(concurring):
I reluctantly concur in the remand called for by our per curiam opinion. This is the first marginal cost pricing rate case to come before us, and it will serve as a guide for further applications of marginal cost pricing to electricity rates. Our remand will undoubtedly allow FERC to supplement its record. But in my opinion, we could just as easily have affirmed the Commission. It has articulated a rational basis for its decision.
*405This case is our first in what will prove to be many applying the principles of marginal cost pricing theory to electricity rate design. As such, it must be expected that the development of a record traditionally thought sufficient to satisfy our review will be difficult. Without the benefit of precedent, it will also prove difficult for the Commission to reconcile the theory to every nuance of actual practice. In my view, we should be more sensitive in our review to the novelty of the proposed design.
The Supreme Court recognized in Permian Basin-Area Rate Cases, 390 U.S. 747, 88 S.Ct. 1344, 20 L.Ed.2d 312 (1968), that weight should be accorded the unusual difficulties of the first proceeding involving a new design. In Permian Basin, the Federal Power Commission was attempting to implement an entirely new concept in determining maximum just and reasonable rates for sales in interstate commerce of natural gas produced in the Permian Basin. The Commission believed that the particular rate structure it adopted would employ price functionally, “as a tool to encourage the production of appropriate supplies of natural gas,” and as an incentive to exploration and production. Id. at 796, 88 S.Ct. at 1375. The Supreme Court found that the rate structure fell within the “zone of reasonableness,” even though price distinctions were unrelated to quality, on the basis that such price discrimination could be employed “to achieve relevant regulatory purposes.” Id. at 797, 88 S.Ct. at 1375.
The Court also examined the method by which the Permian Basin rates were derived. The Commission had adopted prices based on cost factors which it found to serve as an incentive to exploration and production. The court determined that, while it may not produce arbitrary or unreasonable results, “the Commission may employ any ‘formula or combination of formulas’ it wishes, and is free to ‘make the pragmatic adjustments which may be called for by particular circumstances.’ ” Id. at 800, 88 S.Ct. at 1377 (quoting FPC v. Natural Gas Pipeline Co., 315 U.S. 575, 586, 62 S.Ct. 736, 743, 86 L.Ed. 1037 (1942)).
The Court in Permian Basin was acutely aware that it was travelling uncharted waters. See, e.g., id., 390 U.S., at 788, 810-11, 88 S.Ct. at 1382-83. Indeed, even a member of the Commission described its efforts as “admittedly ... experimental.” Id. at 758, 88 S.Ct. at 1355. The Court also recognized that “the breadth and complexity of the Commission’s responsibilities demand that it be given every reasonable opportunity to formulate methods of regulation appropriate for the solution of its intensely practical difficulties.” Id. at 790, 88 S.Ct. at 1372. Thus, in conducting its review, the Court candidly admitted that it was “obliged at this juncture to give weight to the unusual difficulties of this first area [rate] proceeding.” Id. at 792, 88 S.Ct. at 1373. The Court emphasized, however, “that this weight must significantly lessen as the Commission’s experience with area regulation lengthens.” Id. The Court’s review thus appropriately accommodated the uncertainties inherent in the first case.
It is my view that our per curiam opinion ignores the difficulties under which the agency labors. In our zeal to have FERC establish a more sufficient record for affirmance, we require an overly stringent exactness in fitting the theory to the facts of this case. I oppose imposing such a standard of certainty. While pure theory alone is insufficient to support a rate design, our review must be adjusted to accommodate the nature of the decision to be made by the agency. Where theory is so much a part of an initial decision, as here, a reviewing court should not require the same type of evidence as it would in other rate cases.1
*406The per curiam opinion condemns the present “distorted” version of marginal cost pricing because the agency made substantial adjustments to the theory in order to apply it to this case. Because the revenue constraint was derived on an average cost basis, pure theory cannot be implemented. The question then is, what is the most reasonable modification of the theory compatible with the predetermined revenue requirement? According to WEPCO and FERC, the best application calls for implementing pure marginal energy costs and collecting the residue by way of the demand charge. By doing so, the energy charge is calculated according to pure marginal theory. As FERC explains, this is important because such an energy charge would elicit the best response from consumers. J.A. 58. By definition it accurately tracks energy costs. Customers respond proportionately less to changes in the demand charge. Manipulating it therefore diminishes the effectiveness of the marginal cost pricing plan relatively less than would an equivalent change in the energy charge. In the economists’ terminology, the demand for power is more elastic with respect to changes in the energy charge than to changes in the demand charge. In my view, the agency’s application represents a pragmatic adjustment which best achieves the relevant regulatory goals of marginal cost pricing.
The per curiam similarly charges that the modified theory applied here results in a so-called “rate tilt.” In fact, any rate tilt present exists because the overall revenue constraint was derived from average demand and energy costs. Naturally, any design based on marginal cost will be at odds with the revenue constraint. In its effort to retain the benefits of marginal cost pricing, FERC modified its theory. If this modification causes a tilt, it does so only in relation to the average demand and energy charges formerly derived. J.A. 46-48. From the record, it appears that once the theory of marginal cost pricing is found to be satisfactorily supported, certain conclusions inevitably follow. Among these is that marginal cost pricing cannot be reconciled perfectly with the revenue constraint. Accordingly, I cannot agree that FERC has made “absolutely no attempt” to outline its explanation of the so-called rate tilt. This court has recognized that “Congress has carefully eschewed tying ‘just and reasonable’ rates to any particular method of deriving the rates.” American Public Power Assoc. v. FPC, 522 F.2d 142, 146 (D.C.Cir.1975). Rather, “Congress clearly intended to allow the Commission broad discretion in regard to the methodology of testing the reasonableness of rates.” Id.
It is notable that similar proceedings were conducted by the District of Columbia Public Service Commission for large demand users. In approving time-of-day, marginal cost rates, the Commission faced the problem that pure marginal rates would result in exceeding the revenue constraint. The Commission ruled that adjustments should be made to the demand charge:
The Commission is persuaded that, if an adjustment in any of the charges is required because of the revenue constraint, that adjustment should be made to the demand charge and not to the energy charge. There are several reasons for this. In the first place, as people’s counsel witness [] explained, the calculation of marginal energy costs is far more exact than the calculation of marginal demand costs. By adjusting the least exact charge, the commission reduces the chances of designing a rate which fails to send the proper price signal. Secondly, and perhaps more importantly, the ultimate effect of a reduction in kwh consumption will be felt immediately. If a customer reduces its energy consumption on-peak by 100 kwh and the marginal cost has been properly estimated, PEPCO will save exactly that marginal cost in fuel and related operating expenses. This effect contrasts with marginal demand costs when the ultimate *407effect of a saved KW of peak demand may not be experienced for years.
Re Potomac Electric Power Co., 31 P.U.R. 4th 219, 232 (D.C.Commission 1970) (Order No. 7002). The Commission termed this a “pragmatic adjustment.” 31 P.U.R. 4th 249, 254 (D.C.Commission 1979) (Order No. 7034). The District of Columbia Court of Appeals affirmed. Metropolitan Washington Bd. of Trade v. Public Serv. Comm’n, 432 A.2d 343 (D.C.App.1981).
Thus, I reluctantly join in the remand of the record. I am fearful that we require FERC to go to unreasonable lengths to substantiate its decision. There are limits on what an agency can do when approving a rate design that implements a theoretical construct such as marginal cost pricing. Our review should be influenced to a large extent by the nature of the decision confronting the agency. See, e.g., City of Charlottesville v. FERC, 661 F.2d 945, 955 (D.C.Cir.1981) (MacKinnon, J., dissenting in part). The application of marginal cost theory to electricity rates is in its infancy. I doubt seriously that the effects of using a modified form of theory are susceptible to the same degree of proof thought sufficient in other rate making cases. Our per curiam requires an exactness of proof which is unreasonable in this case. I would affirm on the basis of this record, “obliged at this juncture,” as the Supreme Court said in Permian Basin, “to give weight to the unusual difficulties of this first [marginal cost rate design] proceeding.” 390 U.S. at 792, 88 S.Ct. at 1373.
. Sometimes the reason for tolerating a gap either between evidence and findings or between findings and decision has to do with limitations of human intellects or limitations on the magnitude of investigations that may be conducted in particular circumstances.
Not all propositions of fact that are useful and used in the administrative process are susceptible of proof with evidence. Or developing the evidence would be inordinately expensive.
2 K. Davis, Administrative Law Treatise § 16.11, at 473 (1958) (cited in FPC v. Florida Power & *406Light Co., 404 U.S. 453, 465 n. 15, 92 S.Ct. 637, 644 n. 15, 30 L.Ed.2d 600 (1972)).