FERC Issues reactive power order in panda - order 574

​On March 31, 2021, FERC issued Opinion 574 which concerns the reactive power revenue requirement of Panda Stonewall, a generator in PJM.  This case has been pending at FERC for some time - the ALJ previously issued her initial decision on April 26, 2019.  Also on March 31, FERC denied a petition for declaratory order requested by several generator owners as FERC determined the reactive power revenue requirement issues included in their request are best resolved on a case-by-case basis and the decision in Panda Stonewall provides guidance on the issues.  Here are the major findings in Opinion 574 involving Panda:
 

1. Burden of proof:  FERC found that, under section 205 of the FPA, the applicant bears the burden of proof to show that the proposed rate or charge is just and reasonable.  FERC has explained that the applicant must establish a prima facie case, and the test for prima facie evidence is “whether there are facts in evidence which if unanswered would justify [persons] of ordinary reason and fairness in affirming the question which the plaintiff is bound to maintain.”  The burden of going forward then shifts to the opposing parties.  But the ultimate burden of persuasion remains with the proponent, who will prevail only if the preponderance of the evidence supports its position.
2. EPC costs assigned to FERC accounts:  FERC found that Panda failed to demonstrate that its major equipment cost assignments were just and reasonable.  FERC found that the Bechtel spreadsheet provided by Panda for the assignment of the EPC contract costs to FERC accounts was unreliable, as it contains numerous, basic USofA errors.  The primary author of the document, Mr. Young of Bechtel, did not testify and, by Panda’s own admission, had “very limited knowledge” of USofA, and Panda chose witnesses that had limited contact with Bechtel and no contact with the primary author of the spreadsheet.  FERC found that Panda did not attempt to obtain more information from Bechtel or verify and support the information it had received.  Nor did it provide all the information readily available to it.  In fact, the evidence suggests that Panda did everything it could to limit the amount of information from Bechtel that would be introduced into the record.
3. Indirect construction costs:  FERC found that Panda established its prima facie case supporting recovery of its proposed indirect costs, except for those costs related to the switchyard and charitable donations.  However, FERC found that Panda did not demonstrate that its proposed allocation of project indirect costs to only the generator/exciter, GSU and AEE categories and not to other categories included in the balance of plant was just and reasonable.   FERC stated that Panda’s project indirect costs were not only incurred to support the generator and exciter, GSU, and AEE investment categories; rather, as Panda admits, indirect costs “are incurred to build the entire facility.”  Accordingly, FERC found that Panda’s indirect costs should be allocated to all four investment categories, including balance of plant.    
4. Reactive power allocation factor:  FERC affirmed that, because a generator has the ability to produce reactive power up to its nameplate capacity, and because it may be called on to do so, “there is no rationale that would warrant using anything less in determining a generator’s reactive power capability.”[1]  As the Commission has explained on multiple occasions, and in cases involving generators from different regional transmission organizations, under the AEP methodology, generators are compensated for providing reactive service based on a unit’s capability of providing that service.[2]
5. AEE Allocation to supporting generator/exciter:  In Dynegy Midwest Generation, Inc., FERC calculated the AEE allocation factor (AEAF) as the ratio of generator and exciter auxiliary load divided by total production plant auxiliary load.  In more general terms, this means that the AEAF numerator is the amount of electricity consumed by the components of the generator and exciter responsible for reactive power production, and the AEAF denominator is the auxiliary load for the entire plant.  FERC found that Panda’s use of a combination of design and operational data to determine the AEAF to be unjust and unreasonable as maximum design output is generally higher, sometimes significantly so, than actual operational output.  Because the AEAF is a ratio between two numbers, inherent differences in the types of the data would skew the ratio in one direction or another.  In a case where design data (which is typically higher) is used in the numerator, and operational data (which is typically lower) is used in the denominator, as was done by Panda, the result, all other things being equal, would overstate the AEAF. 
6. Balance of plant allocation factor – lagging reactive power:  FERC found that Panda did not show that its proposal to use the higher of lagging or leading reactive power was just and reasonable.  FERC states that the purpose of reactive power compensation is to incentivize investment in equipment related to the production of reactive power, not the power to absorb reactive power.[3] While the Commission has acknowledged that the ability to absorb reactive power provides a reliability benefit to the transmission system,[4] nothing in our precedent suggests that the AEP methodology provides compensation for the ability to absorb reactive power.  Rather, the purpose of reactive power compensation is to incentivize investment in the capability to produce reactive power.[5] 
7. Firm fuel transportation costs:  FERC has previously observed that “[a]s the AEP methodology is a fairly complex, multi-part calculation, it is important for the Commission to assess proposals for new additional reactive power charges to determine whether the charges are already reflected in the AEP methodology”, such as what Panda requested with firm fuel transportation costs (to be included in O&M in the fixed charge factor).  Fuel costs, which include fixed costs (reservation charges) and variable costs (e.g., usage charges), are primarily related to the provision of real power and not reactive power.  Only a small percentage of a generator’s fuel costs is related to the provision of reactive power, and as Panda concedes, the major portion of firm fuel transportation costs will be recovered through the PJM capacity market.  The AEP methodology already has a means in place to provide compensation for the small amount of additional fuel used during the production of reactive power, which is the heating loss calculation based on the MW-hours of actual reactive power production and the usage charges for fuel.  The heating loss component of Panda‘s reactive power revenue requirement already includes the price associated with the transportation of fuel.  To permit Panda to also recover these costs as a fixed O&M expense would result in double recovery.  FERC concluded that heating losses is the proper way to recover these costs as it is meant to compensate the generator for “the added incremental fuel and variable costs, if any, of producing reactive power.”
8. A&G Costs:  FERC found that 100% of A&G costs should not be included in the fixed charge factor, as activities associated with Panda’s variable O&M are supported by the A&G function, so some A&G costs are variable.  FERC clarified that the correct ratio to apply in computing A&G demand-related expenses is the ratio of fixed O&M expenses to total O&M expenses. 
9. Return on equity:  FERC found that the Reference Resource from the 2014 CONE Study, while appropriate for the purpose of determining CONE for the Variable Resource Requirement demand curve, has not been shown to be a just and reasonable proxy for setting Panda’s ROE and capital structure in the context of a reactive power rate case.  In approving the 2014 CONE Study’s cost of capital, the Commission noted that it “is a just and reasonable estimate for the purpose of estimating Gross CONE.”[6]  CONE, by its nature, is a hypothetical estimate of costs used to approximate the revenue that a newly constructed peaking unit would need to recoup those costs; the purpose of the 2014 CONE Study is to calculate CONE for purposes of establishing the curves for the capacity auction.[7]  Here, Panda proposed to use the 2014 CONE Study’s cost of capital for an entirely different purpose, i.e., to serve as a proxy to estimate of the cost of capital for a specific entity—here, Panda—to be included in a reactive power rate.  The methodology used for determining the cost of capital for the hypothetical facility in the 2014 CONE Study differs from the Commission’s standard methodology used in cost-of-service rate proceedings for individual entities.  Cost of service proceedings generally rely on more specific data and financial models, whereas the 2014 CONE Study relied on data points where “there [were] no details provided on how these ranges were developed.”[8]   While the 2014 CONE Study did use a CAPM analysis, it did not consider the other financial models typically used by the Commission in cost-of-service ROE cases, and instead looked to fairness estimates and other estimates from financial advisers and analysts that resulted in increasing the ROE from what the CAPM data itself would produce.  FERC found that Panda has not shown that the 2014 CONE Study’s cost of capital, when carved out of the CONE calculation, is a just and reasonable proxy for Panda’s cost of capital in setting its reactive power rate.  Rather, FERC found that the cost of capital data for Dominion, the interconnected utility, is a just and reasonable proxy for Panda’s cost of capital.  In Chehalis, the Commission observed that when the Commission has used an interconnected utility as a proxy, “it was proposed by the generator and the interconnected utility’s return was a conservative estimate of the generator’s return, as the generator essentially faces more risk than the interconnected utility.”[9]  Merchant generators frequently propose the interconnected utility as a proxy that the Commission accepts as just and reasonable.[10]  Thus, consistent with this precedent, absent a better proxy, using the interconnected utility as a proxy produces a cost of capital that falls within the range of just and reasonable returns.  FERC noted that Panda had the option of performing its own unit-specific ROE analysis but chose not to do so.  Panda also had the option to propose and support the use of a different proxy.  In any case, Panda had the burden of demonstrating that its choice of proxy was just and reasonable, which it failed to do.  Accordingly, FERC determined that the only proxy in the record that falls within a range of just and reasonable results is the interconnected utility.

[1] ATSI, 119 FERC ¶ 61,020 at P 27.

[2] See, e.g., Bluegrass Generation Co., L.L.C., 118 FERC ¶ 61,214, at P 21 (2007) (Bluegrass) (MISO); Calpine Oneta Power, L.P., 116 FERC ¶ 61,282, at P 50 (2006) (Southwest Power Pool, Inc.); Rolling Hills Generating, L.L.C., 109 FERC ¶ 61,069, at P 12 (2004) (PJM). 

[3] See id. (“We agree with AEP (and the judge) that the allocation factor should be based on the capability of the generators to produce VArs . . .”); Va. Elec. & Power Co., 114 FERC ¶ 61,318 at P 3 (citing AEP, 88 FERC at 61,457) (explaining that the AEP methodology is used “to compute the portion of plant investment attributable to reactive power production” (emphasis added)).

[4] See S. Co. Servs., Inc., 80 FERC at 62,080-81.

[5] See AEP, 88 FERC at 61,457; Va. Elec. & Power Co., 114 FERC ¶ 61,318 at P 3.

[6] VRR Curve Order, 149 FERC ¶ 61,183 at P 76 (emphasis added).  We note that NYISO, 158 FERC ¶ 61,028, a case upon which Panda relies, was similarly about establishing a market benchmark and does not support any cost of capital generally applicable in a cost-of-service proceeding.

[7] See 2014 CONE Study at iii.

[8] Id. at 36.

[9] Chehalis, 123 FERC ¶ 61,038 at P 167; see also Dynegy, 121 FERC ¶ 61,025 at PP 54-55.

[10] See, e.g., Bluegrass, 118 FERC ¶ 61,214 at P 86; Calpine Fox LLC, 113 FERC ¶ 61,047, at P 17 (2005).