Docket No. ER23-1067-002
I dissent because the Commission’s order on rehearing, like the underlying order, is both procedurally and substantively unfair to resources that combine energy storage with another technology such as solar.[1]
First, an important part of PJM’s proposal was a transition mechanism for existing interconnection customers, for which PJM imposed an eligibility deadline without proper notice. The Commission’s last-second attempt to remedy this unjust and unreasonable deadline did not satisfy the notice requirements of the Federal Power Act. In direct contrast to the interconnection reforms adopted by the Commission in Order No. 2023, PJM’s proposal approved by the Commission encouraged a last-minute, speculative rush into the interconnection queue by market participants.
Second, for the period between April 10, 2023, and December 11, 2023, the order approves an unjust and unreasonable modeling approach by PJM that undercounted the contributions of Combination Resources[2] and thereby required consumers to buy more capacity than necessary. PJM has since updated its modeling approach, demonstrating that the drawbacks of the proposal approved herein were eminently fixable. Combination resources are harmed by today’s order because it affirms the imposition of the unjust and unreasonable method on them prior to PJM’s update, and consumers are harmed because they paid more than they should have for capacity while PJM’s now-superseded method was in effect.
The deadlines enforced by PJM and the Commission were unjust and unreasonable, and violated the Federal Power Act’s notice requirement
The Federal Power Act requires FERC-jurisdictional utilities to provide “sixty days’ notice to the Commission and to the public” of tariff changes.[3] Changes must be “clearly and specifically” set forth in “full and complete” rate schedules and tariffs.[4] PJM’s proposal here included a transition mechanism that had two important components: (1) substantive requirements detailing how a resource may seek to have additional Capacity Interconnection Rights (CIRs) assigned; and (2) a deadline by which a resource must apply to have those rights increased.[5] The Commission’s Order today re-affirms approval of PJM’s tariff changes even though PJM’s filing plainly did not provide 60 days’ notice of the transition mechanism deadline.
When PJM first filed its tariff proposal in February of 2023, it proposed a March 3, 2023 deadline by which existing Effective Load Carrying Capability (ELCC) resources were required to submit an interconnection request to be eligible for the transition mechanism for any Delivery Year.[6] This deadline came just 23 days after PJM’s filing. Generators, attempting to comply in good faith with PJM’s stated deadline, rushed to submit interconnection requests before March 3, even as PJM continued to issue relevant guidance throughout the month of February.[7] After the deadline passed, PJM stated that it had received more than 400 applications for its transitional study process.[8]
PJM enforced the March 3 deadline by rejecting additional applicants. National Grid Renewables states that it submitted its transitional study request with associated documentation on March 6, one business day after the initial March 3 deadline, and was told by PJM that it would not be eligible.[9]
Over a month later, in an April 7, 2023 decision, the Commission correctly concluded that enforcement of the March 3 deadline did not comport with the FPA’s notice requirements.[10] But rather than requiring proper notice from PJM, the Commission unilaterally modified the deadline. The Commission’s order concluded that the deadline was “not effective or enforceable before April 10, 2023,”[11] but held that this date, April 10, would be the submission deadline even though it fell on the very next business day after the Commission’s order and PJM had never suggested that this would be the deadline.[12] The Commission’s April 7 order was the first time that market participants could have anticipated any deadline other than March 3, the deadline PJM had both publicized and enforced. This deprived existing ELCC resources a meaningful opportunity to make the complex decision of whether to submit an interconnection request.
In brief, market participants rushed to comply with PJM’s illegal deadline. If they filed after that deadline, PJM informed them that they were too late. Then, more than a month after the deadline had passed, this Commission suddenly granted those same entities one additional business day to revisit their decision. Because 61 days passed between the beginning and end of that haphazard process, the Commission today concludes that PJM satisfied its notice obligations under the FPA.[13] It thereby denies market participants faced with the complex task of deciding how many additional CIRs to request[14] the benefit of what the law guarantees absent a waiver of the notice requirements: 60 days’ notice to make a considered choice.[15]
The majority’s conclusion that affected generators were on notice of the April 10 deadline attributes the benefits of hindsight to real-time market participants. In so doing, it warps the Federal Power Act’s notice requirement beyond recognition. Today’s order states that “the parties were on notice that PJM’s proposal could have gone into effect by operation of law on April 10, 2023.”[16] The order essentially reasons that applicants should have disregarded the statement in PJM’s filing that an Interconnection Customer would be eligible to increase its CIRs if it submitted a New Service Request into the queue “prior to March 3, 2023,” its exhortation that “[t]he date of March 3, 2023 is significant because it is latest day [sic] by which PJM staff estimate that they will be able to begin transitional system capability studies in anticipation of the June 2023 BRA,”[17] as well as its subsequent enforcement of that deadline. Market participants are to have understood that each of these statements and actions was meaningless and the real deadline was April 10, 2023, all based on an inference gathered from PJM’s failure to request waiver of the notice requirement and despite the fact that the April 10, 2023 deadline appeared nowhere in PJM’s filing.
The majority “disagree[s] with the suggestion” that anticipating the April 10 deadline required “‘inference,’ ‘prediction’ or ‘intuition’” by the parties.[18] It insists the parties had notice that the real deadline was April 10, all based on the Federal Power Act’s 60-day notice requirement alone, despite PJM’s insistence on a March 3 deadline. But even if the parties knew that PJM’s proposal violated the notice requirement, they could not know prior to the Commission’s order that the response would be to extend PJM’s deadline to April 10 rather than to reject the filing entirely or restart the 60-day clock. Certainly the text of the Federal Power Act alone provides little direction as to the necessary response when the 60-day notice requirement is violated. And the Commission has previously acted in contrary fashion. For example, in Cincinnati Gas and Electric Company, a utility submitted a rate filing on October 4 but then responded to a deficiency letter on December 13.[19] The utility asked for a January 2 effective date, arguing that “affected parties have had more than 60 days’ notice of [its] rate intentions.”[20] But the Commission denied that request, reasoning that “the statutory notice period does not commence until a completed filing has been tendered to the Commission,” and calculating December 13 as the date of that completed filing.[21] Here, PJM never submitted a complete and proper filing to the Commission that included the key term of an allowable interconnection deadline. Even if the parties guessed that the Commission itself would complete the filing by substituting its own date (a dubious proposition without the benefit of hindsight),[22] the parties may reasonably have guessed that this action would reset the 60-day clock, so a date further in the future would be compelled.
Moreover, the Commission’s regulations require utilities to “clearly and specifically” set forth proposed tariff changes 60 days in advance of their operation.[23] The deadline to apply for the transition process was itself a significant aspect of PJM’s proposal, given that it controlled who was eligible and who was not. To hold that PJM “clearly and specifically” set forth an April 10 deadline in its filing is farcical.
The Commission’s regulations are meant to allow market participants to plan for new rate regimes and take action accordingly.[24] Yet PJM’s proposal here made such planning virtually impossible. At great commercial risk, a capacity supplier would have needed to (1) predict that the Commission would reject the initial March 3 deadline, (2) ignore PJM’s subsequent enforcement of that deadline, and (3) intuit (contrary to the statutory requirement of clarity and specificity) that the Commission would unilaterally impose an April 10 due date. The FPA does not demand this much of market participants. Beyond violating the FPA’s notice requirement,[25] it is also neither just nor reasonable to force generators to make the complex business decision of whether to submit a new interconnection request in the span of a single business day.[26]
As I emphasized in my previous statement, the results of repeatedly rushing ELCC resources into the interconnection queue (first at PJM’s March 3 deadline and again before April 10) are costly.[27] Given the halting and confusing process described above, and the difficulty of this decision for generators, affected ELCC resources lacked time to make informed choices. Beyond violating the FPA’s notice requirements, this process also amounted to bad policy. Because I wholeheartedly support the Commission’s efforts to reduce speculative, ill-considered interconnection requests,[28] I object here to FERC’s engineering of a process that is sure to have resulted in many such applications.
PJM’s approach undercounts the capacity value of Combination Resources
I also dissent because the accreditation methodology it approves relies on an irrational modeling assumption that demonstrably undercounts the capacity value of Combination Resources.[29] PJM did not justify its approach as necessary to protect grid reliability, nor as required due to technical challenges. In fact, relatively soon after proposing these changes, PJM fixed this problem by proposing a new methodology.[30] This demonstrated that it was eminently feasible to develop a modeling approach that did not undercount the contributions of Combination Resources.
Understanding how the flaw in the approach approved herein rendered rates unjust and unreasonable requires some background on the PJM market and PJM’s proposal. Each Combination Resource maintains a given amount of CIRs, which represent capacity requested by the Combination Resource and demonstrated to be deliverable through PJM’s interconnection process. [31] A Combination Resource’s CIRs can be no more than the sum of the two components’ nameplate capacities, though most Combination Resources have significantly fewer CIRs. The operator of a Combination Resource generally derives economic value from maximizing the use of the variable resource component. The storage component will generally charge when the variable resource produces excess power and discharge in high-value hours when the variable component is not using the resource’s allotted CIRs.[32]
PJM proposed here to modify the ELCC calculations it used to assess contributions to the capacity market. ELCC reflects the quantity of energy a resource can reliably produce at times when PJM is expected to experience a risk of loss of load, and is used to calculate Accredited Unforced Capacity (UCAP).[33] UCAP, in turn, sets the maximum amount of capacity a resource may offer into the PJM capacity market.[34] For all ELCC resources, PJM’s caps the output of the resource in any modeled hour in the ELCC calculation at the amount of Deliverable MW the resource holds.[35] This ensures that the assumed deliverability of ELCC resources is backed up by credible interconnection analysis.[36] For Combination Resources, PJM models the two components separately and sums the two calculated ELCC values to determine the ELCC value for the Combination Resource.[37] But for the proposal approved herein (i.e. for the period before PJM fixed its modeling approach), PJM capped the modeled output of the variable resource at the Combination Resource’s Deliverable MW (generally, CIRs) minus the effective nameplate capacity of the limited duration component.[38]
This aspect of PJM’s proposal systematically undervalued Combination Resources by implicitly assuming illogical curtailment of the variable resource component. For example, imagine a facility with 125 MW of solar capacity, 50 MW of battery storage capacity, and 80 MW of CIRs. If the solar component were interconnected alone, its ELCC accreditation would flow from a calculation that assumes a maximum injection to the grid capped at the amount of CIRs for the resource. In a sunny hour where the resource is modeled to produce 120 MW, for example, PJM’s ELCC calculation would appropriately credit it for its full 80 MW of CIRs. But where the solar component interconnects together with a storage component, the rules approved herein limited the solar array’s modeled output in all hours of the capacity accreditation calculation to 30 MW (the result of subtracting the 50 MW of storage from the Combination Resource’s 80 MW of CIRs). This means that, for that same sunny hour, PJM assumed that the Combination Resource owner would needlessly curtail the solar component in favor of discharging from the storage. This runs contrary to the anticipated behavior of the resource owner, whose economic incentive, with 120 MW of solar production, would be to allot its 80 MW of CIRs to the solar component.
PJM’s unreasonable cap on the ELCC contribution of the variable component of a Combination Resource flowed in part from its decision to model the two parts of a Combination Resource separately. When modeling storage resources, PJM assumed that they maximize production during the most critical hours.[39] This is a rational assumption for a standalone storage resource because producing during the most critical hours will usually maximize a resource’s value. This approach does not make sense, however, for a Combination Resource limited to its allotted CIRs, where a resource would prioritize injections from the variable component maximize its value.[40] But because in the proposal approved herein PJM chose to model storage and variable components separately and did not track the combined components, its model did not delay the storage resource’s injection to reflect this behavior.[41] The parties requesting rehearing emphasize that this was a deliberate choice by PJM, not a modeling necessity.[42] Rather than using a storage model blind to hours of production from variable resources, they argue that PJM should have tracked the modeled production of the variable component of a Combination Resource in the storage model.[43] Structured this way, the model could shift production from the storage component of a Combination Resource to subsequent hours.
PJM’s failure to model a Combination Resource’s generation shifting behavior artificially lowered Combination Resources’ ELCC ratings. A resource’s ELCC depends primarily on its modeled performance through critical periods of system stress.[44] When such a period extends longer than the duration of the storage resource, PJM’s model will assume that the operator discharges the storage component in the early hours, quickly running the battery out of charge while curtailing the variable component. If the variable component is no longer available in the later hours of the critical period (e.g., if the sun set on a solar array), PJM’s model would not credit the Combination Resource for any production, even though the storage component would be discharging in the real world. Thus, PJM’s modeling approach would directly and arbitrarily reduce the resource’s ELCC.[45]
PJM’s curtailment assumption applied across all hours in the ELCC model, with a bizarre implication which the Commission acknowledged: “there may be certain hours where PJM’s ELCC simulation assumes that the Limited Duration Resource component will be charging or idle, yet the simulation nevertheless reserves Deliverable MW for that component.”[46] In other words, even in periods where PJM’s model assumes that the storage component is charging or idle, it will still assume that the variable component is curtailed. PJM provides no reasonable justification for this significant shortfall in its model, which diminished Combination Resources’ capacity ratings even further.
PJM provides no coherent response as to why it could not have modified its model to more accurately value ELCC resources. The Order finds PJM’s approach reasonable “[b]ecause PJM does not know the precise hours in which the components of a Combination Resource will produce energy at the time PJM conducts the ELCC analysis.”[47] Similarly, the Commission previously concluded that “for the hours most determinative of a Combination Resource’s ELCC accreditation, the Limited Duration Resource component could be discharging at its full Effective Nameplate Capacity, and therefore this amount of Deliverable MW should be reserved.”[48] But these justifications are circular, as they take as a given that the ELCC model for the variable component of the resource is blind to the storage component’s modeled production and vice versa. It is, of course, true that PJM cannot predict the future. But no one has asked PJM to do that—The parties requesting rehearing ask only that PJM make reasonable, rather than unreasonable, assumptions within its ELCC accreditation model. PJM already has year-round production estimates for solar and wind resources which it uses to calculate the ELCC value of those resources. At the time PJM initial made its proposal, it was difficult to see why PJM could not simply import those predictions into its storage model,[49] and neither PJM nor the Commission advanced any reason this could not be accomplished.
The lack of explanation now makes perfect sense. There was no reason PJM couldn’t address the problem and more accurately accredit Combination Resources. Its subsequent fix demonstrates the discretionary nature of its initial decision. In today’s order, the majority emphasizes that there can be more than one just and reasonable approach, so the Commission’s subsequent approval of an approach that more accurately models Combination Resources does not by itself render the prior approach unjust and unreasonable.[50] The problem with this logic, however, is that the only justification given for the inaccuracy of method approved herein is that it was reasonable to take a “conservative approach” given PJM’s uncertainty as to the operation of each component of a Combination Resource.[51] But the subsequent filing disproved the notion that such “conservatism” was reasonable, as the model could simply track the components together and assume more reasonable behavior. Section 205 does not require a utility to propose the best conceivable rate, but it does place the burden on them to demonstrate the rate to be just and reasonable. Where the only purported justification for a rate proposal is shown to be illogical and unnecessary, the Commission cannot conclude that the utility has met its burden.[52]
Far from PJM meeting its burden, protesters here demonstrated that PJM’s approach was unjust and unreasonable. Because of PJM’s arbitrary cap on the variable component of Combination Resources, they received ELCC allocations significantly below their deserved amounts. And because ELCC is used to measure UCAP for the capacity market, the under-measurement of ELCC for Combination Resources forced customers to buy more capacity than they needed. Thus, PJM’s proposal, while in effect, exacerbated the three “significant harms” that the Commission has recognized in capacity markets: “over-procurement of capacity, inflated capacity market prices, and inefficient price signals from the capacity market.”[53] I cannot approve PJM’s inadequately justified proposal that resulted in this outcome.
For these reasons, I respectfully dissent.
[1] PJM Interconnection, L.L.C., 183 FERC ¶ 61,009 (2023) (2023 ELCC Order) (Clements, Comm’r, dissenting).
[2] Under PJM’s rules, a Combination Resource is a resource that includes both an energy storage component, and a variable resource component, such as a solar array. See PJM Interconnection, L.L.C., XXX FERC ¶ XX,XXX, at P 4 (2024) (Rehearing Order).
[3] 16 U.S.C. § 824d(d).
[4] 18 C.F.R. § 35.1(a).
[5] See PJM Transmittal at 18-20.
[6] Id. at 18-19.
[7] Request for Rehearing at 4.
[8] PJM Answer at 8 (“As of the date of this pleading, PJM has received over 400 requests for transitional system capability studies, clearly indicating that PJM Members are well apprised of the significance of this deadline and are acting accordingly.”).
[9] NG Renewables Protest at 3.
[10] 2023 ELCC Order at PP 38-39. As the 2023 ELCC Order and today’s Order explain, PJM was required to seek a waiver of the 60-day notice requirement, 18 C.F.R. § 35.11, if it wished to enforce its March 3 deadline. It did not.
[11] Id. (“[W]e note that the entirety of the proposed tariff revisions, including the March 3, 2023 deadline, are not effective or enforceable before April 10, 2023.”).
[12] Id. at P 38 (“[M]arket participants have until April 10, 2023, to submit a request to increase their CIRs”).
[13] Rehearing Order at P 23.
[14] See Request for Rehearing at 5 (“[A]n interconnection customer[] . . . must consider how many more CIRs, if any, to request, and weigh the potential increase in capacity revenues with potential network upgrades costs. To make these decisions, an interconnection customer must update its financial models with new inputs and assumptions.”).
[15] Id. at 2 (“The transition date set by the Commission provided customers no meaningful notice of a meaningful change in rates, terms, and conditions of service, and allowed PJM to improperly set an effective deadline for compliance without providing the statutory sixty days’ notice.”).
[16] Rehearing Order at P 24.
[17] Transmittal at 18 (emphasis added) & n. 39.
[18] Rehearing Order at P 24 n. 42.
[19] Cincinnati Gas and Elec. Co., 58 FERC 61,131, 61,419 (1992).
[20] Id.
[21] Id. at 61,420.
[22] Indeed, parties may have anticipated that the Commission would be reluctant to modify the proposed rate in light of the D.C. Circuit’s ruling striking down Commission-imposed modifications in NRG v. FERC. See 862 F.3d 108, 116 (D.C. Cir. 2017) (holding that given FERC’s modifications to the rate, “the utility’s customers [did] not have adequate notice of the proposed rate changes”). While that case turned on the Commission having imposed an “entirely new rate scheme,” see id., the parties might have viewed changes to the interconnection deadline through this lens given how critical the deadline was to PJM’s proposal.
[23] 18 C.F.R. § 35.1(a).
[24] See Cargill Power Markets, LLC v. Public Service Company of New Mexico, 132 FERC ¶ 61,079, at P 23 (2010) (explaining that the Federal Power Act “requires all practices that significantly affect rates, terms and conditions of service to be on file with the Commission” so that customers have “proper notice” and “obtain service on a just and reasonable and not-unduly discriminatory basis”).
[25] In support of its decision to unilaterally and after-the-fact extend PJM’s deadline for interconnection queue applications by 38 days, the majority points to prior orders where the Commission corrected ministerial errors in utility filings. See Rehearing Order at P 23 n. 40. For instance, in California Independent System Operator, the Commission provided for a tariff filed by CAISO to go into effect on October 11, rather than October 10, because the utility has miscalculated the 60-day deadline. See Cal. Indep. Sys. Operator, 117 FERC ¶ 61,181, at P 3 (2004). This is an entirely different kettle of fish. Unlike in that case, the date adjusted here is a central feature of PJM’s proposal, disputed by the parties and enforced by PJM. And the Commission makes a very substantive, not ministerial alteration of that date which effectively deprives the parties of adequate opportunity to respond to the proposal.
[26] See Tri-State Generation & Transmission Ass’n, Inc., 173 FERC ¶ 61,015, at P 71-72 (2020) (rejecting a transitional study readiness deadline 10 days after the effective date of approved tariff revisions because “[a]n interconnection customer needs adequate time to meet these significant requirements,” and the deadline “fails to adequately consider and balance the needs of existing interconnection customers.”)
[27] See 2023 ELCC Order (Clements, Comm’r, dissenting at P 12).
[28] See, e.g., Improvements to Generator Interconnection Procs. & Agreements, 184 FERC ¶ 61,054 (2023).
[29] A Combination Resource is one with a limited duration resource component (e.g., energy storage) and a variable (e.g., solar or wind) or unlimited resource component. Rehearing Order at P 11.
[30] See Rehearing Order at P 37 (citing PJM Interconnection, L.L.C., 186 FERC ¶ 61,080, at P 35 (2024)); see also Docket No. ER24-99, PJM Tariff Filing (Rocha-Garrido Aff. at P 27) (“The performance of . . . Combination Resources . . . in the ELCC/RRS model will be based on an hourly-simulated dispatch that depends on other system conditions.” That dispatch approach assumes injection by less available resources, like storage, only after more available resources have injected.). Contrary to the majority’s claim that this statement is “unsupported,” see Rehearing Order at P 37 n. 77, it is clear from PJM’s filing that they did, in fact, fix the irrational modeling assumption they had previously applied. It is irrelevant that PJM also made other changes to its tariff as part of that filing.
[31] 2023 ELCC Order at P 9. PJM defines CIRs as “the rights to input generation as a [capacity resource] into the Transmission System at the Point of Interconnection.” PJM, Intra-PJM Tariffs, OATT, Definitions – C – D (38.0.0).
[32] See Request for Rehearing at 9.
[33] Rehearing Order at P 3.
[34] Id.
[35] Id. at P 10. “Deliverable MW” refers to CIRs in summer months and “winter deliverability MW” in winter months.
[36] 2023 ELCC Order (Clements, Comm’r, dissenting at P 5).
[37] 2023 ELCC Order at P 17.
[38] Rehearing Order at P 11.
[39] See id. at P 38.
[40] See Request for Rehearing at 9 (“[T]he operator of a Combination Resource would have every incentive to maximize injection from its renewable component (rather than the storage component) when available.”). As discussed in my prior statement, it is conceivable that if a resource’s deliverability at times the grid is constrained is in truth greater than its allotted CIRs, a Combination Resource owner might in real-time inject from both the variable component and the storage component. See 2023 ELCC Order (Clements, Comm’r, dissenting at P 24). But in such a hypothetical case, the Combination Resource’s true capacity contribution would be greater than its accredited ELCC even using a rational method that optimizes between the solar and the storage.
[41] See Rehearing Order at P 11.
[42] See Request for Rehearing at 8-9 (PJM’s approach “is only logical if it were true that (1) PJM’s modeling limitations are unchangeable, and (2) storage resources would charge and discharge at illogical times.”).
[43] See Clean Energy Associations Protest at 11 (“PJM should focus on the aggregate facility’s CIRs, and the ability of hybrid resources as a whole—not based upon isolating and diminishing components—to provide reliable capacity to the region.”).
[44] 2023 ELCC Order at P 53.
[45] It bears repeating that nothing about PJM’s modeling approach was justified as required by the real-time operations of Combination Resources. In fact, I object to the modeling approach precisely because it ignores the very different way in which Combination Resources will continue to behave in real time.
[46] 2023 ELCC Order at P 53.
[47] Rehearing Order at P 38.
[48] 2023 ELCC Order at P 53 (emphasis added).
[49] See id. (Clements, Comm’r, dissenting) at PP 23-26.
[50] See Rehearing Order at P 37 n. 76.
[51] Id. at P 37.
[52] In other words, the subsequent design fix proves the majority’s premise false. PJM could design a model that provided for rational behavior of Combination Resources that accounts for the linkage of the two components. This leaves PJM’s choice entirely unexplained, rendering the majority’s approval of that approach arbitrary and capricious.
[53] New York Independent System Operator, 179 FERC ¶ 61,102, at P 39 (2022).