A utility company to whose poles a cable television company attached its lines challenges the charge for that service fixed by the Federal Communications Commission. The Commission’s action is to be judged on the basis of the reasonableness of the method it used to compute the pole attachment rate; we do not look merely to whether the ultimate result fell within the range allowed by statute. In denying the utility company the right to include deferred taxes in its tax base, and thus to normalize its taxes, we find the Commission acted arbitrarily and therefore reverse that determination. We also hold that the utility company is entitled to include a component of its investment in private rights-of-way in its calculation of pole attachment rates and that such component need not be proved on a community-wide per-pole basis, which the Commission arbitrarily required, but may be established by system-wide data. We affirm on this issue, however, because we find that the utility company did not demonstrate the compensable portion of its rights-of-way investment, that is, the amount of its investment in rights-of-way pertaining to its acquisition of rights to erect and maintain poles, in contrast to its acquisition of rights to string and maintain power lines from pole to pole.
I.
Companies furnishing cable television service lease space on existing distribution poles owned by electric utilities and telephone companies in order that they may attach their coaxial cables and related equipment to the poles. As a congressional committee reported, “Use is made of existing poles rather than newly placed poles due to the reluctance of most communities, based on environmental considerations, to allow an additional, duplicate set of poles to be placed.” 1
In 1978, Congress enacted the Pole Attachment Act,2 in part to curb the extraction of monopoly profits by utilities from cable operators in need of pole space.3 The Act confers jurisdiction on the Federal Communications Commission to regulate the rates for pole attachments and to pro*1268vide that such rates are just and reasonable.4 Congress directed the Commission to establish a “simple and expeditious”5 program to regulate pole attachment rates. Paperwork was to be minimized by using rate of return and other accounting factors already in file with other regulatory commissions.6 Pursuant to the Act, “a rate is just and reasonable if it assures a utility the recovery of not less than the additional costs of providing pole attachments, nor more” than a proportionate part of “the sum of the operating expenses and actual capital costs of the utility attributable to the entire pole ... or right of way.”7 The Commission established a rate formula pursuant to which it first calculates the net cost of the pole, “multiplies that amount by a percentage figure that reflects operating expenses ..., and then multiplies the product by the percentage of pole space used by the cable operator.”8
Group W Cable, Inc. operates a cable television system serving three Texas towns: Commerce, Palestine, and Grapevine. In accordance with industry practice and pursuant to a contract with Texas Power & Light Company, Group W attached its distribution lines to approximately 5,645 poles owned by Texas Power. The contract provides for an annual rental of $2.25 per pole for attachments made before October 1, 1974, and $3.50 per pole for attachments made after that date. The contract also provides that Texas Power assumes
no responsibility for securing any franchise, rights-of-way, permits or easements for the making and maintaining of such attachment over, across, or along streets, alleys, roads, or privately or publicly-owned property ... but [Group W] assumes the duty and responsibility of securing the same.
Group W filed a complaint with the Commission in August 1979, which alleged that the rate fixed by the contract exceeded the maximum rate allowable under the Act. In accordance with Commission practice, the action was referred to the Commission’s Common Carrier Bureau. After Texas Power responded, the case languished for more than a year. Group W then amended its complaint, and, on July 14, 1981, the Bureau issued an order granting Group W’s complaint and establishing an annual rental rate of $2.50 for all poles. Texas Power filed a timely application for review with the full Commission. Texas Power asserts in its brief that the FCC staff refused to process its application “as it had refused to process many other such applications.” The Commission does not deny this, but explains that the delay was occasioned by the pressure of trying to handle the new and great volume of work that followed enactment of the 1978 Act. Whatever the reason, no action was taken on Texas Power’s administrative appeal for more than three years. Finally, Texas Power petitioned the District of Columbia Circuit for mandamus. In September 1984, that court ordered the Commission to answer, and the Commission released its decision on December 14, 1984.
In its decision, the Commission affirmed the Common Carrier Bureau’s imposition of the $2.50 per pole rate. Although Group W had indicated its willingness to pay for the power company’s investment in rights-of-way (subject to objections to its method of computation), the Commission continued to exclude that factor and also summarily rejected Texas Power’s argument on the tax normalization issue. Texas Power sought judicial review.
Texas Power first contends that the Commission’s calculation of the power company’s income tax expense was too low because it was based on actual tax paid, after accelerated depreciation, instead of what the tax would have been if “normal*1269ized,” that is, if the tax had been computed on the basis of straight line depreciation. Second, the power company argues that the Commission erred in refusing to allow it to include in the rate calculation any amounts reflecting its investment in rights-of-way for poles. The Commission found that Texas Power had not separated the rights-of-way expenses for those poles to which Group W cable distribution lines were actually attached from its total rights-of-way expenses for all of the poles in its system. Hence, the Commission refused to allow such expenses to be used in calculating the rate to be charged, but added that Texas Power might at any time provide data sufficient to prove the cost of rights-of-way for poles to which attachments had been made and obtain an increase in rates for that item.
II.
The Commission’s decision must be accepted unless it is “arbitrary and capricious, an abuse of discretion, or otherwise not in accordance with law,” a standard fixed by section 10(e) of the Administrative Procedure Act.9 Under this standard, courts exercise final authority on issues of statutory construction and the conformity or nonconformity of agency action with statutory purpose. Although due deference is given to the agency’s findings, the Supreme Court has held that courts are “not obliged to stand aside and rubber-stamp their affirmance of administrative decisions that they deem inconsistent with the statutory mandate or that frustrate the congressional policy underlying a statute.” 10 The agency must articulate clearly its findings and the reasons for its policy choices, so that the court may ascertain whether it engaged in balanced, informed decisionmaking.11
III.
The Commission argues at the outset that any error committed by it in fixing the rate is presumptively harmless because (1) the statute requires only that pole attachment rates fall within a zone of reasonableness, and (2) the total rate fixed in this case has not been shown to be outside that zone, that is, to be either unjust or unreasonable. The same argument was made to the District of Columbia Circuit and was rejected in Alabama Power Company v. Federal Communications Commission.12 The reason the argument fails is obvious. The Commission is not permitted to “luck out” with respect to its decision to set a certain rate; it may not arbitrarily choose any figure within the ephemeral zone of reasonableness and set the rate there. Rather, what the Act requires, read in conjunction with the Administrative Procedure Act, is that the Commission reach a rational decision through rational means. And the Commission must be able to explain that decision, as well as the methods and factors used in reaching that decision, in a coherent and intelligible fashion.13
“The grounds upon which an administrative order must be judged,” the Supreme Court has said, “are those upon which the record discloses that its action was based.” 14 The Commission’s “action must be measured by what [it] did, not by *1270what it might have done.” 15 The Commission’s jurisdiction was invoked by Group W’s complaint that the rate it was being charged by Texas Power was higher than the law permitted. The burden of proof of a statutory violation was on Group W.16 The Commission decided what was the maximum charge, and its “focus on the upper end of the zone of reasonableness was thus a product of both the posture in which this ease arose and the plain language of its own regulations.” 17 As the District of Columbia Circuit Court said in Alabama Power, “The rationality of administrative decisionmaking can only be assessed by reference to the agency’s stated objectives.” 18 Like that court, we will not consider whether the agency might have reached a proper result under a theory upon which it never relied.
IV.
Congress and the Commission, acting pursuant to its mandate, recognize that tax expenses are a legitimate and necessary part of the cost of providing pole attachments and that such expenses should be taken into account in determining the rates for attaching cable television cables to utility poles. The Commission sought to implement this requirement by including in its rate formula a component representing taxes, including income taxes. In calculating Texas Power’s income tax expense, however, the Commission excluded those taxes (known as deferred taxes) that are attributable to the accelerated depreciation provisions of the Internal Revenue code.
This issue was recently presented to the District of Columbia Circuit Court of Appeals in Alabama Power Company v. Federal Communications Commission.19 There the court found that it could not “uphold the Commission’s decision to exclude normalized taxes from the calculation of the maximum,”20 because the Commission had not explained its contradictory treatment of normalization. We agree, but we find in addition that the Commission’s failure to allow Texas Power to normalize its tax expense was arbitrary and capricious.
A.
Two methods have been used by rate-fixing agencies for treating deferred taxes in calculating utility rates. One approach permits the recognition of the debt that the utility has incurred for deferred taxes even though these taxes are not currently paid. Because this method matches the utility’s tax expense with the underlying events that gave rise to the tax liability, it is referred to as normalization. Another method, known as flow-through, permits inclusion in the cost base of only taxes actually paid. The result of the flow-through method is to pass all the benefits of accelerated depreciation through to current year ratepayers, while shifting the burden of the underlying tax liability to future years’ ratepayers. Plow-through proponents assert that it “serves to ensure that rates reflect only those expenses actually borne by the utility, while opponents argue that only through normalization can expenses be matched to the customer that receives the related services.”21
Because Congress had not explicitly forbidden use of the flow-through treatment for all purposes, this circuit held in 1966 that the Federal Power Commission had discretion to use that method in setting *1271natural gas pipeline rates.22 In 1969, however, Congress expressly conditioned a utility’s right to use accelerated depreciation for tax purposes for certain classes of property (i.e., property placed into service after 1969) on that utility’s receiving normalized rate-making treatment from its regulatory commissions.23 Almost immediately, federal regulatory commissions changed their practice to provide for tax normalization. The Commission amended its rules to permit telephone companies to normalize,24 as did the Federal Power Commission with respect to electric utilities and natural gas pipelines.25
The Tax Reform Act of 1969 contained an exception to its normalization provision, however. Under 26 U.S.C. § 167(Z)(2)(C), the continued use of flow-through for property of the same kind as property with respect to which flow-through was used before 1969 was permitted. Because during the 1970’s a significant amount of utility property fell within this category, a number of state regulatory commissions continued to require flow-through of federal tax benefits despite Congress’ expression of contrary federal policy. Congress eliminated this exception in 1981, however, when it enacted the Economic Recovery Tax Act of 1981 (“ERTA”),26 mandating the use of normalization in setting rates as a necessary condition to a utility’s right to use accelerated depreciation under the Code for tax purposes.27
B.
The Commission refused to allow Texas Power to normalize its tax expenses in establishing cable attachment rates, arguing that neither the Tax Reform Act of 1969 nor ERTA affects its policy. It asserts that this is not a common carrier tariff proceeding but a complaint proceeding designed to bring the pole attachment rate within the statutory limits. The Commission, therefore, continued its use of the flow-through method of accounting to calculate the power company’s costs. Texas Power challenges this policy, and contends that the normalization of tax benefits does not render an otherwise acceptable rate unjust and unreasonable.
As noted previously, this tax-method issue was also presented to the District of Columbia Circuit in Alabama Power. Because the Commission had taken contradictory positions on the question,28 that court refused to uphold the Commission’s decision to exclude normalized taxes from the calculation of the maximum pole attachment rate. The case was remanded to the Commission with instructions to allow tax normalization, unless and until a reasoned position for its use of the flow-through method could be successfully reconciled and explained.29
Other considerations support this determination. Congress has expressly provided that regulatory commissions may not exclude deferred taxes in setting rates for regulated utilities. And Congress found that to permit such a practice would frustrate the very purpose of the tax code in permitting utilities to take advantage of accelerated depreciation, and therefore prohibited the practice.30
Even though this is not a common carrier proceeding, the objective of the proceeding *1272is to fix a just and reasonable rate. Increase of the cost base by including taxes that must be paid eventually but have been postponed because of the use of accelerated depreciation, a practice permitted to encourage investment but lacking economic reality, does not, as the Commission asserts, insert a “fictional” component. The deferred tax will some day be paid. The only ultimate net gain the utility obtains is that it has the present use of the amounts paid to compensate for the postponed tax. The Commission’s approach, therefore, results in cable television companies being given a double advantage at the expense of utility ratepayers. The rate-of-return component established for the power company’s rates (both electric and pole attachment) is lower because the rate base is reduced due to accelerated depreciation and deferred taxes. Were it not for these policies, a higher authorized rate of return would have been required. Yet the Commission in setting pole attachment rates utilizes a rate-of-return component that is premised on the recovery of deferred taxes (resulting in a lower overall rate), but utilizes a tax component that fails to permit the recovery of such deferred taxes.
Furthermore, the tax component is inconsistent with the depreciation component under the Commission’s rate formula. The depreciation component is based upon the straight line depreciation method. The tax component, however, assumes that accelerated depreciation has been taken (but refuses to permit its recovery in either the tax or depreciation component). The result is that the Commission’s formula uses inconsistent and arbitrary rate elements.
Group W protests that Texas Power will receive a fair return from utility ratepayers even if it is not permitted to normalize taxes because the burden will be borne by its ratepayers, as if that somehow justifies passing on an increased rate to utility users so that cable television companies can benefit. Texas Power is indeed performing its responsibility to its utility users by minimizing their rates to the extent that other enterprises benefit from the use of property that is in effect paid for by utility users. The question is not, as Group W puts it, how much cable should subsidize Texas Power but how much cable and cable subscribers may benefit from the use of Texas Power’s property without reimbursing that company.
The Commission’s approach is also inconsistent with the intent of Congress in enacting the Pole Attachment Act. Congress directed the Commission to set pole attachment rates based upon simple and expeditious procedures, and in order to do so to deter to the cost determinations made by a utility’s state regulatory commissions. The state regulatory commission to which Texas Power is subject permits the recovery of deferred taxes, and the adjustment that must be made for those rates following a different practice is relatively simple. Finally, that the Commission’s treatment of utility taxes in pole attachment cases has always been the same proves only formal consistency not the correctness of the method or the consistency of the logic by which the conclusion was reached.
V.
The Act requires, as we have noted, that the maximum rate for a pole attachment shall include a component of the “actual capital costs of the utility attributable to the entire pole ... or right-of-way.”31 In determining a utility’s capital costs, a number of items are normally included in the computation, such as the power company’s investment in the poles themselves and items associated with the poles, such as guys and anchors.32 Texas Power obtains rights-of-way from private landowners in order that it may erect poles and string and maintain power lines across them. The issue presented is whether computation of Texas Power’s capital costs should include its investment in these rights-of-way. In considering it, we ignore the red herring dragged into the case in the Group W brief: *1273Texas Power does not seek to recover any part of the franchise fee it pays public bodies for the use of public rights-of-way. Group W itself obtains franchise from public bodies and fully pays for such franchises.
A.
Although Texas Power generally tries to construct its poles within public rights-of-way, it cannot always do so. Access to a public right-of-way may be blocked by a large tree or other impediment, or there may be no public right-of-way for the required route. Public rights-of-way are not available for approximately one-fifth of all of the poles in the state-wide Texas Power System and, in these instances, Texas Power typically buys a right-of-way from the owner of the property before installing its poles. Although the record is not clear, and counsel were not able at oral argument to explain fully, it appears likely the rights-of-way acquired by Texas Power from private landowners include not only the right to erect and maintain poles, but also the right to string and maintain power lines from pole to pole. Counsel for Texas Power airily dismisses the line-maintenance component as “air rights;” but, however designated, the right-of-way charge presumably includes consideration for this “air right.”
Under the contract between the parties, Group W is required to obtain its own independent right-of-way from landowners: Texas Power assumes “no responsibility for securing any ... rights-of-way ... or easements for the making and maintaining of [cable] attachment[s] over, across, or along ... privately or publicly-owned property.” Counsel for Texas Power and Group W both stated in oral argument, however, that the cable company does not, in actual practice, obtain its own rights-of-way from landowners to string and maintain its cable lines. Both agree that, should land owners become aware of .their rights, they might demand payment from the cable company for use of these “air rights.”
The Commission and Group W argue simply that, because the cable company is required by the contract to obtain, at its own expense, its own rights-of-way, the utility’s investment in rights-of-way should not be included in the computation of the pole attachment rate. Essentially, the argument is that because the cable company is, theoretically, not allowed to “use” the utility’s rights-of-way, they do not benefit from the utility’s acquisition of the right. We do not agree. But for the power company’s purchase of these rights-of-way, the poles upon which the cable company attaches its lines would not be standing.33 The cable company benefits from the investment that Texas Power makes in these rights-of-way, but only to the extent the investment pertains to erecting and maintaining poles. A portion of Texas Power’s investment in private rights-of-way undoubtably includes payment for the right to string and maintain its own power lines, and this investment does not inure to the benefit of the cable company. Therefore, if, as the Commission permitted, Texas Power seeks hereafter to obtain compensation for the use of its poles, it should be allowed to recover a proportionate share of Texas Power’s investment in rights-of-way attributable to “land rights” (that is, the right to erect and maintain poles) as opposed to “air rights,” the right to string and maintain lines between those poles. The portion of the rights-of-way investment found to be attributable to the actual poles may then be used in computing the capital costs of the poles, and hence the resulting pole attachment rate.
B.
Texas Power notes, and Group W does not refute, that Group W did not object in the Commission proceedings to the argument that the cost of rights-of-way, at least for poles, is a necessary cost of constructing poles, and that a ratable share of these costs should be included in the determina*1274tion of rates. The Commission apparently agreed, but in its final ruling stated that Texas Power had not proved that its investment in rights-of-way actually related to poles erected in the communities which were the subject matter of this proceeding. It required the power company to submit evidence from public records or private company records showing that it sought reimbursement only for the use of poles “in the Texas communities in this proceeding” that actually supported cable attachments.
In requiring such a showing, the Commission relied upon Williamsburg Cablevision v. Carolina Power & Light, 34 in which it stated:
The Bureau has indicated previously that in some circumstances investment in rights-of-way may be included in the calculation of the cost of a bare pole____ Inclusion of rights-of-way or other accounts, however, is dependent upon a party demonstrating that such accounts are closely related to the particular pole attachments involved.
Group W argues that Texas Power’s rights-of-way investment account is precluded from inclusion in FERC Account 364, the account showing the gross investment in pole plant, because the utility company has not shown that the rights-of-way account is “properly” or “closely related to the particular pole attachments involved.” While Texas Power has not proffered data with respect to the particular pole attachments involved in the three communities at issue here, we do not perceive the rationale behind the Commission’s refusal to permit system-wide average cost data to be used. In many cases in which utilities have attempted to submit data concerning their costs on a community basis, the Commission has refused to consider it, and has instead required system-wide data.35 The Commission does not directly refute this, but argues that use of the community-wide data was rejected in those instances because of insufficient proof, thereby acknowledging its preference for state-wide data. Indeed, the Commission’s rules contemplate use of system-wide cost data in most other respects.36
The Commission rejects the use of system-wide data to establish rights-of-way costs because “it makes no sense” — “there is no discernible nexus between those expenses and the provision of the pole attachment service.” With respect to the computation of the cost of bare poles, however, the Commission allows use of system-wide average cost data because it “makes sense and is fair to the cable operator and utility alike because the average cost of a bare pole must first be determined before the cable operator’s ratable share of the cost of using that pole can be ascertained.” Moreover, the use of system-wide average cost data, the Commission states, “offers an easy and reliable way of determining this essential figure.”
If the acquisition of private rights-of-way by Texas Power were a rare or isolated occurrence, the Commission’s distinction might make more “sense,” to use its own words. But Texas Power must obtain rights-of-way for its poles approximately one-fifth of the time; it has purchased approximately 175,000 rights-of-way for its 849,000 poles. In the communities in question here, Group W has attached its distribution lines to approximately 5,645 poles owned by Texas Power, and the utility company alleges in its brief that about 3,000 unauthorized attachments by Group W exist. If Texas Power were required to sub*1275mit community-specific right-of-way information for each of the eighty communities in its service area with cable attachments, the expense would be truly burdensome. Texas Power would be required to examine each right-of-way land record in its files, determine the exact location of the parcel (many times by reference to metes and bounds), and then physically inspect each parcel to determine whether a cable attachment is on that pole. Such a requirement certainly would not offer “an easy and reliable way of determining” this investment cost, and it does not further Congress’ directive to the Commission to establish a “simple and expeditious” program.
The per-pole standard suggested by the Commission is inconsistent, arbitrary and essentially unworkable. The Commission should permit Texas Power to recover its investment in rights-of-way by allocating the cost of rights-of-way on a system-wide average cost basis just as it allocates other costs found to be sufficiently related to its acquisition of poles.
Counsel for Texas Power has called to our attention the Eleventh Circuit decision in Florida Power Corporation v. FCC, 37 holding that the Pole Attachment Act is unconstitutional because the determination of just compensation was made by an administrative agency rather than by judicial inquiry. The constitutionality of the Act was not, however, attacked in this proceeding.
For these reasons, we VACATE the Commission’s order and REMAND to the Commission for further proceedings consistent with this opinion.