Opinion for the Court filed by Circuit Judge WALD.
WALD, Circuit Judge:This case marks the latest episode in the ongoing saga of the “price squeeze” doctrine in federal utility ratemaking. Several Pennsylvania municipalities1 (“Boroughs”) petition for direct review of decisions by the Federal Energy Regulatory Commission (“FERC” or “the Commission”) approving with minor modifications wholesale *260rates filed by Pennsylvania Power Company (“Penn Power” or “the Company”) under sections 205 and 206 of the Federal Power Act. 16 U.S.C. §§ 824d, 824e. The proceedings before the Commission were bifurcated into a “cost-of-service” phase and a price discrimination phase. Boroughs appeal from several aspects of the Phase I proceedings, in which the Commission set just and reasonable rates for Penn Power aside from any considerations of unlawful discrimination. In Phase II, the Commission considered allegations of price discrimination. Boroughs’ primary claim was that the wholesale rate it was charged by Penn Power was unjustifiably high in comparison with the rate Penn charged certain retail customers for which Boroughs competed, thus creating a “price squeeze.” Although the Commission found non-eostjustified price discrimination by Penn Power against its municipal competitors for nearly a year, it refused to issue a price squeeze remedy. It concluded that the discrimination was not “undue” because it was the result of a company decision to take advantage of a change in the state law governing its retail rates. We vacate the Commission’s decision not to remedy the price squeeze, but we uphold its decision in all other respects.
I.Background
Penn Power supplies electrical power to both retail customers, whose rates are under the jurisdiction of the Pennsylvania Public Utility Commission, and wholesale customers, including Boroughs, whose rates are under the jurisdiction of FERC. Penn Power and its municipal customers are both potential suppliers, and thus in competition, for many retail customers.
Penn Power filed with the Commission a rate increase for its wholesale customers in July, 1977.2 The rates were accepted for filing, suspended for two months, and allowed to go into effect subject to refund on September 11, 1977; they remained in effect until January 23, 1982. The Commission scheduled a public hearing on the lawfulness of the proposed rates under sections 205 and 206 of the Federal Power Act, 16 U.S.C. §§ 824d, 824e, which proscribe rates that are unjust, unreasonable or unduly discriminatory. Boroughs intervened in these proceedings, alleging among other things that the new rates were anti-competitive and discriminatory in relation to the retail rates that Penn Power charged industrial customers for whom Boroughs competed.
The Commission bifurcated the proceedings. In Phase I, the Commission determined the just and reasonable rate, based on Penn Power’s cost of service to the wholesale customers, apart-from any price discrimination issues.3 During the November, 1978, Phase I hearings, Boroughs unsuccessfully contested several elements of Penn Power’s rate filing that are now before us on appeal. They argue here (1) that the Commission erred in permitting Penn to include a forty-five-day working capital allowance that was much greater than Penn’s actual needs; (2) that the test period estimate of revenues from sales of excess reserves had proved to be inaccurate by $2 million and should have been adjusted accordingly; and (3) that the Commission allowed an excessive rate of return on equity of 13.25%.
In the second phase of the ratemaking proceedings, the Commission considered Boroughs’ allegations that the large wholesale rate increase, unmatched by a retail rate increase, resulted in a price squeeze. Boroughs, which are totally dependent on power supplied by Penn, claimed that they had to pay such a high wholesale rate that *261they could not compete with Penn’s lower retail rates for many large customers, especially the industrial customers within the Boroughs’ boundaries. Although the Commission agreed with the AU that a price squeeze had existed for nearly a year— from September 11, 1977, when the wholesale rate increase went into effect, to August 31, 1978, when the retail increase took effect — it found on the basis of additional filings by Penn4 that no price squeeze existed after August 31, 1978. The Commission also overruled the AU’s decision to issue a remedy. It held that the price discrimination was not “undue” and was thus excused because it was the result of Penn’s decision to delay its retail filing until the effective date of a new state law that imposed a definite suspension period.5
Boroughs contend on appeal, first, that the Commission unfairly applied a test of price discrimination that was not developed until after the record was closed, thus precluding them from demonstrating a longer period of price discrimination than the year-long period ultimately found. They contend further that as to the period for which the Commission did find an actual price squeeze, it unjustifiably denied them a remedy.
II. Cost-of-Service Issues
A. Working Capital Allowance
An electric utility may include in its ratebase a cash allowance to permit it to meet current obligations as they arise. As we explained in City of Cleveland v. FPC, 525 F.2d 845, 850 n. 38 (D.C.Cir.1976) (quoting Alabama-Tennessee Natural Gas Co. v. FPC, 203 F.2d 494, 498 (3d Cir.1953), “the need for working capital arises largely from the time lag between payment by the company of its expenses and receipt by the company of payment for service in respect of which expenses were incurred.”
A utility’s actual need for working capital can be most accurately determined by performing a “lead-lag” study of the average number of days that passes between payment of expenses and receipt of revenues for a given service. One part of this calculation is the “revenue lag,” which is in turn made up primarily of the “service lag” — the number of days between the time expenses are incurred for services and the date of billing for those services — and the “payment lag” — the number of days between billing and payment. A utility also experiences “lead time” when it receives payment for services before it pays the expenses associated with those services. The number of lag days minus the number of lead days yields a net lag which represents the utility’s actual need for working capital.6
A reliable lead-lag study, however, is a burdensome undertaking. The Commission therefore long ago established the standard practice of permitting a utility in the absence of a lead-lag study to include in its ratebase an amount equal to forty-five days of operating expenses. See, e.g., Interstate Power Co., 2 F.P.C. 71, 85 (1939); Union Electric Co., 47 F.P.C. 144, 175 (1972).7 The Commission found that the use of this estimate yielded reasonable results while avoiding the costs of regularly performing a reliable “lead-lag” study and of litigating the issue. Carolina Power & Light Co., 6 FERC ¶ 61,154, p. 61,295 (1979). The Commission’s rule of thumb has been approved by at least one court. *262See Union Electric Co. v. FERC, 668 F.2d 389, 397 (8th Cir.1982).
In 1979 the Commission initiated a rule-making to revise the formula, recognizing that it may have become outdated in light of current billing practices.8 Pending that revision, the Commission provided that “[w]here a fully developed and reliable lead lag study is available in the record, we will utilize that study to determine the working capital allowance. When a study meeting these criteria is not available we will continue to apply the forty-five-day convention.” Carolina Power & Light Co., 6 FERC at 61,296.
1. The Commission’s Decision
In this case Boroughs offered a lead-lag study which they claimed demonstrated a need for a working capital allowance of only thirteen days. The study was one which Penn had prepared for its last retail filing and which Boroughs had “adjusted” in various ways that the ALJ found were “troubling, and taint[ed] the use of that study.”9 The ALJ therefore rejected the study. He nevertheless reduced the allowance from forty-five to forty days in recognition of the Commission’s proposed revision of the standard formula. ALJ I, 9 FERC at p. 65,160.
The Commission concurred in the ALJ’s rejection of Boroughs’ lead-lag study, but on a different basis. The Commission pointed to an apparent inconsistency between the study’s estimated revenue lag of forty-five days experienced in paying operating and maintenance expenses and Boroughs’ estimate of thirty days for paying interest on long-term debt, dividends, and federal taxes. Neither estimate, it stated, was supported by a study of actual billpaying practices. For example, the study assumed that the revenue lag was forty-five days for wholesale customers; a Penn Power witness testified, however, that, according to another study of the actual bill-paying practices of one of these customers, the revenue lag was sixty-six days.10 The Commission thus agreed with the ALJ that Boroughs had not presented the kind of reliable study that could be used in place of the forty-five-day convention. FERC I, 12 FERC at p. 61,080.
Unlike the AU, however, the Commission adhered to that convention. It rejected the AU’s use of the forty-day figure appearing in the notice of proposed rule-making because that proposed figure represented an estimate of the lag between the rendition of services and the receipt of revenue rather than the lag between payment of expenses for services and receipt of revenue represented by the forty-five-day figure. Id. Boroughs now challenge both the decision to reject its lead-lag study and the use of the forty-five-day rule.
2. Discussion
(a) The Study. In Cities of Aitkin v. FERC, 704 F.2d 1254 (D.C.Cir.1982), we reversed the Commission’s decision to reject a “lead-lag” study and to apply instead the forty-five-day convention. In that case we found that “[t]he Commission [had] point[ed] to no evidence tending to undermine the study’s validity,” id. at 1258, but *263had relied upon “a formalistic argument that flies in the face of logic.” Id. It had rejected the study simply because it was in part based not on a study of actual payment dates but on the assumption that payment due dates were representative of actual payment dates, an assumption that had been verified by evidence of actual payment practices.
In the case before us, the Commission relied in part on the same flaw — the failure to study actual payment dates. It also noted, however, that the study’s assumption as to the “revenue lag” was expressly contradicted rather than confirmed by Penn’s experience and by the single-customer study, which had yielded a sixty-six-day rather than a forty-five-day revenue lag. On the other hand, the Commission acknowledged that the single-customer study from which the sixty-six-day revenue lag was derived had incorrectly used a thirty-day instead of a fifteen-day “service lag.” 12 FERC at n. 18. Although these two inaccuracies appear in part to cancel each other out, they do tend to undermine confidence in the study’s reliability. We cannot say that the Commission acted arbitrarily in concluding that Boroughs had not presented “a fully developed and reliable lead-lag study.”
(b) The Forty-Five-Day Convention. The Commission’s discussion of Boroughs’ lead-lag study reveals not only the weakness of the study but also the desirability of a standard formula that can be applied, at least in most cases, without burdensome investigations and lengthy administrative proceedings on the details of a given utility’s actual “service lags,” “payment lags,” etc. The forty-five-day convention thus appears still to fulfill an important function.
Another important rationale behind the rule, however, was that it yielded reasonably accurate results. The Commission’s discussion of the lead-lag study here raises serious doubts over the continuing validity of the forty-five-day convention. For example, the single-customer study that the Commission used to support its rejection of Boroughs’ study had, according to FERC, incorrectly used a thirty-day rather than a fifteen-day “service lag.” If the corrected figure of fifty-one days — sixty-six minus fifteen — were plugged into Boroughs’ calculations, it would appear to yield a net lag, and thus a working capital requirement, of about nineteen days. If the bill-paying practices of the single customer are remotely representative, the forty-five-day figure used by the Commission may be seriously defective.
As we have noted, the Commission has recognized this possibility and in 1979 initiated rulemaking proceedings to revise the formula for calculating working capital requirements. Any revision of the formula would seem to require the broader perspective and participation inherent in rulemaking proceedings. In light of this ongoing proceeding, we decline to disturb the Commission’s interim decision to continue to use the forty-five-day convention in the absence of a “fully developed and reliable lead-lag study.” The temporary — and we emphasize temporary — sacrifice of accuracy is outweighed by valid administrative concerns over the implications of permitting ad hoc adjustments based on incomplete data in every individual ratemaking proceeding. We therefore approve the Commission’s decision to apply the forty-five-day convention in this case.11
B. Sale of Excess Reserves
The ratebase is derived from actual cost-of-service and revenue data from the previ*264ous twelve-month period and estimated costs and revenues for a subsequent twelve-month period — the “test year.” Even when the actual data become available during the proceedings, the Commission will not generally adjust the estimates. It inquires first whether those estimates were reasonable when made; if the utility demonstrates that this is so, the Commission will use them unless the party contesting the estimates shows that they are “substantially in error” and would yield “unreasonable results.” Southern California Edison Co., 8 FERC ¶ 61,099, p. 61,375 (1979). We have approved this approach to ratemaking. Cities of Batavia v. FERC, 672 F.2d 64, 74 (D.C.Cir.1981) [hereinafter “Batavia”]; Villages of Chatham v. FERC, 662 F.2d 23, 29-30 (D.C.Cir.1981); American Public Power Association v. FPC, 522 F.2d 142 (D.C.Cir.1975).
1. The Commission’s Decision
In this case, Penn Power estimated that it would earn about $7 million in 1977, the test year, from the sale of excess reserve capacity — i.e., the reserve generating capacity beyond that required to guarantee reliable service to its own customers. In fact it received about $9 million from such sales. The additional $2 million was largely attributable to the addition of new generating units during the latter part of 1977. Boroughs argued that since Penn anticipated the addition of these units and included costs for the additional capacity in its cost estimates, it should have similarly anticipated the resulting increase in reserve capacity sales. They argue that, “[g]iven excess capacity from generating units, it is per se unreasonable not to have projected sale of the same.” Penn pointed out that this assumes a willing buyer.
The ALJ approved the estimate, pointing to additional factors in support of the estimate’s reasonableness. First, Penn did plan more sales and less purchases of firm power during 1977 than in the preceding years, offsetting somewhat the increase in excess capacity. In addition, the ALJ recognized “the inherent limitations in the forecasting process and ... the special problem in accurately predicting sales of excess power in a year when two new generating units are placed in service.” ALJ I, 9 FERC at p. 65,161. Furthermore, he found the $2 million underestimate, only a small fraction of which was allocable to wholesale service, to be not so excessive as to require adjustment of the test year data.12 Finally, he noted that Boroughs “failed to demonstrate that no offsetting costs would occur if the alleged excess capacity were used.” Id. The Commission agreed without further discussion. FERC I, 12 FERC at p. 61,078 & n. 2. Boroughs challenge this conclusion.
2. Discussion
In Indiana & Michigan Municipal Distributors Association v. FERC, 659 F.2d 1193 (D.C.Cir.1981), we recognized that “to require routine revision in the [test year] estimates in light of actual developments would defeat the purpose of the test-year methodology.” Id. at 1197. In that case we approved the use of an estimate of short-term sales amounting to $45 million although the actual sales proved to exceed $81 million. We found the estimate to have been reasonably consistent with past experience, rejecting petitioners’ argument that, because the company knew it would have large excess reserves, it should have necessarily anticipated sales commensurate with these reserves. In particular, we noted that there was no evidence that demand should be expected to increase with supply.13
*265The much smaller disparity in this case between estimated and actual revenues can be more easily justified, not only because it is much smaller but also because of the uncertainty, to which the AU adverted, connected with the addition of two new units. We sense some asymmetry in permitting the inclusion of full capacity costs for such units but the exclusion on grounds of uncertainty of a large portion of revenues that might reasonably be anticipated from sales of additional capacity. Nevertheless, on balance we conclude that the Commission is justified in setting a rather high threshold for parties seeking to adjust test year data, and that it acted reasonably, for all the reasons given by the AU, in refusing to adjust the estimates in this case to account for $2 million in additional revenues.
C. Rate of Return on Equity
The Commission is required to set a rate of return commensurate with other enterprises of comparable risk and sufficient to assure that enough capital is attracted to the utility to enable it to meet the public’s needs. See FPC v. Hope Natural Gas Co., 320 U.S. 591, 603, 64 S.Ct. 281, 288, 88 L.Ed. 333 (1944); see also Bluefield Waterworks v. Public Service Commission of West Virginia, 262 U.S. 679, 43 S.Ct. 675, 67 L.Ed. 1176 (1923).
1. The Commission’s Decision
All parties agreed that Penn’s financial performance had been generally less favorable than other enterprises with comparable characteristics, making it a riskier investment and thus requiring a somewhat higher return on equity in order to attract enough capital to meet the needs of its customers. Intervenors and Staff thus argued for a 12.75% rate of return. Penn emphasized, however, that the recent downgrading of its bonds by Standard & Poor from A to Baa reflected a pessimistic view of its financial prospects. Penn Power proposed a 14.5% rate of return on equity-
The ALJ weighed these factors and arrived at a 13.25% rate of return. He first concluded that Staff, while recognizing that Penn’s return must be “at the high end of the zone of reasonableness,” did not give enough weight to the reduction in the bond rating. Because Penn is a wholly-owned subsidiary of Ohio Edison and is not independently represented in the equity market, he concluded that “it is in the bond market that Penn Power ... is tested the most by the capital market.” ALJ I, 9 FERC at p. 65,164.
Yet the ALJ rejected Penn’s recommendation of 14.5%. The highest rate of return recently granted by the Commission was 14% to Boston Edison Co., 8 FERC ¶ 61,077, p. 61,288 (1979), whose financial situation compared unfavorably to Penn’s in several respects.14 The ALJ settled on a 13.25% rate of return. The Commission adopted the ALJ’s reasoning and conclusion without further discussion. Boroughs argue, first, that the Commission may not simply rely on the ALJ’s reasoning and findings, and that its decision is not supported by substantial evidence. In particular, they argue that the AU/Commission placed undue weight on Penn’s bond rating.
2. Discussion
The Commission is not required to recapitulate the reasoning of the AU if it is satisfied that the initial decision and the reasoning underlying it are sound. We have not been left to guess at the Commission’s findings or reasons, see Greater Boston Television Corp. v. FCC, 444 F.2d 841, 851 (D.C.Cir.1970), cert. denied, 403 U.S. 923, 91 S.Ct. 2233, 29 L.Ed.2d 701 (1971); they are to be found in the ALJ’s decision. See also Borek Motor Sales, Inc. v. NLRB, 425 F.2d 677, 681 (7th Cir.), cert. denied, 400 U.S. 823, 91 S.Ct. 45, 27 L.Ed.2d 52 *266(1970) (agency’s statement that it had considered objections to examiner’s decision but adopted his findings, conclusions and recommendations except as modified adequately preserves reasoning for review under APA). Particularly in light of the elaborate and careful analysis evident in that initial decision, we see no merit in Boroughs’ claim that the Commission failed to give its own reasons for the decision.
On the merits, Boroughs’ chief contention is that the lowered bond rating should have been given less weight, especially since it had no apparent effect on the equity costs of Penn’s parent company, Ohio Edison. We find more persuasive the AU’s reasoning as to the significance of the bond rating in light of Penn’s status as a wholly-owned subsidiary: investor confidence in a particular company is reflected in all aspects of the capital market, including the bond market. Because Penn has no independent presence in the equity market, and because of the impossibility of isolating Penn Power’s contribution to investor confidence in Ohio Edison, it is appropriate to look to Penn’s bond rating as the most direct measure of investor confidence in Penn itself. The significant downgrading of that rating justifies the relatively small upward adjustment that the Commission made to Boroughs’/Staff’s recommended rate of return.
III. Phase Two: The Price Squeeze
Boroughs contend, and the Commission has found, that they were subjected to a price squeeze — they had to pay Penn Power unjustifiably high rates that, in comparison with the rates Penn charged industrial retail customers with similar cost characteristics, handicapped Boroughs’ ability to compete with Penn for those customers. In connection with this price squeeze determination, Boroughs present two issues for our consideration. They contend first that the Commission unfairly applied a test of price discrimination that had not been announced until after the record was closed, and thus denied them an opportunity to present evidence tending to show price discrimination of a greater degree and duration than was found. Boroughs also appeal from the Commission’s decision not to issue a remedy in spite of its uncontested finding of a price squeeze on Boroughs for almost a year.
A. Background
Price squeeze doctrine, which has a long history in antitrust law,15 made its debut in utility ratemaking only recently. Until the Supreme Court’s unanimous decision in FPC v. Conway Corp., 426 U.S. 271, 96 S.Ct. 1999, 48 L.Ed.2d 626 (1976), the Commission denied its jurisdiction to hear price squeeze claims like this one because it had no jurisdiction over retail rates. But in Conway the Court directed the Commission to hear such claims so as to insure that rates over which it does have jurisdiction are not discriminatory in violation of the Federal Power Act. If the wholesale rates, together with the nonjurisdictional retail rates, create a price squeeze on the utility’s wholesale customers/retail competitors, the Court concluded that “[tjhere would be, at the very least, latitude ... to put wholesale rates in the lower range of the zone of reasonableness.” 426 U.S. at 279, 96 S.Ct. at 2004 (quoting Conway Corp. v. FPC, 510 F.2d 1264, 1274 (D.C.Cir.1975)).
1. The Irrelevance of Intent
The Commission, faced with the task of developing a manageable price squeeze doctrine, eventually adopted two central principles, both of which we have considered and approved in our previous encounters with the evolving price squeeze doctrine. First, the Commission unequivocally rejected any inquiry into anticompetitive intent. In Missouri Power & Light *267Co., 5 FERC ¶ 61,086 (1978), decided several months before hearings began on Boroughs’ price squeeze allegations, the Commission declared that “intent is legally immaterial to whether or not we can or must grant a remedy” in price squeeze cases. It continued: “Accordingly, we hereby direct presiding judges ... to preclude any further consideration of intent in price squeeze proceedings.” Id. at n. 17.
We acknowledged and implicitly approved the Commission’s decision in Batavia, 672 F.2d at 87. It seems eminently reasonable in the price squeeze context to preclude lengthy investigations into a utility’s motives and to focus instead on objective criteria. Although Boroughs had alleged, and continue to allege, that Penn imposed the price squeeze as part of a long campaign of discriminatory conduct aimed at ousting them from the most lucrative sectors of the retail market, the proceedings were directed solely to objective indices of a price squeeze.
2. Identifying Non-Cost-Justified Discrimination
This objective focus, however, did not require the Commission to remedy all price differences that hindered the ability of wholesale customers to compete in the retail market. The Commission construed its discretion under Conway as extending only to price differences that were not justified by differences in the cost of service. We approved that reading of Conway in Illinois Cities of Bethany v. FERC, 670 F.2d 187 (D.C.Cir.1981) [hereinafter “Bethany”].
The appropriate method for identifying non-cost-justified price discrimination, however, continued to generate controversy throughout these proceedings. Boroughs argued that the appropriate focus was a direct comparison of rate schedules: if the utility’s costs of serving the wholesale customer were within the range of costs of serving the various members of a particular retail class, then the wholesale customer was entitled to pay the same rate, and a higher wholesale rate demonstrated price discrimination.
The Commission Staff and Penn Power, however, used a “rate of return” test. As described by the AU, this test involved several steps:
(1) developing a cost of service study for the MRS [i.e., municipal wholesale] class, (2) developing a cost of service study for LP and LT [i.e., industrial retail] class, (3) computing the earned rate of return [i.e., price-to-cost ratio] for both wholesale (MRS) and retail (LP & LT) classes at the contemporaneously effective rates, and (4) comparing the earned rate of return for the MRS class with the earned rate of return for the LP and LT classes to determine if there is undue price discrimination.
ALJ II, 11 FERC at pp. 65,048-49. A higher rate of return for wholesale customers than for the retail target of competition would indicate a price squeeze.
The test advocated by Boroughs required only the identification of entire retail rate schedules with which they could not compete.16 The test advanced by Staff and eventually adopted by the ALJ and the Commission, on the other hand, required the identification of the particular target of competition. The target could be an entire retail class or a particular subgroup within the class. We explained the importance of identifying a target of competition in Batavia:
If there [is] a substantial number of customers for whom Cities could compete but for the subsidy those retail customers receive from other retail customers with whom they are grouped and for *268whom Cities could not compete ... the Commission ought to consider whether its Conway discretion ought to be exercised.
672 F.2d at 89-90 (emphasis in original). In other words, the identification of a particular target of competition becomes critical where there is the possibility that one group of retail customers for whom the wholesale customer could not compete is paying a price that yields a high rate of return and thus subsidizing the lower rate of the target of competition.
Penn had urged in these proceedings that, except for those large industrial users served at 138 KV,17 all retail customers, including residential customers, were potential targets of competition and should be included in the rate of return comparison. ALJ II, 11 FERC at p. 65,050. Staff argued for the inclusion of the 138 KV customers but the exclusion of residential customers, reasoning that competition was most active for industrial customers, who were most able to choose among competing suppliers. Id. Exclusion of the class of non-industrial customers, which yielded a high rate of return,18 and inclusion of the class of 138 KV customers, which yielded a low rate of return,19 resulted in a greater disparity between the municipal customers and the comparison group, and thus a stronger showing of discrimination.
B. The Agency Decisions
1. The ALJ Decision
The ALJ in his decision adopted Staffs rate of return approach, which had been approved in the Commission’s decision in Commonwealth Edison Co., 8 FERC ¶ 61,-277 (1979), issued after the record in these proceedings was closed. Using this method, the AU found it undisputed that there was non-cost-justified price discrimination for a period of nearly one year between the effective date of the wholesale rate increase and that of the retail rate increase— September, 1977, to August, 1978. ALJ II, 11 FERC at p. 65,048. He found further that the increase in Penn’s retail rates at the end of this period mitigated but did not eliminate the price discrimination. Id. at p. 65,051. In addition, he found that this price discrimination had an anticompetitive effect in that it took place in a context of competition between Penn and Boroughs for industrial customers currently or potentially located within Boroughs’ boundaries. Id. at pp. 65,051-52. Finally, the AU found that this discrimination was “undue” —i.e., no countervailing public policy considerations or factual circumstances excused or ameliorated the price discrimination — and that it thus required a remedy. Id. at p. 65,055.
In reaching this final conclusion, the AU rejected Penn Power’s primary argument against issuance of a price squeeze remedy. Penn explained that an immediate need for additional revenue dictated its prompt rate filing at the federal level on July 11, 1977, which became effective September 11, 1977. Penn decided to delay its retail rate filing with the state utility commission, *269however, in order to take advantage of an impending change in the state’s regulatory framework. The state law currently in effect permitted an indefinite period of suspension, potentially delaying retail rate relief for up to two years.20 An amendment to the statute, which would take effect after October 7, 1977, imposed a maximum suspension period, from the date of filing, of eleven months.21 Penn claimed that it chose to delay its state filing for several months in order to file under the new law and thus guarantee that its retail increase would not be delayed beyond September, 1978. Penn argued that the resulting price squeeze should thus be excused.
The ALJ rejected this argument as having “strong overtones of subjective intent.” ALJ II, 11 FERC at p. 65,053: Because the reason for Penn’s decision was not to create a price squeeze, Penn argued, but simply to obtain immediate rate relief for both retail and wholesale service as soon as possible, no remedy was warranted. In response to Penn’s argument that the price disparity should be excused as the product of “countervailing factual circumstances” or of “state action,” the ALJ noted that Penn could have delayed its wholesale filing, which would affect only three percent of its revenues, as well as its retail filing: “There is just no way around the fact that the timing of the wholesale and retail rate filings was wholly within Penn’s control, and that the maladroit timing of the filing as made was the cause of the price squeeze.” Id. at p. 65,055.
2. The. Commission’s Decision
The Commission agreed with the AU that the rate of return test was the appropriate method of identifying price discrimination. It also agreed that the appropriate comparison was between the rate of return for industrial retail customers, including the 138 KV customers, and the rate of return for the wholesale customers. According to this test, the Commission found that
a disparity in rates of return which is unexplained by differences in cost factors (i.e., price discrimination) existed during the first IIV2 months [after the wholesale rates became effective], but not thereafter. The price discrimination was of a magnitude and duration that, in the absence of any mitigating circumstances, we would find it undue and require a remedy.
FERC II, 21 FERC at p. 61,818. However, the Commission found mitigating circumstances in Penn’s decision to delay its retail filing until the effective date of the new state statute.
The Commission first dismissed the AU’s characterization of this argument as one going to intent. Although innocent intent was no defense to a price squeeze claim, “there may be factual circumstances that justify the tolerance of the price discrimination.” In this case the Commission found itself confronted with an aspect of “the paradox of dual utility rate regulation.” This “paradox” can give rise to rate disparities for two different reasons. On the one hand, disparities can arise from “differences in state and federal ratemaking policies and differences in the time frame within which the different commissions grant rate relief.” The Commission *270stated “as a general rule” that it “should not lower rates to remedy differences between our ratemaking policies and procedures and those of a state commission.” It was inappropriate, according to the Commission, to allow “the federal interest to yield to competing state interests.” Id. at pp. 61,819-20.22
The Commission vowed to be more sensitive, however, to disparities resulting from differences in the timing of a rate filing, since utilities generally can control their filing dates. Id. at 61,820.. Nevertheless, in this case the Commission found special circumstances amounting to a total defense:
We are persuaded by Penn Power’s testimony that it needed rate relief immediately to maintain its construction program and that its only reason for delaying its filing in the Pennsylvania Commission was that by waiting it could take advantage of a newly enacted statute establishing a suspension period for rates subject to state agency jurisdiction to go into effect. In fact, in view of its need, Penn Power could be viewed as having taken the necessary steps to mitigate the price squeeze for the simple reason that, had it filed at the same time in both jurisdictions while Pennsylvania had no statutory suspension period, the length of the price squeeze period would likely have been much longer than IIV2 months. We believe that the difference in state and federal legislative policies should not be permitted to deprive Penn Power of its right to rate relief at the federal level. There is something to be said for Penn Power’s criticism of the law judge’s opinion that Penn Power should have waited to file its federal rate application until the Pennsylvania statute was effective. It is equivalent to a holding that a finding of no price squeeze depended upon Penn Power foregoing its constitutional right to request permission to recover its cost of service. Such a holding by any administrative agency would clearly be arbitrary and capricious in violation of the Administrative Procedure Act. Accordingly, we find that while there was price discrimination, indeed a price squeeze for IIV2 months, Penn Power has successfully demonstrated mitigating circumstances that make such price discrimination not undue. Thus, no remedy for the price squeeze in this case should be afforded the Boroughs.
One Commissioner flatly rejected the Commission’s reasoning:
While I do not believe state rates or policies should be allowed to “drive” federal rates or policies, the courts have made it clear that this Commission is required to consider anticompetitive effects resulting from dual rate regulation and may place wholesale rates at the lower end of the zone of reasonableness in order to eliminate price squeeze, [citing Batavia and Bethany.]
She nevertheless concurred in the result on the grounds that, due to an error Penn had discovered in its filing,23 Penn had already *271lost revenue equivalent to the refund that would be ordered if the rates were set at the lower end of the zone of reasonableness. FERC II, 21 FERC at pp. 61,822-23.
C. The Price Squeeze Test
Boroughs argue first that they were improperly precluded from demonstrating that the price squeeze lasted beyond the period of nearly one year found by the Commission, and was therefore not traceable to and not excused by Penn’s decision to take advantage of new state law. Although they no longer argue that the rate of return test announced in the Commission’s Commonwealth Edison decision, 8 FERC ¶ 61,277 (1979), and later accepted by this court in Bethany, 670 F.2d at 199 n. 53, and Batavia, 672 F.2d at 90, is an inappropriate test of a price squeeze, they argue that the Commonwealth Edison decision was a significant departure from existing law, and that its application in this case deprived Boroughs of adequate notice of the material facts necessary to make out its ease. Under the rate of return test, they argue, they should have been permitted to discover and present additional evidence pertaining to a specific target of competition — the retail industrial customers located within the Boroughs — that they claim to have identified throughout the hearings as the primary object of competition.
The Commission responded to this argument in its opinion denying rehearing. FERC Ha, 22 FERC at p. 61,317 (1983). The Commission stated that the law in effect during the proceedings permitted designation of a class of retail customers for whom a wholesale customer alleges competition, and that it was therefore too late to designate a narrower class. Furthermore, the Commission found it unlikely that Boroughs were competing solely for industrial customers currently located within their boundaries and not those who might locate there in the future.
We are not fully persuaded by the Commission’s rationale. The price squeeze guidelines that the Commission claims gave adequate notice of the need to specify a target of competition require only designation of a rate schedule and give no indication of the need or possibility of designating a particular subgroup of a rate class such as Boroughs now identify as the target of competition. See 18 C.F.R. § 2.17(a).
Nevertheless we uphold the Commission’s decision not to order further discovery and reopen the proceedings. First, the price squeeze guidelines were, as we have emphasized in the past, merely guidelines for the establishment of a prima facie case. See Bethany, 670 F.2d at 197 n. 39. The Commission Staff’s consistent advocacy of a rate of return test and its attempts to identify the particular target of competition should have put Boroughs on notice of the desirability, even while continuing to advocate a different test, of designating a subgroup if it believed the Staff had not chosen the appropriate target.
In any event, we believe that Boroughs have misconceived the significance of a target of competition. According to our discussion in Batavia, Boroughs have grounds to complain only if they have a reasonable basis for believing that “there are a substantial number of retail customers for whom [they] could compete but for the subsidy those retail customers receive from other retail customers with whom they are grouped and for whom [Boroughs] could not compete.” Batavia, 672 F.2d at 90 (emphasis in original). For Boroughs eventually to prevail, and to demon-*272strate that the class of industrial customers located within Boroughs must be analyzed as a separate target of competition, they would have to show that the industrials outside the Boroughs (for whom Boroughs could not compete) had systematically lower cost characteristics than those inside the Boroughs while paying the same rate; the higher rate of return on the outside industrial customers would then in effect subsidize those inside. Yet at no point in these proceedings — even on appeal to this court — have Boroughs advanced even a theory as to how there could be such a subsidy running from retail customers outside the Boroughs to those within the Boroughs.
Of course we would not require Boroughs to demonstrate the existence of such a subsidy before affording them discovery of the data that they claim would enable them to find out if a subsidy existed. But given the fact that the rate of return test and the notion of a target of competition were prominently featured in the proceedings, we consider it fair to require Boroughs to bring forward some basis — in fact or even in theory — for believing that there could be a hidden subsidized subgroup of industrial customers within the Boroughs. Boroughs have not done so. We therefore uphold the Commission’s application of the rate of return test and its refusal to permit Boroughs additional discovery and further proceedings on a possible.target of competition.
D. The Denial of a Price Squeeze Remedy
The division of jurisdiction over electric utility rates between federal and state regulatory bodies has both made necessary and limited the authority of the Commission to consider and remedy price discrimination between wholesale and retail customers. Before Conway, the Commission denied its authority even to consider such allegations of price discrimination because of its lack of jurisdiction over retail rates. The Supreme Court held that the division of jurisdiction over rates did not preclude consideration by the federal Commission of price squeeze claims. This court in subsequent decisions has made it clear that it is in part “the coordination problems due to the division of regulatory power between federal and state government,” Batavia, 672 F.2d at 91, that give rise to the need for a price squeeze doctrine. See also Bethany, 670 F.2d at 200.
It is also clear, however, that the Commission’s authority to remedy a price squeeze is limited by the division of regulatory authority. For example, because it lacks jurisdiction to adjust retail rates, the Commission’s exercise of its authority to press wholesale rates to the lower end of the zone of reasonableness may sometimes leave unremedied and unremediable a degree of price discrimination.24
In its decision in this case, the Commission considered the extent to which “the paradox of dual utility rate regulation” may wholly excuse a price squeeze that otherwise would require a remedy. In our view, the Commission has in effect converted “the paradox of dual utility rate regulation” from a rationale for the price squeeze doctrine into a fertile source of excuses for the imposition of a price squeeze. While we do not have occasion in this case to consider the validity of the Commission’s announced “general rule” of not remedying a price squeeze that is the product of different policies and procedures at the state and federal level, we are not persuaded that it would necessarily be inappropriate to adjust wholesale rates downward under those circumstances. The conclusion that such an action “would be an abdication of the federal agency’s duties under its statutory charter” appears to rest on a lopsided reading of that statutory charter. The Commission’s statutory duty to ensure that rates are not unduly discriminatory is not so wholly subordinate to its duty to set just and reasonable rates that it may simply ignore the discriminatory ef*273feet of state ratemaking policies and procedures in setting a wholesale rate within the zone of reasonableness.
But the Commission went beyond its “general rule” in this case; while stating that a price squeeze resulting from a utility’s discretionary decision as to when to file retail and wholesale rate increases presents a more compelling case for a price squeeze remedy, the Commission found mitigating circumstances to excuse just such a result in the case now before us. We turn now to an examination of the Commission’s decision to excuse the price squeeze in this case.
1. Penn’s Decision to Delay the State Filing
We begin by observing that the Commission’s reasoning as to why this price squeeze should be excused even though it resulted from Penn’s own decision as to the timing of its filing appears to bootstrap from its analysis of price disparities resulting from different state policies and procedures: immediate filing at the state as well as the federal level may have resulted in a longer delay in rate relief, and thus in a longer period of price squeeze, due to the procedural peculiarities of Pennsylvania ratemaking. Because the Commission would not see fit as a general rule to remedy a price squeeze resulting from simultaneous filings followed by state regulatory delays, it does not make sense, according to the Commission, to remedy the shorter price squeeze resulting from Penn’s discretionary decision to delay filing.
We have already expressed skepticism as to the underlying judgment that a price squeeze resulting from different policies and procedures at the state level ordinarily should go unremedied. But in this case, we have additional problems with the Commission’s reasoning.
(a) The Constitutional Right to File Immediately. The ALJ suggested that Penn could have avoided a price squeeze by delaying its federal filing. This suggestion was made in part simply to rebut Penn’s argument that the price squeeze was a product of “state action” and thus excused. But the Commission inflated the significance of this point, stating that it would “clearly be arbitrary and capricious” to hold “that a finding of no price squeeze depended upon Penn Power foregoing its constitutional right to request permission to recover its cost of service.” The Commission’s reasoning is illogical but illuminating.
The Commission itself acknowledged that a price squeeze resulting from a difference in the timing of state and federal filing — i.e., an earlier federal filing — was ordinarily an appropriate candidate for a price squeeze remedy precisely because the utility could choose its filing dates. Yet in the following paragraph the Commission indicated that if the earlier federal filing was an exercise of the utility’s “constitutional right” to request federal rate relief, then it would be arbitrary and capricious to rely in part on the decision to file the federal increase earlier as the basis for affording a remedy to the wholesale customers.
Even apart from its illogic, the Commission’s reasoning appears to rest on the misconception that a price squeeze remedy is a penalty. Only then could the Commission conclude that it would be unfair to impose the remedy on a utility simply for requesting compensatory federal rates. The utility is free to request whatever rate it wishes and in any event will be entitled to a rate within the zone of reasonableness providing full cost recovery even if a price squeeze is found. The possible prospect of a price squeeze remedy being imposed in the future would therefore not lead any rational utility to forego its “constitutional right” to file for rate relief.25 The Corn-*274mission’s constitutional concerns are simply irrelevant to the decision whether or not to consider the price squeeze as a factor in setting just and reasonable rates.
(b) The Effect of the Decision to Delay. The Commission also attempts to east its decision in terms of the likely effect of Penn’s decision to delay its retail filing:
... Penn Power could be viewed as having taken the necessary steps to mitigate the price squeeze for the simple reason that, had it filed at the same time in both jurisdictions while Pennsylvania had no statutory suspension period, the length of the price squeeze period would likely have been much longer than IIV2 months.
The Commission repeated the point in its decision denying rehearing: “... Penn Power’s decision to delay its retail rate filing may have shortened the duration of the disparities in its wholesale and retail rates that would have resulted from earlier simultaneous filings.” FERC IIa, 22 FERC at p. 61,316. This assessment of the consequences of Penn’s decision is seriously flawed.
First, the Commission in making its initial decision apparently did not read the Pennsylvania statutes at issue closely enough. As Boroughs pointed out to the Commission in a petition for rehearing, the prior statute did have a statutory maximum suspension period of nine months. See supra note 11. It is also true that the prior statute granted the state commission the authority to set temporary rates, and in effect to continue the old rates far beyond the suspension period. It was this authority that was revoked by the statutory change. Id. The Commission’s misstatement of the law, although perhaps merely technical and in any event corrected in its decision denying rehearing, evinces an unseemly readiness to find merit in Penn’s position without a careful analysis of its factual and legal premises.
Second, the Commission paid no heed whatsoever to the gross disproportion between Penn’s claim of an immediate need for revenue and the tiny fraction of its revenues — three percent — that was accounted for by its wholesale customers. Instead it simply accepted without question Penn’s assertion that immediate wholesale relief was necessary to alleviate this need for revenue.
Third and perhaps most disturbing was the Commission’s failure to examine critically Penn’s assessment of the likely result of filing immediately under the old law. Penn’s prediction and the Commission’s conclusion depends on at least three contingencies being resolved against Penn. They assumed that the state commission would (i) suspend Penn’s rates for the maximum period, (ii) impose temporary rates virtually equal to the existing rates, and (iii) refuse any interim rate relief although it was authorized to grant such relief and although Penn claimed it could show a desperate need for it. Instead of examining the plausibility of each of these assumptions, the Commission was satisfied by the fact that Penn’s most recent retail filing had yielded no increase until twenty-two months later. It did not mention the fact that the two prior retail filings each yielded a rate increase within nine months.26
In view of the foregoing observations, it appears to us that Penn found an all-too-sympathetic audience in the Commission. The decision demonstrates little independent analysis or weighing of the evidence concerning Penn’s ability to avoid the price squeeze, but rather echoes uncritically Penn’s own witnesses as to the motivation for its decision. We find the Commission’s decision, on the whole, to be based on a curious mixture of technical inaccuracies, illogic and undocumented assumptions. Even if we were to accept the Commission’s logic — that, in the context of a changing state regulatory framework, a price squeeze inevitably resulting from a decision to delay filing should be excused if it is shorter than the price squeeze that would have resulted from simultaneous filing — we would find the factual support *275lacking. We cannot conclude that Penn explored all available options and chose a course of action that mitigated the price squeeze as much as possible under the circumstances. Penn’s single experience of a twenty-two-month delay does not persuade us that it could not have obtained more prompt relief if it had presented to the state commission the financial exigencies that FERC found so convincing. But our criticism of the Commission’s analysis goes deeper than the sufficiency of the evidence, to the very purpose of the price squeeze doctrine.
2. The Commission’s Restrictive View of the Price Squeeze Doctrine
Even if the Commission had found on the basis of substantial evidence that Penn’s decision to delay its state filing had the probable consequences of reducing the period of price squeeze, we would hesitate to conclude that the resulting price squeeze of almost a year’s duration was therefore automatically excused. The Commission’s reasoning in accepting this justification seems to signal an unduly narrow view of the purpose of the price squeeze doctrine.
(a) The Reemergence of Intent. We agree with the ALJ that the excuse proffered by Penn and eventually accepted by the Commission has “strong overtones of subjective intent.” The Commission found that Penn’s “only reason for delaying its [state] filing was that by waiting it could take advantage of a newly enacted statute establishing a suspension period.” This finding would have been directly contradicted by evidence that anticompetitive intent actually underlay the decision. Yet this is the very evidence that the Commission refused to hear.27 Under these circumstances we find singularly unconvincing the Commission’s attempt to recharacterize the argument as simply offering “factual circumstances that justify the tolerance of the price discrimination.”
(b) The Potential Scope of the Commission’s Reasoning. Furthermore, the factual circumstances underlying the difference in timing of the state and federal filings in this case do not seem particularly unique. The logic of the Commission’s argument as to why it should excuse this price squeeze potentially sweeps broadly indeed. The same logic might bar a remedy in a case where the utility had determined that a later state filing — resulting inevitably in a price squeeze — would result in a larger retail increase and consequently a price squeeze of lesser magnitude over the long run. The Commission’s reasoning might also extend to timing decisions based on other vicissitudes and vagaries of state utility regulation — anticipated changes in personnel, political climate, or policy — that might lead a utility to conclude that it could obtain faster or larger rate increases by deliberately delaying its filing at the state level. In other words, the Commission appears willing to excuse any price squeeze resulting from the utility’s judgment that it could get more money sooner from its retail customers by delaying its state filing than by filing simultaneously at the state and federal levels — a judgment that is of course ordinarily dictated by the utility’s financial self-interest. It is difficult for us to see what price squeezes would be remedied, where the utility had any rational business reason for its time-of-filing decisions. This is, to say the least, an exceedingly narrow interpretation of the scope of the price squeeze doctrine, and one that is at war with the Commission’s avowed aim to eschew discriminatory intent as a prerequisite of a price squeeze.
Throughout the Commission’s decision runs the implicit assumption that a price *276squeeze remedy is essentially a penalty that could not fairly be imposed if the price squeeze was the product either of circumstances beyond the utility’s control or of a “reasonable business judgment.” In accordance with this notion of the price squeeze remedy as a punitive measure, the Commission has declared that it will excuse virtually any price squeeze that can somehow be attributed to “the paradox of dual regulation,” presumably because it is not the result of blameworthy conduct by the utility. Thus understood, the Commission’s logic clashes mightily with the logic underlying Conway and its progeny, which recognize that “the paradox of dual utility rate regulation” is one of the basic problems the price squeeze doctrine was designed to deal with.
The dissonance we detect can be attributed above all to the complete absence from the Commission’s analysis of a concern for the discriminatory effect of the price squeeze on competition and on the municipal competitors, which is the very raison d’etre of the price squeeze doctrine. The Commission appears to have given little weight to these paramount considerations in its decision on the particular facts of this case, its rationale for that decision and its broader pronouncements on the circumstances that warrant a price squeeze remedy. Nor was the Commission’s decision to excuse the price squeeze based on its assessment of the effect of any remedy on Penn’s financial viability.
3. “The Paradox of Dual Regulation” and the Purpose of Price Squeeze Doctrine
The Commission is vested with discretion under Conway to ameliorate or eliminate a price squeeze by lowering rates within the zone of reasonableness. See, e.g., FPC v. Conway Corp., 426 U.S. 271, 279, 96 S.Ct. 1999, 2004, 48 L.Ed.2d 626 (1976); Kansas Cities v. FERC, 723 F.2d 82, 94 (D.C.Cir.1983). But the Commission’s discretion is not so broad that it may decide to ignore the anticompetitive effects of a price squeeze without even considering factors that are critical in light of the purposes of the Conway doctrine. We believe that when the Commission finds that wholesale rates, compared with retail rates, demonstrate non-cost-justified price discrimination with a significant impact on the wholesale customer’s ability to compete in the retail market, it must at least consider this price discrimination and its anticompetitive effect in setting a just and reasonable rate. As we said in Kansas Cities v. FERC, 723 F.2d at 95, “[t]o prescribe rates that are known to be unduly discriminatory or preferential is to prescribe rates that are known to be unlawful, which would be a violation of the Commission’s responsibilities.”
This is not to say that the Commission necessarily must lower the rate it had determined to be just and reasonable apart from allegations of discrimination. For example, a utility’s immediate need for additional revenue, without which a decline in its service capability is seriously threatened, may well outweigh a relatively minor price squeeze and justify a conclusion that the price discrimination is not “undue.” The Commission would necessarily have considerable discretion in weighing the anticompetitive effect of a rate against such factors as the utility’s ability to attract sufficient capital to maintain adequate service to wholesale and retail customers alike, or the immediacy and magnitude of the utility’s need for additional revenue.28 But we are convinced that, in light of the purpose of the price squeeze doctrine, the discretion granted the Commission by Conway when it is faced with a proven price squeeze is in large part the discretion to weigh the effects of a rate and a rate disparity on both the supplying utility and its customers.
This interpretation of the Conway doctrine places far less emphasis than the *277Commission’s on the explanation for a price squeeze, whether it lies in “the paradox of dual regulation” or in some other reasonable business judgment by the supplying utility. Instead it seeks to advance the underlying goal of the doctrine, which is “to assure that genuinely competitive wholesale customers would not be price squeezed out of a retail market because of the imperfections of regulation, an assumed propensity of utilities for unjustified monopoly power or the coordination problems due to the division of regulatory power between federal and state government.” Batavia, 672 F.2d at 91.
We thus hold that the decision not to ameliorate a proven price squeeze ordinarily must be based on the Commission’s determination that the anticompetitive effect of the price squeeze on the wholesale customer/retail competitor is outweighed by the effect of a remedy on the supplying utility’s financial viability and its ability to serve all its customers. Although there may be extraordinary circumstances in which the explanation — whether the cause or the reason — for a price squeeze may mitigate or even excuse price discrimination, we are not persuaded that those circumstances are present here. It is primarily the effects of the price squeeze and its prospective remedy that should guide the Commission’s exercise of discretion, and these effects were completely ignored in the Commission’s decision.
We are forced to conclude, on the basis of both the inconsistencies and inaccuracies in the Commission’s explanation of the particular mitigating circumstances of this case and the irreconcilability of that explanation with the very purpose of the price squeeze doctrine, that the Commission's decision to excuse the price squeeze in this case cannot stand. The Commission has not exercised its discretion under Conway in a manner consistent with the purpose of the Conway doctrine to guard against price discrimination by retail suppliers against their municipal competitors.
IV. Conclusion
We uphold each of the Commission’s Phase One determinations — the forty-five-day working capital allowance, the refusal to adjust the test period estimate of revenues from sales of excess reserves, and the rate of return on equity. We also uphold the Commission’s decision in Phase Two to apply the rate of return test without permitting Boroughs additional discovery and further evidentiary hearings. However, we vacate the Commission’s decision not to issue a price squeeze remedy for the uncontested price squeeze of nearly one year, and we remand for reconsideration of this matter in light of this opinion.
So ordered.
. The Boroughs of Ellwood City, Grove City, New Wilmington, and Zelienople, Pennsylvania.
.Penn filed with the Federal Power Commission, whose powers were transferred to the Federal Energy Regulatory Commission (FERC) on October 1, 1977, pursuant to the Department of Energy Reorganization Act, 42 U.S.C. § 7172(a)(1), and Executive Order No. 12009, 42 Fed.Reg. 46,267.
.The ALJ decision in Phase I is reported at 9 FERC ¶ 63,041 (1979) [hereinafter cited as "ALJ I, 9 FERC at p. -”]. The Commission’s decision in Phase I appears at 12 FERC ¶ 61,049 (1980) [hereinafter cited as “FERC I, 12 FERC at p. —’'].
. The Commission ordered Penn on June 4, 1982, to submit additional filings to ascertain the rate of return comparisons based on the actual retail rates in effect from September 11, 1977, to January 23, 1982, the entire period in which the wholesale rates under consideration were in effect, or the "locked-in” period.
. The ALJ decision in Phase II is reported at 11 FERC ¶ 63,009 (1980) [hereinafter cited as “ALJ II, 11 FERC at p. -”]. The Commission’s decisions are reported at 21 FERC ¶ 61,313 (1982) [hereinafter cited as “FERC II, 21 FERC at p. —”], and 22 FERC ¶ 61,182 (1983) (denying rehearing) [hereinafter cited as "FERC IIa, 22 FERC at p. -”].
. The net lag would be divided by 365 and then multiplied by the annual operating expenses, resulting in a cash allowance equal to some fraction of those expenses.
. See infra note 11.
. The notice of proposed rulemaking was issued on June 7, 1979. "Calculation of Cash Working Capital Allowance for Electric Utilities," 44 Fed. Reg. 33410 (1979). No rule has yet been issued.
. Boroughs had actually urged the Commission to reduce the ratebase; they had subtracted from the working capital amount yielded by the thirteen-day net lag amounts collected from customers to pay interest on long-term debt, taxes, and dividends. The AU rejected these adjustments because he found that these were not working capital items. AU I, 9 FERC at p. 65,160. Boroughs appear no longer to advocate the inclusion of these amounts.
.This disparity is attributable primarily to the use in Boroughs’ study of an estimated 15-day "payment lag” — the time from billing to payment — when the actual payment lag was about 30 days according to Penn witnesses and the single customer study. As we note below, the Commission pointed out that there was also a fifteen day overestimate in this study: it had used a 30-day rather than a 15-day "service lag.” Apart from this error, the Commission found without objection that this single-customer study was not an adequate substitute for the 45-day convention. FERC I, 12 FERC at n. 18.
. Boroughs also objected to the application of the 45-day rule on the grounds that the Commission’s current formulation of the rule’s meaning — the estimated lag from the time of payment of expenditures for services to the time of receipt of revenues for those services — is a misstatement of the rule. It cites the original formulation in Interstate Power Co., 2 FPC 71, 85 (1939), as the correct one: ”[t]he full period between the dates of rendition of services and the payment." The Commission points out in response that since 1949 it has used the current formulation — the period between expenditures and receipt of revenue. See Pennsylvania Water & Power Co., 8 FPC 1, 75 (1949). We agree with the Commission that “Boroughs’ argument ignores 40-odd years of Commission practice,” Brief for FERC at 31, and must be rejected.
. The Commission estimated that Boroughs’ requested adjustment would amount to about 166,400. Brief for FERC at 24.
. In contrast, we concluded that the Commission had failed to justify the use of estimates in Villages of Chatham v. FERC, 662 F.2d 23 (D.C.Cir.1981). The case before us is readily distinguishable from that one, in which we overturned the Commission because it had failed to adjust estimates of wholesale customers’ demand even though the estimates were unsupported by a reasonable explanation, and were "(1) contradicted by actual results, (2) theoretically impossible and (3) based upon data a good deal more impoverished than those relied upon *265by the utility to forecast its retail demand growth.” Id. at 32.
. Boston Edison had a lower common equity ratio and a lower after-tax interest coverage on long-term debt. 9 FERC at p. 65,165.
. See United States v. Aluminum Co. of America, 148 F.2d 416, 436-38 (2d Cir.1945). In general, "a ‘price squeeze’ occurs when a wholesale supplier, who also sells at retail, charges such high rates to its wholesale customers that they cannot compete with the supplier’s retail rates.” Hjelmfelt, A Price Squeeze Theory for Implementation of Federal Power Commission v. Conway, 50 U.Colo.L.Rev. 459 (1979).
. Boroughs claimed, in other words, that a retail customer with precisely the same cost characteristics as Boroughs would pay the lower rates offered under one of Penn’s retail rate schedules. They urged the Commission to find a price squeeze whenever the retail rate, for which Boroughs would qualify but for their identity as wholesale/resale customers, was lower than the wholesale rate. This theory was urged by the municipalities in Batavia, 672 F.2d 64, and rejected in favor of the Commission’s rate of return test, described infra, which we found to be more sensitive to differences in cost characteristics between the two classes of customers. Id. at 89-90.
.Penn had urged the exclusion of those of its customers, six in number, that were served at the highest voltage level because Boroughs did not have and had never requested the capability of serving at 138 KV. In other words, Penn argued, these customers were not properly included in the target of competition because Boroughs could not compete for them. The 138 KV customers yielded a low rate of return — 4.23% compared with 4.64% and 5.20% for other industrial retail customers. Their exclusion, therefore, would raise slightly the overall rate of return for the target of competition, thus reducing the magnitude of the disparity that indicated a price squeeze. See ALJ II, 11 FERC at p. 65,049.
. See ALJ II, 11 FERC at p. 65,049. According to Penn’s own figures, the wholesale customers yielded a rate of return under the proposed rates (i.e., those that took effect after suspension on September 11, 1977) of 9.31%, compared with rates of return between 4.23% and 5.20% for various classes of industrial customers under the retail rates effective until August 31, 1978. Non-industrial retail customers yielded a rate of return of 5.72%; exclusion of those customers thus resulted in a disparity of somewhat greater magnitude than would have appeared if all retail classes had been included in the comparison.
. See supra note 17.
. The prior state law provided that rates must be filed two months before their proposed effective date, and could then be suspended for up to nine months. 66 Pa.Stat. § 1148(b) (repealed). But it also permitted the state commission to impose temporary rates, which could be equal to preexisting rates, for an indefinite period until final rates were determined. 66 Pa.Stat. § 1150. These temporary rates went into effect subject to a refund if greater than the rates ultimately approved, 66 Pa.Stat. § 1153, or a surcharge if less than the final rates. 66 Pa. Stat. § 1150(e). Penn's most recent filing under this statute had resulted in an effective suspension period of 22 months. ALJ II, 11 FERC at p. 65,053.
. The new law, 66 Pa.Cons.Stat. § 1308(b), like the old, provided that rates must be filed not less than two months before the proposed effective date, and could then be suspended for up to nine months. The new law also provided for the imposition of temporary rates but with an explicit exception in the case of a general rate increase such as that in the case before us. 66 Pa.Cons.Stat. § 1310(a).
. The Commission based this conclusion in part on a footnote from our decision in Batavia, 672 F.2d at 90 n. 52:
Further, neither Conway nor Batavia require the Commission to set a wholesale rate so that wholesale customers are guaranteed the ability to compete in the retail market. It may be, for instance, that the state commission is not allowing an adequate rate of return. If so, the Commission is not obliged to follow, and indeed may not follow suit, for the rates it sets must fall within the zone of reasonableness____
(Emphasis supplied by Commission.) The obvious and limited meaning of this passage is that if the rates set by the state commission were so low as to create a price squeeze at the lowest reasonable wholesale rate, the Commission could still not set rates below that level. The Commission brushed aside this interpretation, apparently because it thought the point was too obvious. It interpreted the passage instead as support for the view that the Commission should not, as a general rule, lower rates at all — even within the zone of reasonableness — in response to state regulatory actions. The Commission's interpretation of our decision borders on the disingenuous.
. In preparing materials requested by the Commission in its "Order Requesting Additional Filings,” Penn discovered what it called a "computational error” in its compliance filing in calculating certain fuel adjustment charges it expect*271ed to receive. This alleged error, when incorporated into the "just and reasonable” rates approved by the Commission in Phase I, resulted in a loss, through excessive refunds, of about $300,000. Penn unsuccessfully petitioned the Commission for recovery of this amount, and appealed the Commission’s decision to this court. We agreed with the Commission that Penn’s request ran afoul of both the "filed rate doctrine,” see Montana-Dakota Utilities Co. v. Northwestern Pub. Serv. Co., 341 U.S. 246, 251, 71 S.Ct. 692, 695, 95 L.Ed. 912 (1951), and the principle against retroactive ratemaking, see Public Serv. Co. v. FERC, 600 F.2d 944, 958 (D.C.Cir.1979), and denied Penn’s petition. Pennsylvania Power Co. v. FERC, 725 F.2d 126 (D.C.Cir.1984) (per curiam).
. We have previously indicated that the Commission’s power to remedy a price squeeze was limited by the lower end of the zone of reasonableness. See Batavia, 672 F.2d at 90 n. 52.
. We assume, although the Commission has not found it necessary to decide, that the price squeeze remedy would be limited to the period of the price squeeze — i.e., in this case -11 '/i months — rather than the entire period for which the wholesale rate was in effect. Of course we do not decide that this must be so.
. See ALJ II, 11 FERC at p. 65,051.
. The Commission’s analysis is roughly analogous to a truncated employment discrimination doctrine that permitted the employer to defeat a prima facie case of discrimination merely by articulating a legitimate nondiscriminatory reason for a hiring decision without allowing the victim to show that the articulated reason was actually a mere pretext for discrimination. Cf. Texas Dep’t of Community Affairs v. Burdine, 450 U.S. 248, 254-56, 101 S.Ct. 1089, 1094-95, 67 L.Ed.2d 207 (1981). Of course, we do not suggest that the Commission should adopt the framework of Title VII for its price squeeze doctrine; we mention it only as an analogy to show what we see as a gap in the Commission’s logic.
. For example, the Commission in a case like this one might consider the duration of the price squeeze and special circumstances such as the effect of Penn’s alleged computation error, as Commissioner Sheldon would have done. See supra at 972.