dissenting:
Federal Energy Regulatory Commission (FERC) regulations that govern the grant of preliminary permits for hydroelectric development ostensibly favor the applicant whose plans are better adapted “to conserve and utilize in the public interest the water resources of the region.” 18 C.F.R. § 4.33(g)(1) (1981). If FERC concludes that competing applications are “equally well adapted” to advance these objectives, however, it breaks the tie by granting the preliminary permit to the applicant that filed first. Id. § 4.33(g)(2). Although I agree that FERC’s application of these regulations to particular situations must be upheld if the agency’s decisions are supported by substantial evidence, such evidence is entirely lacking in this case. Instead, the record now before us paints a dismal picture of an agency so understaffed or swamped by work that it appears to be institutionally incapable of giving all but the most obvious problems the attention required by law. Because FERC’s reasons for awarding the preliminary permit to MEAG rather than Dothan smack of a reflexive application of the “first-in-time” rule, I cannot join the majority’s decision.
A brief review of the facts illustrates the difficulties involved in affirming FERC’s action in this case. MEAG filed a permit application in March 1979 for the east side of the Chattahoochie River, basing its project design on a gross miscalculation of the potential elevation of the water behind the dam (the “head”). In December 1979, Dothan filed a competing application for the same site1 that differed from MEAG’s in two material respects. Dothan’s application was for the west side of the river, and assumed an average head of 14 to 16 feet rather than the 24 feet contemplated by MEAG. In May 1980, MEAG implicitly agreed that Dothan’s proposal was superior; it now proposed to locate its powerhouse on the west side of the river, and reduced its calculation of the available head and the energy production capability of the project.2 Notwithstanding these concessions — which meant that almost nothing of MEAG’s orig-
684 F.2d — 6 *166inal proposal remained intact — FERC found that MEAG’s application “as originally filed” was essentially equivalent to Do-than’s. In language strikingly similar to that used in numerous other cases, the Commission stated that because the two proposals were “equally well adapted under the criteria that are appropriate at the preliminary permit stage,” it would grant the permit to MEAG because MEAG’s application was “first in time.” Order, 15 FERC If 61,168 (1981), Joint Appendix (J.A.) 74-75.
There are three independent reasons to doubt that FERC’s “findings” were anything more than empty boilerplate. The first is empirical: in 120 comparable instances canvassed by Dothan, FERC has never granted a preliminary permit based on the “better adapted” approach of 18 C.F.R. § 4.33(g)(1), and has always relied on the mere matter of filing date and the “first-in-time” approach of 18 C.F.R. § 4.33(g)(2).3 When the tie-breaking rule is invoked in each and every ease, one must conclude that FERC’s definition of a “tie” is somewhat extraordinary.
The second reason is the inference suggested by FERC’s recent revision of its regulations governing preliminary permit applications. See Order No. 183, 46 Fed. Reg. 55,245 (Nov. 9, 1981) (17 FERC If 61,-083). Under the new regulation, 18 C.F.R. § 4.35, material amendments to an application for a preliminary permit will change the date of acceptance of the permit to the date on which the amendment is filed. Had FERC applied this regulatory change retroactively, MEAG’s “material amendments” to its application would probably have given the “first-in-time” preference to Dothan.4 The new regulation seems commendable; what is shocking is FERC’s bald explanation for the change:
In the case where an applicant that did not file first-in-time nevertheless has a clearly better plan of development and its competitor then seeks to utilize those ideas to make its plan equally well adapted, fairness militates against allowing the first-filed application to retain its priority in-time for its amended application.
Id., 46 Fed.Reg. at 55,249 (emphasis added). If the second applicant truly has “a clearly better plan of development,” the question of filing date should be absolutely irrelevant; FERC has no reason to invoke the “tie-breaking” rule of section 4.33(g)(2) if one application is “better adapted” under section 4.33(g)(1). The regulatory change essentially constitutes an admission that FERC prefers to grant preliminary, permits by virtue only of filing date. The change *167as to how the filing date will be determined, although apparently motivated by “fairness,” would be unnecessary if FERC made the effort to determine which applications were better adapted in the first place.
Finally, FERC has structured its administrative procedures in such a way that there is every incentive to grant permits based on filing date rather than on the quality of the competing applications. In Order No. 132, 46 Fed.Reg. 14,119 (February 13, 1981), the Commission delegated to the Director of the Office of Electric Power Regulation (OEPR) the authority to grant preliminary permits if competing applicants “do not propose and substantiate materially different plans.” Id. at 14,120. This order does not give the OEPR Director authority to grant permits under more complex circumstances, however; that task is for the Commission. In other words, a finding that projects are essentially equal is purely ministerial, and the permit will be awarded to the applicant that filed first; a finding that the second application is “better adapted” cannot even be made at the OEPR level. As a result, the deck is stacked in favor of initial applicants. Even a passing familiarity with the administrative process suggests the unlikelihood that a field director will burden his overworked home office when he can terminate a matter within his own authority, even if one-sidedly. The biases inherent in FERC’s approach are further demonstrated by the “background” discussion in Order No. 132, which explains that the purpose of this delegation of authority to the OEPR Director is to clear FERC’s backlog of competing preliminary permit applications.
It is true, as FERC noted below when it granted the permit to MEAG, that the Commission’s rules “are grounded in both public interest considerations and a pragmatism necessary in light of the unprecedented increase in permit applications, adding to a case load that must be handled with limited resources.” Order, May 27, 1981, at 4, J.A. 175. Courts must respect and defer to agency efforts to deal with scarce resources and clogged dockets — but not to the extent of permitting agencies to ignore their own regulations.5 Ultimately, of course, these concerns also must raise doubts about how well FERC fulfills its statutory responsibilities under Part I of the Federal Water Power Act, 16 U.S.C. §§ 791a et seq., to promote development of the nation’s water resources in the public interest. The central and still unanswered question here concerns the relevance of a preliminary permit to the Commission’s licensing of water projects under section 7 of the Act, 16 U.S.C. § 800(a). FERC’s statements in this regard have been consistently murky and often contradictory. Commission counsel stated during the oral argument of this case that preliminary permits have little ultimate import because “competition lives on. Dothan can come in with a better license application and ultimately win the license regardless whether the permit is granted to MEAG or not.” This contention, which seems mandated by the competitive licensing scheme contemplated by Congress in section 7 of the Act, has not been proven.6 More important, the conten*168tion ignores the public interest considerations that Congress expected would be used in granting preliminary permits under the Act.7
If the award of a preliminary permit controls the award of the final license, then capriciousness at the first stage means capriciousness at the second. The resulting detriment to the public interest should be self-evident. I would remand this case for FERC to explain (i) why the OEPR Director should be able to grant the initial application but not subsequent ones; (ii) why the new regulations governing “material amendments” to permit applications should not be given retroactive effect;8 (iii) *169whether FERC does automatically apply the “first-in-time” rule; and (iv) the consequences that granting preliminary permits on such an arbitrary basis have for the ultimate licensing of water projects. As this court observed in Connecticut Light & Power Co. v. FERC, 627 F.2d 467, 473 (D.C. Cir.1980), “reciting only nonexplanatory boiler plate” cannot possibly substitute for the reasoned administrative decisionmaking required by law. The majority opinion refers to “the wealth of [FERC’s] experience in hydroelectric projects.” Maj. op. at 164. Based on the record now before us, however, that experience might better be described as bankrupt.
I respectfully dissent.
. Although the majority opinion observes that Dothan filed the competing application “over nine months after [MEAG’s] application,” maj. op. at 3, Dothan had a longstanding interest in the site, and had commissioned engineering and feasibility studies that were completed as early as 1977. See Joint Appendix (J.A.) li-le, 58.
. The majority opinion ignores Dothan’s objections to the way in which MEAG made these concessions. MEAG did not attempt to rebut Dothan’s competing application within thirty days, see 18 C.F.R. § 4.33(e) (1981), but instead filed a protest to Dothan’s application and thus exceeded the thirty-day time limit by several months. Moreover, as Dothan emphasizes and the majority opinion intimates, the literal effect of this end-run procedure is that part of MEAG’s proposal has been abandoned, but not yet replaced with a new proposal. Although it is settled FERC policy that an incomplete application will be rejected, see, e.g., Northeastern Minnesota Municipal Power Agency, 16 FERC H 61,033 (1981); Sun Ventures, Ltd, 15 FERC II 61,015 (1981), at the time this case was argued MEAG had not yet filed the amended project map or revised project description that would seem to be required by 18 C.F.R. §§ 4.31, 4.32 (1981).
. See Brief for Petitioner Dothan (Dothan Brief) at 19, Adden. 43-47. FERC does grant preliminary permits based on statutory preferences, of course. For example, the Commission is required to favor applications from states and municipalities. See maj. op. at 3. The court has been informed of only two decisions, other than those turning on the statutory preference in favor of local governments, in which a preliminary permit was awarded to the second applicant to file. Marsh Island Hydro Associates, 16 FERC fl 61,236 (1981); City of Ukiah, California, 18 FERC If 61,108 (1982). In both cases, however, the second applicant held a license or contractual rights to a related site that could not be disturbed under the statute, and thus these cases also turn on a form of statutory preference rather than analysis of what is “better adapted” to promote the public interest. Indeed, even these decisions reflect FERC’s apparent belief that there should be a presumption in favor of the initial applicant. See City of Ukiah, 18 FERC 1f 61,108 at 4 (explaining why second applicant is better qualified “in this instance”).
. The new regulation defines a material amendment to a permit application as “a change in the location or size of the dam, the location of the powerhouse, or the size and elevation of the reservoir, any of which changes would enlarge, reduce, or relocate the area of the body of water between the farthest reach of the proposed impoundment and the point of discharge from the powerhouse to be affected by the proposed project.” 18 C.F.R. § 4.35(b)(2), published at 46 Fed.Reg. 55,251 (Nov. 9, 1981). FERC has been utterly silent about whether “the outcome would have been different” under this regulation, but see maj. op. at 162 n.2. If MEAG’s revision of the powerhouse location, expected elevation of the water behind the dam, and expected energy output do not constitute “the proposal of a different project,” see id., then what would? I agree with the majority that the Commission and not the court must make this judgment; I simply suggest that the Commission has apparently failed to do so.
. If FERC’s processing of preliminary permit applications is as mechanical as Dothan alleges, then the agency has made a nullity of other regulations as well. 18 C.F.R. § 4.33(d)(2) (1981), for example, requires that a second applicant explain why its plan is better adapted than the proposal of the initial applicant. These applicants too have demands on their time. FERC would do better to explain its administrative processes candidly, rather than requiring applicants to file “better adapted” statements that may never be read. FERC emphasizes that the purpose of the first-in-time rule is to save other applicants the cost and time of preparing a competing application unless they can demonstrate a better plan of development. Brief for Respondent FERC at 20. If Dothan’s allegations are correct, however, the agency’s regulations only create the expensive illusion that a non-statutorily preferred applicant can ever succeed against an initial applicant.
. FERC submitted a post-argument letter to the court stating that although there are now 945 preliminary permit application dockets of which 332 involve competing applicants, there are 208 license application dockets of which only 25 involve competing applicants. Letter of March 19, 1982. This difference alone suggests that competition at the licensing stage is minimal. Moreover, during oral argument *168FERC counsel suggested that sixty percent of all license applications involve situations in which no party held a preliminary permit. As a result, it is possible that none of the 25 contested license applications involves a challenge to the holder of a preliminary permit. Based only on the record before the court, it is entirely possible that granting a preliminary permit is a de facto grant of the final license.
. The Act, with its emphasis on governmental development of hydroelectric sites and limiting the return from such projects earned by private investors, was considered fairly controversial in 1918. Developers of water power projects were viewed as trustees of public resources, and their actions were to be governed by the public interest. Because the Act only authorized the issuance of one permit per site at a time, see 56 Cong.Rec. 8931 (1918) (Representative Sims), Congress expected permits to be issued when there “would be a development upon such reasonable conditions as would be just for the public and for the Government." Id. at 9773 (Representative Dempsey) (emphasis added); see id. (Representative Anderson) (preliminary permit “is for the benefit of the Government and for the benefit of the individual”). FERC does not have any “statutory discretion” in this regard. Compare maj. op. at 164 with 56 Cong.Rec. 9805 (1918) (Representative Hum-phreys) (although statute says commission “may” give preference to better adapted permit applications, “the word ‘may’ is construed to mean ‘shall’ ”).
FERC has consistently minimized its ability to distinguish between projects at the preliminary permit stage. See, e.g., Town of Madison Electric Works Dep’t, 11 FERC fl 61,318 at 61,-673 (1980) (“To determine with any confidence that one of these two [permit] applicants has plans better than the other’s to conserve and utilize in the public interest the available water resources, we would-have to have the results of the detailed studies and analyses to be conducted under the permit itself’); Order No. 183, 46 Fed.Reg. at 55,249 (“the proposals submitted at [the preliminary permit] stage of conceptualization are seldom so altogether different from one another as to permit a clear judgment to be made on the merits, rather than on first-in-time”). Even if this is true, perhaps the applicants should be encouraged to refine their proposals until such a determination can be made. There is no apparent basis in the Act for FERC to solve this problem by arbitrarily closing off one applicant’s efforts merely because the other applicant enjoys an earlier filing date.
. The majority opinion states that this would be “pointless since prospective application of the regulations is directed by statute,” i.e., the Administrative Procedure Act, 5 U.S.C. § 551(4) (1976). Maj. op. at 162 n.2. This suggestion seems unwarranted, for three reasons. First, even so-called “legislative” rules may be applied retroactively, when the regulation simply articulates what is required by reasoned decisionmaking. See, e.g., Greene v. United States, 376 U.S. 149, 162-64, 84 S.Ct. 615, 622-23, 11 L.Ed.2d 576 (1964); Central Power & Light Co. v. United States, 639 F.2d 1104, 1106 (5th Cir. 1981), cert. denied, 454 U.S. 831, 102 S.Ct. 128, 70 L.Ed.2d 108; Hicks v. Califano, 600 F.2d 1048, 1050 (4th Cir. 1979). “[A] court reviewing an agency decision following an intervening change of policy by the agency should remand to permit the agency to decide in the first instance whether giving the change retrospective effect will best effectuate the policies underlying the agency’s governing act.” NLRB v. Food Store Employees, 417 U.S. 1, 10 n. 10, 94 S.Ct. 2074, 2080 n.10, 40 L.Ed.2d 612 (1974). See generally K. Davis, Administrative Law Text § 5.05, at 133 (3d ed. 1972). Second, the promulgation of new regulations does not turn an agency’s application of the underlying statute into retroactive rulemaking. See, e.g., Chrysler Corp. v. EPA, 631 F.2d 865, 889 (D.C.Cir.), cert. denied, 449 U.S. 1021, 101 S.Ct. 589, 66 L.Ed.2d 483 (1980). The adoption of regulations that clarify what the language of the statute was intended to convey “is no more retroactive in its operation than is a judicial determination construing and applying a statute to a case in hand.” Manhattan General Equipment Co. v. Commissioner, 297 U.S. 129, 135, 56 S.Ct. 397, 400, 80 L.Ed. 528 (1936). Finally, notwithstanding the formal rulemaking procedures used by FERC, the regulation is clearly more “interpretive” than “legislative,” because it is a general statement of administrative policy rather than a standard of conduct “issued pursuant to legislatively-delegated power to make rules having the force of law.” Joseph v. Civil Service Comm’n, 554 F.2d 1140, 1153 n.24 (D.C.Cir. 1977); see generally U.S. Chamber of Commerce v. OSHA, 636 F.2d 464, 468 (D.C.Cir. 1980). As an interpretive rule *169that corrects what is arguably FERC’s erroneous prior understanding of the statute, it is manifestly unjust not to give the regulation retrospective effect. The substance of what the agency has done, and not the label it attaches to the rule, is of course dispositive in judging whether a regulation is “interpretive” or “legislative.” CBS, Inc. v. United States, 316 U.S. 407, 416, 62 S.Ct. 1194, 1200, 86 L.Ed. 1563 (1942).