Petitioners, along with the Department of Justice and the Williams Pipe Line Company, challenge an order of the Federal Energy Regulatory Commission (FERC) on a wide variety of grounds. The FERC order in question specified the generic ratemaking methodology to be applied to all oil pipelines pursuant to the Interstate Commerce Act. In its order, the Commission articulated for the first time its belief that oil pipeline rate regulation should serve only as a cap on egregious price exploitation by the regulated pipelines, and that competitive market forces should be relied upon in the main to assure proper rate levels. Furthermore, in devising a specific ratemaking methodology in accordance with these beliefs, FERC retained the rate base formula used in the past in oil pipeline ratemaking, even though this formula had met with severe criticism from this court in Farmers Union Central Exchange v. FERC, 584 F.2d 408 (D.C.Cir.1978), cert. denied sub nom. Williams Pipe Line Co. v. FERC, 439 U.S. 995, 99 S. Ct. 596, 58 L. Ed. 2d 669 (1978). At the same time, the Commission revised its rate of return methodology so that the resulting rate levels would represent ceilings seldom reached in actual practice.

For the reasons set forth below, we find that the Commission's order contravenes its statutory responsibility to ensure that oil pipeline rates are "just and reasonable." In addition, we hold that FERC failed both to give due consideration to responsible alternative ratemaking methodologies proposed during its administrative proceedings, and to offer a reasoned explanation in support of its own chosen ratemaking methodology, and that therefore the FERC order constitutes impermissible "arbitrary and capricious" agency action. Accordingly, we remand this case for further proceedings consistent with this opinion.

Williams Pipe Line Company (Williams),1 an independent common carrier, operates oil pipelines over a large territory in the midwestern United States. Williams entered the pipeline business in 1966, when it purchased its operating assets from the Great Lakes Pipe Line Company. In late 1971 and early 1972, Williams increased its local rates and initiated new joint rates with another pipeline company. Those rates are still at issue today.

Petitioners, various oil producers and refiners that ship their products through Williams' pipeline, challenged the lawfulness of these rates before the Interstate Commerce Commission (ICC) in 1972. After evidentiary hearings, the presiding administrative law judge concluded that the Williams rates were "just and reasonable" within the meaning of the Interstate Commerce Act, 49 U.S.C. Sec. 1(5), and a three-commissioner division of the ICC subsequently adopted in full the administrative law judge's findings. See 355 I.C.C. 102 (1975).2 The full ICC then reopened the proceedings for reconsideration "because of the relative dearth of precedent concerning petroleum pipeline rates, and in view of the substantial sums of money at issue." 355 I.C.C. 479, 481 (1976). Upon reconsideration, the full ICC affirmed the division's decision, ruling that "[c]onsiderations of consistency and fairness require that we adhere to our previously recognized criteria in investigating the rates of particular pipelines," 355 I.C.C. at 484, and that a pending rulemaking was "the [proper] proceeding for considering a change" in the methods for valuating the rate base and for determining the proper rates of return for oil pipelines. See 355 I.C.C. at 485, 487.

Petitioners then sought judicial review in this court. In 1977, during the pendency of the appeal, Congress transferred regulatory authority over oil pipelines to the newly created Federal Energy Regulatory Commission (FERC).3 In 1978, this court remanded the case to FERC for reconsideration, in order "to avail ourselves of some additional expertise before we plunge into this new and difficult area [of oil pipeline regulation], and to allow [FERC] to attempt for itself to build a viable modern precedent for use in future cases that not only reaches the right result, but does so by way of ratemaking criteria free of the problems that appear to exist in the ICC's approach." Farmers Union Central Exchange v. FERC, 584 F.2d at 421 (Farmers Union I ). While at that time this court expressed "unease with the ICC's findings regarding rate base, rate of return, and depreciation costs," id., based as they were upon "weak and outmoded ... products of a bygone era of ratemaking,"4 id. at 418, "[w]hat clinch[ed] our decision to remand [was] the fact that the agency now charged with [ratemaking] responsibility, FERC, ha[d] requested a remand so that it may begin its regulatory duties in this area with a clean slate," id. at 421. Accordingly, we remanded so that FERC could conduct a fresh and searching inquiry into the proper ratemaking methods to be applied to oil pipelines.

In February 1979, after Williams had filed other new rate changes, FERC reopened the remanded case, and assigned an administrative law judge (ALJ) to hold hearings on the consolidated cases.5 At the prehearing conference, the ALJ bifurcated the proceedings. Phase I was to devise generic principles for the setting of just and reasonable oil pipeline rates. Phase II would apply those principles to the Williams case in particular.6 After seventy-six days of hearings in Phase I, FERC directed the ALJ to omit an initial decision and to certify the record directly to the Commission, and instructed the parties to submit briefs directly to the Commission.7 FERC heard oral argument on June 30, 1980. Almost a year then passed without a FERC decision. Accordingly, Farmers Union Central Exchange (Farmers Union) filed a motion in this court to compel agency action, which we dismissed upon receiving assurances from FERC counsel that a decision was forthcoming imminently.8 8] Three months later, however, in October 1981, FERC ordered a reargument by the parties on November 19, 1981.9

Eight months after reargument, FERC had still failed to issue a decision. Upon petition from Farmers Union, the district court, finding that FERC had abrogated its statutory responsibilities under both the Interstate Commerce Act10 and the Administrative Procedure Act,11 ordered FERC to issue a decision within sixty days.12 This court then stayed the district court's order so that FERC would be allowed until November 30, 1982 to issue its decision.13

On November 30, FERC issued Opinion No. 154, the subject of this appeal. See 21 FERC (CCH) p 61,260 (Nov. 30, 1982). The Department of Justice, representing the United States as statutory respondent under 28 U.S.C. Secs. 2344, 2348, joined petitioners in seeking reversal of the FERC opinion.

FERC heralded its Opinion No. 154 (the Williams opinion) as "the longest and most elaborate" decision it had ever issued.14 The Williams opinion announces FERC's intended approach to future oil pipeline ratemaking; thus it is of great importance to oil producers, refiners, and pipeline owners.

FERC's essential conclusion in Williams is that ratemaking for oil pipelines should serve only "to restrain gross overreaching and unconscionable gouging"15 in order to keep rates within the zone of "commercial reasonableness," not "public utility reasonableness."16 As FERC said in a related order issued the same day as Williams:

Williams says that oil pipeline rate regulation should be relatively unobtrusive. It finds competition (both actual and potential) a far more potent force in this industry than in the others we regulate. Accordingly, it proposes to rely in the main on market forces. It views oil pipeline rate regulation as a modest supplement to rather than a pervasive substitute for the market. The supplement, Williams tells us, is in the nature of a check on gross abuse.

Trans Alaska Pipeline System, 21 FERC (CCH) p 61,092, at 61,285 (Nov. 30, 1982). The following summary describes how FERC reached that conclusion, and how it translated that conclusion into a particular ratemaking methodology.

A. The Congressional Purpose in Mandating "Just and Reasonable" Oil Pipeline Rates

In 1906, Congress adopted the Lodge Amendment to the Hepburn Act, which extended the definition of common carrier in the Interstate Commerce Act17 to encompass interstate oil pipelines, and, as a consequence, required pipeline rates to be "just and reasonable."18 In Williams, FERC embarked on a close study of "the climate of opinion" that existed when Congress passed the Lodge Amendment. In doing so, FERC primarily examined the works of Ida Tarbell, a progressivist of the turn of the century, who has been credited with "inflam[ing] the public's long-standing hostility to the [Standard Oil] combination as nothing before had."19 FERC concluded that the Lodge Amendment was motivated by the desire to bust the Standard Oil trust.20

FERC also found that in the early twentieth century the Standard Oil Company maintained its dominance over the entire American oil business by setting its pipeline rates at such extraordinarily high levels that access to the pipelines (and hence to important downstream markets) was cut off. See 21 FERC at 61,597. From this observation, FERC concluded that the Congress, in mandating that oil pipeline rates be "just and reasonable," intended to outlaw only outrageously high rates: "Prohibitive rates were a means to that end [of dominating American oil markets]. Congress wanted to forbid both the use of the means and the attainment of the end. The policy at which it fired was a policy of 'prohibitive' pricing." Id. In the belief that "[t]he phrase in question, 'just and reasonable,' is a high-level abstraction[,] ... a mere vessel into which meaning must be poured," id. at 61,594, and considering numerous differences in the reasons for the establishment of a regulatory scheme over "public utilities," such as electric companies, as opposed to "transportation companies," such as oil pipelines, id. at 61,591-96, FERC determined that:

the authors of the Hepburn Act's oil pipeline provisions did not use the words "just and reasonable" in the sense in which public utility lawyers have used them since the 1940's.

We think that what was meant was not "public utility reasonableness," but ordinary commercial "reasonableness." To be specific, we discern no intent to limit these carriers' rates to barebones cost. What we perceive is an effort to restrain gross overreaching and unconscionable gouging.

Id. at 61,597. Thus, on the basis of this historical survey, FERC interpreted the statutory mandate that oil pipeline rates be "just and reasonable" to require only the most lighthanded regulation, with no necessary connection between revenue recoveries and the cost of service.

FERC next surveyed the changes since 1906 in the economics of the oil pipeline industry, and determined that the modern economic environment does not manifest the same threat of monopolistic practices that bedeviled Congress in 1906.

Comparing the dollars spent in 1981 in America for petroleum products to the dollars spent in the same year for oil pipeline transportation,21 FERC found that pipeline costs are "not very much when viewed in relation to the nation's total oil bill."22 Further, FERC found that any savings created by lower pipeline charges would not necessarily--or even likely--be passed on to consumers. See 21 FERC at 61,601-02. FERC therefore concluded that "[f]rom the consumer's perspective, oil pipeline rate regulation is akin to efforts to do something about the high price of shoes by controlling the pricing of shoe laces [or] to contain the price of food by seeing to it that the price of spice is always 'just and reasonable.' " Id. at 61,601.

FERC also found that, from Congress' perspective in 1906, oil pipeline rates did in fact make a difference to the oil consuming public. Reviewing cost and revenue trends, FERC showed that in the past pipeline charges comprised as much as sixty-eight percent of what the oil producer received for crude oil.23 Thus, FERC concluded that although Congress may in 1906 have reasonably been concerned about oil pipeline prices, today "[p]rohibitive oil pipeline rate structures are now a problem for the economic historian," and the "oil pipeline rate reform crusade is anachronistic ... overtaken by events so that the combatants' rhetoric is no longer in touch with reality." Id. at 61,606-07.

Finally, FERC found that the economic market for oil pipelines has become competitive since 1906. In contrast to the industry during the early part of this century, today "[p]rohibitive pricing has become uneconomic"24 and "[n]o oil company (not even the largest) is wholly self-sufficient."25 Also, FERC appeared to conclude that the significant decline in the price of pipeline transportation from 1931-1969 manifests the existence of competition in the pipeline transportation market.26

In light of all the foregoing considerations, FERC expressed its belief that the consumer's interest in low pipeline rates is "submicroscopic" while the real threat to the public is underinvestment in needed oil pipelines.27 Accordingly, FERC set down as a guiding principle of oil pipeline ratemaking that it is "best to err on the side of liberality" because "the dangers of giving too little vastly outweigh those of giving too much." Id. at 61,613.

FERC then turned to apply this general principle to formulate a ratemaking methodology for oil pipelines.

Under the old ICC method, an arcane formula, comprised chiefly of a weighted average of original cost and cost of reproduction new,28 was used to calculate the pipelines' "valuation rate base."29 While admitting that "[w]ere we beginning afresh on a clean slate we might be inclined to use something different" because the ICC formula contains "anomalies and inconsistencies" that result in an inaccurate picture of the pipelines' cost of service, id. at 61,616, FERC nevertheless concluded that the costs of adopting another rate base formula outweighed the benefits of such a shift. It therefore chose to "adhere to the formula [it] inherited from the Interstate Commerce Commission." Id. at 61,632.

In doing so, FERC expressly rejected two proposed alternatives to the ICC ratemaking formula. First, the Commission eschewed original cost ratemaking in the belief that the chief advantage of such an approach--the facilitation of comparable earnings analysis--was of little use in the oil pipeline context, and that the switch to original cost alternative would create unnecessary regulatory burdens and social costs. See infra at 1511-18. Second, FERC rejected specific alterations to the ICC rate base formula proposed by the Association of Oil Pipe Lines because, in FERCs view, only "relatively insubstantial" amounts of money would be affected, and, in any event, the ICC's methodological errors tend to compensate roughly for one another. See infra at 1518-21.

Thus FERC reaffirmed the ICC rate base method, admitting it to be "much too blunt or too clumsy for close work," but still finding it "pragmatic" and "usable." 21 FERC at 61,616.

Quoting at length from this court's opinion in Farmers Union I, FERC launched its inquiry into rate of return methods from the premise that "[t]he need for reform is plain."30 Finding "the parties' arguments ... so unhelpful and the applicable historical tradition ... so palpably deficient," FERC felt "left to [its] own devices" to fashion a new rate of return methodology.31 It held that a proper rate of return for oil pipelines should be comprised of three elements: (1) debt service, (2) a "full compensatory suretyship premium," and (3) the " 'real ' entrepreneurial rate of return on the equity component of the valuation rate base." See 21 FERC at 61,644 (emphasis in original).

The first component, debt service, represents the amount needed to pay interest on the debt the pipeline has accumulated. The second component, the suretyship premium, represents the additional amount that would have been needed above actual debt service in the absence of a debt guarantee from the oil pipeline company's parent.

The third component, the "entrepreneurial" rate of return, according to FERC, "follows logically from [the] basic concept that what the historical background and contemporary public policy needs call for here is a cap on gross abuse." Id. at 61,645. Accordingly, FERC offered eight different measures for the "entrepreneurial" rate of return. The measures included the nominal rates of return on book equity realized over the most recent one- or five-year period for (1) the oil industry generally, (2) American industry generally, or (3) the parent company or companies, excluding pipeline operations. The remaining two measures of an entrepreneurial rate of return took the total returns (dividends plus capital gains) on a "diversified common stock portfolio" over (1) the past five years or (2) "the long run--25 years, 50 years, or more." Id. Under FERC's method, the pipeline would normally be permitted to choose the applicable rate of return from among these indices.

Once this rate of return is selected, it is adjusted downward "[t]o avoid overcompensation for inflation." Id. at 61,646. FERC's methodology subtracts from the selected rate of return the percentage by which the valuation rate base has increased during "the time period that was looked to in order to derive the appropriate nominal rate of return."32

This adjusted rate of return is applied not to book equity, nor to the percentage of the valuation rate base represented by the proportion of equity relative to debt in the oil pipeline's overall capitalization structure. Rather, this rate is the allowed return on what FERC considers to be the "equity component of the valuation rate base"--the entire valuation rate base, less the face amount of debt. See id. at 61,647-48.

This method, FERC concedes, would result in "handsome rate base writeups," followed by "creamy returns on book equity." Id. at 61,650. FERC, however, believed that such high returns comported with its general ratemaking principles for oil pipelines: "Here we are setting ceilings that will seldom be reached in actual practice."33 Moreover, the Commission allowed generous returns in the belief that oil pipeline equityholders were entitled to the full benefit of appreciation in their leveraged assets, id. at 61,649, and that the more "austere standard of fairness applied in the utility field cannot be divorced from the stringent regulatory controls on abandonment" which, FERC ruled, do not apply to oil pipelines, id. at 61,650.

FERC made three other rulings in Williams that are challenged in this appeal. FERC held that (1) the original cost of transferred pipeline plant--and not its purchase price--should be used in ratemaking, (2) oil pipeline rate regulation should generally take place on a systemwide, rather than point-to-point, basis and (3) the "tax normalization" method of accounting may be employed by the oil pipeline companies if they so wish.

First, FERC set down as a general rule that the "purchase price [for pipeline plant] is not entitled to any recognition at all for any ratemaking purpose."34 There are two ways in which purchase price might have been used in oil pipeline ratemaking: (1) as a substitute for original cost in the rate base, and (2) in calculating the basis for depreciation expenses. FERC rejected the first use of purchase price because to do so would create a systemic incentive for the sale of pipeline plant and the consequent upward push on rates. See id. at 61,634-35. Further, to use purchase price in the rate base would contravene the principle that "a mere change in ownership should not result in an increase in the rate charged for a service if the basic service rendered itself remains unchanged."35 FERC similarly rejected the use of purchase price as the basis from which depreciation would be computed, citing this court's disapproval in Farmers Union I of the practice,36 and finding no valid justification for what it called "this nonchalant, half a loaf, split the difference" policy of using original cost in the rate base, while calculating depreciation by reference to purchase prices. Id. at 61,635.

Second, FERC decided to regulate oil pipeline rates on a systemwide basis. FERC maintained that the alternative--ruling on the reasonableness of particular rates on specific routes--would require cost allocation inquiries that would be "metaphysical inconclusive, and barren." Id. at 61,651. Also, FERC believed that systemwide regulation would give freer play to competitive forces in the oil pipeline industry. FERC restricted its ruling to pipeline systems, in contrast to pipeline companies. The rates of wholly noncontiguous pipeline systems, therefore, would not be computed by averaging companywide costs. FERC further cautioned that a showing that systemwide rates discriminated against nonowners of the pipeline would trigger "strict regulatory scrutiny." Id.

Third, FERC permitted, but did not require, oil pipeline companies to "normalize" their accounts that reflect accelerated depreciation on equipment for tax purposes.37 FERC permitted the use of the tax normalization method because "normalization facilitates the comparable earnings analyses basic to the determination of appropriate rates of return on oil pipeline equity investments." Id. at 61,656. However, because "[c]ompetitive considerations may lead some pipelines to prefer lower rates .... now in return for more later," FERC made the use of the method elective rather than compulsory. Id.

Finally, FERC prohibited pipelines that choose tax normalization from including the resulting tax reserve accounts in their rate bases. Otherwise, "the rate payer who has paid higher taxes reflecting normalization accounting would be paying the carriers for earnings on the tax differential even though it was the rate payer who contributed the differential in the first place." Id. at 61,657 (quoting San Antonio v. United States, 631 F.2d 831, 847 (D.C.Cir.1980)).

The FERC order before us today is an exercise of its general ratemaking authority under 49 U.S.C. Sec. 15(1).38 As such, the Williams proceeding constitutes a rulemaking under the Administrative Procedure Act. See 5 U.S.C. Sec. 551(4) (defining "rule" to include "the approval or prescription for the future of rates"). Although section 15(1) provides that the determination as to the reasonableness of rates shall be made "after full hearing," the resulting decision apparently need not be "on the record," 5 U.S.C. Sec. 553(c), and therefore the standards for formal rulemaking do not apply. See United States v. Florida East Coast Railway Co., 410 U.S. 224, 93 S. Ct. 810, 35 L. Ed. 2d 223 (1973). Accordingly, we review whether FERC's order in Williams was "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." 5 U.S.C. Sec. 706(2)(A).39

Under the "arbitrary and capricious" standard, a reviewing court must conduct a "searching and careful"40 inquiry into the record in order to assure itself that the agency has examined the relevant data and articulated a reasoned explanation for its action including a "rational connection between the facts found and the choice made." Burlington Truck Lines v. United States, 371 U.S. 156, 168, 83 S. Ct. 239, 246, 9 L. Ed. 2d 207 (1962). As the Supreme Court recently elaborated:

Normally, an agency rule would be arbitrary and capricious if the agency has relied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.

Motor Vehicles Manufacturers Association v. State Farm Mutual Automobile Insurance Co., --- U.S. ----, 103 S. Ct. 2856, 2867, 77 L. Ed. 2d 443 (1983). Most fundamentally, our task is "to ensure that the [agency] engaged in reasoned decisionmaking." International Ladies' Garment Workers' Union v. Donovan, 722 F.2d 795, 815 (D.C.Cir.1983); see American Gas Association v. FPC, 567 F.2d 1016, 1029-30 (D.C.Cir.1977), cert. denied, 435 U.S. 907, 98 S. Ct. 1457, 55 L. Ed. 2d 499 (1978).

Agency decisionmaking, of course, must be more than "reasoned" in light of the record. It must also be true to the congressional mandate from which it derives authority. Therefore, a reviewing court must be satisfied that the agency's reasons and actions "do not deviate from or ignore the ascertainable legislative intent." Ethyl Corp. v. EPA, 541 F.2d 1, 36 (D.C.Cir.) (en banc) (quoting Greater Boston Television Corp. v. FCC, 444 F.2d 841 (D.C.Cir.1970)), cert. denied sub nom. E.I. Du Pont de Nemours & Co. v. EPA, 426 U.S. 941, 96 S. Ct. 2662, 49 L. Ed. 2d 394 (1976); see 5 U.S.C. Sec. 706(2)(C) ("The reviewing court shall ... hold unlawful and set aside agency action ... in excess of statutory jurisdiction, authority, or limitations."). Beyond that, however, we are not at liberty to substitute our own judgment in the place of the agency's. In this sense, the "arbitrary and capricious" standard is narrow and restricted. See Small Refiner Lead Phase-Down Task Force, 705 F.2d at 520-21.

The "arbitrary and capricious" standard demands that an agency give a reasoned justification for its decision to alter an existing regulatory scheme. See Motor Vehicle Manufacturers Association, 103 S. Ct. at 2866. We are well aware that changed circumstances may justify the revision of regulatory standards over time. Indeed, our initial remand in Farmers Union I was impelled by our suspicion that prior ICC methods might no longer be useful. See 584 F.2d at 412-20. To acknowledge that circumstances have changed, however, is not to eliminate the burden upon the agency to set forth a reasoned analysis in support of the particular changes finally adopted. Furthermore, in light of the purpose of the remand in Farmers Union I--"to build a viable modern precedent for use in future cases that not only reaches the right result, but does so by way of ratemaking criteria free of the problems that appear to exist in the ICC's approach"41--we believe that FERC's adherence to the old ICC rate base method also demands a reasoned justification. Cf. Food Marketing Institute v. ICC, 587 F.2d 1285, 1290 (D.C.Cir.1978) (courts reviewing agency action after remand should ensure that "genuine reconsideration of the issues" took place).

Thus we take up the task of reviewing the Williams opinion with two objectives in mind. First, we will examine whether FERC's actions and supporting rationale comport with its delegated authority to set oil pipeline rates at a "just and reasonable" level. Second, we then will scrutinize the Williams opinion to see whether FERC considered all relevant factors and demonstrated a reasonable basis for its decision. See Sierra Club v. Costle, 657 F.2d 298, 323 (D.C.Cir.1981).

IV. FERC'S ACTION CONTRAVENES THE STATUTORY DIRECTIVE TO

DETERMINE WHETHER RATES ARE "JUST AND REASONABLE""

Under section 1(5) of the Interstate Commerce Act, all rates charged for oil pipeline transportation "shall be just and reasonable." Similarly, under section 15(1), Congress authorized FERC "to determine and prescribe what will be the just and reasonable" rate for such transportation services.

We find that FERC in the Williams decision failed to satisfy that statutory mandate. We also find unconvincing FERC's attempts at justifying its novel interpretation of "just and reasonable" rates. First, FERC sought to establish maximum rate ceilings at a level far above the "zone of reasonableness" required by the statute. Second, FERC failed to specify in any detail how "non-cost" factors, such as the need to stimulate additional pipeline capacity, might justify its decision to set maximum rates at such high levels. Third, the legislative history of the Hepburn Act betrays FERC's belief that the "climate of opinion" in 1906 shaped a congressional purpose to impose only very lighthanded rate regulation on the oil pipelines. Finally, FERC's reliance on its findings that oil pipeline rate regulation is (1) unimportant to consumers at large, and (2) best left to "regulation" by market forces in most cases, constitutes an improper departure from the basic congressional mandate to ensure that oil pipeline charges are "just and reasonable."

Congress delegated ratemaking authority to FERC in broad terms. Accordingly, "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." Permian Basin Area Rate Cases, 390 U.S. 747, 790, 88 S. Ct. 1344, 1372, 20 L. Ed. 2d 312 (1968). In arriving at a just and reasonable rate, "no single method need be followed." Wisconsin v. FPC, 373 U.S. 294, 309, 83 S. Ct. 1266, 1274, 10 L. Ed. 2d 357 (1963). Indeed, and more specifically, FERC is not required "to adhere 'rigidly to a cost-based determination of rates, much less to one that base[s] each producer's rates on his own costs.' " FERC v. Pennzoil Producing Co., 439 U.S. 508, 517, 99 S. Ct. 765, 771, 58 L. Ed. 2d 773 (1979) (quoting Mobil Oil Corp. v. FPC, 417 U.S. 283, 308, 94 S. Ct. 2328, 2346, 41 L. Ed. 2d 72 (1974)).

On the other hand, the delegation of the power to prescribe rates is accompanied by standards to which FERC, as delegate, must conform. As Judge Leventhal observed, "Congress has been willing to delegate its legislative powers broadly--and courts have upheld such delegation--because there is court review to assure that the agency exercises the delegated power within statutory limits...." Ethyl Corp. v. EPA, 541 F.2d at 68 (Leventhal, J., concurring). Surely, FERC enjoys substantial discretion in its ratemaking determinations; but, by the same token, this discretion must be bridled in accordance with the statutory mandate that the resulting rates be "just and reasonable." See FPC v. Texaco, Inc., 417 U.S. 380, 394, 94 S. Ct. 2315, 2324, 41 L. Ed. 2d 141 (1974); Atchison, Topeka & Sante Fe Railway Co. v. Wichita Board of Trade, 412 U.S. 800, 806, 93 S. Ct. 2367, 2374, 37 L. Ed. 2d 350 (1973).

The "just and reasonable" statutory standard is, of course, not very precise, and does not unduly confine FERC's ratemaking authority. As this court once explained, "[t]he necessity for an anchor to 'hold the terms "just and reasonable" to some recognizable meaning' is plain, for the words themselves have no intrinsic meaning applicable alike to all situations." City of Chicago v. FPC, 458 F.2d 731, 750 (D.C.Cir.1971) (quoting City of Detroit v. FPC, 230 F.2d 810, 815 (D.C.Cir.1955)), cert. denied, 405 U.S. 1074, 92 S. Ct. 1495, 31 L. Ed. 2d 808 (1972). We therefore seek guidance from basic principles developed by the judiciary in furtherance of its task of assuring that ratemaking agencies conform to their duty to prescribe just and reasonable rates.42

We begin from this basic principle, well established by decades of judicial review of agency determinations of "just and reasonable" rates: an agency may issue, and courts are without authority to invalidate, rate orders that fall within a "zone of reasonableness," where rates are neither "less than compensatory" nor "excessive." See, e.g., FERC v. Pennzoil Producing Co., 439 U.S. at 517, 99 S.Ct. at 771; Permian Basin Area Rate Cases, 390 U.S. at 797, 88 S.Ct. at 1375.

When the inquiry is on whether the rate is reasonable to a producer, the underlying focus of concern is on the question of whether it is high enough to both maintain the producer's credit and attract capital. To do this, it must, inter alia, yield to equity owners a return "commensurate with returns on investments in other enterprises having corresponding risks," as well as cover the cost of debt and other expenses.... [W]hen the inquiry is whether a given rate is just and reasonable to the consumer, the underlying concern is whether it is low enough so that exploitation by the [regulated business] is prevented.

City of Chicago, 458 F.2d at 750-51 (emphasis in original). The "zone of reasonableness" is delineated by striking a fair balance between the financial interests of the regulated company and "the relevant public interests, both existing and foreseeable." Permian Basin Area Rate Cases, 390 U.S. at 792, 88 S.Ct. at 1373; see, e.g., FERC v. Pennzoil Products Co., 439 U.S. at 519, 99 S.Ct. at 772; Trans Alaska Pipeline Rate Cases, 436 U.S. 631, 653, 98 S. Ct. 2053, 2066, 56 L. Ed. 2d 591 (1978).

The delineation of the "zone of reasonableness" in a particular case may, of course, involve a complex inquiry into a myriad of factors. Because the relevant costs, including the cost of capital, often offer the principal points of reference for whether the resulting rate is "less than compensatory" or "excessive," the most useful and reliable starting point for rate regulation is an inquiry into costs. See, e.g., Mobil Oil Corp. v. FPC, 417 U.S. at 305-06, 316, 94 S. Ct. at 2344-45, 2349; FPC v. Hope Natural Gas Co., 320 U.S. at 602-03, 64 S.Ct. at 287-88. At the same time, non-cost factors may legitimate a departure from a rigid cost-based approach. See, e.g., Pennzoil Products, 439 U.S. at 518, 99 S.Ct. at 771; Mobil Oil, 417 U.S. at 308, 94 S.Ct. at 2345. The mere invocation of a non-cost factor, however, does not alleviate a reviewing court of its duty to assure itself that the Commission has given reasoned consideration to each of the pertinent factors. On the contrary, "each deviation from cost-based pricing [must be] found not to be unreasonable and to be consistent with the Commission's [statutory] responsibility." Mobil Oil, 417 U.S. at 308, 94 S.Ct. at 2346; see Pennzoil Products, 439 U.S. at 518, 99 S.Ct. at 772. Thus, when FERC chooses to refer to non-cost factors in ratesetting, it must specify the nature of the relevant non-cost factor and offer a reasoned explanation of how the factor justifies the resulting rates.

In Williams, FERC departed from these established ratemaking principles. At the outset, we cannot square FERC's statutory responsibilities with its own, quite novel principle that oil pipeline ratemaking should protect against only "egregious exploitation and gross abuse," 21 FERC at 61,649 (emphasis added), "gross overreaching and unconscionable gouging," id. at 61,597 (emphasis added). Rates that permit exploitation, abuse, overreaching or gouging are by themselves not "just and reasonable." FERC itself overreaches the bounds of its statutory authority when it permits such oil pipeline rates, so long as they are not "egregious," "gross" or "unconscionable." Ratemaking principles that permit "profits too huge to be reconcilable with the legislative command" cannot produce just and reasonable rates. Public Service Commission v. FERC, 589 F.2d 542, 550 (D.C.Cir.1978).

We recognize, of course, that "non-cost" factors may play a legitimate role in the setting of just and reasonable rates. In Williams, FERC invoked the need to stimulate additional oil pipeline capacity as one reason for setting maximum rates at such high levels. See supra at 1494-95. As this court has observed before, "[r]eliance on non-cost factors has been endorsed by the courts primarily in recognition of the need to stimulate new supplies." Consumers Union v. FPC, 510 F.2d 656, 660 (D.C.Cir.1974) (footnote omitted) (discussing Permian and Mobil Oil ). However, in this case FERC failed to forecast or otherwise estimate the dimensions of the need for additional capacity, and did not even attempt to calibrate the relationship between increased rates and the attraction of new capital. See supra note 27.

In the absence of such a reasoned inquiry, we cannot countenance FERC's approval of oil pipeline rates which, by FERC's own admission, ensure "creamy returns" to the carriers, 21 FERC at 61,650, and are "far more generous than those [rates] that [FERC] or other regulators give elsewhere," id. at 61,646. In a similar context, this court explained:

If the Commission contemplates increasing rates for the purpose of encouraging exploration and development ... it must see to it that the increase is in fact needed, and is no more than is needed, for the purpose. Further than this we think the Commission cannot go without additional authority from Congress.

City of Detroit v. FPC, 230 F.2d 810, 817 (D.C.Cir.1955), cert. denied sub nom. Panhandle Eastern Pipe Line Co. v. City of Detroit, 352 U.S. 829, 77 S. Ct. 34, 1 L. Ed. 2d 48 (1956); see San Antonio v. United States, 631 F.2d 831, 851-52 (D.C.Cir.1980) (ICC action, adding seven percent above costs in setting rates, is arbitrary and capricious because it lacks "adequate justification for [the] choice of a particular increment above fully allocated costs"), rev'd on other grounds sub nom. Burlington Northern, Inc. v. United States, --- U.S. ----, 103 S. Ct. 1238, 75 L. Ed. 2d 471 (1983); Public Service Commission v. FERC, 589 F.2d at 553-54 (citing cases). In the Williams proceeding, FERC "made no attempt at all to verify the accuracy of its prediction that granting pipeline [rate] incentives will spur increased investment." City of Charlottesville v. FERC, 661 F.2d 945, 955 (D.C.Cir.1981) (Wald, J., concurring). Indeed, FERC here failed to make its prediction with any specificity beyond the bald statement that "[e]verybody agrees that the nation needs and will need more pipeline plant." 21 FERC at 61,614.

FERC also found another basis for its new and liberal interpretation of "just and reasonable" rates in what it labeled the "climate of opinion," prevalent in the early twentieth century, in favor of dismantling the Standard Oil trust. FERC believed that Congress initiated rate regulation of the oil pipelines out of a desire to eliminate prohibitive pricing practices by the Standard Oil Company, and from this belief concluded that the "just and reasonable" standard requires far less stringent rate regulation than the same statutory standard requires for other regulated industries, including those industries once regulated under the very same section of the Interstate Commerce Act. See supra at 1492-93; 21 FERC at 61,578-99; FERC Brief at 29-44. Accordingly, FERC felt that the Interstate Commerce Act permitted ratesetting at levels so high that they would "seldom be reached in actual practice." 21 FERC at 61,649. We cannot endorse this interpretation of FERC's statutory duties.

In some circumstances, the contrasting or changing characteristics of regulated industries may justify the agency's decision to take a new approach to the determination of "just and reasonable" rates. See, e.g., Permian Basin Area Rate Cases, supra. We find, however, that in this case FERC has not merely developed a new method for determining whether a rate is "just and reasonable"; rather, it has abdicated its statutory responsibilities in favor of a method that, by its own description, guards against only grossly exploitative pricing practices. See supra at 1502. FERC wrongly assumed that the statutory phrase " 'just and reasonable' ... is a mere vessel into which meaning must be poured." 21 FERC at 61,594. While we agree that the statutory phrase sets down a flexible standard, an agency may not supersede well established judicial interpretation that structures administrative discretion under the statute. An agency may not "pour any meaning" it desires into the statute. To accept FERC's view of its own latitude would be tantamount to holding that no standards accompany the delegation of ratemaking authority to FERC, and we think such a delegation would be impermissible. From the outset, however, we noted that the statute prohibits more than grossly abusive rates.

Furthermore, an examination of the relevant legislative history reveals that Congress intended to subject oil pipelines to the same general ratemaking principles that applied to other common carriers. The Hepburn Act of 1906 was enacted primarily to remedy defects in the original Interstate Commerce Act of 1887. Although the Act as passed in 1887 provided that "[a]ll charges made for any service rendered in the transportation of passengers or property ... shall be reasonable and just; and every unjust and unreasonable charge for such service is prohibited and declared to be unlawful," 24 Stat. 379, the Supreme Court ten years later held that the ICC lacked authority to prescribe rates, but instead could only declare whether charges set by the carriers were unreasonable or unjust in the context of granting reparations to injured shippers. ICC v. Cincinnati, New Orleans & Texas Pacific Railway Co., 167 U.S. 479, 17 S. Ct. 896, 42 L. Ed. 243 (1897) (the Maximum Rate Case ); see Trans Alaska Pipeline Rate Cases, 436 U.S. at 639, 98 S.Ct. at 2059. The Hepburn Act remedied this shortcoming by granting to the ICC express authority to set maximum rates to be observed by carriers prospectively. See 49 U.S.C. Sec. 15. In this context, the Congress, by amendment originating in the Senate, adopted the Lodge Amendment, which conferred common carrier status upon oil pipelines, thus subjecting oil pipelines to the ratemaking jurisdiction of the ICC.

It appears evident from the floor debates that oil pipelines were intended to be treated in the same fashion as other common carriers under the Interstate Commerce Act. "It appears to me," Senator Lodge said in support of his amendment, "that it is a plain injustice to the railroads of this country to put them all under the Interstate Commerce Commission, to make the most drastic regulations to control and supervise them, and leave out one of the greatest article of interstate commerce [i.e., oil transported through pipelines]." 40 Cong.Rec. 6365 (1906). "This amendment," he said a few days later, "makes the pipelines and the oil companies subject to all the provisions to the bill." Id. at 7009. Thus Congress chose consciously to regulate oil pipeline rates in accordance with the same principles devised contemporaneously in other provisions of the Hepburn Act, which, as we noted above, augmented the ICC's authority over all common carriers.

The legislative history furthermore evidences that the "just and reasonable" rates prescribed by the Congress in 1906 meant more than a ban on prohibitive pricing. Congress primarily wanted to authorize the ICC to set enforceable rates that would permit the carriers to earn a fair return, while protecting the shippers and the public from economic harm. As Senator Elkins put it:

[T]he present laws are executed and they are being enforced vigorously; but this, as I have said before, is no reason why there should not be the strictest regulation against excessive rates and abuses of every kind .... The aim of wise statesmanship should be to so adjust matters by proper legislation that the shipper and producer can make a fair profit on their products, the [carrier] a fair return for the service rendered, and the consumer get what he buys at a fair price.

Legislative History at 879. Discussions of what constituted a just and reasonable rate focused not upon prohibitive pricing practices, but instead on setting a fair price that would be neither excessive to the shipper nor threatening to the financial integrity of the carrier. See, e.g., id. at 854 (remarks of Senator Clay) (Under the "just and reasonable" standard, ICC must determine "whether or not the rate so fixed is confiscatory or not compensatory for the services performed."); id. at 859 (remarks of Senator Clay) ("Can the [ICC's] power be exercised either to oppress the roads or the shippers? Can this power be exercised either to wrong or injure the carrier or the shipper? .... Can the Commission fix a rate that would prevent the railroads from making operating expenses and denying to them just compensation for the services performed? I answer, 'No.' ... The object and purpose of this legislation is to make [carriers] do right and to make shippers do right."); id. at 880 (remarks of Senator Culberson) ("[T]he Supreme Court has held that the words 'just and reasonable' have relation both to the rights of the public and of the companies, and that the rate must be fixed with reference to the rights of each.").

Additional evidence of congressional intent can be found by examining the decision to delete from the original Hepburn bill the requirement that rates be "fairly remunerative" in addition to "just and reasonable." After quoting the definition of "remunerative" found in a contemporary Standard Dictionary--"Affording, or tending to afford, ample remuneration; giving good or sufficient return; paying; profitable"--Senator Culberson questioned whether the additional phrase served any useful purpose, and worried whether the phrase might "have exclusive reference to the interests of the companies," thus "liberalizing the rule [of 'just and reasonable' rates] rather than narrowing it or keeping it where it is under the common law and under the decisions of the Supreme Court." See id. at 880-81. As Senator LaFollette later elaborated:

The phrase "just and reasonable" has a clear and well defined meaning in the law. It measures what the public must pay. It measures all that the carrier is entitled to receive....

The words "fairly remunerative" are added. What office are they to serve? For what purpose are they introduced? Are they to add something to the rate? If that is the purpose, they should be stricken from the bill. The carrier is entitled to nothing more than a just and reasonable rate. If the words "and fairly remunerative" are not designed to increase the rate, then they serve no purpose and should go out.

Id. at 906. Eventually, the phrase was deleted from the bill, in part because the "fairly remunerative" standard was thought to add nothing to the already established "just and reasonable" standard,43 and in part out of a fear that the courts might wrongly interpret the phrase to permit higher rates.44

If the Congress believed that "fairly remunerative" rates were at best the same as "just and reasonable" rates, and if there was a prevalent concern that "fairly remunerative" rates could exceed the proper ratemaking standard applicable to common carriers, we then find it highly unlikely that Congress aimed its ratemaking provisions solely toward preventing extraordinary exploitation or prohibitive pricing practices. After all, no "fairly remunerative" rate would rise to the level of egregious exploitation. How, then, could a Congress, worried that the "fairly remunerative" standard might permit excessive rates, at the same time be willing to permit rates at any level so long as they are not grossly abusive? We are convinced that the Congress did not intend such a result.

While we recognize that the legislative history of the Lodge Amendment contains a number of references to the Standard Oil Company,45 we do not believe that those references somehow alter the meaning of the language in the ratemaking provisions of the Interstate Commerce Act as applied to oil pipelines. First, the nature of the industry to be regulated is a natural topic for discussion during debate, and at that time Standard Oil dominated the industry. Second, there is nothing else in the legislative history to suggest that the Congress intended the meaning of "just and reasonable" to be transfigured when applied to oil pipelines.46 To rely too heavily on the popular "climate of opinion" in 1906 as evidence of the congressional intent underlying the Interstate Commerce Act would be unwise. See generally Dickerson, Statutory Interpretation: Dipping into Legislative History, 11 Hofstra L.Rev. 1125 (1983). Indeed, the motives of legislators are uniformly disregarded in the pursuit for statutory meaning; it is the purpose or intent behind the statutory provision itself that is relevant. See 2A Sutherland's Statutes and Statutory Construction Sec. 48 (C. Sands 4th ed. 1973 & 1983 Supp.). Thus, even assuming arguendo that it was the popular spirit of trust busting that aroused the 1906 Congress, it does not follow that Congress devised a response directed solely and narrowly toward prohibitive pricing. Congress provided that oil pipelines, as common carriers, could lawfully charge only "just and reasonable" rates; it did not enact a special antitrust or prohibitive pricing provision for oil pipelines. Whatever the historical context of the Hepburn Act, we think that FERC's statutory interpretation overlooks the broad terms of the principal source of legislative intent, the statute itself. Even if the problem Congress addressed was prohibitive pricing, the solution ultimately devised requires that oil pipeline rates be just and reasonable.

Finally, FERC believed that the changes since 1906 in the economics of oil pipelines also justified its novel interpretation of its statutory responsibilities under the Interstate Commerce Act. FERC determined that the cost of pipeline transportation, relative to the price of oil, had become so insignificant that close regulation was not required. See supra at 1493-95. In addition, FERC found that competition in the oil pipeline business had served to keep prices down. See supra at 1494. FERC therefore concluded that oil pipeline ratemaking "can and should rely far more heavily on the market" and that rate regulation should be "peripheral to the pricing process." 21 FERC at 61,649. Accordingly, in FERC's opinion, oil pipeline ratemaking should merely set "ceilings that ... will seldom be reached in actual practice."

We believe that this apologia for virtual deregulation of oil pipeline rates oversteps the proper bounds of agency discretion under the "just and reasonable" standard. First, the fact that oil prices have skyrocketed does not repeal the statutory requirement that oil pipeline rates must be just and reasonable.47 Whether the purpose of oil pipeline rate regulation is "consumer protection" or "producer protection,"48 the statute requires meaningful rate regulation. As the ICC acknowledged, the statutory command controls, despite any dilution in direct impact on the consuming public:

In determination of the question whether rates are lawful, we cannot attach any controlling weight to the fact that [the pipeline] or their beneficial owners [the parent companies] have seen fit to pay charges from one pocket to the other or to operate their common-carrier and industrial property in such a manner that the carrier system is virtually a plant facility of the larger producing, manufacturing, and selling industry. These facts, if they be facts, are immaterial ... whatever the relations between the pipelines and the oil companies which beneficially own them, Congress requires all rates tendered to the public by these common carriers to be just and reasonable, and no more.

Reduced Pipe Line Rates and Gathering Charges, 243 I.C.C. 115, 141 (1940). Despite recent legislative proposals to deregulate the oil pipeline industry, Congress has not as yet altered its command to FERC.49 Accordingly, the fact that the price of oil to the ultimate consumer dwarfs the price of oil pipeline transportation "does not excuse deviation from the just and reasonable standard, for not even 'a little unlawfulness is permitted.' " Consumers Federation of America, 515 F.2d at 358 n. 64 (quoting FPC v. Texaco Inc., 417 U.S. 380, 399, 94 S. Ct. 2315, 2327, 41 L. Ed. 2d 141 (1974)).

Second, we find FERC's largely undocumented reliance on market forces50 as the principal means of rate regulation to be similarly misplaced. It is of course elementary that market failure and the control of monopoly power are central rationales for the imposition of rate regulation. See S. Breyer, Regulation and Its Reform 15-16 (1982). As Representative Knapp expounded in 1906:

It has been stated that rate making is the most complicated and difficult work connected with transportation. Doubtless that has been correctly stated, but whether so or not, it certainly is one of the most important. The contention that competition is a regulator of freight rates is not, in the main, tenable. That, by reason of combinations, has gradually ceased to be a controlling factor, and can not now, except in limited and exceptional cases, be depended upon, as controlling in regulating rates.

Legislative History at 677. The courts have echoed this observation, noting that "[i]n subjecting producers to regulation because of anti-competitive conditions in the industry, Congress could not have assumed that 'just and reasonable' rates could conclusively be determined by reference to market price." FPC v. Texaco, 417 U.S. at 399, 94 S.Ct. at 2327; see, e.g., Tennessee Gas Pipeline v. FERC, 606 F.2d at 1114.

We recognize that the market price of oil could, "in an individual case, coincide with just and reasonable rates" and may "be a relevant consideration in the setting of area rates; it may certainly be taken into account along with other factors." FPC v. Texaco, 417 U.S. at 399, 94 S.Ct. at 2327 (citations omitted). The Williams opinion, however, goes far beyond what we regard as rational or permissible assumptions about the relationship between "just and reasonable" rates and the market price.51

FERC's methodology, by its own admission, merely sets "ceilings seldom reached in actual practice," and permits "creamy returns" to oil pipelines. As we have explained above, such ratemaking does not comport with FERC's statutory responsibilities. FERC's methodology, therefore, exposes a range of permissible prices that would exceed the "zone of reasonableness" by definition, unless competition in the oil pipeline market drives the actual prices back down into the zone. But nothing in the regulatory scheme itself acts as a monitor to see if this occurs or to check rates if it does not. That is the fundamental flaw in the Commission's scheme. See Texaco, Inc. v. FPC, 474 F.2d 416, 422 (D.C.Cir.1972), approved in relevant part and vacated on other grounds, 417 U.S. 380, 94 S. Ct. 2315, 41 L. Ed. 2d 141 (1974).

Congress may indeed have imposed the requirement that rates be "just and reasonable" in order to restore the "true" market price--the price that would result through the mechanism of a truly competitive market--for purchasers of the regulated service or goods. See, e.g., FPC v. Texaco, 417 U.S. at 397-98, 94 S.Ct. at 2326-27; FPC v. Sunray DX Oil Co., 391 U.S. 9, 25, 88 S. Ct. 1526, 1535, 20 L. Ed. 2d 388 (1968). In setting extraordinarily high price ceilings as a substitute for close regulation, FERC assumed that, with the wide exposed zone between the ceiling and the "true" market rate, existing competition would ensure that the actual price is just and reasonable. Without empirical proof that it would, this regulatory scheme, however, runs counter to the basic assumption of statutory regulation, that "Congress rejected the identity between the 'true' and the 'actual' market price." FPC v. Texaco, 417 U.S. at 399, 94 S.Ct. at 2327. In fact, FERC's " 'regulation' by such novel 'standards' is worse than an exemption simpliciter. Such an approach retains the false illusion that a government agency is keeping watch over rates, pursuant to the statute's mandate, when it is in fact doing no such thing." Texaco v. FPC, 474 F.2d at 422.

Moving from heavy to lighthanded regulation within the boundaries set by an unchanged statute can, of course, be justified by a showing that under current circumstances the goals and purposes of the statute will be accomplished through substantially less regulatory oversight. See Black Citizens for a Fair Media v. FCC, 719 F.2d 407, 413 (D.C.Cir.1983). We recognize that this court has sanctioned dramatic reductions in regulatory oversight under, for example, the FCC and ICC licensing provisions, both of which require that the licensee operate in accordance with the "public interest." See id.; National Tours Brokers Association v. ICC, 671 F.2d 528, 531-32 (D.C.Cir.1982). In both cases, this court found that the agency adequately assured meaningful enforcement of the public interest standard. See Black Citizens, 719 F.2d at 413-14; National Tours, 671 F.2d at 533. In other cases, this court has refused to sanction administrative attempts to reduce regulation in the absence of a showing that the goals and dictates of statutes were not being honored. See International Ladies' Garment Workers' Union v. Donovan, 722 F.2d 795 (D.C.Cir.1983); Action on Smoking and Health v. CAB, 699 F.2d 1209 (D.C.Cir.), supplemented, 713 F.2d 795 (D.C.Cir.1983).

In this case, FERC failed to show that the rates resulting from its newly articulated ratemaking principles would necessarily satisfy the "just and reasonable" standard. FERC set rate ceilings which, if reached in practice, would admittedly be egregiously extortionate and then failed to demonstrate that market forces could be relied upon to keep prices at reasonable levels throughout the oil pipeline industry. As a result, we find that FERC's action contravenes its statutory responsibilities under the Interstate Commerce Act.

In the foregoing analysis, we found the general ratemaking principles that guided FERC in the Williams opinion to be "in excess of statutory jurisdiction, authority, or limitations," 5 U.S.C. Sec. 706(2)(C), and "not in accordance with law," id. Sec. 706(2)(A). Because "an agency's action must be upheld, if at all, on the basis articulated by the agency itself," we would remand this case to FERC on the basis of the foregoing considerations alone. Motor Vehicle Manufacturers Association, 103 S. Ct. at 2870; see SEC v. Chenery Corp., 332 U.S. 194, 196, 67 S. Ct. 1575, 1577, 91 L. Ed. 1995 (1947). As independent grounds for our decision today, however, and in light of the apparent need for judicial guidance in this case,52 we further hold that the Williams opinion was not "the product of reasoned thought and based upon a consideration of relevant factors." Specialty Equipment Market Association v. Ruckelshaus, 720 F.2d 124, 132 (D.C.Cir.1983). Accordingly, we now turn to examine the particulars of FERC's oil pipeline ratemaking formula.

In Williams, FERC decided to adhere to the rate base formula it inherited from the ICC. See 21 FERC at 61,632. It gave no rational justification for doing so, however. FERC acknowledged that "rigorous logic and Euclidean consistency are not the system's most striking features," and that the formula is "much too blunt and much too clumsy for close work." It nevertheless concluded that the ICC method is "usable" because oil pipeline ratemaking "is not close work." Id. at 61,616. This is not a sufficient justification.53

It is well established that an agency has a duty to consider responsible alternatives to its chosen policy54 and to give a reasoned explanation for its rejection of such alternatives. See, e.g., Motor Vehicle Manufacturers Association, 103 S. Ct. at 2869-71; International Ladies' Garment Workers' Union, 722 F.2d at 815. This responsibility becomes especially important when the agency admits its own choice is substantially flawed. We find that FERC failed to satisfy this duty with respect to certain proposed modifications in the rate base formula.

Many parties to the Williams proceeding--including the FERC staff, the Department of Energy, the Justice Department, Farmers Union Central Exchange--advocated the calculation of oil pipeline rate bases by reference to original cost.55 These witnesses called for the rejection of the old ICC methodology, because its use of a weighted average of original cost and replacement cost, see supra at 1495, "lacks any economic rationale."56

Despite explicit concessions as to the shortcomings of the ICC rate base formula and the recognized advantages of a rate base formula derived from original cost,57 FERC rejected the original cost alternative. FERC offered four reasons for this decision. First, FERC wished to avoid the "headache" of analyzing the significance of guarantees--given by many parent oil companies to their subsidiary oil pipeline companies--in the estimation of the "true" capital structure of oil pipelines.58 See 21 FERC at 61,620-22. Second, FERC believed that the major regulatory benefit that might be derived from a switch to original cost accounting--the facilitation of comparable earnings analysis in relation to other businesses with a comparable risk to the pipelines--would not be useful in oil pipeline rate regulation, because the oil managers, as "professional risk takers," have ingrained attitudes toward risk and return unlike any other public utility investors. Third, an original cost rate base, without modification for inflation, would result in high initial rates that would decline as the rate base depreciates. FERC believed that competition in the oil pipeline business might prevent the pipelines from collecting the high initial rates, thereby preventing them from reaping their appropriate return on investment. See id. at 61,628-29. Finally, FERC found that any benefits resulting from changes in the rate base formula would not "warrant the social costs entailed," id. at 61,631, specifically, the construction of "transitional rate bases ... for each of the many common carrier oil pipelines," id. at 61,704 n. 376. We find that none of FERC's explanations for its rejection of an original cost rate base satisfies accepted standards of reasoned decisionmaking.59a. Parent Guarantees and Capital Structure

Because of parent companies' debt guarantees and "throughput and deficiencies" agreements, many shipper-owned pipelines are able to obtain debt financing more cheaply and in greater amounts than would be possible in the absence of such agreements. See supra note 58. Further, since cost of equity virtually always exceeds cost of debt, the greater the pipelines' debt ratio, the lower its overall cost of capital. See United States v. FCC, 707 F.2d 610, 613 (D.C.Cir.1983). Accordingly, as FERC recognized in its establishment of a "suretyship premium," see supra at 1496, the "real" cost of capital to a pipeline that benefits from such parent guarantees is greater than its apparent cost of capital.

Regulatory agencies have often assessed a regulated company's true cost of capital by constructing hypothetical capital structures, and then applying the normal costs of equity and debt to the hypothetical mix of securities. See Communications Satellite Corp. v. FCC, 611 F.2d 883, 902-09 (D.C.Cir.1977) (citing numerous cases involving water, gas, electric and telephone utilities). By this method, regulatory agencies ensure that the derived rate is "just and reasonable":

Although the determination of whether bonds or stocks should be issued is for management, the matter of debt ratio is not exclusively within its province. Debt ratio substantially affects the manner and cost of obtaining new capital. It is therefore an important factor in the rate of return and must necessarily come within the authority of the body charged by law with the duty of fixing a just and reasonable rate of return.

Id. at 903 (quoting New England Telephone & Telegraph Co. v. State, 98 N.H. 211, 220, 97 A.2d 213, 220 (1953)). In the case of oil pipelines, the hypothetical capital structure would be approximated by estimating the capacity of the pipeline to support debt in the absence of its parents' guarantees. See 21 FERC at 61,621.

FERC refused to adopt an original cost rate base in part because it believed that the attendant necessity for constructing hypothetical capital structures would be "a laborious exercise in guesswork, a venture 'into the unknown and unknowable.' " Id. at 61,622 (quoting Christiana Securities Co., 45 SEC 649, 668 (1974)). In FERC's view, such an inquiry would be:

a perfect field day for regulatory economists. Professor A would testify that he thinks 70% debt and 30% equity right. Professor B would say 53% debt and 47% equity. Professor C would come on strong for 50-50. Miss D from an eminent Wall Street investment banking firm would testify that her computer tells her that 65% equity and 35% debt are the right mix. Mr. E from an even more eminent investment banking firm would have numbers of his own.

Id. at 61,622. In part to avoid such an inquiry, FERC chose to avoid an original cost rate base.

This explanation runs counter not only to the proven practice of FERC and many regulatory agencies60 but also to FERC's own commentary later in the Williams opinion. As we have explained above, the technique of hypothesizing capital structures for oil pipelines would account for the increased capital costs associated with financing a pipeline in the absence of guarantees from the parents. Later in the Williams opinion, FERC devises its "suretyship premium" to compensate for the parents' guarantees of pipeline debt. FERC, however, appeared confident that any difficulties with estimating the value of this premium could be surmounted:

Credible expert testimony by persons associated with the rating services, the investment banking fraternity, and the credit insurance industry as well as by academics who have made a specialty of the bond market [can] establish[ ] that absent the parents' guarantee [what] the pipeline would have had to pay ....

Id. at 61,644.

We cannot square FERC's apparent confidence in its ability to estimate a pipeline's "suretyship premium" with its extreme skepticism about its ability to construct hypothetical capital structures. After all, the "suretyship premium" represents merely the differential between a pipeline's actual cost of capital and what its cost of capital would have been absent the parent guarantees. Thus the "suretyship premium" measures the same incremental cost of capital to the pipeline as the hypothetical capital structures that FERC felt incapable of estimating. The basis for FERC's preference for its "suretyship premium" approach, and for its aversion to hypothetical capital structures is therefore unclear. The decision to reject original cost accounting on the basis of this preference and aversion appears arbitrary, and, in any event, lacks sufficient explanation.

Moreover, even assuming that FERC's preference for its suretyship premium approach could be explained, its rejection of original cost ratemaking because of that preference relies on the assumption that original cost ratemaking is necessarily tied to hypothetical capital structures and necessarily incompatible with its newly devised "suretyship premium." However, FERC never gave any reason at all why this assumption is valid. Indeed, we see no reason why FERC could not account for the parent guarantees by using a suretyship premium added to an original cost ratemaking formula.

If FERC, in the exercise of informed discretion, decides that the suretyship premium approach is more reliable or easier to administer than hypothetical capital structures, then it should state why.61 As of now, neither FERC nor any of the parties has provided such an explanation. Even if they did so, however, we still would not understand why the hypothetical capital structure method must be used with original cost ratemaking, or why the "suretyship premium" approach cannot be used with original cost ratemaking.b. Comparable Risk Analyses

FERC discerned still "more fundamental problems" associated with the use of original cost ratemaking, beyond the estimation of appropriate capital structures. As typically applied under the "just and reasonable" standard, original cost ratemaking attempts to set the rate of return for a regulated enterprise at the same level as the rate of return of an unregulated enterprise with similar associated risks. See, e.g., FPC v. Hope Natural Gas Co., 320 U.S. 591, 603, 64 S. Ct. 281, 288, 88 L. Ed. 333 (1944) ("By that standard [of 'just and reasonable' rates] the return to the equity owner should be commensurate with returns on investments in other enterprises having corresponding risks."); Bluefield Water Works & Improvement Co. v. Public Service Commission, 262 U.S. 679, 692, 43 S. Ct. 675, 679, 67 L. Ed. 1176 (1923) ("A public utility is entitled to such rates as will permit it to earn a return ... equal to that generally being made at the same time and in the same general part of the country on investments in other business undertakings which are attended by the same risks and uncertainties."); A. Priest, Principles of Public Utility Regulation 191-94 (1969). FERC, however, believed that such a risk inquiry was not useful or relevant to oil pipeline ratemaking. In FERC's view, oil company managers--who own many oil pipelines--are a special breed of risk takers, who demand "a fair chance of earning as much on a pipeline as they would be likely to earn on something else in the unregulated sector" regardless of risk. 21 FERC at 61,623.62 Accordingly, FERC rejected original cost ratemaking in part because the conventional ratemaking inquiry that its use facilitates--the inquiry into risk--was, according to FERC, not helpful in oil pipeline ratemaking.

We think that this argument not only lacks any evidentiary support, it also lacks economic common sense. In neither the Williams opinion nor in its briefs to this court does FERC cite any evidentiary basis for its conclusion that oil managers will invest in only high return enterprises. In fact, the record is chock full of testimony regarding the risks of the oil pipeline business and the corresponding appropriate rate of return.63 Furthermore, major studies of the oil pipeline industry have concluded that the oil company managers decide whether to invest in a particular pipeline only after an examination of whether the expected returns match the associated risks:

When appraising the economic viability of a proposed pipeline venture, the approach taken is similar to that used by investors in general; it is what may be termed as required rate of return analysis. An oil company has widespread operations with numerous investment opportunities bearing different degrees of risk. Because of this, each investment, including pipelines, must be examined individually, and its expected rate of return compared with the opportunity rate of return of other prospective investments with comparable risk characteristics.

G. Wolbert, Jr., U.S. Oil Pipelines, 156 (1979) (footnotes omitted); see Exxon Pipeline Co./Exxon Co., U.S.A., Rates of Return on Petroleum Pipeline Investments, reprinted in Oil Pipelines and Public Policy 261, 268-69 (E. Mitchell ed. 1979) (" 'The required rate of return on an investment opportunity depends on the riskiness of the investment. The greater the riskiness of the investment, the more the return demanded by investors.' ") (quoting E. Solomon & J. Pringle, Introduction to Financial Management 332 (1977)).

ICC oil pipeline ratemaking precedents also belie FERC's novel notions about the relationship between risk and required return in the industry. FERC's notion that the oil companies demand high returns, no matter how low the risk, represents a radical departure from the ICC practice of evaluating risk and estimating the required return accordingly. See, e.g., Reduced Pipe Line Rates and Gathering Charges, 272 I.C.C. 375, 381 (1948); Minnelusa Oil Corp. v. Continental Pipe Line Co., 258 I.C.C. 41, 51 (1944); Reduced Pipe Line Rates and Gathering Charges, 243 I.C.C. 115, 131 (1940). Similarly, in 1978 this court called on FERC to reexamine the "complex of relevant factors" in determining the proper rates of return for oil pipelines, including the hazards prevailing in the pipeline business. See Farmers Union I, 584 F.2d at 419.

We thus find no basis to support, and overwhelming evidence to contradict, FERC's finding that comparable risk analysis has no important role in oil pipeline rate regulation. We therefore believe that FERC's rejection of original cost ratemaking on the basis of that finding is arbitrary and capricious.

c. The "Front-End Load" Problem

FERC next offered another, independent reason for rejecting original cost ratemaking: the "front-end load" problem.64 See supra at 1512. However, FERC itself acknowledged that this problem could be solved by using a trended, inflation-sensitive original cost rate base:

[W]e find the case for an inflation-sensitive oil pipeline rate base strong.

Such a rate base mitigates original cost regulation's income-bunching effect. It does not necessarily follow that the [old ICC rate base formula] is the ideal solution to the front-end load, income-bunching problem. Were we writing on an absolutely clean slate, were we beginning afresh in a brave new world, were pipelines a novelty that had just made their appearance, we would fashion an inflation-sensitive, anti-bunching rate base policy simpler and more logical than the ICC's.

21 FERC at 61,630. According to FERC, this "simpler and more logical" method would "[k]eep[ ] the rate base in tune with the general price level by linking it to the consumer price index or to the gross national product." Id. The trended original cost method of calculating rate bases, as discussed by witnesses in the Williams proceeding and other experts, fits this description. See, e.g., J.A. at 1508-12 (testimony of Stewart C. Myers on behalf of Marathon Pipe Line Co.); J.A. at 1957 (testimony of David A. Roach on behalf of MAPCO); Streiter, Trending the Rate Base, Pub. Util. Fort., May 12, 1982, at 32; cf. J.A. at 1677-1702 (testimony of Michael C. Jensen on behalf of ARCO Pipe Line Co.) (describing "inflation-adjusted original cost" method, the results of which are "equivalent to adjusting the rate base and depreciation by the unprojected inflation"). Indeed, at one point, FERC declared that if it were "beginning afresh on a clean slate [it] might be inclined to use something ... along the lines suggested by Marathon's witness Meyers [sic]." 21 FERC at 61,616. Marathon's witness Myers recommended the use of a trended original cost rate base if the old ICC method were to be abandoned. See J.A. at 1427, 1499. Thus FERC acknowledged that the front-end load problem could be solved, by adjusting an original cost rate base for inflation. Accordingly, FERC could not have reasonably relied upon the "front-end load" problem as a basis for rejecting the admittedly "simpler and more logical" trended original cost alternative.

d. The Social Costs and Benefits of Transition to a New Rate

Although a trended original cost approach would evidently be "simpler and more logical than the ICC's," 21 FERC at 61,630, FERC in the end rejected this alternative because of the "social costs entailed" in a transition from one rate base formula to another. See supra at 1512. FERC specified these "social costs" in an accompanying footnote:

Transitional rate bases would have to be constructed for each of the many common carrier oil pipelines. That would be a formidable, a difficult, and a costly endeavor. The task could be by-passed by using the most recent valuation (or in the alternative the cost of reproduction new less depreciation element of that valuation) as the transitional rate base. But then how much substantive change would there really be for existing pipelines? We conclude the change would be far more costly than it is worth.

Id. at 61,704 n. 376. We are reluctant to sanction the rejection of an admittedly more logical and accurate rate base formula on the basis of the conclusionary statement that the construction of "transitional rate bases" would be too costly. First, FERC failed to give a reasoned basis for its assumption that "[t]ransitional rate bases would have to be constructed" at all. Regulated industries have no vested interest in any particular method of rate base calculation. See FPC v. Natural Gas Pipeline Co., 315 U.S. 575, 586, 62 S. Ct. 736, 743, 86 L. Ed. 1037 (1942). Accordingly, as FERC acknowledged, a switch to a new rate base formula would not disrupt protected pipeline property. So long as the resulting rates are reasonable, the oil pipeline companies should have no difficulty maintaining their financial integrity. We are therefore at a loss to understand FERC's trepidation about a change in its regulatory method. Similarly, when this court granted FERC's request to remand this case "so that it may begin its regulatory duties in this area with a clean slate," Farmers Union I, 584 F.2d at 421, we specifically advised that the pipelines' reliance on an outdated rate base formula should not justify a continuation of the error. Rather, "the solution is not to perpet[u]ate that reliance but to end it prospectively, without allowing reparations based on its occurrence in the past." Id. at 419. We still adhere to that principle today.65

Second, FERC never explained why the construction of transitional rate bases would be so formidable a task. It is not self-evident why the calculation of such rate bases would entail more regulatory costs than the calculation of rate bases under the arcane ICC formula.66 Furthermore, the formulation of a method for calculating transitional rate bases involves questions no more complex than those confronting FERC regularly.

Finally, regardless of the regulatory or social costs entailed, FERC appeared to reject alternatives to the ICC formula because it found "no clear showing" that changing the methodology would "produce substantial social benefits." Id. at 61,626; see also id. at 61,703 n. 373. This finding, however, apparently relies upon FERC's antecedent findings that oil pipeline ratemaking should merely set price ceilings that would seldom be reached in actual practice, and that comparable risk analysis would not be helpful to the ratemaking inquiry for oil pipelines. However, we have found those antecedent findings to be defective. See supra at 1502-03, 1515-16. As a result, we likewise disapprove of FERC's finding that a new rate base formula could not produce any substantial social benefit.

After carefully reviewing the bases put forward by FERC for rejecting the original cost alternative, we hold that FERC failed to "examine the relevant data and articulate a satisfactory explanation for its action." Motor Vehicle Manufacturers Association, 103 S. Ct. at 2866. In our view it did not offer a reasoned explanation for adhering to an admittedly antiquated and inaccurate formula, but rather a host of unconvincing excuses that fail to add up to a rational choice.

2. The Association of Oil Pipelines' Recommendations

The Association of Oil Pipelines (AOPL) endorsed the ICC valuation approach to rate base calculations. See J.A. at 3870 (AOPL Opening Brief to FERC). AOPL, however, did not endorse the ICC approach in all its details. Instead, it asked FERC to make the following alterations to the ICC formula:(1) calculate reproduction costs for current expenses by reference to the current year's price index, or to an average of the indices for the most recent past year, the current year, and the next future year. Under the ICC method, costs are estimated by reference to a five-year "period index" consisting of the current year, one future year and three past years. APOL contended that this method understates actual current costs in times of inflation.

(2) increase the allowance for interest during construction employed in calculating the reproduction cost of pipeline assets. AOPL believed the six percent allowance was far too low to cover the prevailing rates to be paid during construction.

(3) calculate the present value of land and rights-of-way to account for their real appreciation in value over time. The ICC method calculates the "present value" of land at fifty percent of original cost and rights-of-way at original cost less depreciation. The AOPL claimed that such methods seriously undervalue the real present value of land and rights-of-way.

(4) adjust the construction damage allowance to reflect inflation up to the current year. AOPL argued that the ICC method, which adjusted the figures for inflation only from 1947 to 1953, understates actual costs.

(5) adjust the amounts assigned for pipe coating to reflect present prices. AOPL criticized the ICC method, which adjusted such costs for inflation only from 1947 to 1963.

(6) once the foregoing alterations are made, eliminate the six percent "going concern value" escalator to total valuation.

See J.A. at 3915-17 (AOPL Opening Brief). AOPL argued that these modifications "would improve the accuracy of the valuation rate base." Id. at 3917.

FERC rejected AOPL's proposals, finding that (1) only "relatively insubstantial" amounts were at stake, (2) the six percent going concern value roughly compensates for methodological errors elsewhere, and (3) the old ICC method should not be altered without first engaging in a notice and comment rulemaking on the proper method of calculating depreciation. See supra at 1496. AOPL argues to this court that FERC's rejection of its proposals was arbitrary and capricious agency action because it was "not supported by reasoned findings based on the evidence of record." AOPL Brief at 35-39. We agree.

We note at the outset that FERC failed, both in the Williams opinion and in its briefs to this court, to provide any factual basis in the record for its conclusion that "the sums involved are relatively insubstantial." 21 FERC at 61,631. On the other hand, AOPL cites unrebutted testimony in the record that the use of the ICC's "period indices" results in "consistently and substantially understated current valuations." J.A. at 1180 (testimony of John A. Jeter of Arthur Anderson & Co.). This same witness provided further unrebutted testimony that the ICC's allowance for interest during construction should be "much higher" in order to reflect current interest levels. See id. at 1183-85. Furthermore, in its brief, FERC states that the ICC rate base formula "significantly undercounts for interest during construction, several other construction-related elements, and the value of land."67 Indeed, in the Williams opinion FERC conceded that the AOPL proposals "may well be warranted" prospectively. 21 FERC at 61,631.

FERC, however, felt that the need for change was "far from pressing" because it believed that the six percent going concern value in a rough way compensated for the other flaws in the ICC methodology. Thus FERC rejected all of AOPL's objections on the grounds that the over -counting due to the going concern value--which would by itself be "pure water," id.--was in effect cancelled out by the under counting created by the methodological features that gave rise to the rest of AOPL's objections.

In basic terms, FERC reasoned that a series of inaccuracies is permissible because another inaccuracy systematically compensates for the prior errors. Such an approach, of course, assumes that the two errors are in fact predictably related to one another so that the anticipated self-correction will actually take place. In this case, however, FERC failed to make any finding to assure that the errors will offset each other. Especially when, as here, the proposed methodological adjustments appear easy to make, and the methodological defects are discrete, clear and acknowledged, FERC indulged an unreasonable presumption that its two wrongs would in practice render a right result. In the absence of any explanation of what warrants such an assumption, we find FERC's rejection of the AOPL proposals to be arbitrary and capricious.

Neither did FERC explain why its decision on the AOPL proposals should be delayed until it could conduct a notice and comment rulemaking on depreciation methods. FERC merely declared that "it would be wrong to alter the status quo without looking at the whole picture." Id. at 61,632. It is not at all apparent, however, why a decision on the AOPL proposals should be considered so intimately related to depreciation policy. FERC offered no rationale for its assumption that the changes proposed by AOPL should not be made separately from the decisions on depreciation policy. In fact, all of AOPL's proposals would apparently improve the accuracy of the rate base formula, regardless of the particular depreciation method employed. Thus, the adoption of the AOPL proposals would not seem to have any significant bearing on the future consideration of depreciation policy alternatives. FERC also made other similar adjustments to the rate base formula without examining "the whole picture." See FERC Brief at 71 n. 81. Moreover, FERC expressly declined to commit itself to ever conducting a rulemaking on depreciation issues:

To be fruitful, such a rulemaking should be preceded by intensive staff studies. The whole endeavor would be costly and time-consuming. Would it be worth the cost?

This question calls for further reflection. This is neither the time nor the place for that. We can ponder the point on another day.

21 FERC at 61,632. While we recognize that an administrative agency may exercise its informed discretion in deciding whether to proceed on a given issue by way of rulemaking or adjudication, see, e.g., NLRB v. Bell Aerospace Co., 416 U.S. 267, 294, 94 S. Ct. 1757, 1771, 40 L. Ed. 2d 134 (1974); SEC v. Chenery Corp., 332 U.S. 194, 203, 67 S. Ct. 1575, 1580, 91 L. Ed. 1995 (1947), we believe that in this case FERC failed entirely to make any such choice. Instead, FERC decided to delay implementation of the AOPL proposals, which it said were "well taken" and were deserving of "a hard look," id. at 61,631, until it could conduct a seemingly unrelated depreciation rulemaking, which it then said might never take place. Such self-contradictory, wandering logic does not constitute an adequate explanation for its rejection of admittedly valuable proposals.

In sum, we hold that FERC failed to explain adequately its rejection of both the original cost alternative and AOPL's proposed alterations. We emphasize that this holding does not go to the wisdom or efficacy of the ICC rate base formula, although the Williams opinion does not provide a cogent defense of it.68 Rather, our decision here turns on the inadequacies manifest in the decisionmaking process followed by FERC.

Even in the absence of such infirmities in FERC's method of choice among rate base methods, our review would still include scrutiny of the rate of return methodology, to see whether the selected rate of return, applied in combination with the selected rate base, leads to a reasonable result. As FERC observed, the agency must assure that "the combination of rate base and rate of return provides a[n] ... acceptable end result." 21 FERC at 61,616. We now proceed to examine whether FERC engaged in reasoned decisionmaking when it chose its rates of return for use in oil pipelines ratemaking.

FERC divided its rate of return into three components: (1) debt service, (2) the suretyship premium, and (3) the " 'real ' entrepreneurial rate of return on the equity component of the valuation rate base." 21 FERC at 61,644. The debt service element, which represents the cost of interest and repayment of indebtedness, gives rise to no objections from the parties, and need not detain us.

The suretyship premium similarly demands little comment apart from our previous observations that it requires much of the same kind of theorizing involved with the use of hypothetical capital structures. See supra at 1513-14. Farmers Union believes that FERC "erred when it assumed that such a premium is an 'add on' to the cost of capital without comparing pipeline and parent company risk." Farmers Union Brief at 59 n. 1. Our reading of the Williams opinion, and FERC's representations to this court, however, convince us that FERC made no such assumption, and, accordingly, pipelines must show that the guarantees reduce perceived investor risk in order to establish their entitlement to and extent of a suretyship premium. See 21 FERC at 61,621, 61,644, 61,711 nn. 492, 493; FERC Brief at 72-73.

Only the "real entrepreneurial rate of return on the equity component of the valuation rate base" remains. FERC began its discussion of this component from the premise that "[i]t seems obvious to us that allowed real rates of return on oil pipeline equity investments should be appreciably higher than those the Commission awards to natural gas pipelines and to wholesalers of electric energy." 21 FERC at 61,645. Considering that "oil companies [and the owners of the independent pipelines] have lots of places to put their money, ... and that the social need in this field is for returns high enough to induce the construction of new pipelines and to avert the premature abandonment of old ones," FERC enumerated the following eight measures of the rate of return on equity:

(i) Realized nominal rates of return on the book value of shareholders' equity in the oil industry generally over the past 5 years;

(ii) Realized nominal rates of return on the book value of shareholders' equity in the oil industry generally over the past year;

(iii) Realized nominal rates of return on shareholders' book equity in American industry generally over the past 5 years;

(iv) Realized nominal rates of return on shareholders' book equity in American industry generally during the most recent year;

(v) The particular parent or parents' realized nominal rate of return on total non-pipeline book equity over the past 5 years;

(vi) The particular parent or parents' realized nominal rate of return on total non-pipeline book equity in most recent fiscal year;

(vii) Total returns (dividends plus capital gains) on a diversified common stock portfolio over the past 5 years ...; and

(viii) Total returns (dividends plus capital gains) on a diversified common stock portfolio over the long run--25 years, 50 years, or more....

See 21 FERC at 61,645. FERC further held that "it would normally be proper to choose the measure most favorable to the particular carrier or carriers involved." Id.

Although most of these rates of return are expressed in terms of return on the book equity of unregulated companies, i.e., on the basis of original cost,69 FERC's methodology would nevertheless apply them, after an adjustment for "inflation," to the equity component of the ICC valuation rate base. Moreover, under FERC's methodology, the "equity component" is equal to the total valuation rate base, less the face value of the outstanding debt. See supra at ----. By this approach, the entire amount of appreciation in the rate base is allocated to the "equity component," while none of it is allocated to the debt component.

We frankly cannot locate the rhyme nor reason of this rate of return methodology; nor is it based upon a consideration of all relevant factors in oil pipeline ratemaking. To begin with, FERC offered no rational explanation that linked its regulatory purposes with its chosen rate of return indices. FERC made no attempt to estimate the risks involved with oil pipeline operations, and therefore could not reasonably estimate the rate of return required to maintain a viable oil pipeline industry. Moreover, in summary form, with a more elaborate discussion below, the "inflation adjustment" to the selected rates of return does not reliably compensate for the appreciation to the valuation rate base, and, therefore, overcompensation for inflation is not reliably prevented. FERC's willingness to permit the oil pipeline companies to choose among a wide variety of rate of return indices only makes these defects worse. FERC's method of calculating the "equity component" of the rate base further enlarges the allowable returns without good reason. As a result, the total returns allowable under FERC's methodology have no discernible regulatory significance beyond the fact that they are bound to be very large. FERC does not even offer an explanation of why its ratemaking formula sets "a cap of gross abuse," let alone a just and reasonable rate.

As previously discussed, FERC made no effort to study and estimate the risks associated with oil pipeline operations. Accordingly, FERC offered no reason to believe that the risks associated with the unregulated enterprises from which it derived its rates of return were equivalent to the risks of running an oil pipeline.70 Because the level of risk associated with an enterprise determines the returns it requires to attract capital, see supra at 1515-16, FERC never established a reasonable connection between its stated purpose to preserve the financial integrity and economic viability of oil pipelines and its selected rate of return indices.

FERC attempted to establish such a connection by arguing:

If the returns do not exceed those being realized somewhere or other in a roughly comparable segment of the economy's unregulated sector, it is hard to see how they can be branded extortionate or abusive.

Our relative permissiveness makes the risk problem more manageable. Can even the riskiest of pipelines argue that it is so hazardous that it is entitled to more than anyone makes any place else?

21 FERC at 61,645-46 (emphasis in original). The first sentence of this passage lacks any semblance of valid reasoning from the record. FERC never even attempted to establish that the relevant segments of the economy's unregulated sector were in fact "roughly comparable" to the oil pipelines. If the enterprises were "roughly comparable," the reference to them might be justified. FERC, however, assumed, without explanation, the existence of that factual predicate in order to justify its selected rate of return indices. Unfortunately, this assumption is not supported by any sound explanation based on the record, and therefore this attempted justification rests on nothing more than a blind, conclusionary assertion of "rough comparability."

The second paragraph in this passage makes use of a non sequitur. In preceding paragraphs, FERC had permitted the oil pipelines to choose a rate of return for themselves from a buffet bedecked with those found in a wide variety of lucrative unregulated enterprises. It is therefore pure illogic to assume that the "risk problem" is the spectre that the oil pipelines might claim entitlement to even greater rewards. As we have discussed above, the real "risk problem" with FERC's methodology--the problem FERC entirely failed to address--lies in whether FERC's selected indices grossly overestimate the risks and needed returns prevailing in the oil pipeline business.

2. The "Inflation Adjustment" and the "Double Counting" Problem

The problem of "double counting" for the effects of inflation, once in the rate base and again in the rate of return, has plagued oil pipeline ratemaking for some time. See, e.g., Farmers Union I, 584 F.2d at 419, 420-21; Williams Brothers Pipe Line Co., 355 I.C.C. at 487. The ICC rate base formula purports to account for inflation in valuing a pipeline's assets. See 21 FERC at 61,646; see also Farmers Union I, 584 F.2d at 421. If the chosen rate of return also reflected the effects of inflation, then the resulting return might compensate for inflation twice, and so would be excessive.

FERC attempted to eliminate the double counting problem by subtracting an "inflation allowance" from the nominal rate of return before applying it to the "inflation-sensitive" ICC valuation rate base. See 21 FERC at 61,646-47. Because the nominal rates of return are derived from original cost accounting, see supra at 74, they include a premium to compensate investors for the expected future rate of inflation. However, because the ICC valuation rate base is, according to FERC, already "inflation-sensitive," FERC's method should deduct from the nominal rate of return the percentage by which the valuation rate base has been "written up" during "the relevant period." Id. at 61,647. FERC defined "the relevant period" to be "the time period that was looked to in order to derive the appropriate nominal rate of return." Id. at 61,712 n. 511. For example, if the nominal rate of return were set by reference to returns on shareholder book equity over the most recent year, that nominal rate would be reduced by the percentage amount that the valuation rate base had increased over the most recent year. In this way FERC believed it could "avoid overcompensation for inflation." Id. at 61,646.

Farmers Union, among others, objects to this "inflation adjustment" on the ground that it does not compensate for actual inflation. It put forward strong evidence, including calculations made by Commissioner Hughes in his separate statement, to show that the valuation rate base does not track inflation in any predictable manner.71 See 21 FERC at 61,725 (Hughes, Comm'r, dissenting in part and concurring in part) ("A ... serious defect [in FERC's decision], and I believe, an uncorrectable one, is the unstated assumption that the trending of the rate base in the valuation formula approximates or should approximate the course of inflation.");72 see also Farmers Union I, 584 F.2d at 519 & n. 29; J.A. at 2455 (testimony of Thomas C. Spavins) (highlighting "the lack of a clear correspondence between [the ICC] valuation returns and any clear system of indexing returns for inflation").

FERC in a footnote anticipated such a criticism, and responded: "Suppose that [the ICC formula] does lead to an overly generous allowance for inflation in the rate base. What of it? The rate of return on equity is reduced by the precise amount of the overstatement." 21 FERC at 61,712 n. 513. This defense is sound, as far as it goes. Speaking precisely, FERC's "inflation adjustment" does not operate as an adjustment to compensate for the effects of inflation; rather, it operates as an adjustment to compensate for the effects of rate base appreciation, which, if left in the calculus, would lead to "double counting." The important feature of such a scheme is not that the rate of inflation and the rate of rate base "write up" are the same; instead, it is important only to assure that the increase in the rate base--which is affected and indeed justified by the fact that present values reflect inflationary effects--is not counted in calculating rates because expected inflation is already reflected in the level of rates of return. In simple terms, then, the "inflation adjustment" operates to write off the "write-up" in the valuation rate base through a deduction from the nominal rate of return. See 21 FERC at 61,646-47.

Unfortunately, however, and without explanation, FERC decided that the needed adjustment should be determined by reference to rate base appreciation during "the time period that was looked to in order to derive the appropriate nominal rate of return." See supra at 1524. This time period could range from "the most recent year" only, to "the long run--25 years, 50 years, or more." 21 FERC at 61,645; see supra at 1522. The allowable returns to the pipeline, by contrast, reflect the entire appreciation in the rate base over the life of the pipeline's assets. The "inflation adjustment," therefore, will not necessarily reflect the full rate of write up reflected in the rate base. Furthermore, it is likely that the "inflation adjustment" will leave in the final rates significant "double counting," because under FERC's method the oil pipelines are empowered to select for themselves the applicable rate of return index, and, as a corollary, they also select the time period relevant to calculating the "inflation adjustment." Accordingly, the FERC methodology allows the oil pipeline companies to select a time period during which the rate base appreciated at a slower rate than average. In this way, the FERC method permits the regulated companies to select the rate of return index that will result in an adjustment that understates the actual overall rate base appreciation. In Commissioner Hughes' words, the FERC method "invites an enormous amount of gamesmanship. Eight rate of return options are suggested, some with multiple choices of time periods. The inflation/valuation variance gives an exciting new twist to a pipeline's choice among the candidates. Thus a firm might choose to base its return one year on stock market performance after a bull market, and in its next filing switch to a high oil company comparison which might be offset by a small increase in its own valuation." 21 FERC at 61,726 (Hughes, Comm'r, dissenting in part and concurring in part).

3. FERC's "Equity Component" Has No Meaningful Relation to the Rates of Return on Book Equity

Even more capricious was FERC's application of the rates of return, representing revenues on the book equity of unregulated companies, to what FERC called the "equity component of the valuation rate base." As noted above, FERC's notion of the equity component includes the original paid-in equity of the pipeline plus the entire write up in the rate base. See supra at 1522. For example, consider an oil pipeline, originally financed with $900,000 debt and $100,000 equity. The original cost of the pipeline is one million dollars. Over time, the pipeline's valuation rate base increases to, say, $1,500,000. Under FERC's method, the equity component of the rate base amounts to $600,000, six times its book equity, even though the valuation rate base as a whole has appreciated only by half. Thus, FERC's method magnifies the "equity component" of the rate base to spectacular proportions, especially in an industry as highly debt-leveraged as the oil pipelines.73 See supra note 58. At the same time, however, FERC's selected rates of return reflect the revenues of the unregulated companies as a percentage of their book equity. To set allowable revenues for the oil companies, FERC took these rates of return and applied them to a completely different measure of net worth, the "equity component of the rate base." Book equity, unlike FERC's newly devised "equity component," represents the underlying net assets in original cost terms. Because book value of an equity share has no significance as to the present value of the company's assets, the returns on book equity likewise have no significance in relation to the equity component of the valuation rate base. See, e.g., J. Gentry, Jr. & G. Johnson, Finney & Miller's Principles of Accounting 367-68 (8th ed. 1980).

Assuming arguendo that the "inflation adjustment" accurately compensates for the rate of rate base appreciation, which it does not, see supra at 1524-25, such an adjustment would compensate only for the appreciation attributable to the portion of the rate base financed by the paid-in capital of equityholders. It would never compensate for the fact that FERC includes the entire appreciation on the rate base--attributable to both the equity and debt components of the pipeline--in its "equity component." Accordingly, FERC's method ensures that the allowable revenues for oil pipelines will exceed the revenues earned by its selected unregulated companies by the extent to which the pipelines' "equity component" exceeds the portion of the rate base financed through equity investments. Cf. 21 FERC at 61,712 n. 519 (under the "more austere standard of fairness," FERC "would trend only the equity portion of the rate base for inflation"). In most cases, this difference will be very large.74

Indeed, FERC provides no analysis of why its application of its selected rates of return to an unrelated measure of rate base equity should keep "a cap on gross abuse" in the resulting rates, not to mention the lack of any assurance that the resulting rates will be "just and reasonable." Commissioner Hughes appears to have rightly characterized FERC's game as Dialing for Dollars instead of The Price is Right. See 21 FERC at 61,730 n. 4 (Hughes, Comm'r, dissenting in part and concurring in part). We cannot condone such a ratemaking methodology, which assures nothing except that permissible rate levels will be very high.

In an attempt to defend the mismatch between its selected rates of return (on book equity) and its "equity component of the valuation rate base," FERC claimed that its method of calculating the "equity component" gives the equityholders the full benefit of debt leveraging. Just as a seller of a house benefits from the entire appreciation of the value of the house regardless of the amount of debt that financed the original purchase, FERC believed that so, too, should the equityholders in oil pipelines receive an "equity kicker" in their rate base. See 21 FERC at 61,648-50. This analysis overlooks the fact that oil pipeline companies are in fact free to sell their assets, and thereby enjoy the full benefit of debt leveraging in the difference between the sale price and the original cost of the assets. Such an "equity kicker," however, has no significant relationship with the determination of the cost of capital. A rate of return should set the proper rewards for investors in the form of current income, not asset appreciation and sale. FERC's attempted defense of its use of its "equity component" thus fails to meet minimal standards of reason.75

While the determination of a fair rate of return cannot and should not be constrained to the mechanical application of a single formula or combination of formulas, the ratemaking agency has a duty to ensure that the method of selecting appropriate rates of return are reasonably related to the method of calculating the rate base. See, e.g., FPC v. Hope Natural Gas Co., 320 U.S. 591, 605, 64 S. Ct. 281, 289, 88 L. Ed. 333 (1944); Dayton Power & Light Co. v. Public Utilities Commission, 292 U.S. 290, 311, 54 S. Ct. 647, 657, 78 L. Ed. 1267 (1934); NEPCO Municipal Rate Committee v. FERC, 668 F.2d 1327, 1342 (D.C.Cir.1981), cert. denied, 457 U.S. 1117, 102 S. Ct. 2928, 73 L. Ed. 2d 1329 (1982). Our disapproval of FERC's decision to retain the ICC rate base formula, see supra at 1520-21, did not turn on the substantive validity of the rate base calculations. FERC may adopt any method of valuation for rate base purposes so long as the end result of the ratemaking process is reasonable. See, e.g., FPC v. Natural Gas Pipeline Co., 315 U.S. 575, 586, 62 S. Ct. 736, 743, 86 L. Ed. 1037 (1942); NEPCO Municipal Rate Committee v. FERC, 668 F.2d at 1333; Washington Gas Light Co. v. Baker, 188 F.2d 11, 18 (D.C.Cir.1950), cert. denied, 340 U.S. 952, 71 S. Ct. 571, 95 L. Ed. 686 (1951). Rather, our disapproval arose out of the FERC's failure to give a reasoned explanation for its rejection of responsible rate base alternatives. We now find, however, as a result of the foregoing considerations, that the combination of FERC's rate base and rate of return methodologies does not produce an acceptable "end result." Accordingly, we disapprove FERC's ratemaking methodology on this additional basis.

As discussed supra at 1497, FERC rejected the Williams Company's attempts to use the purchase price of its assets in its rate base and depreciation basis calculations. FERC soundly held that the use of purchase price instead of original cost in rate base calculations would engender an undue incentive to trade pipeline assets at a high price, which, under a purchase price regime, would increase allowable rates.76 See 21 FERC at 61,635. Furthermore, in keeping with this court's remarks in Farmers Union I, FERC eliminated the use of purchase price as the basis upon which to calculate depreciation expenses. See id. at 61,635-36.77 As Williams' procedural and substantive objections to these rulings all lack merit, we approve FERC's decision to eliminate purchase price generally from oil pipeline ratemaking.78

B. Systemwide vs. Point-to-Point Rate Regulation

As discussed supra at 1497, FERC decided in Williams to regulate oil pipeline rates on a systemwide, rather than a point-to-point basis. FERC did so by way of a short discussion, on the assumption that the ICC had in the past given "scant attention to particular rates on specific routes." 21 FERC at 61,650. Farmers Union objects to this ruling. It challenges FERC's interpretation of past ICC precedents, citing ICC cases in which rates were determined by reference to specific point-to-point movements and their related costs and valuations. See Farmers Union Brief at 69. Farmers Union also noted that the Interstate Commerce Act requires "every unjust and unreasonable charge ... [to be] prohibited and declared to be unlawful." 49 U.S.C. Sec. 1(5) (emphasis added). Finally, it contends that systemwide rate regulation could shield rate discrimination from proper remedy.

Our review of relevant ICC precedents shows that past oil pipeline proceedings have included attempts to set rates "computed on a detailed allocation of costs to the proper section of the pipe-line system." Petroleum Rail Shippers' Association v. Alton & Southern Railroad, 243 I.C.C. 589, 663 (1941); see Minnelusa Oil Corp. v. Continental Pipe Line Co., 258 I.C.C. 41, 54-55 (1944). In both proceedings, the ICC allocated the operational costs of transportation from each originating station, averaged as to distance and weighted as to volume, to every terminal in the relevant system. Because oil pipelines rates are charged on a point-by-point basis, such cost allocation ensures that the costs of providing service over a given territory will be recovered only from the companies that use that particular service. See Minnelusa Oil Corp., 258 I.C.C. at 53 ("Operating conditions of defendant pipe lines in Rocky Mountain territory are more difficult than those of pipe line in territory east thereof, as hereinabove explained, but these are reflected for the most part in operating expenses."). We also find disturbing the apparent tension between FERC's action and the language of section 1(5). While FERC made assurances in Williams that patently discriminatory tactics will not be immunized from searching regulatory scrutiny, the FERC's systemwide approach would apparently tolerate substantial variance in allowable returns among pipeline segments without any justification, cost-based or otherwise.

However, we need not decide this issue at this time, because FERC made its decision prematurely. The ALJ identified the following issue for consideration during Phase I of the Williams proceeding:

Which unit should the Commission regulate (i.e., should the Commission determine rate base upon a system-wide or upon a segmented basis (e.g., petroleum products pipeline v. fertilizer pipeline))?

J.A. at 242 (Invitation to Submit Comments) (emphasis added). The ALJ designated this question as a "rate base issue." Id. at 241. FERC's ruling, however, went well beyond the determination of the rate base issue, and decided further to abandon all cost allocation to particular pipeline segments, calling the allocation inquiry "metaphysical, inconclusive and barren." 21 FERC at 61,651. Previous ICC cases make clear that the question whether to "determine rate base upon a system-wide or upon a segmented basis" is separate from the question whether costs should be allocated to particular pipeline segments. In those prior ICC cases, the rate base valuation was not broken down into line sections, but the ICC nevertheless proceeded to allocate costs to the proper sections of the pipeline. See Minnelusa Oil Corp., 258 I.C.C. at 54; Petroleum Rail Shippers' Association, 243 I.C.C. at 663. The rate base issue goes to the determination of the proper valuation units upon which a rate of return will be earned, and accordingly constitutes a proper element of the Phase I inquiry, which centered on how to calculate allowable revenue requirements for an oil pipeline. The cost allocation issue, by contrast, determines the fair distribution of the burdens of meeting those revenue requirements among the oil pipeline's customers. See Bonbright, Principles of Public Utility Rates 291-93 (1961). Thus, the cost allocation issue is more properly characterized as a question of rate design. See, e.g., Second Taxing District v. FERC, 683 F.2d 477, 480 (D.C.Cir.1982); Cities of Batavia v. FERC, 672 F.2d 64, 80 (D.C.Cir.1982).

The ALJ, however, expressly deferred rate design issues until Phase II of the proceedings. See J.A. at 243 (Invitation to Submit Comments) ("A number of additional issues, such as 'rate design' ... were suggested .... Those suggestions were not adopted because, in most instances, the issues raised appear to be more appropriate for consideration in Phase II of this proceeding."); id. at 245 (remarks of ALJ at outset of prehearing conference) ("Someone also raised the question of rate design. I consider those Phase II issues. Those issues tend to vary with the particular pipeline."). Accordingly, we find that FERC decided an issue not properly before it.79 On remand, FERC, if it so desired, could consider the cost allocation issue as a part of Phase I, but if it does so it should give adequate notice to the parties so that the issue can be fully debated before determination. In making a decision on cost allocation principles, FERC should be cognizant of the ICC's past cost allocation practices, and should accord appropriate consideration to the mandate of section 1(5).

As discussed supra at 1498, FERC decided in Williams to permit oil pipeline companies to decide for themselves whether or not to use tax normalization accounting, but in any event prohibited companies that choose normalization from including the resulting tax reserve accounts in their rate bases. AOPL challenges the latter ruling in the belief that the exclusion of deferred tax amounts from the rate base "would completely eliminate any benefits that would otherwise result from a carrier's election of accelerated depreciation." AOPL Brief at 42 (emphasis in original).

We think that this challenge misses the mark. Regardless of whether an oil pipeline may include tax reserve accounts in its rate base, tax normalization accounting would permit it to benefit from accelerated depreciation without having to flow those benefits through to its customers. Unregulated companies, of course, do not concern themselves with rate bases, and yet they choose accelerated depreciation solely because it permits them to defer a tax burden. The oil pipeline companies that choose normalization accounting also enjoy the benefit of tax deferral. The amount in the resulting deferred tax account can earn interest even if it is not included in the rate base. Accordingly, we reject AOPL's notion that FERC's ruling "completely eliminates" any normalization benefit.80

For the reasons set forth above, we remand this case to FERC. We hope and expect that FERC will accord to this case the high priority that it deserves. In light of its excessive long pendency, this case should be disposed of in a reasonably speedy manner. FERC may find it necessary to take additional evidence in light of this court's opinion, but in any event, FERC already has the benefit of an extensive record and should be able to issue a new order within the next twelve months.

We emphasize that FERC should give serious and thoughtful consideration to the admittedly difficult problems presented by this case. Throughout this opinion we intended to provide some important and basic guideposts to assist FERC in that mission. Most fundamentally, FERC's statutory mandate under the Interstate Commerce Act requires oil pipeline rates to be set within the "zone of reasonableness"; presumed market forces may not comprise the principal regulatory constraint. Departures from cost-based rates must be made, if at all, only when the non-cost factors are clearly identified and the substitute or supplemental ratemaking methods ensure that the resulting rate levels are justified by those factors. In addition, the rate of return methodology should take account of the risks associated with the regulated enterprise. It should not be forgotten, too, that the choice of a proper rate of return is only part of what should be an integrated ratemaking method, and accordingly FERC must carefully scrutinize the rate base and rate of return methodologies to see that they will operate together to produce a just and reasonable rate.

In all these respects, the original cost methodology, a proven alternative, enjoys advantages that should not be underestimated. FERC should reexamine this alternative, and others, in this proceeding which, after all, was instituted in order to take a fresh and searching inquiry into the proper ratemaking method for oil pipelines. In this way, we hope that FERC can meet its statutory responsibilities without any further undue delay.

So ordered.