Under the Natural Gas Act of 1938, a natural gas producer whose facilities are subject to Federal Energy Regulatory Commission (Commission or the FERC) jurisdiction may not “abandon” those facilities — i.e., “permanently reducef] a significant portion of a particular service,” Reynolds Metals Co. v. FPC, 534 F.2d 379 (D.C. Cir.1976) — unless the Commission first finds, after a hearing, either “that the available supply of natural gas is depleted to the extent that the continuance of service is unwarranted, or that the present or future public convenience or necessity permit[s] such abandonment.” 15 U.S.C. § 717f(b).
Until recently, the Commission employed a “comparative needs test” to determine whether “the present or future public convenience or necessity permit[s] ... abandonment.” Under this test, the Commis*224sion would weigh the needs of the current consumers of the gas against the needs of the proposed new consumers, with “the burden of proof ... on the applicant for abandonment to show that the ... public interest ‘will in no way be disserved’ by abandonment.” Transcontinental Gas Pipe Line Corp. v. FPC, 488 F.2d 1325, 1328 (D.C.Cir.1973) (quoting Michigan Consolidated Gas Co. v. FPC, 283 F.2d 204, 214 (D.C.Cir.), cert. denied, 364 U.S. 913, 81 S.Ct. 276, 5 L.Ed.2d 227 (1960)), cert. denied, 417 U.S. 921, 94 S.Ct. 2629, 41 L.Ed.2d 226 (1974)
In a 1985 opinion granting a natural gas producer’s abandonment petition, the FERC changed its abandonment policy. Instead of comparing the needs of the current consumers against the needs of identified specific new consumers, the Commission announced that it would now compare the needs of the current consumers against the benefit that would accrue to the natural gas market as a whole were the facilities in question released from Commission jurisdiction. Consolidated Edison Company of New York (Con Ed), a local distribution company (LDC) that formerly purchased some of the gas at issue here, petitioned this court to review the FERC’s new abandonment test.1
Although we agree with the FERC that the statutory “public convenience or necessity” language frees it to develop new policies to accommodate a shifting energy marketplace, we nonetheless reverse and remand to the Commission because of its reliance upon an erroneous factual premise as a critical justification for its new policy.
I. Background
A. Legislation and Regulation
During the Depression, the few pipelines that existed exerted double-edged control over the natural gas market. As both monopsonists and monopolists, the pipelines could buy and sell natural gas according to their own whims, with producers and consumers caught in the resultant squeeze, both as to price and supply.
The Natural Gas Act of 1938 changed all that. For the first time, the Federal Power Commission (FPC or Commission) was authorized to regulate the rates at which natural gas could be bought and sold. The FPC’s jurisdiction extended only to the interstate market, but since gas that traveled interstate at any point was included in the Commission’s jurisdiction, see California v. Lo-Vaca Gathering Co., 379 U.S. 366, 85 S.Ct. 486,13 L.Ed.2d 357 (1965), its authority covered a broad span.2
The abandonment provision was one aspect of Congress’ scheme to protect natural gas consumers from exploitation:
No natural-gas company shall abandon all or any portion of its facilities subject to the jurisdiction of the Commission, or any service rendered by means of such facilities, without the permission and approval of the Commission first had and obtained, after due hearing, and a finding by the Commission that the available supply of natural gas is depleted to the extent that the continuance of service is unwarranted, or that the present or future public convenience or necessity permit such abandonment.
*22515 U.S.C. § 717f(b). Under this provision, before a natural gas producer can sell gas outside the Commission’s jurisdiction after its interstate contract expires, it must seek the Commission’s approval. In an energy market in which natural gas supplies were problematic and in which the unregulated intrastate markets threatened to swallow up large portions of those supplies, the abandonment provision served as a brake on the movement of gas out of the interstate stream.
The Commission, of course, had the responsibility to formulate a more precise abandonment test than “public convenience or necessity.” In Michigan Consolidated Gas Co. v. FPC, 283 F.2d 204, 214-16 (D.C. Cir.), cert. denied, 364 U.S. 913, 81 S.Ct. 276, 5 L.Ed.2d 227 (1960), and Transcontinental Gas Pipe Line Corp. v. FPC, 488 F.2d 1325, 1328-29 (D.C.Cir.1973), cert. denied, 417 U.S. 921, 94 S.Ct. 2629, 41 L.Ed.2d 226 (1974), this court approved the FPC’s comparative needs test:
The fundamental tenets established by Michigan Consolidated are that the public interest is the ultimate criterion under § [717f(b)] and that all factors relevant to the determination of which course of action best promotes the overall public interest must be fully considered. The Commission there proposed, and this court accepted, a comparative needs test which balanced the relative needs of the competing pipelines and their ultimate consumers as the principal index of the public interest.
Transcontinental, 488 F.2d at 1328.
In the late 1970’s the Natural Gas Policy Act (NGPA) reshaped the landscape of natural gas regulation. The Act was the product of two forces, an energy shortage, caused in large part by the Organization of Petroleum Exporting Countries’ control over foreign imports, and an emerging view on Capitol Hill that deregulation of many heavily controlled national economic markets was needed to stimulate the nation’s growth and development.
The NGPA itself reflected a compromise between widely divergent views on how best to protect consumers: some legislators believed that moderate price controls were still necessary to ward off residual oligopolistic tendencies in the natural gas industry, while others were convinced that only sweeping deregulation could produce a booming agora of new supplies and lower prices. The lengthy debates that preceded the NGPA’s passage in 1978 focused almost exclusively on the question of what degree of deregulation of “new” gas was appropriate. All of the legislators agreed that the regulatory structure of the NGA would continue to govern pre-NGPA “old” gas; since production-investment decisions with regard to that gas had already been made, there was no need to calibrate the price of that gas to the costliness of its production. Thus, although the NGPA ushered in a deregulated natural gas market for new gas, it explicitly left gas that was already “committed or dedicated to interstate commerce,” 15 U.S.C. § 3301(18), under the watchful eye of the Commission (now dubbed the Federal Energy Regulatory Commission, or the FERC). See 15 U.S.C. § 3431(a)(1)(A) (providing that the NGA shall not apply to natural gas not committed or dedicated to interstate commerce by date of passage). As to this old gas, the abandonment provision remained intact.3
B. The Proceedings in this Case
On April 7, 1959, Felmont Oil Corporation and Esse Offshore, Inc. (Applicants), the producer-sellers who seek abandonment in this case, agreed with other producer-sellers to sell to Transcontinental Gas Pipe Line Corporation (Transco) natural gas produced from Block 86, Vermilion Parish, Offshore Louisiana. On August 23, 1963, the Commission issued a certificate of public convenience and necessity under the Natural Gas Act, § 717f(c), authorizing natural gas produced from Block 86 to be sold in interstate commerce to Transco for resale. When the 1959 contract expired in September, 1981, the parties continued their arrangement first under a one-month and then under a day-to-day contractual *226basis, subject to cancellation on written notice.
Applicants, whose share of the Block 86 gas is just over fifty percent, notified Transco of their intent to cancel this arrangement on October 7,1983, and simultaneously filed an abandonment application with the FERC. Since Transco was taking only about twenty percent of its deliverable gas, applicants sought to increase their revenue by selling this shut-in gas on the spot market; they argued that the natural gas market would be better served if low-cost shut-in gas could be added to the spot market supply stream rather than kept in the ground as reserves for Transco and its customers.
Notice of the abandonment application was published, 48 Fed.Reg. 49,354 (1983), Transco filed a timely motion to intervene, and Con Ed, among others, filed a late notice to intervene, which was subsequently granted by the Administrative Law Judge (ALJ). Con Ed, one of Transco’s customers, argued that if abandonment were allowed Transco’s costs would rise, because it would have to replace the abandoned gas with other, more expensive gas, and its reserves would fall, because the shut-in gas previously available for future use would now be lost immediately to the spot market. Both the cost increase and reserves decrease could be expected to hurt Transco’s customers down the line.
The ALJ rejected the abandonment application, 32 F.E.R.C. ¶ 63,071 (1985), on the ground that the applicants had not specifically identified new customers who would benefit from the released gas if abandonment were permitted, whereas Transco's old customers would be definitely harmed by an increase in costs and a decrease of reserves. Thus, under the Commission’s established comparative needs test, the abandonment application had to be rejected.
The Commission, on review, reversed the AU and granted a limited-term abandonment,4 announcing a new abandonment policy. 33 F.E.R.C. 1161,333 (1985) (Opinion No. 245). Its prior comparative needs test, the FERC said, which was calculated to protect interstate supply and price, no longer made sense in the current deregulated market: “[w]hile the NGA was clearly designed to assure an adequate and reliable supply of natural gas in the interstate market at reasonable prices, the introduction of deregulation and competition into the market has altered the manner in which this can best be accomplished.” Joint Appendix (J.A.) at 292.
Recognizing the dramatic nature of the change in agency policy that it was announcing,5 the FERC detailed its reasons for the shift:
The situation now on many pipelines is that low cost supplies of gas are being shut-in in favor of higher takes of more expensive supplies. The pipelines are taking the more expensive gas because it is generally subject to more onerous take or pay requirements than the low cost gas. The pipeline has an economic incentive to purchase the more costly gas and thus avoid take or pay expense for that gas. The Commission believes that in adopting an abandonment policy which frees low priced gas to enter the market*227place, the marketing of supplies of low-priced gas, especially supplies which are currently either shut-in or subject to substantially reduced takes, will have several salutory effects. It will give other purchasers an opportunity to lower their gas costs by displacing high-cost gas or other fuels with this cheap gas. And to the extent this cheap gas creates more competition in the marketplace of suppliers, it will exert a pressure on all sellers of gas, be it high or low priced gas, to reduce prices or risk being shut-in. If the gas dedicated to an interstate pipeline is likely to be subject to a request for abandonment, that pipeline would be more likely to take this gas, thus lowering its overall cost of gas, or risk losing it forever. Finally, if cheap gas displaces more expensive gas on a pipeline’s system, both the pipeline and high-cost gas producer have an incentive to renegotiate their contracts to reduce take or pay requirements and lower the price of the gas, so that it will make economic sense for the pipeline to continue to buy that gas.
J.A. at 294-95. After giving these economic reasons to support a more liberalized abandonment policy, the FERC announced its new test:
[Wjhere a party can demonstrate that abandonment in a particular instance would have beneficial effects on the market overall, such as increasing competition and causing gas prices to respond to that competition, and the benefits of the abandonment outweigh any adverse effect to the purchaser to whom the gas is presently dedicated, or that purchaser’s customers, the Commission will grant abandonment. On a case by case basis, however, such abandonment may be conditioned in order to mitigate the loss to the dedicated purchasers.
J.A. at 295. The Commission next claimed that only one aspect of its former test was being altered:
By adopting this new policy, the Commission does not reject the prior abandonment policies in toto. The list of factors which the Commission considered in the past, such as environmental and economic consequences of abandonment, the parties’ contract arrangements, and the parties’ comparative needs, will still be weighed. The real change from the past will be a shift in the identification of the public interest, from the interest of only specific customers to the interests of the market as a whole, and in the determination of how the public’s needs are best served.
J.A. at 295.
Applying its new test to this case, the FERC rejected the AU’s finding that an abandonment would increase Transco’s— and hence its customers’ — cost of gas, reasoning that “[i]n view of Transco’s ability to nominate any Block 86 gas it needs, and to continue to take this gas at its relatively low prices, Transco’s [cost of gas] should not be affected by this grant of abandonment.” J.A. at 297-98. In other words, since under the terms of the FERC’s order Transco could fill any of its current needs by nominating the requisite amount of Block 86 gas on a six-month basis, see note 4, supra, Transco would never be forced to purchase more expensive replacement gas. With regard to the problem of declining reserves, the FERC said that “Transco’s ability to nominate and continue to purchase this gas [on a six-month basis] mitigates any problems it might experience because of the effect of this abandonment on its declining reserves-to-deliverability ratio.” J.A. at 298.6 The Commission also found that the abandonment “will have a *228number of positive effects on the market as a whole.”7 J.A. at 298.
On rehearing, the FERC stuck to its initial decision. 34 F.E.R.C. ¶ 61,296 (1986) (Opinion No. 245-A). It cited the various factors that would still be taken into account (“the economic consequences of abandonment, the contractual arrangements between the parties, the parties’ needs for the gas, and the effects of the proposed abandonment on the market as a whole,” J.A. at 357), and it characterized once again the harm to Transco and its customers as “minimal.” J.A. at 357. The economic benefits to the marketplace from the new policy were described, in now familiar fashion, as follows:
Increased purchases of this regulated gas will, by necessity, displace purchases of higher-cost supplies. As purchases of higher-cost gas decline, the producers of this gas will have a strong incentive to decrease their prices in order to preserve their market share. Additionally, while Applicants in this case may not be under a direct compulsion as a result of the abandonment in [the initial decision] to reduce take-or-pay provisions in their existing contracts with Transco, we. believe that the new abandonment policy, on the whole, will have that effect. [Petitioner’s] arguments notwithstanding, we expect reductions in take-or-pay levels to become a quid pro quo for a producer to obtain its pipeline’s concurrence and cooperation in sought-after abandonments.
J.A. at 358. Con Ed filed a petition for review in this court.
II. Decision
A. The Commission’s Grounds for the New Abandonment Policy
At the outset, we agree with the FERC that the NGA itself does not mandate a comparative needs test but only a determination that present or future public convenience and necessity permits the abandonment at issue. In other words, by delegating abandonment power in such broad terms, Congress expected that the Commission would develop an appropriate test to fit the regulatory climate. See Motor Vehicle Manufacturers Association v. State Farm Mutual Automobile Insurance Co., 463 U.S. 29, 42, 103 S.Ct. 2856, 2866, 77 L.Ed.2d 443 (1983) (“[A]n agency must be given ample latitude to ‘adapt [its] rules and policies to the demands of changing circumstances’ ”) (quoting Permian Basin Area Rate Cases, 390 U.S. 747, 784, 88 S.Ct. 1344, 1369, 20 L.Ed.2d 312 (1968)); NAACP v. FCC, 682 F.2d 993, 998 (D.C.Cir. 1982) (“Agencies may modify or repeal existing policies and rules as the conditions which they address change, ... and a court should defer to the [agency] as to whether conditions have in fact changed and what action, if any, is now warranted. But ... the court should be satisfied both that the agency was aware it was changing its views and has articulated permissible reasons for that change, and also that the new position is consistent with the law.”).
As the Commission acknowledges, though, it is the reviewing court’s duty to ensure “that the agency has ... really *229taken a ‘hard look’ at the salient problems, and has ... genuinely engaged in reasoned decisionmaking.” Greater Boston Television Corp. v. FCC, 444 F.2d 841, 851 (D.C.Cir.1970), cert. denied, 403 U.S. 923, 91 S.Ct. 2229, 29 L.Ed.2d 701 (1971). The issue before us, accordingly, is not whether the FERC may change its abandonment policy, but whether its reasons for doing so in its chosen manner are permissible ones. In its critical discussion of the economic factors favoring a liberalized abandonment policy, the FERC relied heavily upon the “several salutory effects” resulting from the release of shut-in gas to the spot market:
1) “It will give other purchasers an opportunity to lower their gas costs by displacing high-cost gas or other fuels with this cheap gas.”
2) “[T]o the extent this cheap gas creates more competition in the marketplace of suppliers, it will exert a pressure on all sellers of gas, be it high or low priced gas, to reduce prices or risk being shut-in.”
3) “If the gas dedicated to an interstate pipeline is likely to be subject to a request for abandonment, that pipeline would be more likely to take this gas, thus lowering its overall cost of gas, or risk losing it forever.”
4) “[I]f cheap gas displaces more expensive gas on a pipeline’s system, both the pipeline and high-cost gas producer have an incentive to renegotiate their contracts to reduce take or pay requirements and lower the price of gas, so that it will make economic sense for the pipeline to continue to buy that gas.”
J.A. at 294-95.8 We review these reasons seriatim:
1) On a theoretical level, the FERC’s first reason is sound: an increase in supply of a product will normally lower the cost of that product to consumers. We do not doubt that if the natural gas spot market were an ideal free market, the release of shut-in cheap gas to that market would permit gas purchasers to switch from higher priced gas. Even in the present imperfect natural gas market, there are some fuel-switchable end-users — such as power plants and large industrial companies — that can go with the flow, as it were, and adjust their purchasing portfolios to take advantage of this influx of cheaper gas to the spot market. But at the same time the FERC has not explained how other, captive consumers in the natural gas market will be able to benefit from the new abandonment policy. These less mobile consumers are geographically bound to pipelines that are in turn contractually bound by take-or-pay contracts that make it uneconomical for them to switch to a new market for natural gas, since the pipelines must still pay for the higher priced contractual gas whether or not they take delivery.9
*230In short, the natural gas market is not a classic free market, where the laws of supply and demand operate ineluctibly for all participants. The FERC’s apparent assumption that such standard laws apply across-the-board — or at least its failure to acknowledge the substantial obstacles to free-flowing gas exchange — is troubling. Nonetheless, at least the FERC’s first reason for its new policy is valid with regard to some consumers, the fuel-switchable end-users, who will be able to lower their gas costs by virtue of the increased supply of cheap gas.
2) The Commission’s second reason— that the competition of the new gas will pressure all sellers to lower their prices or risk being shut-in — appears to be merely a corollary of the first, assuming that all the FERC means by a “risk [of] being shut-in” is the risk of not being able to sell one’s gas if purchasers can buy cheaper gas on the spot market.10 As with the first reason, the implication is that an increase in supply of cheaper gas will lower costs in the end throughout the marketplace. But also as with the first reason, only those consumers who can actually switch to spot market purchases will apply any pressure on producer-sellers. Additionally, as we discuss below with regard to the fourth reason, producer-sellers who benefit from the safety of long-term take-or-pay contracts will feel little pressure from a dip in the spot market price.
3) We have serious doubts about the validity of the FERC’s third reason — that pipelines threatened with abandonment under a more liberal policy will purchase more gas. The FERC’s assumption that a pipeline that has hitherto taken only a small percentage of deliverable gas will decide to take most or all of the gas in order to avoid losing it forever only makes sense if demand increases to accommodate such takes, a scenario that seems likely to occur only if enough fuel-switchable energy buyers move from non-gas supplies to the gas market in response to the influx of previously shut-in gas. However, if demand does not increase to cover the volume of abandoned gas that the FERC assumes the affected pipelines will purchase, then those pipelines would be taking low-cost abandoned gas in lieu of high-cost take-or-pay gas while still incurring take-or-pay liability, thus raising, not lowering, their cost of gas. As we discuss in response to the FERC’s fourth reason, below, this apparent Commission confusion regarding the effect of its new policy on the take-or-pay problem is a fundamental and pervasive flaw that infects the major part of its reasoning on abandonment.
4)As we have indicated at the end of our discussions of the FERC’s second and third reasons, the FERC’s assumption that the take-or-pay problem will be alleviated by the new policy is highly problematic. The Commission's fourth reason lies at the center of our apprehension in this regard. In both its initial and rehearing opinions,11 the FERC relied heavily (see quotations in part I.B. and note 7, supra) on the argument that the existence of more cheap gas on the spot market (which is where the abandoned gas will go) will provide an incentive both to the pipelines and the producers to renegotiate the take-or-pay contracts. The Commission’s logic seems to be that the newly released, cheaper gas will entice the pipelines away from the more expensive take-or-pay contract gas, and that, as a result, the producers will be willing to renegotiate those contracts to keep their customers.
*231This reasoning is difficult to understand. A take-or-pay contract is the product of a deal between a producer and a pipeline whereby the pipeline gains the security of a long-term supply contract in exchange for guaranteeing its supplier that it will pay for a minimum volume of gas whether or not it takes delivery of that gas. One of the major problems facing the natural gas industry today is the burden of such expensive take-or-pay contracts on the pipelines in a market with cheap alternative energy supplies.12 We sympathize with the Commission’s desire to alleviate this problem, but it will not be wished away. It seems counterintuitive to argue that pipelines will stop taking gas that they have to pay for anyway and that as a result producers will have an incentive to renegotiate contracts in which they are guaranteed high payments whether or not the customers take the gas.13 On the contrary: The *232whole purpose of take-or-pay contracts is to give the producers the same benefit whether or not the gas in question actually leaves the ground.
This is not the first time that the FERC has assumed without explanation that in liberalizing one aspect of the natural gas market (here, shut-in gas) it will help resolve the problems of another (take-or-pay contracts). In Order No. 436, 50 Fed.Reg. 42,408 (1985), the FERC expanded greatly the choices available to both producers and customers of pipelines. Having concluded that current gas price distortions have been caused largely by the pipelines’ practice of refusing to transport gas purchased from a third party when it would displace their own sales, the FERC instituted a carrot- and-stick incentive plan. The FERC offered the carrot of blanket transportation certification, permitting a pipeline to obtain generic authorization of transportation services without the encumbrance of an unwieldly individual certification process. As a stick, the FERC attached a number of conditions to such blanket certifications; two are of central importance here: A pipeline would have to agree to provide open-access, nondiscriminatory transportation services, thereby eliminating the price distortion problem. Additionally, a pipeline would have to agree to allow its local distribution company (LDC) customers to convert their contract demand (CD) — that is, their contract commitment to purchase gas — from an obligation to purchase gas to an obligation to use, or pay for, transportation services, thereby increasing the ability of LDCs to diversify their energy portfolios and lower their energy costs. (This condition regarding LDCs was dubbed the “CD conversion-reduction option.”) The combination was irresistible; all parties agreed that in the current deregulated energy market, any pipeline would be under serious competitive pressure to offer blanket certification transportation.
At the same time it established this blanket certification/open-access transportation/CD conversion-reduction option program, the FERC “declined to include in the package any special provision to relieve pipelines from the burden of take-or-pay contracts providing for prices well above current competitive levels.” Associated Gas Distributors v. FERC, 824 F.2d 981,996 (D.C.Cir. 1987). The pipelines argued vehemently that the entire Order No. 436 scheme would be undermined by the FERC’s refusal to tackle the take-or-pay problem head on. For one thing, open-access transportation was one of the items a pipeline could previously offer to a producer in exchange for take-or-pay relief; the new program of conditioning blanket certification on open-access transportation would remove that key weapon from the pipelines’ arsenal. For another, the CD conversion-reduction option program would result in LDCs taking less contractual gas from pipelines, thereby exacerbating pipelines’ take-or-pay liability. See id. at 80; id. at 2 (Mikva, J., concurring in the judgment in part and dissenting in part).
Although upholding “most elements” of Order No. 436, id. at 18, this court in Associated Gas Distributors vacated the order and remanded the record to the Commission. At the core of the panel’s concern was its conclusion that the “FERC’s decision [to take no action on the take-or-pay problem] reflects questionable legal premises and fails to meet the requirement of ‘reasoned decisionmaking.’ ” Id. at 80. Even though it affirmed independently most of the elements of the Order, the Associated Gas Distributors court nonetheless held in the end that
[t]he parts of Order No. 436 are interdependent. The nondiscriminatory access and CD adjustment provisions aggravate the pipelines' jeopardy from take-or-pay liability. When coupled with take-or-pay liability, those provisions may bring about a wasteful imbalance between pipeline sales and unbundled transportation service. Thus the Com*233mission’s apparent insouciance on take- or-pay taints the package.
Id. at 125.
The FERC’s argument in Order No. 436 was strikingly similar to the one it makes here, viz, that the new competitive pressures of the increasingly deregulated natural gas market will lead pipelines to attempt to adjust their gas purchasing practices so as to lower their overall cost of gas. Judge Williams, writing for the panel in Associated Gas Distributors, emphatically rejected the FERC’s assumption that pipelines would be successful in this cost-lowering maneuver:
This reasoning assumes away the problem of the uneconomical contracts to which pipelines are presently bound. All FERC does here is to admit that Order No. 436 dramatically increases the consequences of not getting out from under an uneconomical contract. It seems to confuse the pipelines' incentives to renegotiate contracts with their ability to do so.
Id. at 83 (emphasis in original).
In this case, the FERC’s fourth reason for its new abandonment policy — and implications arising from the FERC’s second and third reasons as well — indicate that the Commission is still confusing “the pipelines’ incentives to renegotiate contracts” — admittedly great — “with their ability to do so” — increasingly slim. In both cases, the FERC has embarked on an essentially deregulatory path to bring the natural gas marketplace into line with 1987 economic reality. In both cases, the FERC claims that the take-or-pay problem not only will not suffer as a result of its maneuvers, but also that it will benefit. In both cases, however, the increased power given to producers and consumers appears at least on first glance to give them fewer incentives to help out pipelines with their take-or-pay worries. Following the lead of the Associated Gas Distributors opinion, which remands an order most elements of which it would otherwise uphold, due to the unexplained take-or-pay rationale, we, likewise, remand this case for the same pervasive defect, although given proper bases we might well uphold it. Similarly, like the Associated Gas Distributors opinion, we acknowledge that the FERC may reach the same result if it explains adequately how it intends to deal with the take-or-pay problem.
Of course, it may be argued that the predicted mitigation of the take-or-pay problem is but one reason given for the new policy, and that therefore we may affirm on the basis of the other, legitimate grounds. The general rule in this area is clear: “an administrative order cannot be upheld unless the grounds upon which the agency acted in exercising its powers were those upon which its action can be sustained.” SEC v. Chenery Corp., 318 U.S. 80, 95, 63 S.Ct. 454, 462, 87 L.Ed. 626 (1943); see Friendly, Chenery Revisited: Reflections on Reversal and Remand of Administrative Orders, 1969 Duke L.J. 199, 222-23 (factual as well as legal error may support a Chenery remand). This black-letter statement of the law, however, does not resolve the familiar dilemma of what to do when an agency has given multiple reasons for a new policy, some of which are acceptable, some of which are not. Here, the FERC has provided what amounts to one acceptable reason and one unacceptable one for its new policy: Because more gas will flow to the spot market, producers of the abandoned gas will make more sales and those consumers able to purchase this newly available gas will benefit from the resultant price dip. However, precisely because producers will benefit from selling the previously shut-in gas, it seems that there will be léss, not more, pressure on them to renegotiate take-or-pay contracts. The critical question is how to tell whether the FERC would have made the same policy shift given a more thorough consideration of the effect on take-or-pay contracts.
Ultimately, we decide to remand the policy for reconsideration because of the nature and implications of the “bad” reason given by the Commission. It is not some minor misstatement of law or fact that can be passed over as an unfortunate lapse. Rather, it reflects a pervasive frame of mind of the Commission about a crucial *234problem in the natural gas industry, repeated in Order No. 436, i.e., a refusal to face the fact that the burdensome take-or-pay contracts will not go away through deregulatory actions that, while aiding much of the industry, actually appear to place more pressure on the pipelines. Because of the FERC's insistence that freeing up cheap gas will alleviate the problems associated with these contracts, we can not be at all sure that the FERC would have adopted its new abandonment policy absent its unexplained take-or-pay reasoning; this justification appears to be infusing much of the Commission’s work these days. In the circumstances, we think the wiser course is to remand for the Commission to take a new, and we hope, harder look at the purported benefits of the looser abandonment policy. Ultimately, of course, the difficult questions of distributional equity that are necessarily implicated in a resolution of the take-or-pay problem are for the FERC, and not the courts, to answer.
B. Other Issues
1. Whether the FERC Should Have Conditioned Approval on Applicants’ Agreement To Renegotiate Take-or-Pay Contracts
The FERC rejected petitioners’ request that the Commission condition abandonment approval on the producer-applicants’ agreement to renegotiate the take- or-pay contracts. The Commission relied on both procedural and policy grounds. With regard to the former, the FERC claimed that petitioners had not raised this issue in the hearings before the AU. The issue was, however, raised in comments on Applicants’ settlement proposal, see J.A. at 267, and since the proposal required a policy determination not hinged to the specific facts of the case, the Commission had a duty to consider it even as late as the petition for rehearing. See Michigan Consolidated Gas Co. v. FPC, 283 F.2d 204, 224 (D.C.Cir.) (proposal that was ineffective as settlement nonetheless “appears prima facie to have merit enough to have required the Commission at some stage of the proceeding to consider it on its own initiative as an alternative to total abandonment”), cert. denied, 364 U.S. 913, 81 S.Ct. 276, 5 L.Ed.2d 227 (1960).
In any event, the FERC responded on policy grounds as well, maintaining that
while we think that a more liberal abandonment policy will contribute to our stated objective to introduce more competition into the gas market, we do not intend to use the abandonment vehicle as the sole or primary mechanism by which we will solve all of the problems that currently exist in the market.
This explanation, however, “utterly fails to address the [problem],” namely, that a “grant of abandonment alone will not intensify competitive pressures on [Applicants] to reduce their high cost gas contracts with Transco, but would [appear] to do just the opposite....” Con Ed Br. at 34. We have already discussed the Commission’s erroneous view of the effect of a liberalized abandonment policy on take-or-pay contracts in part II.A., supra. Along with the Associated Gas Distributors panel, we do not understand the FERC’s argument that increased flexibility in the natural gas marketplace will lead, q.e.d., to renegotiated take-or-pay contracts. At least as a prima facie matter it would seem that the take-or-pay contracts problem will be exacerbated, not mitigated, by the various deregulatory maneuvers in which the Commission is engaged.
The Associated Gas Distributors panel was confronted with a similar claim, namely, that the FERC should have conditioned nondiscriminatory access for producers on their cooperation in solving take-or-pay problems. Because the FERC’s responses there were inadequate, the Associated Gas Distributors panel ordered the Commission to take a closer look at the situation. See id. at 91-94. We do likewise. It may be that the Commission need not condition abandonment on renegotiation of take-or-pay contracts. Our concern, however, is that it seriously address the take-or-pay problem as it affects abandonment policies, something it has refused to do thus far.
The Commission argues that after Panhandle Eastern Pipe Line Co. v. FERC, *235613 F.2d 1120 (D.C.Cir.1979), cert. denied, 449 U.S. 889, 101 S.Ct. 247, 66 L.Ed.2d 115 (1980), the Commission may not use its § 717(e) conditioning power to adjust rates, but must instead convene a § 717d proceeding, and that any conditioning of abandonment approval on producer agreement to grant take-or-pay relief would constitute an illegal § 717f(e) condition. Even if that were so, the FERC may be able to formulate a coherent solution to the take-or-pay problem by utilizing both statutory provisions.
2. Whether the FERC Should Have Required a Pro Rata Distribution of the Abandoned Gas to Transco’s Customers
The FERC gave Transco’s customers a right of first refusal to the abandoned Block 86 gas, but then permitted open bidding, and Brooklyn Union took all the gas. J.A. at 359 & n. 9. Petitioners argue that the Commission was obligated to enforce a pro rata release of the abandoned gas to all interested Transco Block 86 customers.
First, petitioners claim that the FERC’s failure to show appropriate solicitude for all of Transco’s customers reveals that it is no longer conducting a comparison of needs in statutorily required fashion. Since we agree with the FERC, however, that the abandonment provision does not require a comparative needs test, this argument fails.
Second, petitioners contend that a still-pending Commission rulemaking proposal contradicts its failure to pro ration the right of first refusal on the abandoned gas in this case. See Regulation of Natural Gas Pipelines After Partial Wellhead Decontrol, FERC Statutes, and Regulations (CCH), 1132,408 at 33,114-16 (1985). Petitioners cite to a proposed block billing system (originally part of Order No. 436, the subject of the Associated Gas Distributors decision, but then deferred for further consideration) that appears only to ensure that current customers of old gas continue to receive the same economic rent on that gas that they received under the old billing system, proportionately to the amount of old gas in their gas portfolio. Unlike the new abandonment policy, which moves gas out from under federal jurisdiction, the proposal to entitle current customers of old gas to proportionate shares of that gas based on their historic purchase levels from the pipeline affects only gas that is still subject to Commission jurisdiction. It does not seem unreasonable for the FERC to maintain closer control over gas that would otherwise be subject to regulation than over gas that would otherwise be deregulated.
3. Whether the FERC Should Have Required Transco to Transport any Abandoned Gas Taken by its Customers
Petitioners further contend that the FERC should have required Transco to transport any abandoned gas that they bought, arguing that the right of first refusal would be rendered meaningless absent a means of transporting the gas in question. However, since we have approved the absence of pro rationing if the FERC properly promulgates a liberalized abandonment policy, Brooklyn Union may well wind up with all the abandoned gas, so that petitioners’ demand for transportation would be moot. On the other hand, if as a result of this remand no new abandonment policy is promulgated, this issue will again be moot. Accordingly, we defer the question of whether old customers of abandoned gas who have bid successfully for some or all of that gas have a right to transportation of that gas through the former pipeline-middleman.
Conclusion
We agree with the FERC that the “public convenience or necessity” language of the NGA’s abandonment provision, 15 U.S.C. § 717f(b), envisions agency policy-making to fit the regulatory climate. Indeed, the FERC’s new abandonment policy may yet prove to be well justified. However, in articulating the reasons for this new policy the FERC relied heavily upon an unexplained notion of how the shift would affect the take-or-pay contracts *236problem. Because this cryptic view of the take-or-pay problem appears to infect many aspects of the FERC’s natural-gas decision-making, we remand for the Commission to address forthrightly the issue of whether a more liberal abandonment policy can stand given honest and full consideration of its effect on the take-or-pay problem, the case is
Reversed and remanded for Commission consideration in light of this opinion.
. Transcontinental Gas Pipe Line Corporation (Transco), the pipeline involved in this case, also petitioned for review, though it did not challenge the new abandonment test. It presented the question “[w]hether the Commission’s determination ... that Transco is not permitted to purchase only the cheaper Vermilion Block 86 natural gas (ie.. Section [3316(a) ] gas) contradicts the policy in favor of free market competition on the basis of which policy [the opinions in this case] were issued and whether such determination is, therefore, arbitrary and capricious in violation of the law." Transco Br. at 2. Before oral argument, however, Transco reached a settlement with the abandonment applicants, and accordingly its petition is now moot.
. The NGA was designed to ensure an adequate supply and reasonable price for natural gas. See California v. Southland Royalty Co., 436 U.S. 519, 523, 98 S.Ct. 1955, 1957, 56 L.Ed.2d 505 (1978) (“The fundamental purpose of the [NGA] is to assure an adequate and reliable supply of gas at reasonable prices.”); Atlantic Refining Co. v. Public Service Commission, 360 U.S. 378, 388, 79 S.Ct. 1246, 1253, 3 L.Ed.2d 1312 (1959) (‘The purpose of the [NGA] was to underwrite just and reasonable rates to the consumers of natural gas.").
. See generally M. Sanders, The Regulation of Natural Gas (1981).
. The abandonment was "limited" in three ways: First, it was only for three years. Second, under the terms of the FERC’s order, Transco was "required to submit to Applicants at least thirty days in advance an estimate of volumes to be purchased during the succeeding six months.” J.A. at 296. That is, Transco still had first dibs on its needed gas. Third, any gas not nominated by Transco could then be taken by Transco’s customers under a right of first refusal. Although the limited nature of the abandonment order in this case mitigates any hardship to petitioners, it does not alleviate the concerns we express in the text about the absence of reasoned decisionmaking.
. Order No. 436, the subject of this court’s recent decision in Associated Gas Distributors v. FERC, 824 F.2d 981 (D.C.Cir. 1987), and discussed in part II.A, infra, "directs Commission staff to expedite all applications for abandonment of shut-in gas." Griggs, Restructuring the Natural Gas Industry: Order No. 436 and Other Regulatory Initiatives, 7 Energy L.J. 71, 74 (1986); see 50 Fed.Reg. 42,408, 42-466-67 (1985). This aspect of the Order was approved in the Associated Gas Distributors opinion. Id. at 84 n. 28. Order No. 436 does not, however, alter the substantive abandonment test.
. It is not clear how current purchases of gas can mitigate a problem of declining reserves, since reserves are valuable precisely because they do not have to be taken currently. As petitioners maintain, “[t]he real detriment to Transco and its customers will be in the loss of irreplaceable cheaper reserves for future use." Con Ed Br. at 23 n. 12. Indeed, it seems that declining reserves are an ineluctable result of a liberalized abandonment policy, since if more gas is released today, less gas will be available tomorrow. Although this detriment to current captive consumers might be outweighed by some other benefit, the FERC has not addressed this problem head-on.
. The Commission basically reiterated its reasons for the new policy in the context of this case:
While Transco and its customers will continue to have the right to purchase Block 86 gas to the full extent of its deliverability, the excess volumes available for sale under the abandonment will serve to increase competition for sales among producers. This will, in turn, help to mitigate the price distortions which currently exist in the gas market. By placing a downward pressure on gas prices in the open market, we also believe that pipelines, such as Transco, and producers will be induced to renegotiate contracts for high cost gas that do not reflect the true value of the commodity. The result should ultimately be lower prices at the burner-tip. Finally, the Applicants here demonstrated that they do in fact use the cash flow generated from their gas sales to finance their exploration and development programs. Therefore, to the extent that the cash flow increases, either through sales of abandoned gas on the market, or through increased takes of gas by Transco and its customers due to the abandonment, we can expect greater resources available for exploration and development, which leads to increased supplies overall. Thus, we believe there are clear benefits to all segments of the natural gas industry, including consumers, that are sufficient to justify the limited abandonment approved in this order.
J.A. at 298-99. See also note 8, infra.
. Although these were the factors the FERC relied upon to justify its general policy shift, in approving the specific limited-term abandonment under the facts of this case the FERC added that the producers would reinvest the additional cash from the sale of abandoned gas into research and development programs. See note 7, supra. We have no quarrel with this conclusion, but merely note that it does nothing to explain the Commission’s take-or-pay confusion.
. The situation for captive consumers may be changing. Following Associated Gas Distributors v. FERC, 824 F.2d 981 (D.C.Cir. 1987), and Panhandle Eastern Pipe Line Co. v. FERC, 613 F.2d 1120 (D.C.Cir. 1979), cert. denied, 449 U.S. 889, 101 S.Ct. 247,66 L.Ed.2d 115 (1980), captive consumers may be able to escape long-term contracts and take advantage of the newly abandoned shut-in gas ón the spot market, through the combination of the CD conversion-reduction provision of Order No. 436 (discussed below in the text in response to the FERC's fourth reason) and § 717d hearings that would be held to approve the resulting rate shift (discussed in part II.C.l, infra ). See Associated Gas Distributors at 58-75 (remanding CD conversion-reduction provision primarily because the FERC did not link it to the § 717d hearings); see also S. Williams, The Natural Gas Revolution of 1985 3-4 (1985) (long-term contracts not a burden to consumers if they can switch to substitute energy supplies). However, the possibility that such an option might exist in the future is currently quite speculative, given the unsure status of the CD conversion-reduction provision. Moreover, even if the FERC rectifies the CD conversion-reduction provision problems identified by the Associated Gas Distributors panel, it will still have to deal with the resultant pressure placed on pipelines through LDCs’ use of the provision to escape contractual purchase requirements of pipelines’ more ex*230pensive take-or-pay gas. See text, infra, identifying this pressure as a factor that the FERC neglected to deal with in Order No. 436.
. Of course, if the "risk [of] being shut-in" refers to the chance of not being granted abandonment, the second reason would contradict the thrust behind the new abandonment policy. That is, there would be little pressure on sellers to reduce price because the supposed “risk” of being shut-in would be obviated by the very policy the FERC asks us to uphold. It seems more likely, though, that the more general point we discuss in the text is the one the FERC had in mind.
. Although the Commission’s opinions and not its litigation position are under review here, it is interesting to note that its brief also relies heavily upon the fourth reason, which, as we show in the text, is unacceptable. See FERC Br. at 12-13, 17, 21.
. See generally Medina, McKenzie, and Daniel, Take or Litigate: Enforcing the Plain Meaning of the Take-or-Pay Clause in Natural Gas Contracts, 40 Ark.L.Rev. 185, 187-92 (1987).
. We would have thought this proposition unexceptional. See e.g., Pierce, Natural Gas Regulation, Deregulation, and Contracts, 68 Va. L.Rev. 63, 103 (1982) (‘The potential for bargaining out of long-term contracts is limited severely by the existence of take-or-pay clauses in the vast majority of gas purchase contracts.”). Professor Pierce does suggest one scenario in which producers might renegotiate take-or-pay contracts:
Consider a pipeline that buys gas from only one producer. If the contract price reaches a point substantially above market price, the pipeline will begin to experience substantial financial problems because of its inability to recover its costs. Even though he has the protection of a take-or-pay clause, the seller may be willing to decrease the contract price for two reasons. First, at some point, the pipeline will be unable to meet its existing contractual obligations to the seller, and second, the seller presumably has an interest in maintaining the purchaser as an outlet for newly discovered gas supplies. Thus, the seller’s interest in the stability of the relationship may lead to some relaxation of escalation provisions.
Id. at 103-04. But he also explains that ‘‘[b]argaining out of supramarket price contracts is not likely to take place very often through this sequence, however, because most purchasers buy from many sellers." Id. at 104. With a pipeline beholden, to many producer-sellers, it would rarely be to the advantage of one producer-seller to sacrifice his contractual position, because the pipeline's future would still be in the hands of others. A hold-out situation would result. See also Broadman, Natural Gas Deregulation: The Need for Further Reform, 5 J. of Pol’y Anal. & Mgmt. 496, 507 (1986) (noting the need to avoid pipeline bankruptcy and the concurrent hold-out problem).
Thus, the incentives for producer-sellers to renegotiate take-or-pay contracts are minimal. Nevertheless, the FERC argues that Transco’s recent request for blanket abandonment authority, which was approved by the Commission, 36 F.E.R.C. ¶ 61,403 (1986), FERC Supp.Br., indicates that the new liberalized abandonment policy has already had a positive effect on the take-or-pay problem. It appears from the FERC’s opinion in the blanket abandonment matter and the FERC’s supplemental brief in today’s case that Transco’s producer-suppliers did indeed grant Transco some amount of take- or-pay relief in exchange for Transco’s agreement to abandon its shut-in gas. However, we have no way of knowing whether this voluntary bilateral agreement, unopposed by any party, resulted from the new liberalized abandonment policy, occurred simultaneously with it, or occurred in spite of it. Moreover, the FERC’s argument in this case was that the release of low-cost shut-in gas will provide an incentive to both pipeline and producer-seller to renegotiate take-or-pay contracts so that the pipeline would continue to purchase the contractual gas instead of the newly released cheaper gas on the spot market. There is no indication in the Transco blanket abandonment matter, however, that Transco was ready to jump to the spot market. In fact, the problem there was that Transco’s demand for natural gas on the whole had dipped precariously. In sum, the blanket abandonment matter does not provide evidence for the proposition that the new abandonment policy gave the producer-sellers an incentive to renegotiate take-or-pay contracts.
Finally, we think petitioners have convincingly shown that one of the producer-applicants here, Felmont Oil Corporation, refused Transco’s proposal that it grant relief on high-priced take-or-pay contract gas in exchange for Transco’s agreeing to take more gas on the whole from Felmont. That Felmont did not wish to renegotiate the take-or-pay contracts in the face of decreasing takes by Transco is understandable; after all, Transco had to pay under those contracts whether or not it took the relevant gas.
In sum, there is no reason to believe that either the producer-sellers in the blanket abandonment matter or Felmont in the matter before us responded to decreased takes of low-priced gas by voluntarily agreeing to grant take- or-pay relief. Felmont clearly did not grant such relief; the producer-sellers in the blanket abandonment matter did, but in exchange for Transco’s voluntary release of shut-in gas, not as a result of a fear of Transco’s jumping to the *232spot market to gobble up newly released shut-in gas. Such a fear, of course, would be irrational, for under the take-or-pay contracts Transco would still have to pay for the higher-priced gas it was leaving behind.