Opinion for the Court filed by Circuit Judge WILKEY.
WILKEY, Circuit Judge:This action arises on petitions to review orders of the Interstate Commerce Commission (ICC or Commission) served 25 October 1978 and 1 June 1979, which prescribed maximum reasonable rates for unit-train shipments of coal from Cordero Junction, Wyoming to Elmendorf, Texas, a suburb of San Antonio, Texas.1 This case marks the second time that a court has been called upon to review a final order of the ICC setting a maximum reasonable rate for the above described traffic. In a decision served 14 October 1976 (San Antonio I)2 the Commission set a rate of $10.93 per ton, and this decision and order was affirmed by the United States Court of Appeals for the Eighth Circuit.3
At the request of the railroads, the Commission reopened the San Antonio proceeding with a view to modifying this rate. On 25 October 1978 the Commission issued a decision (San Antonio II) which prescribed *252a higher rate of $16.12 per ton.4 San Antonio and the carriers petitioned the Commission for administrative review of its decision, and at the same time San Antonio, the State of Texas, the United States Department of Energy, and the railroads all petitioned this court as well for review of the Commission’s decision.
In response to the petitions for reconsideration, the Commission on 1 June 1979 issued a further decision (San Antonio III) which established a rate of $17.23 per ton.5 Subsequently, this court granted the parties’ motions to amend their petitions for review of the San Antonio II decision to include the Commission’s decision in San Antonio III.6 For reasons that appear below, we vacate the Commission’s orders and remand for further proceedings.
In the interest of convenience and clarity, we include an outline of the opinion to follow:
I.FACTUAL AND PROCEDURAL BACKGROUND
II.ANALYSIS
A. Comparable Rates
B. Cost Determinations
1. Locomotive Costs — Capital Costs
2. Locomotive Costs — Fuel and Maintenance Expenses
3. Allocation of Fixed Costs
4. Fixed Plant Investment Additive
5. Roadway Maintenance Expenses
6. Federal Taxes
C. Revenue Adequacy — Mandate of the 4-R Act
1. Reliance on Cost of Capital Figure
2. Seven Percent Additive Above Fully Allocated Costs
D. Environmental Implications
III.CONCLUSION
I. FACTUAL AND PROCEDURAL BACKGROUND
This matter initially was brought before the Commission by complaint of San Antonio, requesting the Commission to prescribe a maximum just and reasonable rate for the movement of coal from Cordero Junction, Wyoming to San Antonio, Texas. The factual background is set forth in detail in the opinion affirming the Commission’s order in the original proceeding,7 and we may briefly recall those facts here.
San Antonio, acting through its City Public Service Board, owns and operates an electric utility, which before 1972 used natural gas as a primary fuel. In that year, with the sudden, dramatic increase in the cost of natural gas, San Antonio began exploring the possibility of using coal to generate electricity. On 22 May 1974 San Antonio entered into long-term contracts to purchase coal from two suppliers in Campbell County, Wyoming, and at about the same time began construction of two coal-fired generating units. Previously, on 11 March 1974 Burlington Northern, Inc. (BN or Burlington Northern) and Southern Pacific Transportation Company (SP or Southern Pacific) had provided San Antonio with a copy of a proposed tariff setting forth a rate of $7.90 per ton for the transportation of coal by unit-train from the mines in Campbell County, Wyoming to San Antonio, but on 2 May 1974 the carriers revised their quotation to $11.90 per ton as a result of unprecedented cost increases. Approximately one year later, San Antonio, being dissatisfied with the rate proposed by the railroads, filed a complaint with the Commission seeking prescription of a just and reasonable rate.
The Commission, in a decision served 14 October 1976 (San Antonio I),8 established a rate of $10.93 per ton for the San Antonio movement, basing its determination primar*253ily on evidence of rate levels for unit-train movements similar to the San Antonio service and on estimates of the fully allocated costs of the service.9 The Commission stated that its decision was a temporary response to permit the coal movement to begin, and it invited petitions for modification of the rate when actual experience in operating the trains was gained. On petition to review the Commission’s decision and order prescribing the San Antonio rate, the United States Court of Appeals for the Eighth Circuit, affirmed.10
After six months of operation, the railroads petitioned the Commission in June 1977 for a modification of the rate, and in October 1977 the ICC reopened the San Antonio proceeding.11 One year later, on 25 October 1978 the Commission issued a decision (San Antonio II) that prescribed a rate of $16.12 per ton for the San Antonio movement.12 The rate was based on a computation of the full costs of the San Antonio service calculated at $12.65 per ton, and adjusted to include a 10.2% rate of return on the carriers’ investment as a “revenue need” factor.13 Consideration of the latter criterion was intended to reflect adherence to the Railroad Revitalization and Regulatory Reform Act of 1976 (4-R Act), which directs the Commission to aid the railroads in obtaining revenue levels adequate to cover total operating expenses plus a fair return on capital for the purpose of restoring the financial stability of the railway system.14
Both San Antonio and the carriers petitioned the Commission for reconsideration of its decision in San Antonio II, and on 1 June 1979 the Commission issued a further decision (San Antonio III) which established a rate of $17.23 per ton for the San Antonio movement.15 The increased rate reflected certain modifications in the Commission’s calculation of the costs of the service as well as the use of a 10.6% rate of return on capital instead of the 10.2% figure employed in San Antonio II.16 In addition, the Commission determined that for the railroads to achieve adequate revenue levels in accord with the 4-R Act, the carriers must be permitted to set rates exceeding full costs on some services to compensate for the fact that competition forces the carriers to set rates on other services below full costs. The Commission stated that no “simple formula” existed for ascertaining “the extent to which some shippers should *254subsidize others in the interest of producing a financially viable rail system."17 Nevertheless that agency concluded that a rate set at 7% above full costs for the San Antonio movement (including the cost' of capital figure) was reasonable as an interim solution, pending the development of more specific guidelines in another proceeding involving Western coal rates.18
The several parties to this case raise two main challenges to the Commission’s orders in San Antonio II and III: The carriers and San Antonio dispute several of the Commission’s determinations concerning the costs of providing service for the San Antonio movement. All of the parties attack, on various grounds, the Commission’s imposition in San Antonio III of a 7% additive above fully allocated costs.19 We find that certain of the Commission’s cost calculations have no basis in the record and that the selection of the figure seven as a percentage additive above full costs is wholly without supporting rationale or justification. We accordingly set the Commission’s orders aside and remand on these and other issues to be discussed.
II. ANALYSIS
Courts traditionally have applied a deferential standard in reviewing rate determinations by an agency. As the Supreme Court has observed:
The process of rate making is essentially empiric. The stuff of the process is fluid and changing — the resultant of factors that must be valued as well as weighed. Congress has therefore delegated the enforcement of transportation policy to a permanent expert body and has charged it with the duty of being responsive to the dynamic character of transportation problems.20
Bearing in mind that the determination of the reasonableness of rates primarily is the province of the agency charged with that duty, we nonetheless have an obligation to ascertain whether the Commission’s decision comports with applicable law, whether there is evidence in the record to support the Commission’s findings, and whether the agency’s rationale is both discernible and defensible.21
Ordinarily one of the measures that the ICC uses in prescribing rates is evidence of rates from other, comparable movements, and the Commission accordingly considered this factor in its decision in San Antonio I. In establishing the maximum reasonable rate in San Antonio II and III, however, the Commission relied instead on its computations of the carriers’ cost of service and the perceived revenue needs of the railroads. Before turning to a review of the Commission’s cost evidence and revenue need factors, then, we first must determine whether the Commission abused its discretion in dispensing with the comparable rate standard.
A. Comparable Rates
When the Commission reopened the San Antonio ratemaking proceeding in October 1977, it invited the submission of evidence on “any existing rate comparisons which properly reflect similar conditions and transportation characteristics of the movements in question.”22 Both the railroads *255and San Antonio submitted conflicting evidence as to the rates, measured in mills per ton/mile, prevailing on comparable traffic.23 In its decision in San Antonio II, however, the Commission stated that in light of the pendency of a proceeding to establish rate-making guidelines applicable to Western coal rates (Ex Parte No. 347, Western Coal Investigation—Guidelines for Railroad Rate Structure j,24 it would not assign “great weight to rate comparisons which represent only a small portion of a large, developing rate structure. Unless an established rate structure exists, rate comparison is of limited value.”25 In San Antonio III the Commission again rejected evidence of allegedly comparable rates, advancing the same justification.26
San Antonio claims that the Commission’s reasons for failing to resolve the conflicting evidence submitted by the parties and for thus abandoning the comparable rate standard are arbitrary and capricious. We disagree.
Courts properly are reluctant to interfere with an agency’s choice of methodology in ratemaking matters.27 Our review is directed to assuring only that the standards applied are reasonable. In this instance, for reasons sufficiently articulated in both San Antonio II and III, the ICC elected to rely on the costs of the service rather than evidence of rates on other movements in calculating the rate for the San Antonio traffic. The Commission’s approach, provided the underlying cost calculations are substantiated in the record, is a reasonable means of arriving at a determination of the maximum reasonable rate to be prescribed. We thus find no abuse of discretion.
B. Cost Determinations
We next review the Commission’s findings concerning the carriers’ costs of serving the San Antonio traffic. Petitioners allege error with respect to the following cost items: locomotive costs, including capital costs and fuel and maintenance expenses; allocation of fixed costs; fixed plant investment additive; roadway maintenance expenses; and federal taxes.28 We *256are mindful of the fact that the evaluation of evidence of a carrier’s costs of providing a particular service is a function peculiarly adapted to the expertise of an agency;29 nevertheless, we are bound to assess whether the Commission’s cost findings are supported both by sufficient evidence in the record and adequate reasoning.
1. Locomotive Costs — Capital Costs
Power for the San Antonio movement is supplied by the railroads from a pool of locomotives, which are stationed at Alliance, Nebraska and which are assigned to various other unit-train movements as circumstances require.30 To accommodate the San Antonio traffic, the railroads were required to purchase approximately fifty new diesel units. These locomotives were added to the Alliance pool and will be used by the railroads to handle other coal movements in addition to servicing San Antonio. One of the issues in this case arises from the fact that the Commission developed the capital costs of the locomotives to be taxed to the San Antonio movement by using the average cost of the pool locomotives actually utilized in the San Antonio service. The railroads insist that the Commission instead should have used the full cost of the new power units in making this calculation.31
The carriers do not claim here that the newly acquired locomotives will be used exclusively in the San Antonio service; nor do they claim that only new locomotives will be used for that traffic. Rather, their position rests on the fact that, even though the entire pool of coal-train locomotives is not used to power the San Antonio service, the addition of that traffic necessitated the acquisition of the new locomotives. The railroads argue that San Antonio therefore should bear the entire cost of the new equipment irrespective of the fact that such equipment is placed in a pool and used interchangeably with other traffic.
We find that the Commission’s approach in developing the capital costs of the San Antonio service was reasonable. The Commission had before it evidence furnished by the railroads of the average costs of the actual locomotives used to move the coal to San Antonio as well as evidence of the purchase price of the new locomotive.32 Although the new power units were purchased to service San Antonio, clearly other coal traffic handled by Burlington Northern and Southern Pacific also reaps the benefit from this expansion of equipment. The Commission’s election to compute capital costs based on the average costs of the actual locomotives used, so as not to charge San Antonio with the entire expense of the new power equipment, thus is entirely logical.33
*2572. Locomotive Costs — Fuel and Maintenance Expenses
The railroads next argue that the ICC erred in rejecting their proposed locomotive fuel and maintenance cost adjustments. To support the proposal the railroads did not submit data on amounts actually expended; rather, they adjusted systemwide average costs upward to the extent that the weight of the coal locomotives exceeded the systemwide average weight.34 The Commission ruled the railroads’ evidence insufficient, finding that “weight is not the only factor to be considered in adjusting the fuel and maintenance expenses.”35 The calculations were found inadequate because they did not account for other “factors such as horsepower, age of locomotives and type of locomotives, which would also influence the locomotive unit mile costs.”36 In the absence of actual cost data, the Commission used the systemwide average for fuel and maintenance in calculating variable costs.
The railroads contend that the ICC’s determination is arbitrary because it is contrary to the Commission’s disposition of the issue in Incentive Rate on Coal — Hayden, Colorado to Kings Mill, Texas (Kings Mill)37 and Annual Volume Rates on Coal— Wyoming to Flint Creek, Arkansas (Flint Creek),38 both decided after San Antonio II but before San Antonio III. In Kings Mill and Flint Creek the ICC held that the utilities had not demonstrated that the railroads’ adjustment based on the locomotives’ heavier weight was invalid, and therefore the Commission accepted the railroads’ figures. The railroads argue that substantially the same evidence was presented in San Antonio, Kings Mill, and Flint Creek, that the same locomotive pool was involved in both San Antonio and Flint Creek, and that therefore the Commission should have accepted the railroads’ figures in San Antonio or explained the difference in treatment.
We affirm the Commission’s decision on this point. We do not believe that the Commission's decisions in Kings Mill and Flint Creek obligated it in San Antonio III to alter the disposition of the fuel and maintenance issue previously made in San Antonio II. That the Commission’s decisions in Kings Mill and Flint Creek differed from its earlier disposition in San Antonio II indicates, if anything, that the Commission does not have a firm policy with respect to such locomotive costs,39 but treats the question on a case-by-case basis. We do not have before us the records in Kings Mill and Flint Creek and do not intimate a view as to the correctness of the ICC’s decisions there.40 But, in the present case, we find *258that the Commission’s treatment of fuel and maintenance costs is logical and that substantial evidence in the record supports it.
San Antonio’s cost expert testified that fuel, constituting thirty percent of locomotive expenses, is consumed more efficiently in unit-coal train service because power is assigned on the basis of known trailing gross tonnage. He demonstrated that the annual mileage of locomotives used in the San Antonio service was double that of the system average, so that expenses that are a function of time (e. g., inspections) would be lower. The witness also pointed out flaws in a study that the railroads had submitted in Incentive Rate on Coal— Cordero, Wyoming to Smithers Lake, Texas 41 to justify an adjustment of costs based on weight.42 Thus, there is adequate evidence in the record to support the conclusion that fuel and maintenance costs for the unit-coal train locomotives here are not necessarily greater in proportion to their weight.
San Antonio II was initiated, inter alia, to allow the parties to request rate modifications supported by “new cost evidence based on the actual operating experience” of this traffic.43 The railroads, as the proponents of the upward adjustment of fuel and maintenance costs, had the burden to produce actual operating evidence that these costs exceeded systemwide averages.44 Because no such evidence was presented, and in view of record testimony that the proposed adjustment was incorrect, we find no error in the Commission’s resolution of this issue against the railroads.
3. Allocation of Fixed Costs
The railroads also challenge the Commission’s decision to use the ratio method rather than the ton/ton-mile method of allocating constant costs. In response to the carriers’ objection that the ratio method produced a lower allocation of constant costs than the ton/ton-mile method, the Commission in San Antonio II explained its treatment of fixed costs as follows:
The mere fact that one method produces higher or lower costs than the other does not support the use of either basis. Rather, our purpose is to determine the costs which are properly attributable to the traffic at issue.
In docket No. 34013, Rules to Govern [the] Assembling & Presenting [of] Cost Evidence, 337 I.C.C. 298, 323 [& n.10] (1970), the Commission stated:
[T]he determination that one method is preferable to another method of allocating expenses in a particular situation, is a ratemaking function * * * [a logical and reasonable allocation of expenses in one ratemaking situation might not be so in another * * *.] For example, the allocation of constant costs on a ton-mile basis, might (for ratemaking purposes), in certain circumstances, result in an inordinate portion of such costs being borne by extremely heavy-loading traffic, as compared to relatively light-loading traffic.
Consequently, we do not suggest that any single method is always proper. Since coal is characterized by heavy loadings, however, the dollar or ratio basis was, *259and is, a reasonable method of apportionment of constant costs to the affected traffic.45
The railroads contend that this allocation decision is based on either no finding or on contradictory reasoning. They argue that if the selection of the ratio method is based on the method's lower apportionment of costs to heavy-loading traffic, that finding contradicts the Commission’s statement that “[t]he mere fact” that a method produces higher or lower costs “does not support the use of either basis.” If the determination is not made on that basis, then it is said to have no basis at all.46
We disagree. The Commission’s rationale is not that the ton-mile method allocates to heavy loading coal higher costs, but that it allocates an “inordinate portion” of fixed costs. The reasoning is supported by the Rules to Govern Assembling and Presenting Cost Evidence, quoted by the Commission in its decision 47 Further, use of the ratio rather than the ton-mile method appears to be the general course followed by the Commission in Western coal cases.48 We conclude that the Commission’s method of allocation is reasonable and adequately explained in its decision.49
4. Fixed Plant Investment Additive
Normally, fixed plant investment costs apportioned to new rail traffic are calculated using Rail Form A.50 Rail Form A, a formula published by the Commission’s Bureau of Accounts to determine rail freight service costs, divides among all rail traffic, inter alia, the railroad’s fixed plant investment costs. The fixed plant investments entered in Rail Form A include such items as track structure, land value, shops, and engine houses.51 Ordinarily, the amounts invested to handle the San Antonio traffic would be included with other investment in Rail Form A, and San Antonio, as any other customer, would bear its share of the total investment costs.
Fifty percent of fixed plant investment costs are assigned as variable costs of service; the other fifty percent are attributed to constant costs.52 The amount assigned as variable costs to particular traffic may be said to be the Rail Form A estimate or “prediction” of the amount of fixed plant investment caused by that traffic.
In this case, the ICC believed that actual investment costs caused by the San Antonio movement would exceed the amount predicted by Rail Form A.53 Believing it more fair to allocate these coal-related invest*260ment costs directly and exclusively to coal shippers (rather than requiring all traffic to share them),54 the ICC developed a fixed plant investment additive. The additive was computed by tabulating the amounts spent by the railroads in 1976 and 1977 for capital improvements to handle new coal traffic.55 Included in the additive were investment costs for items such as rails, ties, rail anchors, and a new diesel shop56 — precisely the kind of investment costs that Rail Form A allocates.57 These investment costs were divided exclusively among the coal shippers58 using the improvements59 and not entered in Rail Form A.60 The additive, so computed, was added to the ICC’s variable cost-of-service calculation, which already reflected the Rail Form A estimate of San Antonio’s share of fixed plant investment costs.61
San Antonio argues that the use of both the Rail Form A prediction and the additive results in a double count of investment costs.62 We agree that the ICC’s methodology results in something approaching a double count. Before explaining why we reach this conclusion, we pause to express our understanding of the proper relationship of the additive and Rail Form A.
As the Commission noted, Rail Form A is flexible enough to allow substitution of actual costs for predicted costs.63 To substitute costs, however, one must remove as well as replace. For example, in allocating directly coal locomotive capital costs to coal shippers, the ICC first removed from Rail Form A all locomotive capital costs.64 Logically, the ICC should have treated coal-related fixed plant investment costs the same way. In calculating San Antonio’s rates, the ICC should have cleanly removed from Rail Form A all costs of the nature apportioned by the additive before including the additive. For instance, since the additive apportions costs for rail, ties, and rail anchors, all investments in rail, ties, and rail anchors should have been removed from Rail Form A in calculating San Antonio’s share of Rail Form A investment costs. Labelling Rail Form A costs so adjusted “Rail Form A ( — ),” this process of substitution may be expressed mathematically as:
*261Costs = [Rail Form A ( — )] + [fixed plant investment additive]
In this fashion, the coal shipper pays the actual cost of coal-related fixed plant investment instead of (and not in addition to) the Rail Form A amount reflecting the same type investment. Having elucidated a logical manner of using the additive, we now examine the Commission’s disposition of the matter.
The Commission’s reasoning and methodology respecting the additive are murky at best. It appears the Commission discerned that summing the unadjusted Rail Form A costs with the additive would double count investment costs, and therefore the Commission purported to exclude from Rail Form A65 the specific investment amounts divided among the coal customers. While this exclusion is helpful,66 it does not go far enough.67 The Commission should have excluded not only the specific coal-related investments allocated by the additive, but should have entirely removed from Rail Form A all other investment amounts of the same character.68 By way of illustration, normally Rail Form A would include investment in ordinary rail as well as heavier coal-related rail, this total amount to be shared by all traffic. Because here the additive directly allocates the coal-related rail costs to San Antonio, in calculating its rates an adjustment must be made in Rail Form A. The ICC excluded from Rail Form A only those coal-related rail costs apportioned by the additive; it should have removed all rail investment costs to compute San Antonio’s rate. As a result of this failure, both Rail Form A and the additive include amounts invested in rail. The additive includes costs for heavier coal-related *262rail; Rail Form A includes costs for all other rail in the system. The costs for other rail are not actual costs brought about by coal shipper use of such rail; they are predicted system average costs. The coal shipper, therefore, pays twice for the rail it uses: it pays the actual costs of heavier rail through the additive, and the system average predicted cost of rail via Rail Form A.
The Commission apparently believed this result justified because coal shippers to some extent use ordinary rail (and other fixed plant) whose costs are not included in the additive. This may be true. But non-coal traffic may use the heavier rail (and other coal-related fixed plant) whose costs are no longer included in Rail Form A. This means that, pursuant to the ICC’s methodology, coal shippers pay all costs for the coal-related fixed plant and share in other fixed plant costs as if they were non-coal shippers,69 while noncoal shippers pay nothing for their use of coal-related fixed plant and have their costs for fixed plant partially defrayed by coal shippers.
This amounts to cross-subsidization of noncoal shippers by coal shippers.70 While cross-subsidization is legitimate when differential rates are set, it is impermissible to calculate variable costs of service so that one customer subsidizes another. Variable costs by definition are only the costs caused by the relevant service and should not include costs caused by other services. We recognize that costing is not a particularly exact science, but by occasioning cross-subsidization in variable cost calculation, the Commission’s decision is not merely imprecise, but arbitrary and unreasonable71
If on remand the Commission again elects to use Rail Form A and the additive in apportioning fixed plant investment costs, it must adjust Rail Form A costs as we have indicated to avoid the possibility of a double count of investment costs and cross-subsidization by San Antonio of other shippers’ variable costs.
5. Roadway Maintenance Expenses
Another of the cost items contested in this proceeding is the amount necessary to *263compensate the carriers for maintaining the rail lines traversed by the San Antonio coal trains. In San Antonio II and III, the Commission allowed an expense based on the system average maintenance costs drawn from Rail Form A estimates. San Antonio argues that the ICC’s reasons for rejecting its evidence — which allegedly established that the roadway maintenance expenses for the San Antonio traffic are lower on a per ton basis than the system average expenses — are arbitrary and capricious. We decline to upset the Commission’s disposition of this cost item.
After the Commission reopened the San Antonio proceeding in October 1977, San Antonio argued for a downward adjustment to system average costs for actual maintenance expenses incurred on the San Antonio movement.72 San Antonio contended that while the absolute costs of maintaining the rail lines travelled by the San Antonio trains were indeed high, the actual per ton costs were lower than the system average costs because of the enormous volume of coal handled. In support of its position, San Antonio submitted actual maintenance expenses, furnished by the railroads, for the year 1977. Petitioner San Antonio had attempted to compel the carriers to produce maintenance expenses, by line segment, for a five-year period, but the Commission found this request too burdensome given the time constraints of the proceeding and ordered the railroads to produce only the expenses for the year 1977.73
In its decision in San Antonio II the ICC rejected San Antonio’s proposed adjustment, stating that one year’s data was inadequate to give an accurate indication of normalized expenditures on the San Antonio traffic.74 In San Antonio III, the Commission again refused San Antonio’s proposed adjustment to the roadway maintenance costs for the same reasons that it had rejected this same adjustment in its prior decision.75 This time, however, the ICC, acknowledging the magnitude of this particular cost item,76 ordered the carriers to file status reports with the Commission divulging the actual maintenance costs incurred for each of the line segments for the years 1973 through 1978.77
San Antonio’s quarrel with the Commission in this case is not with the Commission’s finding per se that one year’s data was not a reliable indicator of normalized maintenance expenditures — San Antonio’s own cost expert noted that figures based on one year’s expenditures did not constitute a representative sample and that a five-year data base would be more accurate 78 Rather, San Antonio asserts that the Commission’s rejection of its “computation as based on only one year’s data, when that is all it was able to obtain from the railroads de*264spite diligent and repeated efforts, borders upon the perverse.” 79 Underlying San Antonio’s complaint, then, is the Commission’s denial of San Antonio’s request to compel the railroads to produce data for each line segment for the years 1973 through 1978.80
There is no question that the maintenance cost item comprises a substantial portion of the entire costs of the service. San Antonio therefore should be permitted to obtain accurate information, if available, on this item. The Commission, however, has taken reasonable measures to provide San Antonio with the information it seeks: in San Antonio III the Commission ordered the carriers to file status reports revealing maintenance data for each line segment for the years 1973 through 1978.81 Thus it would appear pointless for this court to take any further step at this time. Once the carriers have produced the requested information, San Antonio presumably will be free to appear before the Commission to seek an adjustment if the figures so indicate.82
6. Federal Taxes
The final cost determination disputed by San Antonio is the Commission’s treatment of federal taxes. In San Antonio III the Commission applied the statutory tax rate to adjust after-tax returns to the pretax level based on its conclusion that not only must the investment tax credit and actual taxes paid be reflected in the rate charged, but that deferred taxes should be accounted for as well. Thus the Commission computed the cost of service as if the carriers were using a straight-line method of depreciation even though the carriers actually are using an accelerated depreciation schedule. Under this method (normalization) the railroads receive the benefit of the tax deferral and the advantage of charging rates as if the higher tax actually had been paid.
San Antonio first claims that the Commission’s election to normalize taxes for ratemaking purposes is unsupported by reasoned analysis.83 We disagree.
In San Antonio III the Commission succinctly stated its reasons for electing to normalize taxes as follows:
Our prescribed accounting system considers deferred taxes as part of the tax expense considered in the computation of net income. [See 49 C.F.R. § 1204, 1-12 (1978).] Deferred taxes, to a large extent, must eventually be paid and cannot be ignored as an element of the carriers’ liability. In addition, allowing the carriers to retain the tax benefit of accelerated depreciation (the primary cause of deferred taxes) provides an incentive for investment.84
The Commission’s approach also is in accord with its policy announced in Ex Parte No. 338, Standards and Procedures for the Establishment of Adequate Railroad Reve*265nue Levels,85 In that proceeding, the Commission noted its intent to adopt the statutory tax rate in railroad ratemaking proceedings, stating that “[w]hile there are some advantages of using the actual tax rate, use of the statutory tax rate will assure that accelerated depreciation and the investment tax credit remain an incentive to capital spending.”86 We therefore find the Commission’s election to normalize taxes supported both by adequate reasoning and established policy.
San Antonio next argues that even if normalization is proper, the Commission neglected to deduct the deferred tax account from the net investment base prior to calculating the rate to be assessed, contrary to the policy expressed in Ex Parte No. 338. The Commission responds by asserting only that its decision to normalize taxes for rate-making purposes was proper. While we believe as we have indicated that it is permissible for the Commission to apply the statutory tax rate in its ratemaking proceeding, the Commission’s response does not address the distinct issue of an adjustment to the net investment base for deferred taxes once normalization is elected.
As this court has recognized on more than one occasion, the principle of excluding a deferred tax reserve from the rate base, as such reserve comes into existence, is an essential component of an agency’s election to normalize taxes for ratemaking purposes.87 Otherwise the rate payer who has paid higher rates reflecting normalization accounting would be paying the carriers for earnings on the tax differential even though it was the rate payer who contributed the differential in the first place.
The Commission’s own procedures set forth in Ex Parte No. 338 also underscore the necessity of subtracting the deferred tax reserve from the net investment base:
We are cognizant of the fact that this treatment [normalization] confers a benefit on the carriers, in that they are receiving and retaining revenue which is not accounted as income. As indicated by some of the parties, the capital funds arising from deferred taxes have been contributed by the ratepayers rather than by investors in the company. Thus, it is appropriate to deduct the deferred tax account from the net investment rate base prior to any calculation of rate or return.88
There is no indication in the record that the Commission adjusted the investment base to reflect the deferred tax reserve prior to calculating the rate for the San Antonio traffic. As the Commission totally ignored this point in its brief, we may only assume that the Commission also neglected to make this deduction. On remand therefore the Commission should make the required adjustment if in fact it has not already been made.89
C. Revenue Adequacy — Mandate of the 4-R Act
Section 205 of the Railroad Revitalization and Regulatory Reform Act of 1976 (4-R *266Act) instructs the Commission to make a “continuing effort” to “assist [rail] carriers in attaining revenue levels” that are “adequate, under honest, economical, and efficient management, to cover total operating expenses . . . plus a reasonable and economic profit or return (or both) on capital employed in the business.”90 Such revenue levels should “provide a flow of net income plus depreciation adequate to support prudent capital outlays, assure the repayment of a reasonable level of debt, permit the raising of needed equity capital, cover the effects of inflation,” and should ensure retention and attraction of capital “in amounts adequate to provide a sound transportation system in the United States.” 91
To effectuate this statutory mandate, the Commission in San Antonio III included in its rate prescription a 10.6% return on investment equal to the carriers’ cost of capital plus a 7% increment above full costs (including the cost of capital figure). Having concluded our review of the Commission’s cost findings, we now must assess the Commission’s reasons for selecting first the cost of capital figure and then the 7% additive as revenue need factors.
1. Reliance on Cost of Capital Figure
In prescribing the maximum reasonable rate in San Antonio II the ICC first allowed a 10.2% rate of return on investment equal to the carriers’ overall cost of capital.92 In San Antonio III, the Commission permitted a higher rate of return on investment to coincide with its findings in Ex Parte No. 353, Adequacy of Railroad Revenue,93 that the weighted cost of capital for the railroads currently equalled 10.6%.94 San Antonio objects to the Commission’s decision to rely solely on the rate of return standard, arguing that the Commission’s own regulations and decisions obligate it to consider other financial criteria, in particular a funds flow analysis, in arriving at a determination of revenue adequacy. The Commission concedes that other financial indicia “may be appropriate” to determine revenue adequacy in the future but contends that its approach in San Antonio II and III is a reasonable, interim one until more precise standards can be developed.
A brief review of the Commission’s regulations and revenue adequacy proceedings demonstrates that the Commission has eschewed sole reliance on a cost of capital figure for assessing the revenue needs of the carriers. While it is not our intent or purpose to require the Commission to choose one methodology over another, we believe that further consideration of this issue by the Commission is appropriate.
On 31 January 1978 the Commission promulgated Ex Parte No. 338, Standards and Procedures for the Establishment of Adequate Railroad Revenue Levels.95 The notice of proposed rulemaking suggested that a determination of the adequacy of railroad revenues would be derived from a consideration of the following: (1) whether such revenues provide a return on net investment equal to the carriers’ cost of capital, (2) whether such revenues enable the carriers to maintain financial ratios indicative of a sound financial condition, and (3) whether such revenues make possible a sufficient flow of funds to allow the railroads’ capital *267spending requirements to be met.96 In its decision in Ex Parte No. 338, however, the Commission stated that because a funds flow approach was still in a developmental stage,97 it was unable to determine at that time how these various financial indicia would coalesce to indicate the revenue needs of the carriers and which, if any, measure would predominate. It did assert that a return on net investment equal to the carriers’ cost of capital was one criterion to be considered in ascertaining revenue adequacy but emphasized that it would not place “sole reliance on this measure.”98
Shortly after the issuance of Ex Parte No. 338, the Commission adopted regulations pertaining to the procedures and standards to be followed in establishing adequate revenue levels for the railroads.99 Section 1109.25 of the Commission’s rules provides that the ICC shall make a yearly determination of revenue adequacy for Class I railroads on a national, district, and individual basis. That section specifically directs that such determinations shall be conducted upon consideration of “all pertinent financial indicators, including a rate of return on net investment equal to the cost of capital, other financial ratios, and the flow of funds ...[;] and other evidence as to ability to make needed investment.” 100
The first proceeding to be conducted under the revenue adequacy regulations was Ex Parte No. 353, Adequacy of Railroad Revenue, issued on 5 December 1978.101 The Commission’s findings in that proceeding were limited to settling upon 10.6% as the overall cost of capital for use as a reference in judging returns on net investment. The ICC repeated, however, that “a rate of return on net investment equal to the cost of capital is only one of several factors to be considered in determinations of overall revenue adequacy” and concluded that “[t]he amount of weight accorded to this factor, and conclusions derived from consideration of other factors, are matters to be discussed in (a] subsequent decision.” 102
Thus far the Commission has issued no subsequent decision to deal with these issues. In the absence of other developed financial criteria, the Commission in San Antonio III included within the prescribed rate a 10.6% rate of return on capital to meet the carriers’ revenue requirements.103
We appreciate the Commission’s reasons for relying on the cost of capital criterion. Nevertheless, considerable time has elapsed since the ICC last indicated that it would undertake an analysis of other criteria to be applied in determining the overall revenue needs of the railroads, and presumably the Commission has progressed somewhat to that end. Given the mandate of Ex Parte No. 353, Ex Parte No. 338, and the rules promulgated thereunder,104 we think that *268the Commission on remand should reassess the revenue requirements of the carriers using other appropriate financial measures, to the extent that these have been refined sufficiently, in addition to the cost of capital indicium.105 We do not mean to suggest a view as to the exact means that the Commission shall use in ascertaining the revenue needs of the railroads — that is not our function. Instead, our concern here is limited to ensuring that the Commission follows its own published standards and regulations.106
2. Seven Percent Additive Above Fully Allocated Costs
To assist the carriers in attaining adequate revenue levels in accord with the mandate of the 4-R Act, the Commission in San Antonio III also determined that a 7% increment above fully allocated costs (including the 10.6% rate of return on capital) was necessary. This conclusion was based on the theory of differential pricing, which postulates that rates on some services must be set at levels higher than full costs to compensate for those services on which the railroads are required to price below full costs to remain competitive.
The Commission cautioned that in the application of this theory the railroads should not be permitted to recover their entire shortfall by extracting monopoly profits from captive shippers. Stating that no “simple formula” existed for judging the .extent to which some shippers should subsidize others in the interest of producing a financially viable rail system,”107 the Commission asserted that before it would impose a “substantial burden on some shippers,” the railroads would be required to submit data on the following:
(1) specific identification of the traffic that must be subsidized by other traffic and the reason why rates cannot be increased on that traffic; (2) the extent to which the railroads provide service on unprofitable branch lines and the reason^) why such service cannot be made profitable or abandoned; (3) identification of commodities other than coal which could also make substantial contributions to the railroads’ system revenue needs; *269and (4) identification and quantification of excess capacity on a carrier’s system.108
Until the railroads could demonstrate that they were unable to increase revenues on competitive traffic, the Commission further asserted, it would “be reluctant to approve or prescribe rates which are significantly higher than fully allocated cost[s].”109
Apparently concluding that the contemplated additive did not fall within the latter category, the Commission, without benefit of the data identified above, set the San Antonio rate at seven percent above fully allocated costs. Admitting that this choice depended on no more than a “policy judgment,” the Commission nonetheless asserted that its approach in the San Antonio proceeding was a reasonable measure pending the conclusion of Ex Parte No. 347, Western Coal Investigation — Guidelines for Railroad Rate Structure, which presumably would lead to the adoption of more precise guidelines.110 Two Commissioners dissented to what one termed the adoption of a “seven percent solution” as plainly lacking in reasoned analysis. All of the parties to this case oppose the Commission’s so-called seven percent solution for various reasons. Briefly, San Antonio and the State of Texas attack the Commission’s whole theory of differential pricing, contending that the ICC may not require a particular movement to subsidize other traffic.111 Expanding on this notion, Texas also asserts that the Commission’s determination to require the San Antonio service to subsidize other traffic violates section 10741 of the Interstate Commerce Act, which proscribes unreasonable discrimination against particular regions or types of traffic.112 For their part the carriers complain that the prescribed rate113 is lower than rates approved by the Commission on comparable services prior to the passage of the 4-R Act.114 With the enactment of that Act, the carriers assert, Congress intended the railroads to have greater, rather than less, flexibility in setting rates to permit a revitalization of the railway system, and the imposition of a seven percent “ceiling” above full costs thereby frustrates congressional intent.115 The Department of Justice complains that the Commission’s decision fails to indicate how the adoption of a seven percent additive comports with both the national policy favoring conversion to coal as a fuel source and with the Commission’s own reservation concerning the extent to which captive shippers should subsidize other traffic.
Uniting these various arguments is the common objection that the Commission’s decision is arbitrary and capricious in that it provides no defensible rationale for the inclusion of the seven percent increment. We base our conclusion to set aside the Commission’s action on precisely this ground.
*270As this court acknowledged in Houston Lighting & Power Co. v. United States, differential pricing is “pertinent to the objective [set forth in the 4-R Act] of providing an adequate overall level of earnings.” 116 Even though differential pricing thus may be a legitimate criterion for the ICC to consider in the San Antonio rate-making proceeding, the Commission still must provide adequate justification for its choice of a particular increment above fully allocated costs. In San Antonio III, however, the ICC did no more than make the general assertion that it could not find that the railroads had achieved revenue adequacy. There is nothing in the record in the way of findings, evidence, or rationale to support the seven percent solution or any percentage solution. The Commission’s general allusion to the need to consider the revenue requirements of the carriers and the economics of differential pricing is so broad as to be meaningless as a standard— this rationale could be put forth just as readily in an attempt to justify a 1%, 21%, 45%, or even a 99% additive. The Commission here defends its action on the ground that adoption of the appropriate additive involves a policy judgment that is not susceptible to precise quantification. Concededly the problem is a difficult one, but that does not excuse the Commission from articulating “fully and carefully the methods by which, and the purposes for which, it has chosen to act.”117
In San Antonio III, the Commission listed four criteria that it stated would be used to assess the extent to which particular movements should be required to subsidize other services,118 but the ICC then completely ignored this standard in its indiscriminate selection of the seven percent additive. Apparently the Commission assumed that because its solution assertedly was a temporary one, it was relieved from its responsibility to follow any standard whatsoever.119 That notion provides little comfort to those affected by the ratemaking proceeding and is irreconcilable with established principles of reasoned decisionmaking.
Finally, the 4-R Act evinces a congressional policy to “balance the needs of carriers, shippers, and the public.”120 “[N]either the interest of electric consumers in lower rates nor that of the public in coal conversion dictates that the railroads must receive compensation for their services inadequate to maintain financial soundness”;121 still, the Commission’s decision must reflect adequate consideration of all the pertinent interests. As should be obvious from our discussion, the Commission has not met this obligation. From the carriers’ standpoint, the ICC made no effort to determine *271whether the revenues that would be produced by a seven percent additive would be adequate. From the shipper’s and ultimately the consumer’s standpoint, the Commission professed concern that the railroads should not “make up their entire shortfall by extracting monopoly profits from captive shippers,” 122 but then failed to explain how its choice of a seven percent additive was consistent with that concern. It is undisputed that San Antonio falls into the category of a “captive shipper”: it is locked into long-term contracts for the shipment of coal with no alternative means at present of transporting this coal. And with respect to the public interest in coal conversion, although the Commission in San Antonio II discussed the effect on national energy policy of including a return on equity capital as a revenue need factor, in San Antonio III it neglected to discuss what effect the imposition of an additional seven percent subsidy would have. On remand, then, the Commission is directed to explain more fully its reasons for selecting a particular increment, if any, above fully allocated costs, making sure that its decision is based on a thorough analysis of the relevant interests.
D. Environmental Implications
Lastly, the State of Texas argues that the Commission was required by the National Environmental Policy Act (NEPA) to prepare a detailed environmental impact statement (EIS) before approving the rate increases on the San Antonio traffic.123 Texas’s major concern in this respect is that the Commission’s decisions in San Antonio II and III will not only result in higher electricity rates for San Antonio area customers, but that it will also set the pattern for future rate increases which may inhibit subsequent conversion to coal by the Texas utility companies.
Although in San Antonio II the Commission first characterized the effect predicted by the State of Texas as an economic impact not within the scope of NEPA, the Commission nevertheless assured petitioner that the environmental consequences of its ratemaking policy would be considered in its proceeding in Ex Parte No. 347, Western Coal Investigation — Guidelines for Railroad Rate Structure.124 On appeal Texas claims that the challenged rate should be set aside until an environmental impact statement is completed.
As a product of its investigation of Western coal rates in Ex Parte No. 347, the Commission already has issued a draft environmental impact statement to which both the State of Texas and the carriers have filed extensive critical comments.125 In light of this fact, no purpose would be served by ordering the Commission to duplicate its efforts by exploring these questions again in the context of the San Antonio ratemaking proceeding.126 Instead it suffices that the Commission on remand integrate the tentative conclusions derived from the comments filed to the draft EIS into its consideration of the maximum reasonable rate to be established for the San Antonio coal traffic.
III. CONCLUSION
With the foregoing guides to the Commission, we vacate the orders before us and remand for proceedings not inconsistent with this opinion.
So ordered.
. The coal is hauled to San Antonio by joint routing over four railroads: Burlington Northern, Inc. (BN) and its subsidiaries, the Colorado and Southern Railway Company (C&S), Fort Worth and Denver Railway Company (FW&D); and the Southern Pacific Transportation Company (SP).
. San Antonio, Texas v. Burlington Northern, Inc., 355 I.C.C. 405 (1976) (San Antonio I) aff’d, 555 F.2d 637 (8th Cir. 1977).
. Burlington Northern, Inc. v. United States, 555 F.2d 637 (8th Cir. 1977).
.San Antonio, Texas v. Burlington Northern, Inc., 359 I.C.C. 1 (1978) (San Antonio II).
. San Antonio, Texas v. Burlington Northern, Inc., 361 I.C.C. 482 (1979) (San Antonio III).
. The petition of the Department of Energy was withdrawn upon notice to the court that the Department’s interests would be represented by the United States Department of Justice.
. See Burlington Northern, Inc. v. United States, 555 F.2d 637, 639 (8th Cir. 1977).
. San Antonio I, 355 I.C.C. 405 (1976), affd, 555 F.2d 637 (8th Cir. 1977).
. The term “fully allocated costs” is defined as “that level of expenses which represents the sum of the ‘variable costs’ plus an appropriate allocation of fixed expenses, i. e., an allocation which assigns an adequate portion of fixed expenses to the movement of traffic on which the rate in issue applies.” Rules to Govern the Assembling and Presenting of Cost Evidence, 337 I.C.C. 298, 308 (1970).
. Burlington Northern, Inc. v. United States, 555 F.2d 637 (8th Cir. 1977).
. The Commission ruled that it would consider the following matters in the reopened proceeding: (1) errors in the Commission’s costing methodology in its prior decision, (2) new cost evidence based on actual operating experience, (3) evidence concerning the question of what constitutes a fair and reasonable profit or return, and (4) any appropriate rate comparisons reflecting changed circumstances in similar Western coal movements. See San Antonio, Texas v. Burlington Northern, Inc., I.C.C. Docket No. 36180, at 2 (27 Oct. 1977), reprinted in Joint Appendix (J.A.) at 36, 36a.
. San Antonio II, 359 I.C.C. 1 (1978).
. In response to the Commission’s order reopening the proceeding, see note 11 supra, both the carriers and San Antonio submitted evidence of rates on allegedly comparable shipments, but the Commission elected not to attach “great weight” to these comparisons on the ground that coal rates in the West were in flux, see San Antonio II, 359 I.C.C. at 7; notes 22-27 infra and accompanying text.
. Pub.L.No. 94-210, § 205, 90 Stat. 39, 41, as amended by Pub.L.No. 95-473, 92 Stat. 1373 (1978) (codified at 49 U.S.C.A. § 10704(a)(2) (West Supp.1978)). The San Antonio proceeding was initiated prior to the passage of the 4-R Act: thus the Eighth Circuit excused the Commission’s failure to consider the impact of that statute in the San Antonio I decision.
. San Antonio III, 361 I.C.C. 482 (1979).
. The Commission’s computation of the fully allocated costs of the San Antonio service, adjusted both to reflect a 10.6% rate of return on capital and to reflect the 1978 general rate increase for all traffic, equalled $16.10 per ton.
. San Antonio III, 361 I.C.C. at 495-96.
. Id at 496 (Ex Parte No. 347, Western Coal Investigation — Guidelines for Railroad Rate Structure (instituted 17 May 1978».
. The carriers, San Antonio, the State of Texas, and the United States Department of Energy all petitioned for judicial review of the Commission’s decision in San Antonio II, and, following the Commission’s decision in San Antonio III, these parties amended their petitions to include that decision. The United States Department of Energy, however, notified this court that its interests would be represented by the Department of Justice.
. Atchison, Topeka & Santa Fe Ry. v. Wichita Bd. of Trade, 412 U.S. 800, 806, 93 S.Ct. 2367, 2374, 37 L.Ed.2d 350 (1973) (quoting Board of Trade v. United States, 314 U.S. 534, 546, 62 S.Ct. 366, 372, 86 L.Ed. 432 (1942)).
. See, e. g., National Ass’n of Recycling Indus., Inc. v. ICC, et al., 627 F.2d 1328, at 1334 (D.C.Cir.1980) (NARIII).
. San Antonio, Texas v. Burlington Northern, Inc., I.C.C. Docket No. 36180, at 2 (27 Oct. *2551977), reprinted in J.A. at 36, 36a. In San Antonio I the Commission identified rates of two movements, measured in mills per ton/mile, that the Commission considered most comparable to the San Antonio movement. These rates coincided to movements of coal by unit train from the same origin point as the San Antonio traffic—Belle Ayr, Wyoming—to the destinations of Pueblo, Colorado; and Amarillo, Texas. See San Antonio I, 355 I.C.C. 405, 414-15 (1976), aff'd, 555 F.2d 637 (8th Cir. 1977). In reviewing the reasonableness of the prescribed rate, the Eighth Circuit concluded that the rate comparisons were the “strongest evidence in [the] case to support the Commission’s conclusion that the prescribed rate is the maximum reasonable rate.” Burlington Northern, Inc. v. United States, 555 F.2d 637, 641 (8th Cir. 1977).
. The carriers offered comparisons of rates from six other movements, see San Antonio II, 359 I.C.C. at 37, App. B; San Antonio presented evidence of comparisons from eleven other movements, see id. at 38-39, App. C.
. Ex Parte No. 347 was instituted on 17 May 1978. See id. at 7.
. Id. (footnote omitted). The Commission’s only other reference to the evidence submitted by the parties was the assertion that the evidence adduced suggested that the present rate was below a maximum reasonable rate. See id. At the time of the ICC’s decision, the prescribed rate for the San Antonio traffic, expressed in mills per ton/mile, was 7.25 mills. The comparisons submitted by San Antonio reflected average earnings of 7.66 mills, see id. at 38-39, App. C; those submitted by the carriers reflected average earnings of 11.33 mills, see id. at 37, App. B.
. See San Antonio III, 361 I.C.C. at 497. The rate prescribed by the Commission in San Antonio III is equivalent to 10.4 mills per ton/mile.
. See, e. g., FPC v. Hope Natural Gas Co., 320 U.S. 591, 602, 64 S.Ct. 281, 287, 88 L.Ed. 333 (1944); Houston Lighting & Power Co. v. United States, 606 F.2d 1131, 1148 (D.C.Cir.1979), cert. denied, 444 U.S. 1073, 100 S.Ct. 1019, 62 L.Ed.2d 755 (1980).
. The Department of Justice expresses no view on any particular cost calculation. The State of Texas agrees with San Antonio in general that the Commission’s cost computations overstate the actual costs of the San Antonio unit-train movement, but takes no specific position on any cost item.
. See, e. g., Houston Lighting & Power Co. v. United States, 606 F.2d 1131, 1146 (D.C.Cir. 1979), cert. denied 444 U.S. 1073, 100 S.Ct. 1019, 62 L.Ed.2d 755 (1980).
. The coal cars furnished by San Antonio consist of six sets of 100 cars each, weighing approximately 10,000 tons when loaded and requiring heavy duty, diesel locomotives.
. In San Antonio II, the railroads submitted evidence on the capital costs of the locomotives, based on the purchase price of 50 new locomotives acquired in 1976 — a cost of $490,-000 per unit. See Verified Statement of Marion L. Hall 28-32 & Exhibit MLH-5 (8 Dec. 1977), reprinted in J.A. at 383, 410-14, 458-59. In addition the carriers submitted calculations on this issue based on the average costs of the locomotives actually used on the San Antonio service. This evidence was derived from a three-month study conducted by the railroads, which indicated that the actual locomotives used for the San Antonio traffic were built between 1969 and 1976 with an average cost of $398,906 per unit. The Commission adopted the $398,906 figure as a base for developing the capital costs of the service, although it did not question the carriers’ need to acquire additional locomotives to serve San Antonio. See, e. g., San Antonio ¡I, 359 I.C.C. at 33-34, App. A.
. See note 31 supra and accompanying text.
. The railroads also argue that the Commission’s averaging method of computing capital costs leads to the impractical result of necessitating an increase in the rates to all other coal shippers whenever new equipment is purchased to service an additional unit-train movement. The Commission responds that the cost of new equipment may also be recouped through general revenue increases and increases in average Rail Form A costs to the extent that systemwide averages are augmented. It is our duty to review, not the wisdom, but the reasonableness of the Commission’s action. *257We cannot say that the Commission’s position is unreasonable. Nor would it be unreasonable to make individual rate adjustments requiring the cost of each service ultimately to be calculated on the basis of equipment actually employed.
. Thus the maintenance and fuel costs were adjusted upward 71% for the C&S, 72% for the FW&D, 24% for the BN, and 15% for the SP. Verified Statement of Marion L. Hall, Exhibit MLH-4 at 7, 14, 20, 26 (8 Dec. 1977), reprinted in J.A. at 383, 430, 437, 443, 449.
. San Antonio II, 359 I.C.C. at 33, App. A.
. Id.
. 359 I.C.C, 749, 808 App. B (1979).
. I.C.C. Docket Nos. 36970, 36980, App. C, at 18-19 (21 May 1979).
. Even if Kings Mill and Flint Creek represent some sort of fuel and maintenance cost policy, we do not think the Commission erred by not explaining the difference in approach here. The ICC could not have made such an explanation in San Antonio II because Kings Mill and Flint Creek had not been decided yet. The issue was not raised in the papers submitted prior to San Antonio III, and we cannot fault the Commission for not dealing with the question then. We also note that in San Antonio, Texas v. Burlington Northern, Inc., 362 I.C.C. 161 (1979) (San Antonio IV), not reviewed here, the Commission did list reasons for the differing results in the three cases. Id. at 163-64.
.The railroads contend that “the same expert witness testified for the complainants, offering virtually the same evidence in the same language” in San Antonio, Flint Creek, and Kings Mill, Reply Brief of Railroad Respondents-Intervenors at 18 (emphasis in original), and copies of the testimony in each case are appended to their reply brief. Our perusal of the expert’s testimony indicates that different evidence was presented in the three proceedings. Absent in *258Kings Mill and Flint Creek was his testimony quantifying the increased utilization of the pertinent unit-coal locomotives; absent in Kings Mill was his testimony regarding increased fuel efficiency of locomotives in unit-coal trains. Whether the absence of this evidence justified the decisions reached in Kings Mill and Flint Creek is an issue not before us. But the inclusion of this testimony along with other evidence presented in this record is sufficient to uphold the Commission’s conclusion here.
.I.C.C. Docket No. 36608 (30 Nov. 1977), aff'd sub nom. Houston Lighting & Power Co. v. United States, 606 F.2d 1131 (D.C.Cir.1979), cert. denied, 444 U.S. 1073, 100 S.Ct. 1019, 62 L.Ed.2d 755 (1980).
. Verified Statement of Leroy E. Peabody, Jr. (13 Feb. 1978), reprinted in J.A. at 947, 969-73.
. San Antonio, Texas v. Burlington Northern, Inc., I.C.C. Docket No. 36180, slip op. at 2 (27 Oct. 1977), reprinted in J.A. at 36, 36a.
. See, e. g., 5 U.S.C. § 556(d) (1976); San Antonio III, 361 I.C.C. at 483-84; San Antonio II, 359 I.C.C. at 4, 7-8.
. San Antonio II, 359 I.C.C. at 9 (emphasis added by Commission).
. Brief of Railroad Petitioners-Intervenors at 41.
. San Antonio II, 359 I.C.C. at 9 (quoting 337 I.C.C. 298, 323 & n.10 (1970)).
. E. g., Bituminous Coal, Hiawatha, Utah to Moapa, Nevada, I.C.C. Docket No. 36938 (31 July 1979); Unit Train Rates on Coal — Burlington Northern, Inc., I&S Docket No. 9199, I.C.C. Docket Nos. 36944, 37021, 37029, App. A at 122 (13 July 1979); Arkansas Power & Light Co., I.C.C. Docket No. 36719, slip. op. at 12-13 (14 June 1979); Annual Volume Rates on Coal — Wyoming to Flint Creek, Arkansas, I.C.C. Docket No. 36970, 36980, App. C at 44 (21 May 1979).
. The railroads also argue that use of inconsistent methods of cost allocation could result in a revenue shortfall. This might well be the case if the Commission always selected the apportionment method that resulted in lower cost allocation. No evidence is presented that the Commission has engaged in such a practice.
Allocation of fixed costs is a ratemaking function particularly within the Commission’s expertise, and we are reluctant to interfere with its exercise of discretion in this area. In the absence of a showing that the Commission has embarked on a policy sure to prevent recovery of fixed costs, we decline to disturb its determination here.
. ICC Bureau of Accounts, Rail Form A (Statement IFI-73) Formula for Use in Determining Rail Freight Service Costs (1973).
. ICC Bureau of Accounts, Explanation of Rail Cost Finding Procedures and Principles Relating to the Use of Costs (Statement No. 7-63) 110-11 (1963).
. ICC Bureau of Accounts, Rail Carload Cost Scales 1976 (Statement No. ICI-76) 156, 184.
. San Antonio III, 361 I.C.C. at 486; San Antonio II, 359 I.C.C. at 11.
. San Antonio III, 361 I.C.C. at 489.
. See Verified Statement of Marion L. Hall 32-34, Exhibit MLH-7 (8 Dec. 1977), reprinted in J.A. at 383, 414-16, 461-62.
. See id. at 33, reprinted in J.A. at 415; Verified Statement of Jerry R. Masters 2-11 (5 Dec. 1977), reprinted in J.A. at 325, 326-35.
. See note 51 supra.
. We note that as part of the status reports ordered in San Antonio III the ICC requested “[t]he basis for and the percentages used to allocate fixed plant investment costs among coal and noncoal shippers.” 361 I.C.C. at 503, App. B. This implies that the Commission contemplates that noncoal shippers may eventually pick up part of the cost of new investment through their own additive. At least in this case, however, the record indicates the fixed plant investment costs are shared only by coal unit trains. Verified Statement of Marion L. Hall 33 (8 Dec. 1977), reprinted in J.A. at 415; Verified Statement of Jerry R. Masters 9-10 (5 Dec. 1977), reprinted in J.A. at 333-34.
. Verified Statement of Marion L. Hall 33, Exhibit MLH-7 (8 Dec. 1977), reprinted in J.A. at 383, 415, 461-62. The ICC’s determination of San Antonio’s allocable portion of the costs is not challenged in this petition.
. See San Antonio III, 361 I.C.C. at 487; San Antonio II, 359 I.C.C. at 11. Because Rail Form A calculations were based on 1976 data, only coal-related investments for 1976 were excluded from input data. See Verified Statement of Marion L. Hall 21-22 (8 Dec. 1977), reprinted in J.A. at 383, 403-04.
. San Antonio III, 361 I.C.C. at 492.
. San Antonio concedes that there is “nothing economically unsound with the direct ascertainment of the new plant expense assignable to San Antonio’s coal traffic for service cost purposes,” but contends that it is error to include both the Rail Form A predictions and the fixed cost investment additive in the same calculation. That is, “you can use the Rail Form A method, or the direct method, or you could possibly make both computations and divide by 2, but it is improper to include two (2) estimates of the same expense in a single computation.” Brief for San Antonio, Texas at 28 (emphasis in original).
. San Antonio II, 359 I.C.C. at 11.
. See id. at 21, 33-34, App. A.
. Our review of the record discloses that the ICC figures may not reflect an exclusion from Rail Form A averages of the incremental fixed plant investment at issue. Although the railroads did make such an exclusion in their rate submission, San Antonio did not. This is because San Antonio believed it appropriate to apportion the investment costs by using Rail Form A figures without including an additive. See Bureau of Accounts Section of Cost and Valuation, Analysis of Cost Data in Docket No. 36180, San Antonio, Texas v. Burlington Northern, Inc., at 20 (1 May 1978), reprinted in J.A. at 2202. In calculating variable costs, the Commission used San Antonio’s figures, see San Antonio III, 361 I.C.C. at 499, from which no exclusion had been made for the incremental investment. See Railroads’ Petition for Administrative Review at 11 note (14 Nov. 1978), reprinted in J.A. at 2010 note. While this point is not present on appeal, the ICC’s apparent use of figures not reflecting an exclusion of investment costs in conjunction with its claim that the exclusion has been made casts doubt on the validity of its conclusion that no double count has occurred.
. Had the expenses been entered in Rail Form A, then the railroads would recover costs for the identical items twice — first, through the additive, and second through the Rail Form A averages shared by all traffic.
. The Commission reasoned in its decision in San Antonio II and III that the existence in Rail Form A of investment costs unrelated to San Antonio traffic would result in a slight, though unquantifiable overstatement of investment costs. See San Antonio III, 361 I.C.C. at 487; San Antonio II, 359 I.C.C. at 11-12. In view of the overstatement problem, and its conclusion that the railroads had not shown that the entire amount of the additive was attributable to the San Antonio traffic, the ICC in San Antonio II reduced the additive by 30%. See San Antonio II, 359 I.C.C. at 12. The Commission reversed its position in San Antonio III, however finding that the amount of the additive was adequately supported and that the overstatement was not significant enough to warrant a cutback in the increment. See San Antonio III, 361 I.C.C. at 487, 489.
We are puzzled by the I.C.C.’s focus on inclusion of “unrelated” costs in Rail Form A. As we understand it, Rail Form A almost always includes investment unrelated to the particular movements whose costs are being computed. It is because of the difficulty of assigning particular investment costs to particular movements that the shared average cost concept of Rail Form A is used. If inclusion of “unrelated” investments in Rail Form A here results in an “overstatement,” then that is inherent in Rail Form A and is neither exacerbated nor ameliorated by use of an additive. No one challenges use of the shared cost methodology of Rail Form A. Challenged is the use of Rail Form A and the additive to compute cost for the same categories of investments.
.We leave it to the ICC on remand (if it again chooses to use both Rail Form A and the additive to recover fixed plant investment costs) to define the categories of investment included in the additive and determine which investment costs presently entered in Rail Form A fall within these categories.
. We understand the ICC’s policy to be that once investment costs are apportioned by an additive, they may never be entered in Rail Form A. This means that the Commission excludes such coal-related expenses from Rail Form A when rates are calculated for noncoal as well as coal traffic. Thus, coal shippers are charged the same Rail Form A costs as noncoal shippers, in addition to the fixed plant investment additive.
. Even if no noncoal shippers used the coal-related investment San Antonio’s costs would still be significantly overstated. This is so for the following reasons: Fixed plant investment costs are apportioned by Rail Form A principally as gross ton mile (GTM) costs. San Antonio is charged for each GTM by Rail Form A the same amount charged a noncoal shipper. That is, San Antonio pays for each GTM as if all its fixed plant were ordinary and noncoal-related. Through the additive, San Antonio is charged for each GTM as if some of its investment is coal-related. Using both methods together compels San Antonio to pay twice for the same coal-related investment — it pays once as if the investment is coal-related, and once as if non-coal-related. For example, for each stretch of heavy coal-related track, San Antonio pays twice, once as if noncoal-related, and once the actual (coal-related) cost.
.The Commission also maintains that the additive represents only the difference between actual and Rail Form A predicted costs, so that summing the additive and Rail Form A costs yields the actual investment costs, not a double count. We disagree. The additive was not computed by subtracting Rail Form A predicted costs from actual costs; rather, San Antonio’s entire portion of coal-related investment costs was added to that predicted by Rail Form A.
Nor can the sum of the additive and Rail Form A costs even be said to approximate actual costs. As we have noted, the sum of both significantly overstates investment costs — whether that overstatement be labelled “cross subsidization” or a “double count.”
If Rail Form A is adjusted as we suggest, however, then the sum of Rail Form A ( — ) and the additive yield a close approximation of actual costs. This is because the cost categories in Rail Form A ( — ) and the additive by definition and construction are mutually exclusive. Since the categories do not overlap, no longer will the shipper make contributions for the same cost category twice. In a sense, the additive is the “difference” between approximate actual costs and Rail Form A ( — ) costs.
. The carriers claimed that the maintenance costs attributable to the San Antonio coal traffic were higher than the system average because of the greater physical stress associated with the heavy coal cars. In San Antonio II and III the Commission rejected the carriers’ proposed adjustment as unsupported by any evidence of actual operating experience. The railroads do not challenge the Commission’s findings here.
. See San Antonio, Texas v. Burlington Northern, Inc., I.C.C. Docket No. 36180, at 3 (3 May 1978), reprinted in J.A. at 1761, 1763.
. See San Antonio II, 359 I.C.C. at 34, App. A. In addition, the Commission noted that the figures submitted by San Antonio represented only an estimate; San Antonio computed its per ton figure, intended to reflect actual maintenance expenses for the year 1977, by apportioning the 1977 expense figure on the basis of estimates of 1978 tonnages rather than the actual tonnages for the year 1977. See id. San Antonio defended its methodology claiming that it was unable to obtain the 1977 tonnage data from the carriers.
. See San Antonio III, 361 I.C.C. at 489-90.
. San Antonio estimates that the maintenance expense, as amplified by the 7% additive, represents nearly $4.00 per ton of the entire rate. See Brief for Petitioner San Antonio, Texas at 34 n. 32.
. See San Antonio III, 361 I.C.C. at 489-90 & 502-03, App. B.
. See San Antonio’s Petition for Reconsideration at 5-6 (16 May 1978), reprinted in J.A. at 1785, 1790-91.
. Brief for Petitioner San Antonio, Texas at 34.
. See San Antonio, Texas v. Burlington Northern, Inc., I.C.C. Docket No. 36180, at 3 (3 May 1978), reprinted in J.A. at 1761, 1763.
. See San Antonio III, 361 I.C.C. at 502-03, App. B.
. We note that in a further decision, not reviewed here, the Commission again stated that more refined data based on actual operating experience was desired and again ordered the carriers to file the status reports described in San Antonio III. See San Antonio IV, 362 I.C.C. 161, 164, 168 (1979).
. San Antonio does not dispute the Commission’s election to account for actual taxes paid and the investment tax credit in the prescribed rate; its complaint lies only with the Commission’s treatment of deferred taxes.
.See San Antonio III, 361 I.C.C. at 493. In San Antonio II the ICC indicated in a footnote that it was using an effective tax rate for the purpose of the ratemaking proceeding, see San Antonio II, 359 I.C.C. at 17 n. 32, but the cost computations appended to that decision corresponded to an approximation of the statutory tax rate. In its petition for administrative review of San Antonio II, San Antonio complained of this discrepancy, see San Antonio’s Petition for Administrative Review at 3-5 (8 Dec. 1978), reprinted in J.A. at 2131, 2134-36, and in San Antonio III the Commission made clear its intent to normalize taxes.
. 358 I.C.C. 844, aff'd as modified, 359 I.C.C. 270 (1978).
. Id. at 893-94.
. E. g., Public Sys. v. FERC, 606 F.2d 973, 975-76 (D.C.Cir.1979); Memphis Light, Gas & Water Div. v. FPC, 500 F.2d 798, 800 (D.C.Cir. 1974). See Warren, Tax Accounting in Regulated Industries: Limitations on Rate Base Exclusions, 31 Rutgers L.Rev. 187, 189-94 (1978).
. Ex Parte No. 338, Standards and Procedures for the Establishment of Adequate Railroad Revenue Levels, 358 I.C.C. 844, 894, aff'd as modified, 359 I.C.C. 270 (1978).
. The carriers argue that because San Antonio’s own witness calculated the cost of service by using the railroads’ figure for the investment base, without making any adjustment for deferred taxes, San Antonio now should be precluded from seeking a modification in the investment base. We disagree. That portion of the witness’s statement which discusses the Commission’s treatment of federal taxes was devoted to arguing that the ICC should have used the effective rather than the statutory tax rate on the equity portion of capital. See Verified Statement of Leroy E. Peabody, Jr. (13 Feb. 1978), reprinted in J.A. at 947, 956-62. Because the calculations of the witness reflected the effective tax rate, see id. at 1058, he had no need to make any adjustment for deferred taxes.
. Pub.L. No. 94-210, § 205, 90 Stat. 39, 41, as amended by Pub.L. No. 95-473, 92 Stat. 1373 (1978) (codified at 49 U.S.C.A. § 10704(a)(2) (West Supp.1978)). Section 205 originally was added as § 15a(4) of the Interstate Commerce Act, 49 U.S.C. § 15a(4) (1976), and is now revised and codified at 49 U.S.C.A. § 10704(a)(2) (West Supp.1978).
. 49 U.S.C.A. § 10704(a)(2)(A), (B) (West Supp.1978).
. San Antonio II, 359 I.C.C. at 17. This figure reflected an embedded debt cost of 8.1%, an equity cost of 12%, with a capital structure of 35% debt and 65% equity.
.(5 Dec. 1978), reprinted in J.A., Tab 5.
. San Antonio III, 361 I.C.C. at 491. This figure is based on a 13% cost of equity capital, a 7.0% cost of embedded debt, with a capital structure of 40% debt and 60% equity.
. 358 I.C.C. 844, aff'd as modified, 359 I.C.C. 270 (1978).
. See id. at 845, 856.
. A funds flow model is intended to project the needed capital outlays and other fund requirements of the railroads, to determine the amount of funds available to them from operations and capital sources, and to predict the amount by which the available funds will fall short of the projected requirements. The dollar amount of this shortfall is added to the projected revenue level to find the amount of revenue needed, and this figure is then translated into a percentage return on net investment. See id. at 859.
. Id. at 858.
. See 43 Fed.Reg. 5836 (1978) (codified in 49 C.F.R. § 1109.25).
. 49 C.F.R. § 1109.25(a)(1), (2) (1979). The regulation directs the Commission, inter alia, to publish on or before 30 June of each year a “funds flow projection for comment by participants.” Id. § 1109.25(b)(3).
. See J.A., Tab 5.
. Id. at 40.
. The Commission determined that the carriers’ rates of return were low. See San Antonio III, 361 I.C.C. at 496; San Antonio II, 359 I.C.C. at 14. Indeed the railroads submitted evidence that their rates of return on net investment hovered around 2%. See Verified Statement of Bruce G. McPhee (6 Dec. 1977), reprinted in J.A. at 279.
. San Antonio also claims that the statutory language and the relevant portions of the legislative history of the 4-R Act indicate that other financial techniques, in particular a funds flow *268analysis, must be used to determine whether railroad revenues are adequate. While section 205 of the 4-R Act directs that revenue levels must “provide a flow of net income” “adequate to support prudent capital outlays, assure the repayment of a reasonable level of debt, [and] permit the raising of needed equity capital . . . 49 U.S.C.A. § 10704(a)(2) (West Supp.1978), that language specifies not at all the exact methodology that the Commission shall use in accomplishing this mandate. Likewise, although the relevant portion of the Senate Report cited by San Antonio does indicate that the Commission should not focus solely on a rate of return analysis but instead should adopt a prospective view of carrier revenue needs, it does not prescribe specific standards by which to determine the adequacy of revenue levels. Instead, the report states only that “the Commission will be expected to utilize the most modern financial analysis techniques available.” S.Rep. No. 499, 94th Cong., 1st Sess. 52 (1975), U.S.Code Cong. & Admin.News 1976, pp. 14, 66.
.We note that San Antonio presented its own projections in San Antonio II of the funds necessary to meet the railroads’ revenue requirements, see, e. g., Verified Statement of George H. Borts (13 Feb. 1978), reprinted in J.A. at 1177, 1191-98, and in neither San Antonio II nor San Antonio III did the Commission even comment on these figures. We think that on remand the Commission at least should consider San Antonio’s projections.
. The railroads assert that San Antonio’s actual grievance is that the Commission has not yet completed a funds flow analysis, which it indicated in Ex Parte No. 353 that it would undertake. The carriers claim that because San Antonio never petitioned to review the Commission’s decision rendered in Ex Parte No. 353, it is now precluded from challenging the ICC’s failure to use a funds flow analysis in the San Antonio ratemaking. The short answer to the carriers’ objection is that San Antonio had no need to challenge the Commission’s decision in Ex Parte No. 353 concerning the standards to be used in judging revenue adequacy. The Commission specifically stated in that proceeding that the cost of capital criterion was only one of several factors to be considered in determining overall revenue adequacy. See, e. g., note 102 supra and accompanying text. In the San Antonio ratemaking proceeding, the Commission disregarded this position as well as its decision in Ex Parte No. 338 and the regulations promulgated thereunder.
. San Antonio III, 361 I.C.C. at 495-96.
. Id. at 496.
. Id.
. See id.
. This claim is based on their theory that differential pricing may only be applied to the allocation of fixed “common” costs — a contention that is completely wrongheaded. By definition, fixed costs are not associated with any particular traffic. See American Commercial Lines v. L&N R.R., 392 U.S. 571, 586 nn. 14 & 16, 88 S.Ct. 2105, 2113 nn. 14 & 16, 20 L.Ed.2d 1289 (1968). San Antonio also claims in the alternative that before the ICC may impose a subsidization burden it is required to make a specific finding concerning the level of revenues necessary to achieve “revenue adequacy.” See note 122 infra and accompanying text.
. See 49 U.S.C. § 10741 (West Supp.1978). See note 116 infra and accompanying text.
. The rate set by the Commission with the 7% increment is equivalent to 162% of variable costs.
. The carriers cite cases decided prior to the enactment of the 4-R Act in which the Commission permitted the railroads to charge rates in excess of 200% of variable costs. See Brief of Railroad Petitioners-Intervenors at 19 n.12. The rate advocated by the railroads is equal to 178% of variable costs; the carriers claim that this figure is well within the zone of the maximum reasonable rates.
. See, e. g., Pub.L.No.94-210, § 101(b)(3), 90 Stat. 31, 33 (1976) (4-R Act enunciates congressional policy to “permit railroads greater freedom to raise or lower rates for rail services in competitive markets”) (emphasis added). We note that the San Antonio market is not a competitive one. See text accompanying note 122 infra.
. 606 F.2d 1131, 1148 (D.C.Cir.1979), cert. denied, 444 U.S. 1073, 100 S.Ct. 1019, 62 L.Ed.2d 755 (1980). Texas’ claim that any cross-subsidization of traffic violates section 10741 of the Interstate Commerce Act is thus at odds with our conclusion in Houston Lighting that differential pricing may play a role in railroad ratemaking. That does not dispose, however, of the argument that the choice of a specific increment above full costs may work an unreasonable burden on the San Antonio traffic and thus constitute a violation of section 10741. The discrimination argument was not addressed directly by the Commission. It would be appropriate therefore for this court not to attempt to resolve it here, even if we had an adequate record to do so. In addition, Texas’ objection is merely one aspect of the larger concern that the seven percent solution is unsupported by any rationale or evidence. Texas’ interests will be adequately protected when the Commission on remand is required to explain more fully its reasons for selecting any increment above fully allocated costs.
. Permian Basin Area Rate Cases, 390 U.S. 747, 792, 88 S.Ct. 1344, 1373, 20 L.Ed.2d 312 (1968).
. See note 108 supra and accompanying text.
. We note the skepticism expressed by all parties that the Commission’s approach is a temporary one. The Commission has employed the 7% factor in several decisions involving Western coal rates subsequent to San Antonio. See, e. g., Brief of Railroad Petitioners-Intervenors at 25 n.18.
. Pub.L.No.94-210, § 101(b)(1), 90 Stat. 31, 33 (1976).
. Houston Lighting & Power Co. v. United States, 606 F.2d 1131, 1149 (D.C.Cir.1979), cert. denied, 444 U.S. 1073, 100 S.Ct. 1019, 62 L.Ed.2d 755 (1980).
. San Antonio HI, 361 I.C.C. at 495.
. See 42 U.S.C. § 4332(2)(C) (1976).
. See San Antonio II, 359 I.C.C. at 18 & n.33.
. See Ex Parte No. 347, Western Coal Investigation — Guidelines for Railroad Rate Structure, Draft Environmental Impact Statement.
. Cf. Aberdeen & Rockfish R.R. v. Students Challenging Regulatory Agency Procedures, 422 U.S. 289, 324-25, 95 S.Ct. 2336, 2357-58, 45 L.Ed.2d 191 (1975) (SCRAP II) (in general revenue proceeding, permissible to defer preparation of detailed environmental impact statement to “more appropriate proceeding”).