Opinion for the Court filed by Circuit Judge STARR.
STARR, Circuit Judge.In 1982, Congress adopted a comprehensive statutory response to problems spawned by nuclear wastes generated by nuclear power plants in the production of electricity. These consolidated cases raise a single issue of statutory interpretation under that statute, the Nuclear Waste Poli*129cy Act of 1982, 42 U.S.C. §§ 10101 et seq. (1982). That statute, among other things, establishes a system of fees to be levied on utilities which generate nuclear power in order to finance the establishment and operation of a fund to pay for the transportation, storage and disposal of spent nuclear fuel and high-level radioactive wastes. 42 U.S.C. § 10222 (1982).
The specific question before us is whether a statutorily prescribed fee levied on nuclear generation of electricity applies to ■all nuclear-generated electricity (gross generation) or, instead, only to that nuclear-generated electricity which is sold by utilities, thus excluding electricity that the generating plant itself consumes (net generation). Pursuant to informal rulemaking, the Department of Energy imposed a fee at the statutorily prescribed level (1.0 mil per kilowatt hour) on all nuclear-generated electricity, as opposed to that nuclear-generated electricity which was sold by the utilities. On this distinction between gross generation, embraced by DOE, and net generation, championed by various electric utilities, turns a sizeable sum, estimated by the DOE to be approximately $16 to $20 million for the year 1984 alone.1 Based upon our analysis of the statute and the legislative history, we conclude that the pertinent statutory provision imposes the fees only on net generation of electricity. Accordingly, we grant the utilities’ petition for review.
I
At the outset, we address briefly a threshold jurisdictional question in the wake of a controlling, recent decision by this court. Due to uncertainty over the proper forum for seeking review of DOE’s rule, the utilities filed both a petition for review in this court and a complaint, based on federal question jurisdiction, 28 U.S.C. § 1331 (1982), in the United States District Court for the District of Columbia. The proceedings in the Court of Appeals were stayed by order of this court dated December 27, 1983, pending the outcome of .the federal district court action. On July 18, 1984, the District Court dismissed the complaint for lack of jurisdiction; in a careful memorandum opinion, the trial court concluded that the exclusive avenue for judicial review law in the Court of Appeals. Memorandum Opinion at 5-6.2
After the briefing cycle in this case had been completed but prior to oral argument, this court handed down its decision in General Electric Uranium Management Corp. v. United States Department of Energy ("GEUMCO”), 764 F.2d 896 (D.C.Cir.1985). In that case, industry petitioners challenged a separate but related provision of the Nuclear Waste Policy Act, codified at 42 U.S.C. § 10222(a)(3), which imposes a one-time fee on electric utilities.3 The court held that “the court of appeals has *130original and exclusive jurisdiction to review DOE’s one-time fee rule and that the District Court lacked subject matter jurisdiction to consider the instant case.” 764 F.2d at 904. That decision, as the parties before us fully agree, is controlling as to the present challenge as well, because the instant case involves an attack on the DOE’s ongoing fee imposed under 42 U.S.C. § 10222(a)(2), a companion subsection of the same provision of the Nuclear Waste .Policy Act. Thus, we hold, based on GEUMCO, that the district court correctly concluded that it lacked jurisdiction. Moreover, because the utilities’ petition for review was consolidated with the appeal from the district court’s decision, we reach the merits of the utilities’ petition, to which we now turn.
II
The statutory provision before us consists of a single sentence: “For electricity generated by a civilian nuclear power reactor and sold on or after the date 90 days after January 7, 1983, the fee under paragraph (1) shall be equal to 1.0 mil per kilowatt-hour.” 42 U.S.C. § 10222(a)(2). Paragraph 1, which immediately precedes the pivotal subsection which is before us for analysis, consists of two sentences.4 It states, in summary, that the Secretary is authorized to enter into contracts with “any person” generating or owning “high-level radioactive waste, or spent nuclear fuel, of domestic origin” for the transportation and disposal of such materials. These contracts must provide for the payment of an ongoing fee as prescribed in paragraph (2) of section 10222(a) and a one-time-only fee as prescribed in paragraph (3) of the same section, i.e., the provision at issue in GEUMCO. The purpose of these two types of fees, as more fully described in our decision in GEUMCO, 764 F.2d at 898-900, is to provide the wherewithal for a DOE-administered “Nuclear Waste Fund.” That fund is designed to finance, among other things, the construction and operation of facilities for the storage and disposal of nuclear waste. See 42 U.S.C. § 10222(d) (enumerating the specific uses to which the Nuclear Waste Fund can be dedicated).
The dispute in this case turns on the meaning of the phrase “and sold on or after the date 90 days after January 7, 1983.” The utilities maintain that this language by its terms identifies nuclear-generated electricity which is sold; thus, under this reading, the statute excludes from the ongoing fee described in subsection (a)(2) that portion of commercially generated electricity consumed by utilities in the operation of their reactors. DOE argues, in contrast, that this phrase does not in fact require the nuclear-generated electricity to be “sold”; instead, in the agency’s view, the entire phrase was intended by Congress merely to establish a temporal benchmark for imposition of the fee. That is to say, by including the two words, “and sold,” in the phrase, “and sold on or after the date 90 days after January 7, 1983,” Congress meant merely to fix a date certain when the ongoing fee would go into effect.
A
The utilities’ reading of the statute comports with the plain language of the measure. In contrast, by the agency’s interpretation, the two words “and sold” could just as readily haye been left out of the statute in the first instance; indeed, the practical effect of DOE’s interpretation is *131to blue pencil out two words in an already brief one-sentence provision. The statute, as DOE would have us interpret it, would provide as follows: "For electricity generated by a civilian nuclear power reactor on or after [April 7, 1983], the fee under paragraph (1) shall be equal to 1.0 mil per kilowatt-hour.”
To be sure, this rather different wording is consistent with the adjacent provision in the statute. The companion fee provision establishing the one-time or subsection (a)(3) fee, 42 U.S.C. § 10222(a)(3) (1982), which was before us for decision in GEUMCO, imposes that fee as to “fuel [which] was used to generate electricity in a civilian nuclear power reactor prior to [April 7, 1983].” Not a word relating to “sale,” much less the words “and sold,” is to be found in this companion provision. But this different wording demonstrates that Congress well knew how to craft — in the very next sentence of the same statute — a fee regime that said naught whatever about electricity that is or was “sold.”5
The Secretary’s interpretation would thus have the unhappy result of obliterating express language from the subsection (a)(2) provision, in contravention of long-settled principles of statutory construction. See, e.g., Reiter v. Sonotone Corp., 442 U.S. 330, 337-38, 99 S.Ct. 2326, 2330, 60 L.Ed.2d 931 (1979); United States v. Oregon, 366 U.S. 643, 648, 81 S.Ct. 1278, 6 L.Ed.2d 575 (1961); Community Nutrition Institute v. Young, 757 F.2d 354, 357-58 (D.C.Cir.1985), petition for cert. filed, 54 U.S.L.W. 3311 (U.S. Nov. 4, 1985) (No. 84-665). Where the language Congress chose to employ is clear, the duty of the judiciary is likewise clear. We must follow that language and give it effect. As the Supreme Court has instructed us in terms that could scarcely be plainer, when we have before us language crafted by the Article I branch, “[a]bsent a clearly expressed legislative intention to the contrary, that language must ordinarily be regarded as conclusive.” Consumer Product Safety Commission v. GTE Sylvania, Inc., 447 U.S. 102, 108, 100 S.Ct. 2051, 2056, 64 L.Ed.2d 766 (1980). That language, moreover, is to be interpreted as plain English; the framers and enactors of language on Capitol Hill presumably meant what one would mean in everyday parlance. Malat v. Riddell, 383 U.S. 569, 571-72, 86 S.Ct. 1030, 16 L.Ed.2d 102 (1966).6
B
The Secretary does not directly quarrel with these general principles nor with the fact that DOE itself, in the context of the proposed rulemaking, once read the language as the utility petitioners do. See note 6, supra. While sketching an unfavorable legislative scene suggesting that Congress “never scrutinized or commented on” this language and that the pivotal words, “and sold,” “may even have been introduced into the statute by petitioners,” Appellees’ Brief at 10-11 & n. 2,7 the Secre*132tary ultimately rests his case on loftier ground. In its final rule, the agency eschewed its initial literalism and embraced the gross-generation approach, setting forth the reason for this volte face as follows:
The Department now believes this [the gross generation approach] to be the proper interpretation ... for the contrary interpretation [net generation] would unfairly subject future users of electricity (in the rate charges passed on by the Purchaser utility) to increased charges.
48 Fed.Reg. 16,591 (1983).8 This single sentence provides the sole pre-litigation rationale for construing the words “and sold” as merely part of the longer phrase, “and sold on or after [April 7, 1983],” designed narrowly — the Secretary argues — to establish the trigger date for ushering in the era of ongoing fees.9
Since the litigation began in late 1983, however, additional theories have been advanced by DOE to buttress its case. We need not linger long over these latter-day inspirations since they constitute post hoc rationalizations of counsel.10 As such, under settled principles, we need not entertain them. See, e.g., Securities Industry Association v. Board of Governors, 468 U.S. 137, 104 S.Ct. 2979, 2983, 82 L.Ed.2d 107 (1984); Camp v. Pitts, 411 U.S. 138, 142, 93 S.Ct. 1241, 1244, 36 L.Ed.2d 106 (1973); National Black Media Coalition v. Federal Communications Commission, 775 F.2d 342, 354, (D.C.Cir.1985); Secretary of Agriculture v. Interstate Commerce Commission, 551 F.2d 1329, 1331 (D.C.Cir.1977). But in any event, we can *133find no support whatever in this record for the principal after-the-fact rationale, namely that a gross-generation rule would have the inequitable result of penalizing efficient operators that generate the least unsold electricity. The utilities represent, and the DOE confirms in its brief, that the industrywide experience reflects approximately a modest five percent “diversion” of electricity for utility consumption purposes. The utilities further represent that any variations from plant to plant are, in the first place, insignificant and, in all events, related to plant design, not to the efficiency of plant operations. In its filings and argument in this court, the DOE does not directly confront or dispute those representations, arguing instead broader, more theoretical points. But we need not and choose not to wade fully into these highly technical waters; not a word in the Federal Register explication relies on this efficiency-penalty approach. We will therefore let the record stand as the agency created it.
C
We thus turn to the DOE’s principal contention here, namely that embracing the literal language of the statute would work an inequity to future ratepayers. That is to say, tomorrow’s ratepayers would have to pay for today’s nuclear wastes. This result would fly in the teeth, as the Secretary sees it, of clearly stated legislative intent that those who benefit from the generation of nuclear power should pay in full for the costs of transporting and disposing of the resulting nuclear waste. The Secretary captures this thought in the colorful pronouncement that no nuclear waste should get a “free ride,” with the ratepayers of the future being saddled, inequitably, with the tab for earlier waste generation. The DOE points in this respect to the companion one-time fee provision, subsection (a)(3), addressed in GEUMCO, 764 F.2d at 896, which was designed to assure equity and fairness in the allocation of the considerable costs of disposing of already extant wastes. Congress clearly intended, the DOE argues, to create a pay-as-you-go regime; adoption of the utilities’ fundamentalist approach to statutory interpretation would, the Secretary suggests, needlessly exalt literalism at the expense of Congress’ overriding intent.
The Secretary thus draws us to the unexceptional principal that it is the intent of Congress that provides the polestar for analysis, and that in the particular setting before us Congress could not have intended the inequitable regime that would reign under petitioners’ interpretation.
But upon analysis, we are unable to find any inconsistency with Congress’ intent that would be created by strict fidelity to the express terms of the statute,11 as petitioners urge. In the first place, the existing fee level is not carved in stone, but to the contrary can be increased (or decreased) by the DOE. In seeking to assure the Nuclear Waste Fund’s adequacy, the statute itself calls upon the Secretary to review annually the level of the ongoing fee “to evaluate whether collection of the fee will provide sufficient revenues to affect the costs [of administering the Nuclear Waste Fund].” 42 U.S.C. § 10222(a)(4) (1982). If on the basis of that review the *134Secretary determines that insufficient revenues are being collected, then he is obliged to “propose an adjustment to the fee to insure full recovery.” Id.12
Recognizing the availability of this review-and-adjustment procedure, the Secretary in this court calls into question the wisdom of the utilities’ litigating this question. As DOE views all this, if the utilities prevail and the subsection (a)(2), ongoing fee is imposed only upon net generation of electricity, then the Secretary could simply propose an upward adjustment of the fee under subsection (a)(4). But whether this litigation is prudent and wise is not the question; the relevant point is that the Secretary essentially concedes that the ultimate statutory purpose of assuring adequate funding levels can be accomplished through the vehicle of annual reviews and adjustments.13 The mechanism, which is not even alluded to in the Secretary’s Federal Register explanation for adopting a gross-generation approach, provides by Congressional design an effective antidote to the perceived evil of saddling the ratepayers of the distant future with the costs of transporting and disposing of nuclear waste generated currently.14
As additional points to justify departure from the clear statutory language, the agency argues (1) that the utilities’ interpretation of the statute would similarly have the effect of reading specific language out of the statute, and (2) that even if the court disagrees with DOE’s interpretation, settled principles require judicial deference to the reading of the expert agency. As to the first point, the Secretary suggests that the utilities’ emphasis upon the words, “and sold,” would have the effect of rendering pointless the statutory term, “generated,” prominently featured in the subsection (a)(2) provision. See p. 3, supra. Thus, as the Secretary sees it, neither party’s interpretation would retain the efficacy of all parts of the provision.
We cannot agree. The word “generated,” as employed in the statute, does in fact have a definitional function that is vouchsafed by the utilities’ reading. The term, as used in the statute, introduces a phrase that identifies the category of electricity which is subject to the ongoing fee, namely electricity “generated by a civilian nuclear power plant.” Without this qualifying phrase, the fee would be imposed on electricity from other sources, even that generated by, say, military nuclear facilities or even that generated by non-nuclear facilities. The language, in context, thus singles out nuclear-generated electricity from civilian power plants as the category *135of electricity which will be subject to the fee.15
As to the general principles of agency deference, we, of course, accept fully the Secretary’s proposition that “an agency’s interpretation of the statute it has been directed to administer is entitled to ‘great deference.’ ” Appellees’ Brief at 23 (quoting Udall v. Tollman, 380 U.S. 1, 16, 85 S.Ct. 792, 801, 13 L.Ed.2d 616 (1965)).16 But in our system of government, an agency is not at liberty to override the clearly expressed intent of Congress. See Association of American Railroads v. Costle, 562 F.2d 1310, 1318-19 (D.C.Cir.1977) (no deference accorded to agency interpretation that misconstrues statutory mandate). Like the judiciary, the agency is bound to follow the law as Congress passed it and the President signed it. Here, the law is clear. What is more, for reasons we have already mentioned, faithfully following the law’s express dictates would not work a result incompatible with what Congress intended. The five percent “gap” between the gross-generation and net-generation approaches is modest; it is not so wide as to defeat Congress’ intent with respect to the adequacy of the Nuclear Waste Fund or as to fairness to future ratepayers, particularly when any revenue shortfall can be overcome by means of the annual review-adjustment mechanism which Congress framed.17
We thus take the statute as Congress and the President have given it to us. We hold that the-ongoing-fee-applies only to electricity generated by a civilian power plant and.sold on or after April 7, 1983. The judgment of the District Court, holding that it lacked jurisdiction over this challenge, is affirmed; for the reasons stated, the petition for review is granted.
So ordered.
. Declaration of William M. Sprecher, filed October 7, 1983, ¶ 3 at p. 2. Mr. Sprecher is the Team Leader for Fee Administration in the DOE’s Office of Civilian Radioactive Waste Management. His affidavit accompanied the Government’s memorandum in support of its motion to dismiss or in the alternative for summary judgment before the district court.
. After concluding that it lacked jurisdiction, the District Court went on in dicta to opine on the substantive issue of statutory interpretation that is now before us. The court observed, in brief, that the DOE's interpretation was reasonable, citing the Supreme Court’s decision (rendered shortly before the district court reached its determination that it was without jurisdiction) in Chevron, U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984).
.The provision at issue in GEUMCO establishes a one-time fee on spent nuclear fuel, and other nuclear wastes, used to generate electricity in a civilian nuclear power reactor prior to the date on which the ongoing fee went into effect. The subsection in relevant part provides as follows:
For spent nuclear fuel, or solidified high-level radioactive waste derived from spent nuclear fuel, which fuel was used to generate electricity in a civilian nuclear power reactor prior to the application of the fee under paragraph (2) to such reactor, the Secretary shall, not later than 90 days after January 7, 1983, establish a 1 time fee per kilogram of heavy metal in spent nuclear fuel, or in solidified high-level radioactive waste. Such fee shall be in an amount equivalent to an average charge of 1.0 mil per kilowatt-hour for electricity generated by such spent nuclear fuel, or such solidified high-level waste derived *130therefrom, to be collected from any person delivering such spent nuclear fuel or high-level waste, pursuant to section 10143 of this title, to the Federal Government.
42 U.S.C. § 10222(a)(3) (1982).
. The section provides as follows:
In the performance of his functions under this chapter, the Secretary is authorized to enter into contracts with any person who generates or holds title to high-level radioactive waste, or spent nuclear fuel, of domestic origin for the acceptance of title, subsequent transportation, and disposal of such waste or spent fuel. Such contracts shall provide for payment to the Secretary of fees pursuant to paragraphs (2) and (3) sufficient to offset expenditures described in subsection (d) of this section. ____
42 U.S.C. § 10222(a)(3) (1982).
. The Secretary suggests that the two provisions should be read in the same manner, notwithstanding their differing language, because of indications in the legislative history that the two provisions were intended to be "comparable.” But comparability is not fairly to be equated with identity, and the language of the two provisions is, of course, not identical. Moreover, in the GEUMCO litigation the agency itself stressed the distinctions between the subsection (a)(2) ongoing fee provision and the subsection (a)(3) one-time fee provision; it can scarcely be heard to contend now that the two provisions are in fact precisely the same. See Appellee's Brief in GEUMCO at pp. 34-36.
. What is more, DOE itself had no difficulty, initially, in reading and understanding this language. In February 1983, the Secretary published a proposed rule expressly providing that the post-April 7, 1983, on-going fee would be based on the principle of net generation. The model contract set forth in the proposed regulations provided that "[effective April 7, 1983 Purchaser shall be charged a fee in the amount of 1.0 mi! per kilowatt hour (MKWH) on electricity generated and sold by Purchaser’s nuclear power reactors.” 48 Fed.Reg. 5464 (1983). Further, the model contract defined "Kilowatt Hours Generated and Sold" as "electricity generated at a civilian nuclear power reactor as measured by the station busbar, net of all station uses.” Id. at 5463.
.If anything, an industry pedigree would tend strongly to suggest that the language embodied a pro-industry meaning, such as the interpreta*132tion urged by petitioners. But we need not rely on this for our analysis and do not; we pause merely to observe that the Government's observations in this respect scarcely assist the agency’s interpretation.
. The Federal Register notice does not provide any further explanation of this concern. The agency apparently reasoned, however, that under the net generation approach there are no current payments into the fund to provide for future disposal of the nuclear fuel that generates the power which the utilities presently use in their own plants. When the time arrives to dispose of that fuel, in DOE's view, the cost will fall on those who happen to be the utilities’ customers at that time.
. The entire explication for the switch from the net-generation approach of the proposed rule to the gross-generation approach of the final rule is as follows;
The definition of "kilowatt hours generated” (Article 1.14) is a substantive change from the definition of the term "Kilowatt Hours Generated and Sold” in the proposed rule. Some commenters noted that the former definition did not account for or adequately reflect the Congressional intent regarding the method of establishing waste disposal charges. The definition in the proposed rule was based upon the premise that the language in Section 302(a)(2) of the Act "for electricity generated by a civilian nuclear power reactor and sold on or after the date 90 days after the date of enactment of this act,” meant to establish disposal charges on the basis of electricity "generated and sold,” i.e., the “net” electricity. These commenters pointed out the [sic] Section 302(a)(2) of the Act should be read to require establishment of disposal charges based upon the "electricity generated by a civilian nuclear power reactor,” i.e., the total or "gross” electricity generated. The Department now believes this to be the proper interpretation of Section 302(a)(2), for the contrary interpretation (i.e., “net" rather the [sic] gross”) would unfairly subject future users of electricity (in the rate charges passed on by the Purchaser utility) to increased charges. The words "and sold on or after a date 90 days after enactment,” in our view, should be read to establish a date certain for calculation of the Act’s differing methods of establishing fees for disposal charges, (i.e., past 90 days after enactment and prior to 90 days after enactment). Accordingly, the contract has been modified to reflect this interpretation.
48 Fed.Reg. 16,591 (1983).
.One theory which did not find its way into the written briefs but was articulated at oral argument is that "and sold” is, even in petitioners’ theory, laden with ambiguity since the petitioners would, to be pristine in approach, have to exclude nuclear generated power which is lost during the course of transmission, or sent to another grid in exchange for power at some future time, and is thus never "sold." This argument need not detain us; the agency never articulated this reason in promulgating the rule. Moreover, petitioners represented both to the district court and to us on appeal that the "and sold” language should, in theory, exclude transmission losses, but that petitioners simply chose not to assert the point. See also Transcript of Hearing before District Court, June 28, 1984, at 19.
. The parties argue inconclusively about the legislative history of the subsection (a)(2) provision, see also supra note 7, with respect to the agency suggestion that the “and sold” language may have stemmed from the electric utility industry itself. The Secretary emphasizes portions of the legislative history which employ the term, "generated,” as evidencing Congress’ focus upon that term. See, e.g., 128 Cong.Rec. 4261 (1982); id. at 4252; id. at 4158. This, of course, would suggest that the words, "and sold," were mere surplusage. The utilities, in contrast, quote from one colloquy during the Senate debate in which Sen. McClure, a principal sponsor of the Senate bill, stated again and again that the fee would apply to electricity that was generated and sold. Id. at 4162. We find neither side’s reading of the legislative record to be of especial force; certainly we find nothing in the history of this measure to suggest that the language employed by Congress should be read in any fashion other than in its everyday meaning. See United States v. Oregon, 366 U.S. 643, 648, 81 S.Ct. 1278, 1281, 6 L.Ed.2d 575 (1961) (statutory provisions "clear and unequivocal on their face” make resort to legislative history unnecessary).
. The statute provides that the proposed adjustment is to be transmitted to Congress, and will take effect "after a period of 90 days of continuous session have [sic] elapsed”' unless “either House of Congress adopts a resolution disapproving the proposed adjustment. Id. The validity vel non of this one-house veto provision, adopted prior to the Supreme Court’s decision in Immigration & Naturalization Service v. Chadha, 462 U.S. 919, 103 S.Ct. 2764, 77 L.Ed.2d 317 (1983), is, of course, not before us for review.
. The Secretary puts it this way in his brief: DOE has authority to raise or lower its nuclear waste charges to utilities if either insufficient or excess revenues are being collected. Consequently, if it is assumed that DOE’s revenues are deemed sufficient at present, then a victory by petitioners' [sic] in this case will simply make revenues insufficient and result in a decision by DOE to increase the 1.0 mil per kilowatt-hour charge to all utilities.
Appellees’ Brief at 11.
.To be sure, some lag time will occur between the generation of waste and any subsequent adjustment of the fee charge. But this gap need not be long, by virtue of the annual review-adjustment procedure, and is not of such duration as to create a fundamentally inequitable system that Congress could not have intended. Thus, a regulatory scheme founded on the net generation principle would not be impermissibly at odds with the statute’s purpose. See generally United States v. Shimer, 367 U.S. 374, 382-83, 81 S.Ct. 1554, 1560, 6 L.Ed.2d 908 (1961).
It bears noting that the Secretary has not seen fit, at least as of the time of the briefing and argument of this case, to seek to raise the ongoing fee. Moreover, the utilities argue with some force that the statutorily prescribed fee level was designed by Congress to be sufficiently conservative as to ensure adequate revenues. Appellants’ Brief at 36-37.
. In contrast, the agency's reading of "and sold" as merely providing a measuring point for initiation of the on-going fee could equally be achieved without those two words. That is to say, the April 7, 1983 inauguration date would have been clear had Congress not included the "and sold" language in the first instance.
. The Secretary also relies on the fact that increased deference is due by virtue of the fact that the agency’s interpretation was contemporaneous with the statute’s passage. Appellees' Brief at 23. The strength of this proposition, which would otherwise be unexceptional, is markedly dissipated in the circumstances before us by the fact that the agency initially construed the language in the very manner asserted by the utilities. See Morton v. Ruiz, 415 U.S. 199, 237, 94 S.Ct. 1055, 1075, 39 L.Ed.2d 270 (1974).
.We observe in this respect, as DOE itself did initially, that adoption of the net generation approach would not create an enormous gap or “exemption,” as the Secretary would now put it. The difference at issue here, as we have seen, is approximately five percent of total nuclear electric output; thus the ongoing fee remains intact as to the lion’s share of electricity generated from nuclear facilities.