Petitioner Sithe/Independence Power Partners ("Sithe") challenges two orders of the Federal Energy Regulatory Commission ("FERC" or "Commission"), dismissing Sithe's complaint against Niagara Mohawk Power Corporation ("Niagara") and denying rehearing of that dismissal. Pursuant to a 1991 agreement, Sithe, an independent power producer, receives transmission services from Niagara, an electric utility. The challenge in this case centers on the loss component of the overall rate that Sithe pays Niagara for this service. Sithe argues that, in computing loss rates on an incremental basis, Niagara is violating established Commission policy, which requires that base and loss rates be computed using a consistent methodology. Sithe also argues that Niagara's method of assigning loss calculation priorities unfairly discriminates on the basis of a customer's contract date. In the orders on review, FERC dismissed Sithe's challenge, finding that Niagara's rate methodology was not inconsistent with Commission policy, and that the rates charged were not unduly discriminatory. Because we are unable to ascertain the basis for the Commission's ruling, we remand this matter to the Commission for further proceedings.

Sithe owns and operates a cogeneration facility in New York state. Pursuant to a 1991 power purchase agreement, Sithe sells the bulk of the electric energy produced at this facility to Consolidated Edison Company of New York ("Con Ed"). Niagara, in turn, owns and operates electric generation, distribution, and transmission facilities in New York. In order to enable the transmission of energy from the cogeneration facility to Con Ed, Sithe entered into a Transmission Services Agreement ("TSA") with Niagara on November 5, 1991. Under the agreement, which extends for a twenty year period from the initiation of service, Niagara provides Sithe with roughly 805 megawatts ("MW") of firm transmission service and up to 250 MW of interruptible transmission service. In exchange, Sithe pays Niagara a base transmission rate and also compensates Niagara with in-kind energy for transmission line losses. FERC accepted the TSA for filing on February 5, 1992, and accepted a revised version for filing on July 6, 1994. Service under the TSA commenced on November 15, 1994.

Under the TSA, the total charge for transmission service consists of a base charge for Niagara's embedded costs--i.e., construction and operation expenses and a reasonable return on its investment--plus an additional charge for Niagara's transmission loss costs--i.e., the costs to the utility of making up the power loss that inevitably occurs during transmission. There are at least two basic methodologies that a utility may use to compute its base and loss rates: the "rolled-in" rate method and the incremental rate method. The rolled-in rate methodology charges a rate based on the average cost to the utility of serving all of its customers on an integrated system. "[E]ach transmission customer [is treated] not as using a single transmission path but rather as using the entire transmission system," and, accordingly, "each transmission customer pays its share of the capital costs of the entire system" based on the amount of electricity it has transmitted over the system. Northern States Power Co. v. FERC, 30 F.3d 177, 179 (D.C.Cir.1994). On the other hand, the incremental rate methodology charges a rate based on how a given customer's individual use affects the system. Each transmission customer pays an amount that compensates the utility for the marginal costs that its use imposes, which is a function of such factors as distance and direction.

The methodology for computing rates in this case was not clearly specified in the TSA. With respect to the base transmission rate, the rate schedule that is part of the TSA identified a firm service rate of $1.49 per kilowatt ("kW") per month, which was later modified to $1.76 per kW per month. By agreement, this rate was based on the formula that had governed Niagara's earlier arrangements with New York State Electric & Gas Corporation ("NYSEG") and Central Hudson Gas & Electric Corporation ("CHGE"). Yet, the parties' characterizations of this underlying formula are at odds. It appears from the record that the $1.76 figure represents an average of the costs associated with a substantial portion--but not the entirety--of Niagara's transmission system. In other words, the starting point from which averages are taken excludes investment in a particular substation, a flat sum of $43.5 million, and various other expenses. In Sithe's view, then, the formula reflects an average, rolled-in methodology, even if it does not produce a fully embedded rate. In Niagara's view, however, the formula reflects a "hybrid," modified cost-of-service approach that is not "fully rolled-in."

With respect to the transmission loss rate, the TSA expressly provides for compensation, in the form of in-kind energy, for losses resulting from Niagara's service to Sithe. Section 9.1 of the agreement states that:

[Sithe] shall compensate NIAGARA MOHAWK for losses incurred by NIAGARA MOHAWK in its control area and NIAGARA MOHAWK shall compensate [Sithe] for losses avoided by NIAGARA MOHAWK in its control area as a result of NIAGARA MOHAWK's provision of transmission services hereunder. The determination of such losses and the procedure for compensation thereof shall be determined by NIAGARA MOHAWK's Power Control Department in accordance with NIAGARA MOHAWK's practices relating to other similar transactions and in accordance with GOOD UTILITY PRACTICE.

Transmission Services Agreement § 9.1, reprinted in Joint Appendix ("J.A.") 57. There is a dispute among the parties as to whether this provision explicitly identifies an incremental loss methodology. Before the Commission, Niagara argued that this conclusion was dictated by the language of the agreement, and that Sithe was fully aware of this when it entered into the agreement. Sithe, for its part, contended that the provision did not specify any methodology, and that it was unaware that Niagara employed an incremental approach until Niagara eventually supplied it with information underlying the rates it was charging, well after execution of the TSA. The Commission did not make an express finding on this point.

In any event, there is little dispute that Niagara does, in fact, compute Sithe's transmission loss charge on an incremental, rather than average, basis. Niagara's approach can be roughly summarized as follows. First, using a load flow model, Niagara takes two "snapshots" of its system each month, one during peak use and one during off-peak use, in order to determine the amount of energy being transmitted over the system and the amount of energy lost in transmission. Niagara then extrapolates from this data to calculate the total transmission losses incurred for each month. Next, Niagara removes its customers from the model one at a time and re-calculates losses at each step, thereby arriving at a marginal loss figure for each customer. To establish the order of removal, Niagara assigns its customers "priorities" based on the dates of their firm transmission contracts. Because transmission losses are a function of the current flowing over the system, incremental losses decrease exponentially as each load is removed from the calculation--i.e., the assigned priority substantially affects the amount of the loss for which a given customer will be charged. According to Sithe, "Niagara's use of an incremental loss factor to assign losses to Sithe instead of using an average, rolled-in loss factor costs Sithe approximately $10,000 per day." Amended Brief of Petitioner at 9.

On March 29, 1995, Sithe filed a complaint with the Commission pursuant to § 206 of the Federal Power Act ("FPA"), 16 U.S.C. § 824e (1994), challenging Niagara's method of computing transmission losses under the TSA as unjust, unreasonable, and unduly discriminatory. Sithe contended, inter alia, that Niagara was violating the Commission's longstanding "matching policy" by using an incremental method to calculate Sithe's transmission loss charges, while at the same time using an average rolled-in method to calculate Sithe's base transmission rate. Sithe also contended that Niagara's method for determining incremental losses discriminated unfairly among its transmission customers by prioritizing transactions based on contract dates, and by reserving the highest priority for its own uses of the system. Sithe sought an order from the Commission finding Niagara's calculation of transmission losses to be unlawful and directing Niagara to calculate transmission losses on an average, system-wide basis.

In its answer, Niagara raised multiple arguments in support of its current method of computing Sithe's rates. First, Niagara suggested that Commission policy recognizes the benefits of flexibility in rate structuring and does not strictly require the use of matching methodologies. Next, Niagara represented that it does not use a rolled-in methodology to compute Sithe's base transmission rate, but rather a hybrid methodology that includes only a portion of Niagara's fixed costs. On this basis, Niagara argued that it was not required to compute losses using system-wide averages even if the matching policy applied. Finally, Niagara attempted to demonstrate that the total rate it charges Sithe is lower than the total rate it could charge Sithe if it computed both base and loss rates using a fully rolled-in methodology, and that Sithe had bargained for this benefit. On this alternative basis, Niagara asserted that its rates were just and reasonable and, accordingly, that Sithe had not been harmed.

On September 16, 1996, FERC summarily dismissed Sithe's complaint. See Sithe/Independence Power Partners, L.P. v. Niagara Mohawk Power Corp., 76 F.E.R.C. p 61,285 (1996) ("Initial Order"). After acknowledging that "Commission policy requires consistency in pricing a utility's transmission capacity charge and transmission losses," the Commission found Niagara's "computation of Sithe's transmission losses to be consistent with this directive." Id. at 62,458. The Commission agreed with Niagara's characterization of the base transmission rate as reflecting a "hybrid" methodology. See id. The Commission also found that Niagara charges Sithe a base transmission rate that is "discounted below its average system cost," and that "[t]he amount of discount each customer receives is based on the incremental transmission losses it imposes on [Niagara's] transmission system." Id. Finally, the Commission reasoned that Niagara's system "allows customers that impose lower variable costs than other customers to receive a commensurately larger discount," and, therefore, that Niagara's methodology does not produce unduly discriminatory or preferential results. Id. at 62,458-59. Although it found Sithe's complaint to be "without merit," id. at 62,457, the Commission ordered Niagara to provide Sithe with "sufficient information for Sithe to verify [Niagara's] calculations of Sithe's transmission losses." Id. at 62,459.

On October 21, 1997, the Commission denied Sithe's request for rehearing. See Sithe/Independence Power Partners, L.P. v. Niagara Mohawk Power Corp., 81 F.E.R.C. p 61,071 (1997) ("Rehearing Order"). In this order, FERC affirmed its earlier conclusion that Sithe "enjoys a discount from the maximum legal rate [Niagara] could impose for the transmission service it provides," id. at 61,301, purporting to base this conclusion on an "independent analysis" by which it confirmed that "Sithe's overall transmission rate under its agreement with [Niagara] is less than it would be if both the transmission rate and the loss rate were calculated to include a full allocation of embedded average costs." Id. at 61,301 n. 1. FERC also reiterated that, "because [Niagara] reduces the price to a greater extent for those customers imposing lower losses, any price disparity that results from using incremental cost ratemaking to calculate the loss factor portion of the rate does not amount to undue discrimination. All customers are treated comparably." Id. at 61,301. After finding that Sithe enjoys a discount, the Commission went on to deem Sithe's arguments on this score "irrelevant," as FERC's "ratemaking practices allow [Niagara], because it used a hybrid method to compute Sithe's transmission rates, to consider incremental costs in computing transmission losses" without violating the "matching" rule. Id. at 61,302. As additional support for its finding of no undue discrimination, the Commission noted that Sithe had "freely negotiated" its rates, and that FERC does not consider rate disparities inappropriate where they result from arm's-length negotiations. Id. Finally, the Commission concluded that Sithe's argument concerning a violation of the matching policy failed because Sithe incorrectly presumed that its base rate incorporated Niagara's full average costs. See id.

This petition for review followed.

We review FERC orders under the Administrative Procedure Act's ("APA") arbitrary and capricious standard. See Union Pac. Fuels, Inc. v. FERC, 129 F.3d 157, 161 (D.C.Cir.1997); 5 U.S.C. § 706(2)(A) (1994). Because " '[i]ssues of rate design are fairly technical and, insofar as they are not technical, involve policy judgments that lie at the core of the regulatory mission,' our review of whether a particular rate design is 'just and reasonable' is highly deferential." Northern States, 30 F.3d at 180 (quoting Town of Norwood v. FERC, 962 F.2d 20, 22 (D.C.Cir.1992)). Nonetheless, such review is not an "empty gesture." Id. "[T]he Commission must be able to demonstrate that it has 'made a reasoned decision based upon substantial evidence in the record,' " id. (quoting Town of Norwood, 962 F.2d at 22), and the "path of [its] reasoning" must be clear. Id. at 182.

Because Sithe is challenging Niagara's existing rates, its claim must be brought pursuant to § 206, rather than § 205, of the FPA. Section 206(a) provides, in relevant part, that:

[w]henever the Commission, after a hearing had upon its own motion or upon complaint, shall find that any rate, charge, or classification, demanded, observed, charged, or collected by any public utility for any transmission or sale subject to the jurisdiction of the Commission, or that any rule, regulation, practice, or contract affecting such rate, charge, or classification is unjust, unreasonable, unduly discriminatory or preferential, the Commission shall determine the just and reasonable rate, charge, classification, rule, regulation, practice, or contract to be thereafter observed and in force, and shall fix the same by order.

16 U.S.C. § 824e(a). As complainant, Sithe bore the burden in this proceeding of demonstrating that Niagara breached the statutory just and reasonable standard. See id. § 824e(b); Alabama Power Co. v. FERC, 993 F.2d 1557, 1571 (D.C.Cir.1993). By dismissing Sithe's complaint in the orders on review, then, the Commission implicitly found that Sithe did not satisfy its burden--either that it failed to establish its prima facie case, or that Niagara had effectively rebutted that case.

The core of Sithe's complaint in this case is that the transmission loss component of Niagara's rate violates the Commission's long-standing policy requiring that the base and loss elements of an overall rate be computed using consistent methodologies. Sithe's position rests on its contention that Niagara's rate components do not "match," because its base rate is calculated on a rolled-in basis, while its loss rate is calculated on an incremental basis. In its brief to this court, Sithe goes so far as to suggest that these characterizations are "undisputed", see Amended Brief of Petitioner at 16, but that is a stretch. To be sure, there does not appear to be much dispute that Niagara computes Sithe's loss rate on an incremental basis. See 81 F.E.R.C. at 61,301. As noted earlier, however, there is considerable disagreement concerning the appropriate characterization of the methodology by which Niagara computes Sithe's base rate: Sithe argues that it is an "average, rolled-in" approach, whereas Niagara argues that it is a "modified cost-of-service" approach that is not "fully rolled-in." The record, as far as we can tell, suggests that this calculation makes use of an average formula, but does not incorporate the full complement of Niagara's costs, as it excludes investment in certain facilities, a flat sum, and various expenses.

Whatever the appropriate description of this disputed methodology (fully rolled-in, partially rolled-in, or not rolled-in at all), it is clear that Niagara does not use the same approach to calculate both the base and loss components of its rate. This would appear to defy FERC's well-established policy of requiring utilities to compute the base and loss components of their transmission rates using consistent methodologies. See, e.g., Northern States Power Co., 59 F.E.R.C. p 61,100, at 61,369, reh'g denied, 60 F.E.R.C. p 61,076, at 61,252-53 & n. 25 (1992), clarification denied, 64 F.E.R.C. p 61,111, at 61,920 (1993), aff'd sub nom. Northern States Power Co. v. FERC, 30 F.3d 177 (D.C.Cir.1994). In the Northern States proceeding, FERC stated that it "does not allow the use of incremental loss factors when the transmission charge is developed on a rolled-in basis." 59 F.E.R.C. at 61,369. It further explained that:

[t]his policy flows directly from the Commission's longstanding practice of requiring that the demand and energy components of a rate be calculated on the same basis. Where, as here, the customer pays for average fixed costs rather than the fixed costs of only certain incremental facilities, logic dictates that the customer pay for average variable costs in the energy charge.

Id. In upholding the Commission's application of this "matching" policy, this court recognized that, because rolled-in costing assumes that each customer in an integrated system "enjoys the benefits of the transmission system as an integrated whole," it makes sense to hold "each customer ... responsible for an indivisible portion of the transmission system losses imposed upon the system by the configuration of the group of customers using it at any one time." Northern States, 30 F.3d at 182. Where, then, is the Commission's explanation for not requiring "matching" methodologies in this case?

The Commission purported to re-affirm its "matching" policy in the orders on review, see 76 F.E.R.C. at 62,458, but concluded that Sithe had failed to establish a violation thereof. In so holding, however, FERC neither provided a clear explanation of its rationale nor revealed the data and assumptions underlying its findings. On the record before us, therefore, we are unable to satisfy ourselves that the Commission engaged in reasoned decision making and reached conclusions supported by the record. Indeed, we are unable to penetrate the logic of FERC's orders to ascertain whether FERC in fact departed from established policy and precedent and, if so, whether it justified that departure. Cf. Northern States, 30 F.3d at 180 ("[T]he Commission must be able to demonstrate that it has 'made a reasoned decision based upon substantial evidence in the record.' ") (quoting Town of Norwood, 962 F.2d at 22); Hatch v. FERC, 654 F.2d 825, 834 (D.C.Cir.1981) ("[A]n agency must provide a reasoned explanation for any failure to adhere to its own precedents.").

The Commission's position could be, as both FERC's appellate counsel and Niagara urge before this court, that the "matching policy" is not implicated in this case, because Niagara does not compute its base transmission rate on a fully rolled-in basis. See Brief for Respondent at 19; Brief of Intervenor at 9. Indeed, if we stretch our imagination, we might even conclude that a single sentence in the Rehearing Order, indicating that Sithe's "matching" argument failed because it "incorrectly presumes that [Niagara] developed Sithe's transmission rates using full average costs while Sithe's transmission losses were based on incremental costs," 81 F.E.R.C. at 61,302 (emphasis in original), lends some support to the view pressed by FERC's appellate counsel and Niagara. And this argument is tempting, for it avoids the absurdities that are faced in any attempt to employ a "matching" rule in this case--the notion of "matching" is utterly baffling when the methodologies themselves are impure, as with Niagara's base rate in this case. What, we wonder, would "match" this rate, which apparently resulted from a negotiated compromise?

The problem, however, is that FERC did not articulate the rationale that appellate counsel and Niagara now propound. Thus, we cannot accept this articulation of the agency's judgment. As we explained in Hatch:

Without any explicit recognition by the Commission that the standard has been changed, or any attempt to forthrightly distinguish or outrightly reject apparently inconsistent precedent, we are left with no guideposts for determining the consistency of administrative action in similar cases, or for accurately predicting future action by the Commission. The failure to admit or explain such a basic change in the interpretation of a statutory standard to be applied to conduct of the public undermines the integrity of the administrative process.

654 F.2d at 834-35 (footnote omitted).

The bottom line is that, if the asserted conclusion were so obvious, it would have been a simple task for FERC to clearly state and support the view that its matching rule does not, or even should not, apply unless the base rate is fully rolled-in. FERC did not do this, so we cannot rely on the theory now advanced by its appellate counsel. See SEC v. Chenery Corp., 318 U.S. 80, 87, 63 S. Ct. 454, 87 L. Ed. 626 (1943). The simple fact here is that FERC purported to apply the "matching rule" in a situation in which the disputed rate methodologies concededly do not match.

Alternatively, the Commission's ruling arguably could rest on the finding that the rate paid by Sithe to Niagara reflects a discount below Niagara's average system costs. See 76 F.E.R.C. at 62,458, 81 F.E.R.C. at 61,301 & n. 1. Yet, FERC did not elaborate on what significance, if any, it ascribed to this purported discount. Compare 76 F.E.R.C. at 62,458 (apparently relying on the existence of the discount in deeming Niagara's methodology consistent with the matching rule), with 81 F.E.R.C. at 61,302 (describing as "irrelevant" Sithe's argument that it did not receive a discount from Niagara). For instance, FERC may have considered this discount in deciding that Niagara's base rate was hybrid, rather than purely rolled-in. Or, FERC may have accepted an inference that Sithe pays an overall rate--both cost and loss components--that is less than it would pay if Niagara computed both costs and losses using a fully rolled-in method and, accordingly, that the rate is reasonable notwithstanding any noncompliance with the matching policy.

Niagara urges this second view, stating that "[b]y requiring consistent application of rolled-in or incremental methodologies, the Commission has created a short-hand means for preventing excessive rates." Brief of Intervenor at 9. In other words, "[i]f the 'end result' of the combined rate is less than the just and reasonable rate produced by consistent methodologies, then the rate is ipso facto just and reasonable." Id. at 11. Lending some support to this view, we have, in the past, suggested that the use of a so-called "comparison rate"--comparing the rate charged to the rate that could lawfully have been charged--seems "eminently reasonable." City of Holyoke Gas & Elec. Dep't v. FERC, 954 F.2d 740, 742 (D.C.Cir.1992). FERC, however, did not state this rationale in the orders below.

It is also noteworthy that the Commission itself has, in the past, interpreted the § 206 burden scheme to require a customer seeking an investigation into existing rates to "provide some basis to question the reasonableness of the overall rate level, taking into account changes in all cost components and not just [the challenged component]." Houlton Water Co., 55 F.E.R.C. p 61,037, at 61,110 (1991); see also City of Hamilton, Ohio and Am. Mun. Power-Ohio, Inc. v. Kentucky Power Co. and Ohio Power Co., 72 F.E.R.C. p 61,158, at 61,785-86 (1995) (dismissing, without prejudice to re-filing, a customer's § 206 complaint because it failed to satisfy the threshold of providing a basis to question the overall reasonableness of the utility's rates). Although it appears that Sithe never made a proffer to show net harm under Niagara's existing scheme, FERC did not articulate this as its basis for dismissing Sithe's complaint. In short, we cannot uphold FERC's reliance on the discount without a clearer explanation of the basis for that reliance.

We find, moreover, that the Commission's reasoning was cryptic with respect to its finding that Sithe receives a discount. The Commission offered no indication of what exactly its "independent analysis" entailed or what issues it considered. Cf. Maine Pub. Serv. Co. v. FERC, 964 F.2d 5, 9 (D.C.Cir.1992) (remanding in part because "the Commission's analysis ... contains no explanation of how it derived the [relevant] figure"); City of Holyoke, 954 F.2d at 743 ("The Commission must support its decision with enough data to enable an adversely affected party and by extension a reviewing court, to understand its calculation of the comparison rate upon which it would rely, as well as the underlying assumptions."). Because the Commission in this case failed to disclose its calculations, we do not know whether it simply adopted Niagara's figures and, if so, why. We also have no basis to confirm FERC's assumptions--for instance, we are left to wonder if it considered whether Niagara's variable costs, such as transmission line losses, would remain constant and, therefore, whether the alleged discount would actually exist over the long term. If the Commission does, in fact, intend to rely on the existence of a discount to support its disposition of this case, it must clarify its reasoning and support its findings in this regard.

As a final matter, we note that Sithe also raised an undue discrimination claim in this case, based on Niagara's method of prioritizing its customers for the purpose of assigning transmission losses. Under Niagara's system for calculating losses, the order in which customers are removed from the model substantially affects the losses for which they are charged. Sithe argued, on rehearing and before this court, that the Commission failed to address this argument. We are not so sure. Throughout both orders, FERC made reference to the issue of undue discrimination, apparently rejecting Sithe's claim because Niagara's customers are treated "comparably," and because Sithe "freely negotiated" its rates. See 81 F.E.R.C. at 61,301-02. FERC did not, however, explicitly discuss contract dates per se, leaving us simply to infer that a customer with a later contract date imposes greater losses on Niagara's system by the sheer fact of added volume. On remand, FERC should clarify its ruling on this point. In so doing, FERC must be careful to eschew any assumption that, where aggregation of load increases average costs, the costs a customer "imposes" on a system depend on the sequence in which the customers are added. As we noted in Town of Norwood,

Each customer's contribution to the coincident peak load "causes" the costs associated with the peak, regardless of whether that contribution comes from the customer's increasing, or its failing to diminish, its historic consumption.

962 F.2d at 24 n. 1 (emphasis added). In the meantime, we express no view on the merits of Sithe's undue discrimination claim.

In sum, although the Commission's dismissal of Sithe's complaint may be justifiable, we cannot find "the path of [FERC's] reasoning" in the orders on review. Northern States, 30 F.3d at 182. FERC did not discuss the precedent, if any, upon which it relied or chose not to rely; nor did it disclose in any meaningful way the underlying data and assumptions that supported its factual findings. Because we are unable to ascertain whether the Commission's basic conclusion is that Niagara complied with the matching policy, that the matching policy does not apply in this case, or that for some other reason Sithe has failed to satisfy its burden under FPA § 206, we remand to the agency for further proceedings.

So ordered.