end of the four-year resource commitment period. But revenues might not quite cover costs, depending on results in intervening auctions during the four years.)
Consider this small sample of some of the most difficult issues posed by PJM’s RPM construct.
• Too Few Zones. PJM’s phase-in plan calls for two large LDAs during the first year (6/1/06 to 5/31/07), then four LDAs in year two. Critics say that will socialize costs by allocating “locational price adders” (LPAs—the constraint premium) over too broad an area in early years.
• Too Many Zones. PJM implements the full complement of 23 LDAs in years three and four (June 2008 through May 2010), though it admits that not all zones will exhibit constraints, and thus not all LDAs will carry LPA pricing premiums, or exhibit capacity pricing differentials between zones. Critics say this unnecessary granularity will balkanize markets and reduce the number of sellers in any one zone.
• Resource Delivery Period. RPM plans a four-year span between the initial “base residual auction” and the resource delivery period (sprinkled with intervening catch-up or true-up auctions), so that developers have four years from the offer/solicitation to construct and deliver the resource. Yet many want a 10-year period, saying that large-scale grid projects can take that long to complete.
• Market Power Mitigation. The RPM plan will mitigate all bid offers and clearing prices in any LDA if the number of pivotal suppliers fails to exceed three (the so-called N3PS test).
• VRR Capacity Cost. PJM’s design of the sloping VRR demand curve begins with the basic gross capacity cost (“Cost of New Entry,” or CONE) of a proxy capacity resource (a 2x7FA GT gas-fired peaker). CONE is then adjusted, as in a typical ICAP/UCAP/LICAP plan, to reflect operations and maintenance costs, offsets for energy and ancillary service sales revenues, plus special locational factors (differences in labor, land, taxes, etc.) Critics say RPM understates CONE at $59/kW-yr. (based on an installed capacity cost of $466/kW). By contrast, ISO-NE’s proposed capacity cost for LICAP runs about $87/kW-yr. A competing consulting study offered by PEPCO calculates CONE at $78.76/kW-yr (reflects installed capacity cost of $610/kW).
• VRR Revenue Offset. Opponents question whether PJM should rely on six years of past historical data of natural-gas prices to calculate gen-plant performance and likely energy sales revenues as an offset to the capacity cost. Reliant Energy’s Neal Fitch urges use of forward cost data, such as NYMEX gas futures. “Choosing historical revenues,” he notes, “is tantamount to an individual investor saying, ‘I will invest in Delta Airlines based on its performance over the past six years.’”
• Opt-Out Option. PJM proposes to allow self-supplying LSEs to opt out of RPM auctions, but to do so they must procure a reserve that includes an added 3-percent “uncertainty factor” above and beyond the RPM’s nominal auction target of IRM=15 percent plus 1 percent. West Virginia PSC says that penalizes and discriminates against large, vertically integrated electric utilities that operate in areas without retail choice, and which supply their own capacity.
Perhaps the most novel aspect of PJM’s RPM construct concerns the proposed right of merchant transmission developers to bid grid upgrade projects in the capacity auction as the equivalent to a power plant—the typical capacity resource—if the upgrade will increase transfer limits into an LDA. However, this feature brings with it a whole new set of unique problems and issues. (Note: Only merchant grid projects can bid into the RPM auction, because the RPM model limits bidding rights to grid projects to be funded through a rate “specific to the facility.”)
Bowling Pins. Commenting for National Grid USA, counsel Joel deJesus notes that grid projects bidding into the RPM auction are asked to state an offer price in terms of the price differential between capacity resources located outside and inside the LDA, but that this rule appears “unworkable.”
As deJesus observes, any price differential between resources in different LDAs will be reduced significantly by the construction of the transmission project.
“Because of the rent-destroying nature of such lumpy investment,” notes deJesus, “basing payments to transmission owners on residual congestion (i.e., congestion that the resource in question failed to eliminate) provides a poor investment signal.
“It would be akin to scoring a bowler for the pins he left up; it would incentivize developers of transmission to leave uneconomic congestion on the system, rather than remove it.”
Moving Targets. What happens, asks deJesus, if a grid project clears the base residual auction at the start of the four-year commitment period, but then fails to clear when submitted into subsequent intervening auctions (in which the developer is also required to submit bids).
“In such event,” deJesus notes wryly, “the transmission owner would need to disconnect its ‘out-of-merit’ transmission facilities to prevent access by transmission customers.”
DeJesus speculates that a would-be grid developer could submit a zero-cost supply bid in all those subsequent intervening auctions, but believes such a bidding strategy would introduce “significant revenue uncertainty for sponsors of merchant grid projects.
In a very similar observation, lawyers from Spiegel & McDiarmid, representing Blue Ridge Power Agency and municipal utility associations in Virginia and Illinois, note that PJM fails to explain how it will calculate LPAs (price premiums for constrained capacity zones).
“Will PJM apply existing transfer capabilities [those prevailing before the auction], or those anticipated for the delivery year if all [grid upgrade] projects are timely completed?”
In other words, do the ground rules change during the delivery-commitment period, governing calculation of constraints, delivery capability, and prices to reflect anticipated and eventual completion of promised grid resources?
In Spiegel’s opinion, PJM’s filing “is far too murky on this crucial point.”