Opinion for the court filed by Circuit Judge TAMM.
TAMM, Circuit Judge:Income taxes are one of the costs of service that the Federal Energy Regulatory Commission must consider in setting wholesale rates for public utilities and natural gas pipeline companies. Determining the proper tax allowance is complicated by disparities between the policies of the Commission and the Internal Revenue Service. For example, ratemaking rules may require that a current utility expense not be borne entirely by current ratepayers but rather be charged to customers over time, while tax rules may permit the utility to deduct the entire expense in the current year. The issue then is whether the utility’s tax deduction for the expense should reduce rates only in the current year, the flow-through method, or whether the benefit of the tax deduction should be spread over the period that ratepayers are charged for the expense, the normalization method. Over the past twenty years the Commission has several times changed its position on the normalization versus flow-through issue. In 1976 the Commission adopted a general policy of permitting normalization. This court in Public Systems v. FERC, 606 F.2d 973 (D.C.Cir.1979) [hereinafter Public Systems I], held that the Commission had failed to provide a reasoned basis for its action. On remand the Commission addressed the concerns of the court and the issues raised by commenters and ruled that generally normalization must be employed in computing the income tax component of cost of service. We find that the Commission has adequately answered the concerns expressed by the court in Public Systems I and that the tax normalization policy is the product of reasoned decisionmaking. Accordingly, we affirm.
I. Background
Under the Federal Power Act, 16 U.S.C. § 824 et seq., and the Natural Gas Act, 15 U.S.C. § 717 et seq., the Federal Energy Regulatory Commission1 is charged with setting just and reasonable wholesale rates for interstate, investor-owned public utilities and natural gas pipeline companies. The rates are based on cost of service, which comprises “all expenses incurred, including income taxes, plus a reasonable return on capital.” Public Service Co. v. FERC, 653 F.2d 681, 683 (D.C.Cir.1981).
Determining the cost of service tax allowance is difficult because of timing difference transactions. Timing differences arise from differences between tax regulations and regulatory pricing policies. Because their policy objectives are not always the same, taxing authorities and regulatory commissions do not always allocate costs and revenues to the same time period. A timing difference occurs when an expense or revenue is recognized for tax filings in periods before or after it is recognized in rates. For example, the interest expense incurred in constructing a plant yields the utility a tax deduction in the year the expense is incurred, but the expense is capitalized and deferred for ratemaking purposes until the plant is constructed and placed in service. See Notice of Proposed Rulemaking, Docket No. RM80-42, Appendix A, Staff Study on Comprehensive Tax Normalization (Staff Study), at 6-10 (March 31, 1980), Joint Appendix (J.A.) at 85-89.2
Flow-through and normalization are methods of dealing with the tax effects of *250timing difference transactions. Under flow-through, ratepayers realize the tax benefits of expenses when those expenses are used as tax deductions by the utility.3 The entire benefit arising from a tax deduction passes through to the company’s current customers. The primary rationale for flow-through is the actual taxes paid principle, i.e., in each year customers are charged no more than the current tax liability of the utility. Under normalization, the ratepayers who are charged with an expense receive the benefit of the tax deduction.4 Normalization matches the tax effect of an expense with the ratemaking treatment for recovery of that expense. The primary rationale for normalization is the matching principle, i.e., equity demands that the customers who pay the expense receive the tax benefit associated with that expense.
By definition, timing differences do not result in permanent differences in the rate-making and tax books. Every difference that arises in one period will exactly reverse itself in one or more subsequent periods. Regardless of whether flow-through or normalization is employed, over time the rate-making and tax accounting procedures will recognize the. same amount of revenue or expense with respect to each transaction. In contrast, permanent differences are not offset by reversals in other periods. Permanent differences arise when the tax law places a different dollar value on a revenue or expense than does the ratemaking law. An example of a permanent difference is municipal bond interest, which is excluded from taxable income but receives no special treatment in ratemaking. See generally Staff Study at 7-8, J.A. at 86-87. Permanent differences are not within the scope of the rules under review in this case.
The flow-through versus normalization dispute focuses primarily on the possibility that normalization will lead to utilities’ enjoying continual tax deferrals. Under normalization, a deferred tax account records the difference between taxes actually paid by the utility and taxes charged to ratepayers. Because the predominant timing differences are those that arise from expenses that are recognized as tax deductions prior to their collection in rates, during the first year of normalization customers are usually charged for more taxes than the utility actually pays.5 If the utility produces the same revenues and expenses for a number of years, the utility will eventually be paying the same amount of taxes that it charges ratepayers. Utilities, however, generally grow and produce more revenues and expenses each year. Therefore, the deferred tax account tends to grow each year.6 Although each timing difference will reverse itself, the deferred tax account may increase for the foreseeable future. Petitioners argue that this continual tax deferral results in ratepayers’ being charged for taxes that are never paid by the utility.
The normalization versus flow-through issue first arose after the 1954 enactment of section 167 of the Internal Revenue Code, which authorized the use of accelerated depreciation for tax purposes. The Commission initially permitted utilities to normalize the benefits of rapid depreciation, and the courts upheld this policy. E.g., El Paso Natural Gas Co., 22 F.P.C. 260, 267 (1959), aff’d 281 F.2d 567, 573-74 (5th Cir.1960), cert. denied, 366 U.S. 912, 81 S.Ct. 1083, 6 L.Ed.2d, 236 (1961). The Commission changed its position in the mid-1960’s and *251required the use of flow-through with accelerated depreciation. Alabama-Tennessee Natural Gas Co., 31 F.P.C. 208 (1964)7 The courts again affirmed the Commission’s choice. E.g., Alabama-Tennessee Natural Gas Co. v. FPC, 359 F.2d 318, 339 (5th Cir.), cert. denied, 385 U.S. 847, 87 S.Ct. 69, 17 L.Ed.2d 78 (1966). The Tax Reform Act of 1969 partially reversed Commission policy by allowing a utility to use normalization with respect to expansion property. Tax Reform Act of 1969, § 441(a), Pub.L. No. 91-172, 83 Stat. 487, 625-28 (codified at I.R.C. § 167(l)).8 Again changing its position, the Commission in 1970 extended nor-, malization to utility property acquired before 1970 and replacement property. Texas Gas Transmission Corp., 43 F.P.C. 824 (1970). The Supreme Court upheld the change by finding that the Tax Reform Act had not limited the authority of the Commission to permit utilities to change from flow-through to normalization. FPC v. Memphis Light, Gas & Water Division, 411 U.S. 458, 472-73, 93 S.Ct. 1723, 1731, 32, 36 L.Ed.2d 426 (1973).9 The present controversy began in 1975 when the Commission adopted a general policy of permitting normalization with respect to all timing difference transactions not covered by prior Commission orders.
II. Case History
In June 1975, the Commission issued Order No. 530, 5 Fed.Power Serv. (MB) 5-1017 (FERC June 18, 1975), announcing a general policy of permitting normalization upon an appropriate factual showing. Id. at 5-1024. The Commission found that utilities faced problems in generating needed capital and that normalization would improve regulated firms’ cash flow and capital financing. Id. In January 1976, the Commission modified its rule in Order No. 530-A, 8 Fed.Power Serv. (MB) 5-224 (FERC Jan. 19, 1976). It stated that a utility seeking normalization would have to demonstrate that only tax deferral, as opposed to tax savings, would result from the change. Id. at 5-227-28. In July 1976, the Commission again amended its rule by issuing Order No. 530-B, 10 Fed.Power Serv. (MB) 5-187 (FERC July 6, 1976), which adopted a general policy of permitting normalization as long as only timing differences are involved. Id. at 5-188-89. The Commission held normalization the preferable method for ratemaking and accounting purposes. Id.
Several municipally-owned utilities challenged the Commission’s orders as permitting regulated companies to receive payments for tax costs that they may never incur. In Public Systems I, this court held that the Commission had failed to “assess the consequences of its action for the industry” and to “indicate ‘fully and carefully’ the purposes behind the order,” 606 F.2d at 980, and remanded the orders to the Commission, id. at 983. More specifically, the court found: (1) that there was no assessment of the impact of the orders on consumers or utilities and no indication of the financial resources at issue, id. at 980; (2) *252that the orders did not specify all of the expenses that would be subject to normalization, id. at 981; (3) that most of the expenses that were specified were inadequately discussed, id.; (4) that there was nothing to support the Commission’s conclusion that there would be no tax savings, id.; (5) that the orders did not indicate the goals of the Commission’s policy, id.; and (6) that the Commission gave no reason for its decision to consider the price squeeze implications of normalization on a case-by-case basis, id. at 982-83. Shortly after the court’s decision, the Commission issued an Order Establishing Interim Procedures in Docket Nos. R-424 and R-446 (June 8, 1979) 44 Fed.Reg. 34,471 (1979), that provided that Order No. 530-B would remain in effect pending the outcome of the new rulemaking and that all rate decisions would include a provision permitting utilities to retain amounts related to normalization subject to refund. See Cities of Batavia v. FERC, 672 F.2d 64, 78-79 (D.C.Cir.1982).
In response to Public Systems I, the Commission on March 31, 1980, issued a notice of proposed rulemaking that would permit utilities and pipelines to use normalization for all timing difference transactions not covered by prior Commission orders. Notice of Proposed Rulemaking (Notice), Docket No. RM80-42 (March 31, 1980), J.A. at 17.10 The notice addressed the concerns expressed by the court in Public Systems I and contained a staff study of the impact of normalization on rates, cash flow, cost-of capital, efficiency, and rate stability. Comments on the notice were filed by 38 electric utilities, 5 electric utility groups, 4 natural gas pipeline companies, 8 customer groups, and others. Order No. 144, Appendix A (May 6,1981), J.A. at 740-41. The Commission also received reply comments. J.A. at 569-604. On May 6, 1981, the Commission issued the final rule. It differed from the proposed rule in that it required, rather than permitted, tax normalization. Order No. 144 at 2-3, 129, J.A. at 608-09, 735. Also, instead of requiring the continuation of flow-through for timing differences arising prior to the adoption of normalization, the Commission required utilities to make some provision in future rate cases to account for deficiencies in deferred tax accounts as a result of the prior use of flow-through. Id. at 3, 130-31, J.A. at 609, 736-37. Order No. 144 responded in detail to the concerns raised in Public Systems I and to the arguments raised by commenters.
Applications for rehearing were filed by several participants in the rulemaking, and on August 13, 1981, Congress passed the Economic Recovery Tax Act of 1981 (ERTA), Pub.L. No. 97-34, 95 Stat. 172. The Commission then requested and received comments on the impact of the new Tax Act on the final rule. J.A. at 843-50. *253On February 22, 1982, the Commission issued Order No. 144-A, J.A. at 851, which rejected the arguments challenging the decision to adopt normalization and declined to reopen the record because of the passage of ERTA. The Tax Act created a new depreciation system, Accelerated Cost Recovery System (ACRS), that replaces all prior systems for property placed in service after 1980. ERTA § 201(a) (codified at I.R.C. § 168). ACRS abandons the useful life concept and instead prescribes annual depreciation percentages by assets class and prescribes the treatment of asset salvage value. Id. Under ACRS, a utility must use normalization in ratemaking for depreciation timing differences attributable to the use of ACRS or lose all associated tax benefits. ERTA § 201(a) (codified at I.R.C. § 168(e)(3)).11 The Commission reasoned that because it had no option to permit flow-through of ACRS timing differences, there was no need to reopen the record to analyze the impact of ERTA on the final rule. Order No. 144UA at 9, J.A. at 861.
Representatives of publicly-owned municipal electric distribution systems that purchase electricity at wholesale from investor-owned public utilities now challenge Order Nos. 144 and 144-A. Various investor-owned public utilities subject to the Commission’s ratemaking jurisdiction have intervened to support the Commission in this appeal.
III. Discussion
The ultimate issues are whether the Commission has adequately answered the concerns expressed by this court in Public Systems I and whether the tax normalization policy is the product of reasoned decision-making. In reviewing the Commission’s decision, we keep in mind the Supreme Court’s statement that ratemaking bodies are not bound to “any single formula or combination of formulas” and that if the Commission’s order “produces no arbitrary result, our inquiry is at an end.” FPC v. Natural Gas Pipeline Co., 315 U.S. 575, 586, 62 S.Ct. 736, 743, 86 L.Ed. 1037 (1942).12 We first examine the questions raised in Public Systems I and then turn to petitioners’ other arguments.
*254A. Issues Raised in Public Systems I
1. Goals of the Commission’s Policy
The Commission’s primary rationale for normalization is the matching principle: as a matter of fairness, customers who pay an expense should get the tax benefit that accompanies the expense.13 Under normalization the tax effect of an expense is recog- ' nized in the cost of service in the same time period as the expense itself. To do otherwise would subsidize present customers at the expense of future ones. Order No. 144 at 10-12, J.A. at 616-18.
Underlying the matching principle is the concept of “used and useful” property. Property or equipment can usually be included in the rate base only if the property is providing service to current ratepayers.14 For example, construction work in progress is generally not included in the rate base.15 Normalization is consistent with the used and useful property concept in that it allocates the tax benefits of expenses to the periods when ratemaking policy recognizes the expenses. Conversely, flow-through is inconsistent with general ratemaking policy and gives special treatment to taxes. When tax law permits the interest expense from construction work in progress to be immediately deducted, and regulatory policy requires the expense to be capitalized, flow-through follows the tax law’s interperiod allocation rather than normal regulatory policy and gives current ratepayers the tax benefit. Id. at 27-28, J.A. at 633-34.
Petitioners contend that the matching principle is inconsistent with the ratemaking treatment of rate base. Since the entire investment of a plant is included in the rate base when it is placed in service, consumers pay more in rates in the early years when the plant is undepreciated than in later years even though all consumers benefit equally from use of the plant in each year. Therefore, the argument runs, the' Commission treats rate base without any concern for the matching principle. The Commission responds first that leveling rates is not the principal objective of normalization; rather, normalization seeks to give customers who pay an expense the tax benefits generated by the expense. Moreover, flowing through the tax effects of the predominant types of transactions covered by the rule, construction-related interest, tax, and pension costs, would not produce more level rates because flow-through would reduce rates only during the plant’s construction. Further, rates are higher in the early years because the plant is more valuable before it depreciates, and the Commission cannot deny a utility a reasonable return on its investment. See Bluefield Water Works & Improvement Co. v. Public Service Comm’n, 262 U.S. 679, 690, 43 S.Ct. 675, 678, 67 L.Ed. 1176 (1923). Because the Commission generally requires utilities to use straight-line depreciation, customers are charged the same depreciation expense each year over the life of the plant. Nevertheless, utility rates are higher in the early years because the plant is more valuable then and its higher value raises the rate base. Suggesting that the method of depreciation be changed to even out payments involves complicated cash flow and financial questions and far exceeds the scope of the orders at issue. Order No. 144 at 98-100, J.A. at 704-06.
Petitioners also argue that the rule-making has gone only part of the way toward achieving true matching because it does not normalize the tax-on-tax effect. Tax normalization generally increases revenues in the first years and thereby increases current taxes. This increase in taxes caused by normalization, petitioners contend, should be deferred and recovered in rates over the period of the original normal*255ization. The Commission responds that any increase in taxes resulting from normalization is not a timing difference transaction subject to normalization because the increased taxes are not an expense item for tax purposes. Also, petitioners’ argument incorrectly assumes that flow-through is the proper method and that the tax-on-tax effect is incurred for the benefit of future ratepayers. Normalization, however, does not create timing differences but rather is a method of equitably dealing with them, and normalizing the tax-on-tax effect would work to defeat the matching principle. Order No. 144 at 97-98, J.A. at 703-04. We find that the Commission has adequately indicated and explained the rationale for adopting normalization.
2. Permanent Tax Savings
Ratemaking is generally cost-based, and it is usually improper to consider “costs” that a utility does not actually incur. See Public Systems I, 606 F.2d at 978 n. 24. Petitioners argue that normalization will lead to permanent tax savings because utilities will charge customers for taxes that the utilities will never pay. Under normalization, a deferred tax account records the difference between taxes actually paid by the utility and taxes charged to ratepayers. Because the predominant timing differences are those that arise from expenses that are recognized as tax deductions prior to their collection in rates, during the first year of normalization customers are usually charged for more taxes than the utility actually pays. Because utilities generally grow and produce more revenues and expenses each year, the deferred tax account tends to grow each year. This continual tax deferral, the argument runs, is a permanent tax savings. The Commission responds that only timing difference transactions qualify for normalization and, by definition, every individual timing difference turns around in later years. The growth in the deferral tax account simply reflects the growth of the company. Also, there cannot be a continual deferral of tax payments by the utility unless there is a corresponding deferral in the collection of the associated non-tax costs by the utility. Assuming a fifty percent tax rate, for every dollar of accumulated deferred taxes recorded on the company’s books, there must be two dollars of unrecovered investment in assets. When the utility has recovered in rates all of the investment associated with the timing difference, there will be no deferred tax benefits that have not been reflected in lower rates. Order No. 144 at 44 — 46, J.A. at 650-52.
The courts have never held that the Commission must disallow normalization if there is a possibility of a growing balance in the deferred tax account. In El Paso Natural Gas Co. v. FPC, 281 F.2d 567 (5th Cir. 1960), cert. denied, 366 U.S. 912, 81 S.Ct. 1083, 6 L.Ed.2d 236 (1961), the Fifth Circuit affirmed the Commission’s decision to permit normalization of the tax effects of liberalized depreciation. The court stated, “The one exception [to the actual taxes paid doctrine], if it be an exception, is the treatment to be afforded the tax saving or deferral resulting from the use by the taxpayer of the declining balance method of depreciation of new equipment. In practical effect this works a tax deferral rather than a tax savings.” Id. at 573. In Cities of Lexington v. FPC, 295 F.2d 109 (4th Cir.1961), the Fourth Circuit, without indicating that a finding on continual tax deferrals was necessary, upheld the Commission’s decision to permit tax normalization: “It is obvious that we are confronted in this case with a problem in a special field which requires the exercise of expert skill that has been generally entrusted by state and Federal legislation to administrative regulatory tribunals whose decisions, in the absence of abuse of discretion, should be followed by the courts.” Id. at 114. In the mid-1960’s, the Commission found that-accelerated depreciation resulted in true tax savings and ruled that utilities could no longer normalize accelerated depreciation. The Fifth Circuit affirmed the policy change in Alabama-Tennessee Natural Gas Co. v. FPC, 359 F.2d 318 (5th Cir.), cert. denied, 385 U.S. 847, 87 S.Ct. 69, 17 L.Ed.2d 78 (1966): “In final analysis, putting to one side the issue of *256congressional intent, the petitioner’s challenge to the FPC order addresses itself to the competency of the Commission to choose between competing accounting theories and competing economic guesses. We see nothing irrational in the Commission’s choice.” Id. at 339. In 1970 the Commission, finding that no tax savings would result, extended normalization to the depreciation of utility property acquired before 1970 and replacement property. The Supreme Court upheld the policy change by finding that the Tax Reform Act of 1969 had not limited the authority of the Commission to permit utilities to change from flow-through to normalization. FPC v. Memphis Light, Gas & Water Division, 411 U.S. 458, 472-73, 93 S.Ct. 1723, 1731-32, 36 L.Ed.2d 426 (1973). These cases teach that a reviewing court should accept the Commission’s analysis of the tax savings versus tax deferral issue if the Commission has engaged in reasoned decisionmaking. We find the Commission’s conclusion that tax savings will not result from the use of normalization reasonable and adequately supported.
3. Impact of Normalization
The Commission’s staff undertook a comprehensive study of the impact of tax normalization,16 which the Commission generally adopted. Order No. 144 at 15, J.A. at 621.17 First, the Commission found that although tax normalization may initially result in higher rates than flow-through, the increase in rates will not be substantial. Id. The impact analysis indicated that for 1980 normalized rates exceeded flow-through rates by 2.2 percent for electric utilities and 0.1 percent for gas pipelines. Staff Study at 98, J.A. at 177. In the year 2000 normalized rates are projected to be less than flow-through rates in the pipeline industry and to be a maximum of 0.9 percent above flow-through rates in the electric utility industry. Id.18 The Commission noted, however, that these figures probably overstate the difference in rates. Notice at 39 n. 47, J.A. at 55.19 Although the Commission found that normalization may have a favorable impact on cost of capital, Order No. 144-A at 30, J.A. at 882, the study assumed no cost of capital differential between normalization and flow-through because of uncertainty about the magnitude of the differential. Staff Study at 108, J.A. at 187. If one assumes a 0.4 percent cost of capital differential favoring utilities using normalization, normalized rates are projected to fall below flow-through rates by 1990. Id. at 110, J.A. at 189. Also, the study did not take into account plant removal costs, which would work to decrease the differ*257enee in rates,20 because of the speculative nature of the costs of offshore plant removal and nuclear decommissioning. Id. at 113-15, J.A. at 192-94.
The study demonstrates that even though continual tax deferrals are likely to exist for the foreseeable future, normalized rates may fall below flow-through rates. This is possible because of the Commission’s policy of deducting accumulated deferred taxes from the rate base. The Commission reasons that a utility should not earn a return on funds contributed by ratepayers. Customers’ rates may be lower under normalization because the utility need not borrow as much money due to the deferred tax account and because the utility earns no return on the deferred taxes. Notice at 41-42, J.A. at 57-58.
This court in Public Systems I stated that the “pro-normalization policy must comport with the spirit of federal utility regulation by ensuring that consumers at least will suffer no detriment under [it].” 606 F.2d at 979. This statement does not mean that the Commission must adopt the method that produces the lowest current rates or that normalization cannot be adopted if it results in continual tax deferrals. Because protecting the consumer includes maintaining the financial integrity of utilities and includes considering future rate levels as well as current ones, “normalization [is] permissible even though it might result in an increase in rates.” Memphis Light, Gas & Water Division v. FPC, 500 F.2d at 807; see Public Systems I, 606 F.2d at 979 n. 27.
Second, the Commission found that normalization tends to improve a utility’s cash flow situation. For those expenses that are paid by the utility before they are collected in rates, normalization provides funds (deferred tax debits) in the periods when the expenses are paid, and thereby lessens the impact of the expenses on the utility’s cash flow. Subsequently, when the utility collects the expenses in rates, normalization reduces the utility’s cash flow by requiring a lower than actual tax allowance in rates. In general, normalization tends to increase a utility’s internal cash flow when the need is greatest and to decrease it when the need is less. Order No. 144, at 15, J.A. at 621.
Third, the Commission found that normalization may have a favorable impact on cost of capital. Order No. 144 at 16, J.A.' at 622. Because it has an improved cash flow, a utility’s external capital needs are less under normalization than under flow-through, and creditors may perceive the utility as a better risk. Various studies support this proposition. See Staff Study at 60 n. 1, J.A. at 139; Brief for Respondent at 25 n. 25. Because of a lack of evidence on the magnitude of the cost differential, however, the Commission found only that there is support for the position that normalization lowers the cost of capital. Order No. 144 at 16, 89-96, J.A. at 622, 695-702; Order No. 144-A at 29-30, J.A. at 881-82.
Fourth, the Commission found that the rate of return earned on common equity is the same under either flow-through or normalization. Deferred taxes do not accrue to the benefit of utility stockholders. The policy of subtracting accumulated deferred taxes from the rate base denies investors a return on deferred taxes but provides ratepayers a “return” in the form of lower rates. Also, to the extent normalization lowers cost of capital, customers are charged lower rates because of the utility’s interest savings. Order No. 144 at 86-89, J.A. at 692-9521
We find the Commission’s conclusions on the impact of adopting normalization adequately explained and reasonable.
4. Price Squeeze
“Price squeeze” results when a wholesale supplier charges higher prices to its whole*258sale customers than to its retail customers for which the wholesale customers also compete. See FPC v. Conway Corp., 426 U.S. 271, 96 S.Ct. 1999, 48 L.Ed.2d 626 (1976). This anticompetitive situation may be caused by a disparity between the normalization policies of the Commission and those of state regulatory agencies. The Commission decided to deal with price squeeze on a ease-by-case basis because no single rate-making principle necessarily creates price squeeze. Order No. 144 at 67-69, J.A. at 673-75. There are many differences between the principles the Commission utilizes in setting wholesale rates and those that the states utilize in setting retail rates. A given state may treat rate of return, cost allocation, rate base valuation, and normalization differently than the Commission. It ■is possible that the divergent policies have offsetting effects that minimize or eliminate any rate disparity otherwise created by conflicting normalization practices. Therefore, price squeeze can be shown to exist only by examining the circumstances in a particular rate proceeding. If it finds undue discrimination and anticompetitive impact, the Commission may consider the denial of normalization as one of the remedies for reducing or eliminating the discrimination. Order No. 144 at 67, J.A. at 673.
The Commission also decided that, subject to the discretionary modification by the presiding administrative law judge in a given case, a determination concerning the otherwise reasonable rate level will precede an examination of price squeeze allegations. Notice at 48, J.A. at 64. Petitioners argue that the Commission has failed to meet its responsibility to guard against anticompetitive impact because irreversible damage may occur in the period between the rate-making and the price squeeze examination. The Commission responds that petitioners are wrong to assume that all disparities between the normalization policies of the Commission and those of the states lead to price squeeze. Normalization’s possible impact on price squeeze is best examined along with all the other factors in the case and after an otherwise reasonable rate has been set. Moreover, the Commission has made it clear that presiding judges have discretion to proceed immediately with the price squeeze discovery or hearing portions of a case when they believe such action is appropriate. See Arkansas Power & Light Co., Docket No. ER79-339 (FERC Aug. 6, 1979). We find the Commission’s price squeeze policy rational and adequately explained.22
B. Other Issues Raised By Petitioners
I. Make-up Provision
The final rule requires rate applicants to make up deficiencies or eliminate excesses in their deferred tax reserves so that within a reasonable period of time to be determined on a case-by-case basis, utilities will be operating under full normalization. Order No. 144 at 3, 130-31, J.A. at 609, 736-37.23 To the extent that its past timing difference transactions were given flow-through treatment, a utility’s deferred tax account may be less than that needed to provide for future associated tax liabilities.24 The make-up provision, by spreading over time the liability for prior flow-through of tax benefits, prevents an undue hardship from falling upon a portion of future ratepayers. Because some previous *259tax benefits were given flow-through treatment, future rates must be higher than they would have been if normalization had always been used. The make-up provision attempts to spread the burden of the higher rates fairly and to shift to full normalization in an administratively efficient manner. Order No. 144 at 119-20, J.A. at 725-26.
Petitioners argue that the make-up provision is illegal retroactive ratemaking. Unlike the agency action in the cases cited by petitioners,25 however, the provision does not adjust for shortfalls in prior rates. It only adjusts future rates so that tax costs will not fall disproportionately on one ratepayer generation. Ratepayers are not charged for a greater tax allowance under the provision than they otherwise would be; they merely incur the cost over a different time period. Because the Commission deals with make-up on a case-by-case basis, we merely find, without considering any specific methodology,26 that a general make-up policy is not retroactive ratemaking and is reasonable.27
2. Loan Analogy
In its notice of rulemaking the Commission utilized a loan analogy to contrast the rate consequences of normalization and flow-through. Notice at 29 n. 40, J.A. at 45. One can view the early higher rates under normalization as a loan from the customers to the utility and the later lower rates, which result from deducting accumulated deferred taxes from the rate base and from the utility’s having to borrow less money, as interest payments. The customers’ “return” on the “deferred tax loan” would be equal to the utility’s after-tax cost of capital.28
Petitioners argue that the loan analogy reveals that the deduction of deferred taxes from the rate base does not provide a sufficient return on consumer investment. They contend that most consumers must borrow or pay off old loans more slowly in order to pay higher utility bills and that consumers must pay interest rates much higher than the utility’s cost of capital. Therefore, the return on the loan is far less than their increased costs. The impact is especially severe on lower-income groups.
In response to petitioners’ arguments, the Commission in its order held that the loan analogy was never meant to be more than illustrative and that it does not imply that ratepayers have an ownership claim to the “loaned monies.” Order No. 144 at 59-62, J.A. at 665-68. The Commission rejected the proposition that under normalization today’s customers pay tomorrow’s customers’ tax costs; each generation of customers pays its own costs. The difference in the two methods is that flow-through gives current customers benefits that belong to future periods. Normalization does not charge ratepayers for any *260costs in excess of those associated with the service they receive. For example, under normalization current rates are generally unaffected by the utility’s construction activity.29 Moreover, the Commission found the loan analogy to have little merit even for illustrative purposes because it created confusion by emphasizing the importance of the fictitious return. Id. We find the Commission’s discussion of the loan analogy rational.30
IV. Conclusion
The Commission has conducted a lengthy and thorough rulemaking. It addressed the concerns expressed by the court in Public Systems I, undertook a sophisticated study of the impact of normalization on rates, and responded in detail to the arguments raised by commenters. Finding the Commission’s positions rational, we hold that the normalization policy is the product of reasoned decisionmaking. “The choice between normalization and flow-through is for the Commission.” Public Systems I, 606 F.2d at 986 (Robb, J., dissenting). Accordingly, the Commission’s orders are
Affirmed.
. The Federal Energy Regulatory Commission was created in 1977 and succeeded to the Federal Power Commission’s ratemaking authority. 42 U.S.C. § 7172(a)(1)(B), (C) (Supp. IV 1980). In this opinion we refer to both agencies as “the Commission.”
. On March 10, 1983, the Commission voted in principle to issue a final rule that would allow utilities to file for rates that include a return on up to 50% of their investment in plants under construction. The inclusion, however, will be limited to amounts that increase rates by not more than 6% per year. Fact Sheet in RM8138.
. Similarly, ratepayers realize the tax burden of revenues when those revenues are recognized as taxable income by the utility.
. Similarly, the ratepayers who benefit from a revenue are charged with the tax burden generated by that revenue.
.- Expenses that are recognized as tax deductions later than when they are recognized in rates work towards the utility’s paying more taxes in the first year than are charged to customers. Revenues that are recognized for tax filings in periods prior to when they are recognized in rates have the same effect. Revenues that are recognized for tax filings in periods later than when they are recognized in rates work towards the utility’s paying less taxes in the first year than are charged to customers.
.A firm that declines in size will usually have a declining deferred tax account.
. In an extreme application of flow-through, the Commission even required companies using straight-line tax depreciation on their tax returns to determine their tax allowance as if they were using liberalized tax depreciation. Midwestern Gas Transmission Co., 36 F.P.C. 61 (1966), aff'd, 388 F.2d 444 (7th Cir.), cert, denied, 392 U.S. 928, 88 S.Ct. 2286, 20 L.Ed.2d 1386 (1968).
. More specifically, the Act provided that for post-1969 expansion property the utility had the right (1) to elect to change to a straight-line method of tax depreciation or (2) to elect, if permitted by the regulatory agency involved, to continue to use liberalized tax depreciation if it used a normalized method for both accounting and ratemaking purposes. To allow utilities and their customers to receive the benefits of accelerated depreciation, the Commission permitted normalization for accelerated tax depreciation as to post-1969 expansion property for both accounting and ratemaking purposes. Order No. 404, 43 F.P.C. 740 (1970); Order No. 404-A, 44 F.P.C. 16 (1970), aff’d sub nom. Memphis Light, Gas & Water Div. v. FPC, 462 F.2d 853 (D.C.Cir.), cert. denied, 409 U.S. 941, 93 S.Ct. 234, 34 L.Ed.2d 192 (1972).
.More precisely, the Supreme Court held that the Act did not limit the authority of the Commission, and on remand this court affirmed the Commission’s decision. Memphis Light, Gas & Water Div. v. FPC, 500 F.2d 798 (D.C.Cir.1974).
. The Commission listed the following timing differences as having been covered by prior orders:
(1) Differences that result from the use of accelerated depreciation and-Class Life Asset Depreciation Range (ADR) tax provisions. (Order No. 404, May 15, 1970; Opinion No. 578, June 3, 1970; and Order No. 504, February 11, 1974.)
(2) Differences that result from the use of accelerated amortization provisions on certified defense and pollution control facilities. (Order Nos. 203, May 29, 1958; 204, May 29, 1958; and 432, April 23, 1971.)
(3) Differences that arise from recognition of extraordinary property losses as a current expense for tax purposes but as a deferred and amortized expense for book purposes. (Order No. 504, February 11, 1974.)
(4) Differences that arise from recognition of research, development, and demonstration expenditures as a current expense for tax purposes but as a deferred and amortized expense for book purposes. (Order Nos. 408, August 26, 1970; 504, February 11, 1974; and 566, June 3, 1977.)
(5) Differences that result from different tax and book reporting of deferred gains or losses from disposition of utility plant. (Order No. 505, February 11, 1974.)
(6) Differences that result from the use of the Asset Guideline Class “Repair Allowance” provision of the Internal Revenue Code of 1954. (Order No. 504, February -11, 1974.)
(7) Differences that result from recognition of purchased gas costs as a current expense for tax purposes but as a deferred expense for book purposes. (Order No. 13, October 18, 1978.)
Notice of Proposed Rulemaking (Notice), Docket No. RM80-42, at 3-5 (March 31, 1980), J.A. at 19-21.
. I.R.C. § 168(e)(3) provides:
(3) Special rule for certain public utility property.—
(A) In general. — The term “recovery property” does not include public utility property (within the meaning of section 167(f)(3)(A) if the taxpayer does not use a normalization method of accounting.
(B) Use of normalization method defined. —For purposes of subparagraph (A), in order to use a normalization method of accounting with respect to any public utility property—
(i) the taxpayer must, in computing its tax-expense for purposes of establishing its cost of service for ratemaking purposes and reflecting operating results in its regulated books of account, use a method of depreciation with respect to such property that is the same as, and a depreciation period for such property that is no shorter than, the method and period used to compute its depreciation expense for such purposes; and
(ii) if the amount allowable as a deduction under this section with respect to such property differs from the amount that would be allowable as a deduction under section 167 (determined without regard to section 167(1) using the method (including the period, first and last year convention, and salvage value) used to compute regulated tax expense under subparagraph (B)(i), the taxpayer must make adjustments to a reserve to reflect the deferral of taxes resulting from such difference.
(C)Public utility property which is not recovery property. — In the case of public utility property which, by reason of this paragraph, is not treated as recovery property, the allowance for depreciation under section 167(a) shall be an amount computed using the method and period referred to in subparagraph (B)(i).
. This court in Public Systems I interpreted the Natural Gas Act and Federal Power Act as establishing a substantial evidence test for review of Commission orders, including informal rulemaking orders. 606 F.2d at 979-80; see American Paper Inst., Inc. v. American Electric Power Serv. Corp., - U.S. -, - n. 7, 103 S.Ct. 1921, 1927 n. 7, 76 L.Ed.2d 22 (1983). Several courts, however, have suggested that the arbitrary and capricious test and the substantial evidence test may not, in practical effect, impose different standards of review for informal rulemaking. ECEE, Inc. v. FERC, 611 F.2d 554, 565 n. 22 (5th Cir.1980); American Public Gas Ass’n v. FPC, 567 F.2d 1016, 1028-29 (D.C.Cir.1977), cert. denied, 435 U.S. 907, 98 S.Ct. 1456, 55 L.Ed.2d 499 (1978); Associated Indus, of N. Y. State, Inc. v. United States Department of Labor, 487 F.2d 342, 349-50 (2d Cir.1973).
.Timing differences can arise from either expense or revenue transactions that are prepaid or deferred in rates. Since expense transactions that create tax deferrals predominate, however, we will refer primarily to those transactions in our discussion.
. Recognizing important national interests, the Commission has made exceptions for pollution control and fuel conversion facilities. Notice at 28 n. 39, J.A. at 44.
. But see note 2.
.The staff’s quantitative evaluation covered the most important timing differences, which are
1. Differences that result from recognition of the following items as current expenses for tax purposes and as capitalized and amortized expenses for ratemaking purposes:
(a) the interest component of the Allowance for Funds Used During Construction (AFUDC);
(b) construction-related pensions; and
(c) construction-related taxes.
2. Differences that result from the use of different depreciation lives for taxes and ratemaking.
3. Differences that result from recognition of regulatory commission cost as a current expense for tax purposes and as an expense amortized over a number of years for rate-making.
Notice at 35, J.A. at 51. Also, the study discussed, but did not attempt to quantify, the impact of normalizing plant dismantling costs.
Notice, Appendix A, Staff Study on Comprehensive Tax Normalization (Staff Study) at 113-15, J.A. at 192-94.
. Intervenor East Tennessee attacks the methodology used by the Commission in the impact study on several grounds. We extend substantial deference to the Commission in this highly technical area and find none of East Tennessee’s arguments convincing.
. For electric utilities, the analysis indicates that in 1980 the required revenues were $149 to $183 million higher under normalization than under flow-through and will be $58 to $148 million higher under normalization than under flow-through in the year 2000. Staff Study at 101, J.A. at 180.
. The final rule and order adopted and incorporated by reference statements made in the notice. Order No. 144 at 7-8, J.A. at 613-14.
. Removal cost is an expense that is recovered in rates prior to its availability as a tax deduction to the utility. With normalization of the tax effects of the expense, customers realize the tax benefit before the utility recognizes the tax deduction.
. The Commission also found that normalization will not promote inefficient plant construction, Order No. 144 at 103, J.A. at 709, and, at worst, will not affect rate stability. Order No. 144-A at 31, J.A. at 883.
.The court’s finding in Public Systems I that the expenses that would be subject to normalization were inadequately specified and discussed has not been raised in the present case. The Commission carefully defined “timing difference transaction,” listed the transactions that were within the scope of the rule, and undertook an impact study of the most important transactions subject to the rule. The Commission thereby preempted any argument that the scope of the rule is uncertain or inadequately discussed.
. In the proposed rule the Commission had indicated its intent to continue flow-through treatment of timing differences that had received such treatment unless it established a method for adjusting the deferred tax account. Notice at 50-51, J.A. at 66-67.
. On the other hand, a utility may have over-collected amounts in its deferred tax accounts by charging on the basis that it would have to pay future income taxes at a 48% rate rather than a 46% rate. See Order No. 144 at 118, J.A. at 724.
. See Montana-Dakota Utilities Co. v. Northwestern Public Serv. Co., 341 U.S. 246, 71 S.Ct. 692, 95 L.Ed. 912 (1951); FPC v. Hope Natural Gas Co., 320 U.S. 591, 64 S.Ct. 281, 88 L.Ed. 333 (1944); United States v. Corrick, 298 U.S. 435, 56 S.Ct. 829, 80 L.Ed. 1263 (1936); Arizona Grovery Co. v. Atchison, T. & S.F. Ry. Co., 284 U.S. 370, 52 S.Ct. 183, 76 L.Ed. 348 (1932).
. See, e.g., Natural Gas Pipeline Co. of America, Opinion No. 108, 22 Fed. Power Serv. (MB) 5-172 (Dec. 23, 1980), where the Commission concluded that in the circumstances of that case the deferred tax cost arising from prior flow-through should be amortized in equal installments over about ten years.
. Petitioners also contend that the Commission erred in failing to prepare an impact analysis on the effect of the make-up provision. The Commission’s answer, which we find reasonable, is that the impact is too uncertain to quantify because the provision will be implemented on a case-by-case basis, but the Commission believes the provision will equitably spread over time the liability for prior flow-through of tax benefits.
.The staff study noted several problems with the analogy. First, the “loan” is involuntary. Second, the “return” realized by customers is tax free, thereby increasing the real rate of return of those customers who are taxpayers. Third, those who provide the loan may not regain the principal because some customers will change power sources or change their level of consumption. Staff Study at 38-41, J.A. at 117-20.
. But see note 2.
. Some petitioners argue that it was an abuse of discretion for the Commission to hold that deferred taxes are not customer-contributed capital in which ratepayers have an equitable interest. The Commission’s answer, which we find reasonable, is that customers who pay normalized taxes are simply paying the full tax cost of the service they receive, whereas customers who pay flow-through taxes are receiving tax benefits that belong to other time periods. Order No. 144 at 59-62, J.A. at 665-68. Cf. Memphis Light, Gas and Water Div. v. FERC, 707 F.2d 565, 571-72 (D.C.Cir.1983). While the Commission agrees with these petitioners that deferred taxes should be deducted from the rate base, it declines to go further and hold that each customer has an interest in the funds he has “contributed.”