Public Service Co. v. Federal Energy Regulatory Commission

Opinion for the Court filed by Circuit Judge MacKINNON.

MacKINNON, Circuit Judge:

Petitioner electric companies1 seek review of ten orders of the Federal Power Commission (hereinafter “Commission”).2 In the challenged orders, the Commission refused to allow the electric companies to impose rate surcharges to compensate for alleged uncompensated fuel costs to which they allege they became entitled when they filed new rates based on fuel costs incurred in the billing month rather than for some prior period. The Commission held that the proposed surcharges amount to retroactive ratemaking, which is prohibited by the Federal Power Act. Petitioners counter that the surcharges are necessary because of changes in the companies’ rate structure, and that they are not retroactive ratemaking.

The First, Third and Fourth Circuits have already ruled on this issue. The First Circuit 3 largely agreed with the position taken by the petitioners here. The Third4 and Fourth5 Circuits, on the other hand, held that it was proper for the Commission to reject the surcharges as prohibited retroactive ratemaking. We agree with the decisions by the Third and Fourth Circuits and affirm the Commission’s orders, except those involving petitioner Appalachian Power Company.

Approved rates for electric service usually have two components. The “demand charge” covers the utility’s fixed (capacity related) costs. The “energy charge” is designed to recover variable costs, primarily the cost of fuel, which recently has been increasing.

The energy charge also has two elements. The first element is the “basic energy rate.” It recovers the “base cost” of fuel, which is an estimate of what fuel will cost. The basic energy rate must be approved in advance by the Commission. The second element is the “fuel adjustment” charge. This charge is based on a formula designed to recover the difference between the base cost of fuel and the actual cost of fuel. A utility’s fuel adjustment formula must be approved by the Commission, for it is considered part of the utility’s filed rate. The monthly charge calculated under the fuel adjustment formula, however, is not subject to Commission approval. Thus, fuel adjustment clauses enable utilities to keep their rates in line with the current cost of their fuel without continually having to file for rate increases and decreases.

*134While a utility’s fuel adjustment clause must fall within certain boundaries, much of the design is left to the utility. Basically, utilities have used one of two types of fuel adjustment formula. “Cost of service” tariffs are designed to reimburse the utility for its actual fuel expenditures. These tariffs have the advantage of being accurate, but the inclusion of current costs in the monthly billings must be deferred while the utility collects and assimilates current data. “Fixed rate” tariffs incorporate a charge of a “predetermined price per unit based on costs incurred during a past test period, subject to some adjustments.”6 Fixed rate tariffs are less accurate than cost of service tariffs, but monthly billings need not be postponed until information on fuel costs is tabulated. Each type of fuel adjustment formula has its advantages and disadvantages, and the Commission allows a utility to use either one.

Until late 1975, the fuel adjustment clauses employed by the petitioners used fuel costs from up to six months earlier to compute current fuel adjustment charges.7 For example, Public Service’s January billing incorporated a fuel adjustment charge which reflected the difference between the base cost and the actual cost of fuel in the preceding November. Petitioners term this a “billing lag.”

Petitioners were forced to amend their filed rates in late 1975 because of a change in Commission regulations.8 Some of the petitioners chose this occasion to eliminate the “billing lag” in their fuel adjustment charges. That is, rather than computing current fuel adjustment charges on the basis of costs in a period prior to the billing period, the utilities switched to formulas that incorporated the actual cost of the fuel in the billing month. Thus, where under Public Service’s old formula the January fuel adjustment charge was based on the cost of fuel in the preceding November, under the new formula, the January charge is computed on the basis of the actual January cost.

In addition to the changes just mentioned, petitioners requested approval for temporary rate surcharges. The surcharges are designed to recover “deferred charges” related to the “lag period” which petitioners allege are still owed by customers under the old (superseded) fuel adjustment clauses. Petitioners maintain that adoption of the new (non “billing lag”) fuel adjustment clauses wiped out the “deferred charges” and made them uncollectable.

Public Service’s experience illustrates how elimination of the “billing lag” arguably makes “deferred charges” unrecoverable. Public Service’s superseded fuel adjustment formula was based on the cost of *135fuel in “the second preceding month.”9 The January fuel adjustment charge reflected the cost of fuel in the preceding November. Therefore, recovery of the difference between the basic energy charge and November costs was “deferred” until the following January.10

Public Service’s new fuel adjustment clause, which eliminated the “billing lag,” became effective on January 1, 1976. Under the new “cost of service” clause the cost of fuel in January is accounted for in the January charge. The January charge no longer reflects the difference between the actual and estimated costs in November. Thus, the “deferred charge” from November, 1975, and December as well, became uncollectable under the new fuel adjustment formula. (Instead the higher January cost was collected). Petitioners argue that the surcharges are necessary to recoup costs for such periods, which were never included in any billings.11

The Commission refused tó approve the surcharges. It held that (1) the proposed surcharges would be retroactive rate increases, and (2) the Federal Power Act prohibits retroactive ratemaking.12

Ill

The primary issue in this case is whether imposition of the proposed surcharges would be retroactive ratemaking. Petitioners argue that the surcharges are not retroactive rate increases. They contend that the fuel adjustment clauses which were superseded were cost of service tariffs with deferred billing and therefore that they (petitioners) were entitled under the superseded clauses to be compensated for all incurred fuel costs;13 and that since the surcharges merely enable petitioners to recover costs to which they were already entitled, but for which billing had been deferred, imposition of the surcharges would not constitute retroactive ratemaking.

*136The Commission disagrees with petitioners’ characterization of the superseded fuel adjustment clauses. It contends that the old clauses were fixed rate rather than cost of service tariffs. That is, the superseded clauses established a “formula, based on past experience, i. e., costs incurred during a prior test period, to” approximate the actual cost of fuel in the billing month.14 In Public Service’s case, for example, the costs in the second preceding month were a proxy for the costs in the billing month. Under this view, once a utility collected the fixed rate charge based on the costs in an earlier month it was fully paid, according to the rationale of its rate, for the service provided in the billing month. There were no “deferred charges” still owed by the customer. Under this analysis the requested surcharges would be retroactive rate increases.

Whether approval of the proposed surcharges would be retroactive ratemaking depends upon one’s characterization of the superseded fuel adjustment clauses. If those clauses are viewed (as petitioners do) as cost of service tariffs with deferred billing, then the requested surcharges — which merely assure that the utilities recover their fuel costs — would not be retroactive rate increases. But if the superseded clauses are viewed (as the Commission does) as fixed rate tariffs which used past costs as a proxy for the actual current cost, then the proposed surcharges would indeed be retroactive rate increases. We believe that the superseded fuel adjustment clauses were fixed rate tariffs, and therefore, we agree with the Commission’s finding that approval of the surcharges would be retroactive ratemaking.15

Petitioners make four arguments to support their claim that the superseded clauses were cost of service tariffs with deferred billing. First, they argue that the language of the clauses supports their interpretation. That is not true. While the clauses do not state that past costs are a proxy for current costs,16 neither do they state that the formulas are a method by which billing is deferred. Construed most favorably to the petitioners, the clauses are ambiguous. In light of this ambiguity, other evidence must be considered.

Petitioners’ second argument is based on their use of “deferred fuel accounting.” Public Service’s deferred fuel accounting method “reflect[s] the amount to be collected for [fuel adjustment] charges in the month in which the increased fuel costs are incurred.” 17 For example, the money to be received from a January billing is “entered in current revenues” for the preceding November.18 Pennsylvania Electric and Metropolitan Edison use a similar (though slightly different) accounting system. According to their counsel:

*137[The] accounting method defers recognition of actual expense . . . until such time as those actual dollar amounts are billed under the fuel clause. Thus, the portion of the fuel cost that is included in Penelec’s base rates is charged as an expense each month and the portion in excess of that base amount is deferred. The deferred costs are carried forward and recognized as expenses in the month in which the related fuel adjustment revenues are billed. Thus deferred fuel cost accounting properly matches costs and revenues.19

Petitioners contend that their use of deferred fuel accounting “would not make sense” 20 unless the superseded clauses were cost of service tariffs with deferred billing.

Petitioners’ use of deferred fuel accounting is consistent with their claim that one month’s fuel adjustment charge was merely deferred billing for an earlier month’s costs. Nevertheless, petitioners’ choice of this accounting method does not establish that the superseded clauses were cost of service tariffs with deferred billing. First, at least with respect to Public Service, the deferred fuel accounting method was not adopted until over one year after the superseded fuel adjustment clause became effective.21 Thus, contrary to Public Service’s counsel’s contention, this accounting method is not “tangible evidence of the Company’s intent” 22 at the time the superseded clause became effective.23 Second, a company’s accounting method may or may not evidence its intention with respect to fuel adjustment charges. Deferred fuel accounting may have been adopted because the companies believed that their fuel adjustment clauses were cost of service tariffs with deferred billing. But adoption of this accounting method may have been for entirely different reasons. The switch to this accounting method resulted in an increase in assets on the books of the utilities and could have helped in sales of their securities at a time when they were hard hit by the foreign oil embargo.24 Thus, a utility’s choice of an accounting method does not necessarily control the Commission’s decision as to the nature of the filed rate. Also, there is nothing in the Commission’s prior approval of the filed rate that requires a conclusion that it is a cost of service tariff. We are not persuaded, therefore, that petitioners’ accounting method establishes that the superseded clauses merely deferred billing.25

Petitioners’ third argument is based on Commission regulation 35.14(a)(1). This regulation provides that a fuel adjustment formula must reflect “the difference between the fuel cost ... in the base period and in the current period. ,”26 Petitioners contend that this regulation “scrupulously require[s] that only actual fuel costs enter into this calcula*138tion [of the fuel adjustment charge].”27 Since only “actual” costs can be reflected in the formula, petitioners maintain that the reference to past costs must have been a way to defer billing, rather than as a proxy for current costs.

The flaw in this argument is the assumption that “current costs” can not be estimated. If this assumption were correct, then the regulation would have the effect of prohibiting fixed rate tariffs entirely. Petitioners cite no language in the regulation that supports their assumption and there is no basis for us to conclude that the Commission would have instituted such a major change from its past practice in less than explicit language. A much more reasonable conclusion, which is consistent with Commission practice, is that “current costs” is a term of art. In cost of service tariffs, “current costs” are indeed the actual costs. But in fixed rate tariffs, “current costs” means an estimate of costs based on past experience. Therefore, we reject the argument that because of regulation 35.14 the superseded clauses must be viewed as cost of service tariffs.

Finally, petitioners contend that in light of the Commission policy of making utilities whole the superseded fuel adjustment clauses must be viewed as cost of service tariffs. We acknowledge, of course, that the goal of fuel adjustment clauses is to accurately compensate utilities for their fuel costs, and that cost of service tariffs are the best way of doing this. The Commission, however, does not require that utilities use cost of service tariffs. Within certain boundaries, utilities are permitted to design their own fuel adjustment clauses. The risk involved in the selection of a fuel adjustment clause is on the utility.28 Thus, when a utility chooses a fixed rate tariff, a kind of fuel clause that has long been permitted, Commission policy does not require that courts transform the clause into a cost of service tariff.

In short, none of petitioners’ arguments warrants the conclusion that the superseded fuel adjustment clauses were cost of service tariffs with deferred billing. We therefore turn to the arguments in support of the Commission’s decision.

We concur, for three reasons, in the Commission’s judgment that the superseded fuel adjustment clauses were fixed rate tariffs. First, the scheme of the superseded fuel adjustment clauses belies petitioners’ claim that the clauses were cost of service tariffs with deferred billing. Petitioners acknowledge that the superseded clauses were based on formulas. Their argument is that those formulas merely postponed recovery of the actual cost of service. An examination of the superseded clauses, however, indicates that was not the case. If the clauses were cost of service tariffs with deferred billing, then the formulas would have applied the fuel adjustment factor based on a particular month’s fuel costs to the amount of energy used by the customer during that month.29 But instead, the clauses involved in the tariffs filed in this case (except Appalachian) applied the fuel adjustment factor (which was based on fuel costs in a prior month or months) “to all kilowatt-hours supplied during the billing month.”30 Thus, the fuel adjustment *139charge that was used was not based on the cost of fuel that had been used to generate the power for which users were billed. Since customer usage varies from month to month, this mismatch — applying a fuel adjustment factor based on one month to kilowatt-hours used during another month— would not lead to exact recovery of actual fuel expenses.31 Thus, the superseded clauses did not simply defer recovery of the cost of fuel. The clauses brought about a charge that differed from actual cost. This result is inconsistent with petitioners’ claim that the clauses were cost of service tariffs with deferred billing. But the result squares with the Commission’s view that the clauses were fixed rate tariffs which used fuel costs in a prior period as a proxy for actual current costs.

Second, petitioners’ billing practice is inconsistent with their claim that the superseded fuel adjustment clauses were merely a means of deferred billing. If these clauses were a method of deferred billing, then the utilities would not have begun billing under the superseded clauses immediately after the clauses became effective. Rather, they would have waited until the “billing lag” had elapsed. Yet the petitioners32 began billing under the superseded fuel adjustment clauses the first month they were effective.

Public Service’s billing practice illustrates this inconsistency. Its superseded fuel adjustment formula became effective in January, 1973. Public Service began billing in accordance with that formula immediately. Under petitioners’ analysis, the January bill was really a deferred charge for the “second preceding month” — November, 1972. But as the Commission notes, Public Service could not have intended to charge for November, 1972, because its customers had already paid for service received in November, and therefore the January charge “would have constituted retroactive rate-making and would have been barred.”33 If Public Service really intended its fuel adjustment clause to be a means of deferred billing, then the proper procedure would have been to start billing under that clause in March, 1973. The only reasonable explanation for Public Service’s decision to bill under the superseded clause the first month it was effective is the one arrived at by the Commission: the superseded clauses were fixed rate tariffs which used costs incurred during a prior test period as a proxy for actual current costs.

Petitioners have no real answer to this point. They lamely contend that “the initial billings show no more than a lack of refined thinking on the correct application of the clause.” However, we cannot ascribe such carelessness in applying the clauses as their application is in conformance with their wording and they were obviously drafted in a very careful manner.34 Petitioners have generously offered to refund the extra money billed during the first months the superseded clauses were effective in exchange for the much greater amounts they would receive if their arguments about deferred billing were accepted. We are not persuaded. One’s thinking need not be very “refined” to understand that if the purpose of a fuel adjustment clause is to defer billing, then billing must be deferred.*14035 Petitioners’ decision to bill during the first months the superseded clauses were effective is a clear indication that petitioners did not consider the clauses to be cost of service tariffs with deferred billing.

Finally, the language and tone of the superseded fuel adjustment clauses is consistent with the Commission’s conclusion that they were fixed rate tariffs. It would have been very easy for petitioners, who wrote the clauses, to have included explicit language indicating that the fuel clauses merely deferred billing for one month’s fuel costs until a subsequent month. But the clauses do not contain such language. Instead they provide, for example, that a “fuel cost adjustment factor shall be applied ... in accordance with the formula set forth . . . ”36 This emphasis on a formula, which as we indicated above, did not result in an exact recovery of fuel costs, belies petitioners’ claim that the clauses merely deferred billing for a few months.

For these reasons, we believe that the Commission properly characterized the superseded fuel adjustment clauses as fixed rate tariffs. Since we agree with the Commission, it is not necessary to address the niceties between various standards of review. Associated Press v. Federal Communications Commission, 146 U.S.App.D.C. 361, 366-67, 452 F.2d 1290, 1295-96 (1971). But even if we did not agree with the Commission, reversal on this point would be inappropriate. The Commission’s interpretation of fuel adjustment clauses is entitled to a considerable amount of deference. As Judge Leventhal wrote for the court in Gulf-States Utilities Co. v. Federal Power Commission, 171 U.S.App.D.C. 57, 64, 518 F.2d 450, 457 (1975):

where the decision involves the interpretation of the parties’ intent as revealed in the language of a contract, it is proper to defer to the Commission’s expertise if its decision is “amply supported both factually and legally.” United Gas Pipe Line Co. v. Memphis Light, Gas & Water Division [358 U.S. 103, 79 S.Ct. 194, 3 L.Ed.2d 153] supra, 358 U.S. at 114, 79 S.Ct. at 201.

Accord: City of Kaukauna, Wis. v. Federal Energy Regulatory Comm’n, 581 F.2d 993, 997 (1978) (approached agency interpretation with deference but finally rejected it); New England Power v. Federal Energy Regulatory Comm’n, 187 U.S.App.D.C. 264, 270, 571 F.2d 1213, 1219 (1977); Appalachian Power Co. v. Federal Power Comm’n, 174 U.S.App.D.C. 100, 109 n.67, 529 F.2d 342, 351 n.67, cert. denied, 429 U.S. 816, 97 S.Ct. 58, 50 L.Ed.2d 76 (1976). In the instant case, even if one were somewhat persuaded by petitioners’ arguments, there is no question that the Commission’s finding is “amply supported.” 37

*141c

The Commission held hearings before it concluded that the surcharges proposed by Public Service, Pennsylvania Electric and Metropolitan Edison are retroactive rate increases. No such hearings were held involving Appalachian Power. Instead, relying on its decision in Public Service No. 79038 the Commission “summarily disallowed] the proposed fuel surcharge . ,”39 Appalachian Power contends that even if the Commission is correct with respect to the other petitioners, rejection of Appalachian Power’s surcharges without a hearing was error. We agree.

The Commission may reach decisions without holding evidentiary hearings only when there are no material facts in dispute. Citizens for Allegan County, Inc. v. Federal Power Commission, 134 U.S.App.D.C. 229, 232, 414 F.2d 1125, 1128 (1969); Independent Bankers Association of Georgia v. Board of Governors of the Federal Reserve System, 170 U.S.App.D.C. 278, 292, 516 F.2d 1206, 1220 (1975). On this record, material facts are in dispute and Appalachian Power’s formula as drafted does not determine the issue.

The central factual issue is whether the two superseded fuel adjustment clauses used by Appalachian Power were cost of service tariffs with deferred billing or fixed rate tariffs. Our decision that the clauses used by the other petitioners were fixed rate tariffs is based primarily on two facts. First, the formulas mismatched the costs from an earlier month with the power consumption in the billing month. Second, petitioners’ billing was not deferred. In Appalachian Power’s case these facts are not established in the record.

1. Mismatch. The fuel adjustment clauses that were used by Public Service, Pennsylvania Electric and Metropolitan Edison plainly applied the fuel adjustment factor (which was based on fuel costs in a prior month) “to all kilowatt-hours supplied during the billing month.” 40 Thus there was a mismatch of costs from one month with fuel usage in a later month.

The record does not conclusively establish that Appalachian Power’s superseded fuel adjustment clauses prescribed such a mismatch. For example, the clause in Appalachian Power’s filed rates for its wholesale customers provided that if the cost of fuel exceeded the base cost, then “an additional charge . . . equal to the product of the actual kiloWatt-hours used and a fuel clause adjustment factor . . . shall be made . . . .”41 Appalachian Power acknowledges that the fuel adjustment factor was based on the costs in the second preceding month.42 Its fuel adjustment factor, therefore, was essentially the same as that used by the other petitioners.

But while the fuel adjustment factor was similar, the fuel consumption that this factor was applied to may have been different. Appalachian Power’s clause (see Appendix 149 of 195 U.S.App.D.C., 963 of 600 F.2d) provides that the factor should be applied to “the actual kiloWatt-hours used,” but the clause does not indicate which month’s “actual kiloWatt-hours” is intended. Perhaps the factor was applied to the kilowatt-hours consumed in the billing month, just as in the clauses used by the other petitioners. But it is also possible that the factor was applied to the kilowatt-hours used in the second preceding month— the same month that was the basis for the fuel adjustment factor. If the latter were true, then the fuel adjustment charge *142would have accurately recovered the actual cost of service on a deferred basis.

The testimony of Appalachian Power’s witnesses supports its contention that there was no mismatch in Appalachian Power’s fuel adjustment clauses. One witness stated:

The terms of the [superseded] clause reflected, as a practical matter a billing lag; i. e., if fuel cost in one month were above (or below) the base rate level, then this increased fuel cost would be recovered in the rates for the next succeeding month.43

Similarly, another Appalachian Power witness indicated that under the superseded fuel adjustment clause the exact difference between the base cost and actual cost of fuel in one month was recovered in a subsequent month.44

Since the fuel adjustment clause is ambiguous, and since Appalachian Power’s witnesses support the company’s claim that there was no mismatch in billing under the superseded clauses, we are unable to agree with the Commission that there are no material facts in dispute. At the same time, the testimony provided by Appalachian Power does not establish that there was not a mismatch. On remand, therefore, the Commission should determine the actual billing practice used by Appalachian. If the Commission finds that the fuel adjustment factor was applied to fuel consumption in the billing month, then this would support the Commission’s belief that the clause was a fixed rate tariff. But if the superseded clause was applied in a way that resulted in exact recovery of actual costs (e. g., no mismatch), then this would support the company’s claim that the clause was a cost of service tariff with deferred billing.45

2. Immediate billing of fuel adjustment charges. With respect to billing practices, Public Service, Pennsylvania Electric and Metropolitan Edison all started billing customers for fuel adjustment charges in the first month that their superseded clauses were effective. This billing practice is inconsistent with their claim that the fuel adjustment clauses were a method of deferred billing.46 But since no hearings were held on Appalachian Power’s billing practice, we have no record evidence to review. Thus, as with the mismatch question, this factual issue is still in dispute as to Appalachian.

In summary, it is not certain, on this record, that Appalachian Power’s superseded clauses mismatched costs and consumption. Similarly, there is no evidence as to Appalachian Power’s billing practice. In short, the primary reasons that the other petitioners’ superseded clauses were found to be fixed rate tariffs do not appear on this record to be true as to Appalachian Power. Neither are such reasons disproved. The Commission’s rejection of Appalachian Power’s proposed surcharges based solely on the Commission’s holding with respect to the other petitioners, therefore, was erroneous. There is still a dispute as to whether Appalachian Power’s superseded clauses were fixed rate tariffs. A remand on that question is therefore necessary.

Thus far, we have determined that the superseded clauses were fixed rate tariffs, and therefore that the proposed surcharges are retroactive rate increases.47 The second major issue to be resolved is whether this finding absolutely precludes approval of the *143surcharges. The Commission held that it does, relying on the rule against retroactive ratemaking. The First Circuit, however, suggests that the rule need not be strictly applied, and that in some cases retroactive ratemaking is permissible.48 Since we disagree with the decision of the First Circuit, and since the rule against retroactive rate-making was barely mentioned in the opinions by the Third and Fourth Circuits,49 we will address this issue.

This court stated in Nader v. Federal Communications Commission, 172 U.S.App.D.C. 1, 21, 520 F.2d 182, 202 (1975) that “[i]t is ... a cardinal principle of rate-making that a utility may not set rates to recoup past losses, nor may the Commission prescribe rates on that principle.” Though this rule is applied in a variety of settings,50 it is most prominent in ratemaking under the Federal Power Act, 16 U.S.C. § 824d (1977) and the Natural Gas Act, 15 U.S.C. § 717c (1977).

[a]ll rates and charges made, demanded, or received by any public utility for or in connection with the transmission or sale of electric energy subject to the jurisdiction of the Commission .

16 U.S.C. § 824d(a) (1977). If the Commission finds that existing rates are unreasonable, then it may establish reasonable rates “to be thereafter observed and in force . . . .” 15 U.S.C. § 717d(a) (emphasis added). The provisions of the Natural Gas Act are “virtually identical.” As a result of the similarity in the two Acts, cases applying the rule against retroactive ratemaking to ratemaking under one Act have been viewed as authority for applying the rule to ratemaking under the other Act as well. Indiana & Michigan Electric Co. v. Federal Power Commission, 162 U.S.App.D.C. 384, 342, 502 F.2d 336, 344 n.2 (1974), cert. denied, 420 U.S. 946, 95 S.Ct. 1326, 43 L.Ed.2d 424 (1975).51

Perhaps the leading case dealing with ratemaking under the Federal Power Act is Montana-Dakota Utilities Co. v. Northwestern Public Services Co., 341 U.S. 246, 254, 71 S.Ct. 692, 696, 95 L.Ed. 912 (1951). Therein, the court cited the legislative history of the Act for the proposition that “Congress withheld from the Commission power to grant reparations.” Justice Frankfurter’s dissent (on other grounds) elaborated on this same point:

Despite the unqualified statutory declaration that unreasonable rates are unlawful, we think it clear that Congress did not intend either court or Commission to have the power to award reparations on the ground that a properly filed rate or charge has in fact been unreasonably high or low.

341 U.S. at 258, 71 S.Ct. at 698. Similarly, in Federal Power Commission v. Sierra Pacific Power Co., 350 U.S. 348, 355, 76 S.Ct. 368, 372, 100 L.Ed. 388 (1956), the court stated:

while it may be that the Commission may not normally impose upon a public utility a rate which would produce less than a fair return, it does not follow that the public utility may not itself agree by contract to a rate affording less than a fair return or that, if it does so, it is entitled to be relieved of its improvident bargain.

*144Finally, this Circuit stated in Indiana & Michigan, supra, that the “Supreme Court has interpreted [the Federal Power Act] as precluding the Commission from ordering refunds, even when it determines that the rates in effect are unjust and unreasonable.” 162 U.S.App.D.C. at 343, 502 F.2d at 345.

The rule against retroactive rate-making is forcefully stated in cases involving the Natural Gas Act. The Supreme Court wrote in Federal Power Commission v. Tennessee Gas Co., 371 U.S. 145, 152-53, 83 S.Ct. 211, 215, 9 L.Ed.2d 199 (1962):

[A] rate for one class or zone of customers may be found by the Commission to be too low, but the company cannot recoup its losses by making retroactive the higher rate subsequently allowed . . . The company having initially filed the rates and . . . failed to collect a sufficient one must, under the theory of the Act, shoulder the hazards incident to its action including ... its losses where its filed rate is found to be inadequate.

Accord: Atlantic Ref. Co. v. Public Serv. Comm’n, 360 U.S. 378, 389, 79 S.Ct. 1246, 3 L.Ed.2d 1312 (1959), Federal Power Comm’n v. Hope Natural Gas, 320 U.S. 591, 618, 64 S.Ct. 281, 88 L.Ed. 333 (1944), Gillring Oil Co. v. Federal Energy Regulatory Comm ’n, 566 F.2d 1323, 1325 (5th Cir. 1978), and State Corporation Comm’n of Kansas v. Federal Power Comm’n, 215 F.2d 176, 184 (8th Cir. 1954). In summary, cases under both the Federal Power Act (which applies here) and the Natural Gas Act establish the principle that only prospective ratemaking is allowed. A rate designed to recoup past losses is retroactive and illegal.

Basically, petitioners do not contest the validity of the rule against retroactive rate-making. The force of their argument is, as Appalachian Power stated, that the “proposed surcharge is not a reparations award.”52 But to a certain extent, petitioners pick up on the theme enunciated in the First Circuit’s Maine Public Service opinion: that the “Commission has more latitude in shaping a lawful rate than it has recognized.”53

The First Circuit’s reasoning appears to be that while retroactive ratemaking is generally forbidden, in light of the unique circumstances involved in fuel adjustment clause cases, the rule should not be applied inflexibly.54 Judge Campbell explained:

the validity of the surcharge presents a unique policy question for the Commission rather than a black or white legal issue. The surcharge falls somewhere between a species of the forbidden “retroactive rate-making” and an acceptable adaptation of the earlier fuel clause to special conditions. .

The difficulty in classification stems not only from the inflationary rise in fuel costs that triggered the problem, but from the fact that fuel cost adjustment clauses are themselves unique animals that are not easily assimilated to classical rate-making principles. We believe that the Commission erred insofar as it regarded itself as automatically precluded by prior judicial decision such as Montana-Dakota or Hope Natural Gas from approving the surcharge.

579 F.2d at 668. Thus, the First Circuit held that the Commission erred when it rejected the surcharge proposal solely on the basis of the rule against retroactive ratemaking. In so ruling the Court stated that the Commission should have paid more attention to whether the surcharge was “just and reasonable,” and it ordered a remand on the reasonableness issue.

The First Circuit offered four reasons for its conclusion that the rule against retroactive ratemaking was not controlling in cases involving fuel adjustment clauses. First, *145“fuel cost adjustment clauses are . unique animals”55 designed “ ‘to make utilities whole,’ and the surcharge can now be defended as simply implementing the same policy.”56 Second, a Commission regulation permits “deviation from the prescribed operation of fuel adjustment clauses . for good cause shown . . . .”57 Third, “the Commission does allow in other circumstances for ‘after the fact matching of actual costs and revenues,’ . . . .”58 Finally, “the Commission . . . has permitted a type of refund outside of the ordinary rate-suspension context.”59 We are not persuaded by any of these arguments.

1. “Unique Animals.” The purpose of fuel adjustment clauses is to enable utilities to recover their fuel costs without constantly having to file amendments to their tariffs. According to the Commission, “[t]he philosophy underlying the design of fuel adjustment clauses is to increase or decrease the charge to the customer by an amount reflecting the actual increase or decrease in the cost of fuel.” New England Power Co., 48 F.P.C. 899, 921 (1972). Following the lead of the First Circuit in Maine Public Service, supra, petitioners argue that since fuel adjustment clauses are “unique animals” intended to ensure accurate recovery of fuel costs, the goal of accurate recovery should take precedence over the rule against retroactive ratemaking. Therefore they maintain that reasonable retroactive rate increases, which merely permit a utility to recover fuel costs missed by the prior filed rate, should be allowed.60

If we were writing on a clean slate this argument would have some attraction. But there is nothing new about the goal that utility rates should accurately reflect costs and ensure a reasonable return on investment. And despite this goal, courts have consistently held that while the Commission can not require a utility to adopt an unfavorable rate, it has no authority to retroactively allow an increase when the utility has voluntarily adopted a non-compensatory rate.61 Neither petitioners nor the First Circuit cite any cases to the contrary.

In addition, although fuel adjustment clauses are intended to enable utilities to recover the actual cost of fuel, the particular design of a clause is left to the utility. The Commission permits both fixed rate and cost of service tariffs. The goal of accurate recovery of actual fuel costs is best met by cost of service tariffs, but petitioners opted instead for fixed rate formulas.62 Having made that choice, they now argue that they should be exempt from the rule against retroactive ratemaking because the rate they chose did not recover their costs. Perhaps the Commission should, in the interest of accurate recovery, permit only cost of service tariffs. The fact is, however, that fixed rate tariffs are permitted, and petitioners chose to use them. We simply can not accept the argument that petitioners’ improvident choice entitles them to an exemption from the well established rule against retroactive ratemaking.

2. Commission Regulation. The First Circuit implies that 18 C.F.R. § 35.-14(a)(10) (1974) permits retroactive rate-making. Section 35.14(a) delineates filing requirements for fuel adjustment clauses. Section 35.14(a)(10) provides that:

Whenever particular circumstances prevent the use of the standards provided for herein, or the use thereof would result in an undue burden, the Commission may, upon application under § 1.7(b) of the rules of practice and procedure and for good cause shown, permit deviation from these regulations.

*146We do not understand this provision as allowing retroactive ratemaking. Since nothing in the regulation relates to the rule against retroactive ratemaking, it is hard to see how allowing a deviation from its filing requirements could have any substantive impact on the rule. Moreover, the Supreme Court has interpreted the Federal Power Act as prohibiting retroactive ratemaking, and we do not believe that the Commission in its regulations had either the intention or the authority to exceed this statutory limitation. Therefore, section 35.14(a)(10) merely permits deviation from the filing requirements for prospective rather than retroactive rate changes.

3. After the Fact Matching. Petitioners maintain that Commission approval of after the fact matching in other areas belies the agency’s claim that retroactive ratemaking is uniformly prohibited. Petitioners cite two examples of authorized after the fact matching: purchased gas cost adjustment (PGA) clauses and research demonstration and development (RD&D) expense adjustment clauses.63 PGA clauses permit utilities to “recover or return” the difference between actual fuel costs and the amount recovered by the fuel charge.64 RD&D expense adjustment clauses recover prior research and development expenditures.65 According to the First Circuit

Though these adjustment provisions differ from fuel cost adjustment clauses in several respects their retrospective nature is similar to that of the adjustment clause, and their existence indicates that the Commission may not be so devoid of the power to match past costs and present revenues as it here maintains.

Maine Public Service, supra, 579 F.2d at 668 n.13.66

The existence of PGA and RD&D clauses conclusively establishes that the Commission permits after the fact matching. The First Circuit errs, however, by equating after the fact matching with retroactive ratemaking. All cost of service clauses involve after the fact matching of costs and charges, and there is no question that the Commission can, under the Federal Power Act, approve cost of service clauses. The Commission’s approval of both Public Service’s and Appalachian Power’s new cost of service fuel adjustment clauses is an example of this. Both clauses allow the utility to conform the fuel adjustment charge to fuel costs after those costs have been incurred. They are not retroactive rate increases, however, because both clauses were prospectively approved.

The question before us is not whether cost of service clauses with their after the fact matching can be approved. Rather, it is whether the Commission may retroactively change a fixed rate clause to a cost of service clause. The prospective application of the previously approved cost of service PGA and RD&D clauses can not support an affirmative answer to this question.67 Therefore we reject the argument that *147PGA and RD&D clauses are precedent for permitting retroactive rate increases.

4. Extra-ordinary Refunds. In Indiana & Michigan Electric Co. v. Federal Power Commission, 162 U.S.App.D.C. 334, 341 — 42, 502 F.2d 336, 343-44 (1974), cert. denied, 420 U.S. 946, 95 S.Ct. 1326, 43 L.Ed.2d 424 (1975), this court held that the Commission unlawfully “delayed the effective date of I&M’s rate filings . . The court ordered “I&M’s customers ... to pay retroactive rate increases for the five-month period during which the Commission’s unlawful suspension was in effect.” The First Circuit states that this is an example in which the rule against retroactive ratemaking was ignored.

The Commission persuasively responds that its power to correct its own mistakes does not establish the existence of a power to remedy insufficient rate filings by utilities. The Supreme Court specifically addressed the difference between these two powers. Justice Douglas wrote for the Court in United Gas v. Callery Properties, 382 U.S. 223, 229-30, 86 S.Ct. 360, 364, 15 L.Ed.2d 284 (1965):

We also conclude that the Commission’s refund order was allowable. We reject, as did the Court of Appeals below, the suggestion that the Commission lacked authority to order any refund. While the Commission “has no power to make reparation orders,” Federal Power Comm’n v. Hope Natural Gas Co., 320 U.S. 591, 618 [64 S.Ct. 281, 295, 88 L.Ed. 333], its power to fix rates under § 5 being prospective only, Atlantic Refining Co. v. Public Service Comm’n, supra, [360 U.S.] at 389 [79 5.Ct. 1246, 1254], it is not so restricted where its order, which never became final, has been overturned by a reviewing court. Here the original certificate orders were subject to judicial review; and judicial review at times results in the return of benefits received under the upset administrative order. See Securities & Exchange Comm’n v. Chenery Corp., 332 U.S. 194, 200-201 [67 S.Ct. 1575, 1579-1580, 91 L.Ed. 1995]. An agency, like a court, can undo what is wrongfully done by virtue of its order. Under these circumstances, the Commission could properly conclude that the public interest required the producers to make refunds for the period in which they sold their gas at prices exceeding those properly determined to be in the public interest.

Thus, as United Gas and Indiana & Michigan indicate, the Commission may “undo what is wrongfully done by virtue of its order.”68 But the Commission may not permit a utility to retroactively increase its filed rate.69

In summary, then, we are not persuaded by any of the reasons cited by the First Circuit for its conclusion that the “Commission has more latitude in shaping a lawful rate than it has recognized.”70 On the contrary, precedent establishes that the Commission is prohibited under the Federal Power Act from approving retroactive rate increases.

In summary, we affirm the Commission orders involving Public Service, Pennsylvania Electric and Metropolitan Edison. The Commission correctly determined that their proposed surcharges are illegal retroactive *148rate increases. We remand the orders dealing with Appalachian Power Company. The facts that established that the superseded fuel adjustment clauses used by the other petitioners were fixed rate tariffs have not been conclusively proven with respect to Appalachian Power. Therefore, on this record we can not hold that Appalachian Power’s proposed surcharges are retroactive rate increases.

Judgment Accordingly.

APPENDIX

EXHIBIT B

5th Revision

February 1, 1975

PUBLIC SERVICE COMPANY OF NEW HAMPSHIRE

RATE SCHEDULE FOR TRANSMISSION SERVICE AT WHOLESALE FOR RESALE

MONTHLY RATE APPLICABLE TO SALES AT EACH DELIVERY POINT

Demand Charge: $2.95 per kilovolt-ampere of maximum demand.

Energy Charge: 0.73<t per kilowatt-hour.

Fuel adjustment: There shall be an upward or downward adjustment applied to all energy billed under the resale service rate based upon the difference between (i) the actual costs of fossil fuels burned in generating stations owned by the SELLER and in other generating stations from which the SELLER purchases capacity and energy under agreements to enable it to fulfill its power requirements, and (ii) the fixed base costs of such fuels computed at a cost per fossil fuel generated kilowatt-hour of $0.00457045. The charge or credit resulting from this difference shall be adjusted to reflect energy losses incurred in the delivery or resale service sales. The adjusted charge or credit shall be applied to kilowatt-hours billed pursuant to meter readings taken in any calendar month and shall be in an amount determined monthly based upon costs and quantities of fuels burned and all other required data taken from or applicable to the second preceding month. The adjustment shall be computed to the nearest one-thousandth cent ($0.00001) per kilowatt-hour.

Minimum Charge: The demand charge but not less than $200.00 per month.

Low Voltage Delivery Charge: For any delivery point on the low voltage side of a distribution substation of the SELLER, the demand charge shall be increased by $0.25 per month per kilovolt-ampere of maximum demand, and the maximum demand and kilowatt-hour energy use recorded at the delivery voltage shall each be increased by 2.5 percent. Such delivery at low voltage may be terminated by the SELLER at any time by giving the PURCHASER not less than one (1) year’s written notice specifying a date for termination.

Pennsylvania Electric Company

FPC Electric Tariff

Original Volume No. 1.

Revised Sheet No. 15

Superseding Second Revised Sheet No. 15

Revised — 12/17/73

RESALE POWER SERVICE FOSSIL FUEL COST ADJUSTMENT CLAUSE

A fossil fuel cost adjustment factor shall be applied to each kilowatt-hour supplied under this tariff. This fuel cost adjustment determined to the nearest one-thousandth of 1 mill per kilowatt-hour in accordance with the formula set forth below, shall be applied to all kilowatt-hours supplied during the billing month.

(Fc Fb) Gc 1 (Gc Gb) X Sc X1 - T

Where A = Adjustment factor in mills per kilowatt-hour to be applied to each kilowatt-hour supplied under this tariff.

F = The cost of fossil fuel determined as the sum of Accounts 501 and 547 (less handling costs) in the current (c) and base (b) periods.

*149G = Net kWh generation by fossil fuels in the current (c) and base (b) periods.

S = Sales shall be all kWh’s sold excluding intersystem sales in the current (c) period.

Fb = Base fossil fuel cost (less handling Gb costs) of 4.254 mills per kilowatt-hour.

T = The applicable Pennsylvania gross receipts tax.

The “Fc”, “Gc” and “Sc” factors are to be determined as the three month totals for the period ending with the second calendar month preceding the billing month.

The “T” factor is to be determined for the billing month.

Minimum bills shall not be reduced by reason of this fuel cost adjustment clause. This clause shall be applied to all kilowatt-hours supplied, and such charge shall be an addition to any mínimums applicable.

Metropolitan Edison Company

FPC Electric Tariff

Original Volume No. 1

Second Revised Sheet No. 17

(Superseding First Revised Sheet No. 17)

And Orig. Sh. No. 17

5-8-75

RESALE POWER SERVICE FOSSIL FUEL COST ADJUSTMENT CLAUSE

(Fc Fb) Gc 1

A=---— x — x-xL

(Gc Gb) Sc 1 —T

Where A = Adjustment factor in mills per kilowatt-hour to be applied to each kilowatt-hour supplied under this tariff.

F = The cost of fossil fuel determined as the sum of Accounts 501 and 547 (less handling costs) in the current (c) and base (b) periods.

G = Net kWh generation by fossil fuels in the current (c) and base (b) periods.

S = Sales shall be all kWh’s sold excluding intersystem sales in the current (c) period.

Fb = Base fossil fuel cost (less handling Gb costs) of 6.035 mills per kilowatt-hour.

T = The applicable Pennsylvania gross receipts tax.

L = A factor for adjustment of losses to the delivery voltage.

The “Fc”, “Gc” and “Sc” factors are to be determined as the three month totals for the period ending with the second calendar month preceding the billing month.

The “T” factor is to be determined for the billing month.

The “L” factor is 0.94 for delivery at transmission voltages and is 0.97 for delivery at distribution voltages.

Effective May 10, 1974

APPALACHIAN FUEL CLAUSE FOR SERVICE TO VIRGINIA AND WEST VIRGINIA WHOLESALE CUSTOMERS AS FILED AND IN EFFECT WHEN THE $1,587,000 OF UN-BILLED FUEL EXPENSE WAS INCURRED

No. 77-2004

Fuel Clause

When the unit cost of fuel (Fm/Sm) used to meet Appalachian Power Company’s Net Energy Requirement less losses (Sm) is above or below the base unit cost of 3.2710 mills per kiloWatthour (Fb/Sb), an additional charge or credit equal to the product of the actual kiloWatthours used and a fuel clause adjustment factor (A) shall be made, where “A”, calculated to the nearest 0.0001 mill per kiloWatthour, is as defined below:

*150Adjustment Factor (A) = Fm — Fb Sm Sb

In the above formula “F” is the expense of fossil and nuclear fuel in the base (b) and current (m) periods; and “S” is the kWh sales in the base and current periods, all as defined below:

Fuel Costs (F) shall be the cost of:

(a) fossil and nuclear fuel consumed in Appalachian Power Company’s plants, and Appalachian Power Company’s share of fossil and nuclear fuel consumed in jointly owned or leased plants;

(b) the actual identifiable fossil and nuclear fuel costs associated with energy purchased for reasons other than identified in (c) below;

(c) the net energy cost of energy purchases, exclusive of capacity or demand charges (irrespective of the designation assigned to such transaction) when such energy is purchased on an economic dispatch basis (including therein shall be such costs as the charges for economy energy purchases and the charges as a result of scheduled outage, (all such kinds of energy being purchased by Appalachian Power Company to substitute for its own higher cost energy); and less

(d) the cost of fossil and nuclear fuel recovered through inter-system sales including the fuel costs related to economy energy sales and other energy sold on an economic dispatch basis.

Sales (S) shall be equated to the sum of (a) generation, (b) purchases, (c) interchange-in, less (d) energy associated with pumped storage operations, less (c) inter-system sales referred to in (d) above less (f) total system losses.

Sales (S) shall be modified to reflect losses of 4.67% associated with Appalachian Power Company’s deliveries to customers served under this schedule.

The adjustment factor developed according to the preceding paragraphs may be further modified to allow the recovery of gross receipts or other similar revenue based tax charges occasioned by the fuel adjustment revenues.

The cost of fossil fuel shall include no items other than those listed in Account 151 of the Commission’s Uniform System of Accounts for Public Utilities and Licensees. The cost of nuclear fuel shall be that as shown in Account 518, except that if Account 518 also contains any expense for fossil fuel which has already been included in the cost of fossil fuel, it shall be deducted from this account. All reference to the Commission’s Uniform System of Accounts for Public Utilities and Licensees shall be to such Uniform System of Accounts for Public Utilities and Licensees as is in effect as of December 1, 1975.

APPALACHIAN FUEL CLAUSE FOR SERVICE TO KINGSPORT POWER COMPANY AS FILED AND IN EFFECT WHEN THE $736,000 OF UN-BILLED FUEL EXPENSE WAS INCURRED

9.02 The energy rates in the rate schedule specified under Section 8.01 above are based upon a Weighted Average Fuel Cost of eight cents ($0.08) per million Btu. In the event the Weighted Average Fuel Cost for any month exceeds eight cents ($0.08) per million Btu by at least one mill, said energy rates for the next succeeding month shall be increased by an adjustment factor of 0.0125 mills for each such one mill of such excess Weighted Average Fuel Cost. In the event the Weighted Average Fuel Cost for any month is less than eight cents ($0.08) per million Btu by at least one mill, said energy rates for the next succeeding month shall be decreased by 0.0125 mills for each such one mill that the Weighted Average Fuel Cost is less than eight cents ($0.08) per million Btu; provided, however, that said adjustment factor of 0.0125 mills shall be subject to adjustment as hereinafter provided.

. Public Service Company of New Hampshire (Public Service), Pennsylvania Electric Company, Metropolitan Edison Company, and Appalachian Power Company.

. None of the orders has been officially reported. The Appendix to the Commission’s brief contains the relevant orders. See Comm’n Br., App. 49-84.

. Maine Public Service Co. v. Federal Power Comm'n, 579 F.2d 659 (1st Cir. 1978).

. Jersey Central Power & Light Co. v. Federal Energy Regulatory Comm’n, 589 F.2d 142 (3d Cir. 1978).

. Virginia Electric and Power Co. v. Federal Energy Regulatory Comm'n, 580 F.2d 710 (4th Cir. 1978).

. Public Service Co. of New Hampshire, No. 790, 3 (March 21, 1977) (hereinafter Public Service No. 790) in Comm’n Br., App. 53.

. Public Service’s fuel adjustment charge was based on fuel costs from two months preceding the billing month. Public Service No. 790, supra at n.6, App. 62. The clause used by Pennsylvania Electric and Metropolitan Edison focused on “the three month totals for the period ending with the second calendar month preceding the billing month.” “Resale Power Service Fossil Fuel Cost Adjustment Clause” Comm’n Br., App. 49 (Pennsylvania Electric), 50 (Metropolitan Edison) [hereinafter “Tariff of Pennsylvania Electric/Metropolitan Edison”]. Appalachian Power used two different fuel adjustment clauses. For most of its customers, the charge was based on the second preceding month. But the clause applicable to its affiliate the Kingsport Power Company was based on the cost of fuel in the preceding month. Comm’n Br. 18, App. Power Br. 8. Each of these fuel adjustment clauses is contained in an appendix to this opinion.

.The new regulation required utilities to include nuclear fuel costs in the base cost of fuel. “Order [No. 517] Amending Section 35.14 of the Regulations under the Federal Power Act,” 52 F.P.C. 1304 (1974), 18 C.F.R. 35.14 (1977). Under the old regulations, nuclear fuel costs were not included in the “base cost” and therefore were not recovered by the “basic energy charge.” The requirement that nuclear fuel be figured into the base cost resulted in some recovery of nuclear fuel costs by the basic energy charge. Of course, the difference between the actual cost of nuclear fuel and the “base cost” would still be reflected in the fuel adjustment charge.

. Rate Schedule of Public Service, Comm’n Br., App. 36 (see Appendix, infra).

. The January charge would not recover the exact difference between the base cost and the actual cost of fuel in November because the quantity of fuel used differs from month to month, and the fuel adjustment factor was applied to kilowatt-hours used in the billing month. See discussion at III.B., infra.

. Pennsylvania Electric, Metropolitan Edison, and Appalachian Power (in its tariff to its affiliate Kingsport Power) did not eliminate the billing lag in their fuel adjustment clauses. Despite this, they contend that the new Commission regulation requiring utilities to include the cost of nuclear fuel in the “base cost” made the “deferred charges” for nuclear fuel uncollectable. See, e.g., Testimony of M. E. McCrary, Appendix of Appalachian Power A 64- 66.

The Commission disagrees. It argues that since the “billing lag” was not eliminated, the utilities “would not be precluded by the raised base fuel cost from reflecting those past costs in future charges.” Comm’n Br. 46.

We are skeptical of petitioners’ argument on this point. For the purposes of this opinion, however, we will assume arguendo that inclusion of nuclear fuel costs in the “base cost” will cause some “deferred charges” to become unrecoverable.

. The Commission stated:

It is clear that the [superseded] fuel clause which became effective in January, 1973, was never intended to permit PSNH to charge its January customers for fuel costs incurred in November, 1972. The intent was to use the November experience, the most recently available data, as a measure of the January fuel expense which would be recoverable from customers using power in January.

During the thirty-six months the superseded fuel clause was in effect each customer was billed monthly based on the rate determined by the fuel adjustment formula then in effect. After a customer had paid that bill the company had received everything it was entitled to receive under the then existing rate schedule. The fact that this rate schedule may not have adequately compensated the company for its costs incurred at that time does not permit us to retroactively increase that rate. It is well established that the Federal Power Commission has no authority to order reparations and can only set rates for the future.

Public Service No. 790, supra n.6, at 12, Comm’n Br., App. 62.

.Counsel for Appalachian Power state in their brief: “The earlier fuel clauses recovered actual fuel expense incurred in any one month, but in a later billing month, whereas the updated fuel clauses recover actual fuel expense on a more current basis." Appalachian Power Br. 7. Similarly, counsel for Pennsylvania Electric wrote: “Penelec was expecting to recover the actual cost of fuel several months after those costs were incurred. . . . ” Pennsylvania Electric/Metropolitan Edison Br. 32.

. Jersey Central, supra at n.4, at 146.

. This conclusion does not apply to the superseded clauses of Appalachian Power Co. See III.C., infra.

. Petitioners’ briefs focus on this point. Public Service’s entire argument is as follows:

[Public Service] vigorously denies that the record demonstrates that it had any intent to use the second preceding month’s fuel costs in a proxy relationship. For evidence of such a relationship, one would look in the first instance to the language of the fuel clause itself. There is nothing in that language to imply such a relationship.

Public Service Br. 17. Similarly, counsel for Pennsylvania Electric and Metropolitan Edison states:

The key language in Penelec’s old fuel clause is the definition of fuel costs and sales for purposes of calculating the change from the base cost: “The three-month totals for the period ending with the second calendar month preceding the billing month.” . . .

Absolutely no language in this clause even hints that the past three-month period constitutes an “estimate” or “proxy” of what current month fuel costs will be. Had an “estimate” or “proxy” been intended, the clause would have so stated.

Pennsylvania Electric/Metropolitan Edison Br. 30. Thus, while petitioners’ counsel claim that the language of the superseded clauses supports their position, petitioners only argue that this language is not explicitly contrary to their view.

. Public Service Br. 18, quoting from order of New Hampshire Public Utilities Commission.

. Public Service Br. 18.

. Pennsylvania Electric/Metropolitan Edison Br. 31. Appalachian Power also uses deferred fuel accounting. Testimony of M. E. McCrary, Appalachian Power App. A-67.

. Public Service Br. 18.

. Public Service Reply Br. 13. Comm’n Br. 38. Similarly, Appalachian Power did not use this accounting method until 1974, long after its superseded fuel clause became effective. Testimony of M. E. McCrary, Appalachian Power App. A-67.

. Public Service Br. 19.

. The companies’ billing practice immediately after the fuel clauses became effective is relevant, and it strongly supports the Commission’s interpretations. See 139-140 of 195 U.S.App.D.C., 953-954 of 600 F.2d supra.

. Cf. Detroit Edison Company v. Michigan Public Service Comm’n and Attorney General, 82 Mich.App. 59, 266 N.W.2d 665 (1978), at Comm’n Br., App. 44. See Comm’n Br. 39.

. We note also that the Securities and Exchange Commission has refused to unequivocally endorse deferred fuel accounting because “uncertainties exist in some cases as to whether public utility commissions will permit the recovery of these deferred costs at a time when full new rate schedules are adopted.” Series Release No. 166, Dec. 23, 1974, 20 SEC Docket 773 (Jan. 6, 1975), quoted in Intervenor New Hampshire Electric Cooperative Br. 14.

. 18 CFR § 35.14(a)(1) (1977).

. Pennsylvania Electric/Metropolitan Edison Br. 30.

. See discussion at 143-144, 145 of 195 U.S.App.D.C., at 957-958, 959 of 600 F.2d, infra.

. For example, Public Service’s superseded clause was based on the costs in the “second preceding month,” e. g., November fuel costs were reflected in the January fuel adjustment charge. If the clause were intended to defer recovery of November’s costs until January, then the January charge would reflect both (1) November costs and (2) November power consumption. See Appendix, A-l.

.Tariff of Pennsylvania Electric/Metropolitan Edison, supra n.7, Appendix, 148, 149 of 195 U.S.App.D.C., 962, 963 of 600 F.2d; Rate Schedule of Public Service, Comm’n Br., App. 36, Appendix 148 of 195 U.S.App.D.C., 962 of 600 F.2d. But see discussion of fuel adjustment clauses used by Appalachian Power, infra at III.C.

.For example, the Commission states in its brief:

[Public Service’s] own interpretation of its fuel clause would result in substantial imprecision in fuel cost recovery. Because past fuel costs are applied to current kilowatthours under the clause ., there is a mismatch of fuel costs and kilowatthour usage. As a result, the record shows that [Public Service] would have overrecovered fuel costs by as much as $288,565 for two months even under its present theory of its prior fuel clause.

Comm’n Br. 42. Public Service’s comptroller, Elroy Littlefield, acknowledged that this mismatch produces imprecise recovery of costs. JA 26-27.

. Public Service No. 790 12, at Comm’n Br., App. 62 (Public Service); Pennsylvania Electric Co., Docket No. ER76-607, at 12 (December 19, 1977), at Comm’n Br., App. 82 (Pennsylvania Electric and Metropolitan Edison). There is no record on Appalachian Power’s billing practice.

. Public Service No. 790 12, at Comm’n Br., App. 62.

. Public Service Br. 20.

. Essentially, petitioners’ argument is that the fuel adjustment bill received by a Public Service customer in March is actually the fuel adjustment charge for January. Billing of the January charge was deferred for two months. If this were really Public Service’s intent, then it is obvious that customers should not have been billed for a fuel adjustment charge in January, which was the first month the superseded clause was in effect.

. Tariff of Pennsylvania Electric/Metropolitan Edison, supra n.7, Appendix, 148, 149 of 195 U.S.App.D.C., 962, 963 of 600 F.2d.

. The First Circuit reversed the Commission’s determination that a surcharge was impermissible retroactive ratemaking. Maine Public Service, supra. Apparently, one reason for that decision was the court’s belief that whether the surcharge was a retroactive rate increase was a close question. Judge Campbell wrote:

To be sure it can be argued, as does the Commission, that the surcharge is merely a belated attempt to collect in a later year sums that were overlooked by the earlier fuel clause. But it can also be argued with some force that the prior fuel formula anticipated that customers would eventually have to pay the very sum which the surcharge has now assessed in a different form. . . . [I]t may be contended that the past formula did not set a rate that was too low. The trouble was more one of deferred collection, a question of timing ....

579 F.2d at 667.

We disagree with this novel approach. Commission decisions are not subject to reversal merely because they decide close questions. On the contrary, as was indicated above, the Commission’s interpretations are entitled to a considerable amount of deference.

. Cited at n.6, supra.

. Appalachian Power Co., Docket No. ER77325 (June 30, 1977) at Appendix of Appalachian Power A-18 (municipal customers); Appalachian Power Co., Docket No. ER77-426 (June 30, 1977), at Appendix of Appalachian Power, A-72 (Kingsport affiliate).

. See n.30. The language in Public Service’s superseded clause (Comm’n Br., App. 36) is not clear on this point, but the record establishes that there was a mismatch in Public Service’s billing. JA 26-27.

. “Appalachian Fuel Clause for Service to Virginia and West Virginia Customers,” Appalachian Power Br., App. 2, see Appendix 149 of 195 U.S.App.D.C., 963 of 600 F.2d, infra.

. Appalachian Power Br. 8.

. Testimony of Ronald H. Hively, No. 77-426, Appalachian Power App. 53.

. Testimony of M. E. McCrary, No. ER77-426, Appalachian Power App. 63, Table A, column D.

. The other Appalachian Power fuel clause involved here, which was part of the rates to its Kingsport Power affiliate, does not appear to be capable of exact recovery of past costs. But in view of the direct testimony by one of Appalachian Power’s witnesses to the contrary, we believe that a fact issue is raised here as well. See Appalachian Power App. 63.

. See III.B., supra.

. This statement does not apply to the fuel adjustment clauses that were used by Appalachian Power. See III.C., supra.

. See Maine Public Service, supra at n.3.

. See cases cited at n.4 & 5, supra.

. Nader involved a proposed increase in telephone rates. The rule was also stated in Payne v. Washington Metropolitan Area Transit Comm’n, 134 U.S.App.D.C. 321; 330-31, 415 F.2d 901, 910-11 (1968), which involved bus fares. The Federal Power Act gives the Commission authority to review.

. Though we have no doubt about the applicability of this rule to ratemaking under both statutes, the case for applying the rule to rate-making subject to the Federal Power Act is even stronger than the case for applying it to ratemaking under the Natural Gas Act. Congress specifically eliminated a section of the Federal Power Act that would have authorized “the issuance of reparation orders.” S.Rep. No. 621, 74th Cong., 1st Sess. 20 (1935). Cases applying the rule against retroactive ratemaking in the context of the Natural Gas Act have not referred to the legislative history of that statute.

. Appalachian Power Reply Br. 10; Pennsylvania Electric/Metropolitan Edison Reply Br. 6.

. 579 F.2d at 669.

.The First Circuit also stated that the surcharge might not be a retroactive rate increase. 579 F.2d at 667. The argument is discussed supra at n.37.

. 579 F.2d at 668 (see block quote in the preceding paragraph).

. 579 F.2d at 667-68.

. 579 F.2d at 668.

. Id. at n.13.

. Id. at 668.

. Public Service Reply Br. 4-8; Pennsylvania Electric Reply Br. 18-20.

. See cases cited in text at IV.A., supra.

. See III.B., supra.

. Petitioners also point to the surcharge permitted in Commission Opinion 699-G, 52 F.P.C. 1551 (Nov. 29, 1974). We agree with the Commission, however, that the surcharge permitted in Opinion 699-G was not a retroactive rate increase. According to one of the opinions authorizing such a surcharge, it merely allows a pipeline company to “recover producer increases resulting from Opinion 699 more expeditiously than would be the case under the normal operation of its PGA clause. . United Gas Pipe Line Co., 52 F.P.C. 1578, 1580 (Nov. 29, 1974) (emphasis added).

. PGA clauses are authorized in 18 C.F.R. § 154.38(d)(iv)(a) & (b), which apply to natural gas companies.

. 18 C.F.R. § 35.22(e)(3); 42 Fed.Reg. 30156 (June 13, 1977).

. Only PGA clauses were mentioned in Maine Public Service. This argument, however, is equally applicable to RD&D clauses.

. Petitioners argue that PGA and RD&D clauses and the proposed surcharges involved here are similar in that each is designed to match costs and services. Pennsylvania Electric/Metropolitan Edison Reply Br. 9-10. This is true. The difference between PGA and RD&D clauses on the one hand, and the proposed surcharges on the other, is that the former prospectively provided for after-the-fact matching, whereas imposition of the latter (the surcharge) would be a retroactive shift from a “fixed rate” to a “cost of service” tariff.

. 382 U.S. at 229, 86 S.Ct. at 364.

. It is somewhat ironic that Indiana & Michigan is cited for the proposition that retroactive rate increases are permissible. After rehearing, the court modified its original order and provided that

1. I&M may not collect retroactive rate increases for the five-month period from the effective date of its rate filing [during which the rate was unlawfully suspended],

162 U.S.App.D.C. at 342, 502 F.2d at 344. The court adhered to its view that the rate had been

illegally suspended. Nevertheless, the court withdrew its order permitting retroactive rate increases because such increases “would create considerable hardship for I&M’s customers [who] . . are precluded by federal law from retroactively raising their rates ..” Id. Indiana & Michigan, then, turns on the rule against retroactive ratemaking — the very doctrine that the First Circuit claims it undercuts.

.579 F.2d at 669.

Public Service Co. v. Federal Energy Regulatory Commission

Related Cases