Barasch v. Pennsylvania Public Utility Commission , 507 Pa. 496 ( 1985 )


Menu:
  • 507 Pa. 496 (1985)
    491 A.2d 94

    David M. BARASCH, Consumer Advocate of Pennsylvania, Appellant,
    v.
    PENNSYLVANIA PUBLIC UTILITY COMMISSION and Pennsylvania Power Company, Appellees.

    Supreme Court of Pennsylvania.

    Argued May 15, 1984.
    Decided April 9, 1985.

    *497 *498 *499 Daniel Clearfield, Harrisburg, for appellant.

    Julian Suffian, Charles F. Hoffman, Harrisburg, for appellee Pa. Public Utility Com'n.

    Thomas Gadsen, Alan L. Reed, Albert D. Brandon, Philadelphia, for Pa. Power Co.

    Before NIX, C.J., and LARSEN, FLAHERTY, McDERMOTT, HUTCHINSON, ZAPPALA and PAPADAKOS, JJ.

    OPINION OF THE COURT

    HUTCHINSON, Justice.

    I.

    The Consumer Advocate appeals by allowance Commonwealth Court's order affirming the Public Utility Commission in a rate case. In dealing with the rates proposed in Pennsylvania Power Company's Tariff Supplement No. 15 *500 the Commission allowed the Company to "normalize"[1] both its federal and state taxes for three separate classes of property: (1) property placed in service between 1971 and 1980, (2) property placed in service in 1980 and (3) property placed in service after January 1, 1981. The Commission's order approved the setting of rates on the basis of the higher income taxes the utility would hypothetically have paid if it used straight-line depreciation, an historical method of depreciation related to the asset's actual useful life.

    Specifically, the Commission allowed "normalization" of federal and state taxes for: (1) the difference between depreciation by asset class life (ADR) and straight-line depreciation for property placed in service between 1971 and 1980; (2) the difference between depreciation by the Double Declining Balance method (DDB) and straight-line depreciation for property placed in service in 1980; and (3) the difference between depreciation by the Accelerated Cost Recovery System (ACRS)[2] provided in the Economic Recovery *501 Tax Act of 1981 (ERTA), United States Internal Revenue Code (IRC) Section 168, 26 U.S.C. § 168, and straightline depreciation for property placed in service after 1980.

    Commonwealth Court, in affirming the PUC order, held (a) that the decision to approve normalization of both state and federal tax expense in all three respects was supported by the Commission's short general analysis of that concept and (b) that the application of normalization to all of these assets did not violate Pennsylvania's longstanding doctrine of allowing only "actual taxes paid" to be considered in the rate base.

    The Consumer Advocate challenges those parts of the Commission's order permitting normalization of the state income tax expense for assets placed in service in all three of these periods. He also challenges normalization of the federal income tax expense for assets placed in service for the period 1971 through 1980. He concedes the propriety of normalizing federal taxes for assets placed in service after 1980. The Commission gave no indication that it either recognized any distinction among the situations presented, or that it considered the "actual taxes paid" doctrine.

    On the record before us normalization of this utility's state taxes violates the "actual taxes paid" doctrine[3] in all instances and offers no offsetting advantage to ratepayers. The decision to permit normalization of state taxes is therefore reversed.

    With respect to normalization of the reduced federal taxes attributable to rapid depreciation allowable on pre-1981 assets we remand to the Commission for further *502 consideration of the "actual taxes paid" doctrine. In so doing we direct the Commission to determine whether the distinctions traced in IRC § 167(l) between assets placed in service at different times require separate calculation of the annual tax reductions attributable to those assets earlier placed in service. On examination the Commission may find that the decline in those tax deductions, concomitant with the ever-shrinking deductible base of such older assets, can be offset by the continuing tax reductions a going concern will obtain from rapid depreciation on the newer assets which it must from time to time provide. On the other hand, the Commission may find, after considering trends in both inflation and energy demand, that this continuing provision of new assets is not likely to produce such an offset. In such case the Commission is further directed to determine when the turn around to higher tax expense is likely to occur. Based on that last determination the Commission must finally determine whether it is reasonable to charge current ratepayers for those future tax expenses. Such findings and conclusions are necessary to enable the judiciary to exercise its reviewing function and to deal separately with the rate-making effect of normalization of federal tax expense for each of the three asset categories covered in the Commission's order.

    II.

    Commonwealth Court described the procedural and factual history of this case as follows:

    Proceedings in this case began on April 15, 1981, when Pennsylvania Power Company filed. . . . Supplement No. 15 requesting an increase in total annual operating revenues of $32,735,000.00 based on a future test year ending December 31, 1981. By order adopted June 11, 1981, the PUC instituted a formal investigation of the lawfulness and reasonableness of the proposed rates. . . . The matter was assigned for hearing and six complaints against the proposed rate increase were filed. . . . The complaints were consolidated with the PUC's investigation for hearing *503 and disposition. . . . [T]he Administrative Law Judge, on December 3, 1981, issued his recommended decision. Exceptions and Replies to Exceptions were timely filed by the Power Company, the PUC trial staff, and the Consumer Advocate.
    On January 22, 1981, the PUC entered the Order appealed here and approved rate levels which would produce total annual operating revenues of $164,238,000.00, an increase in total operating revenues of $24,915,000.00. The order also expressly approved the Power Company's claim for normalization of deferred federal and state income taxes.

    Cohen v. Pa. PUC and Pennsylvania Power Company, 76 Pa. Commonwealth Ct. 353, 355, n. 1, 463 A.2d 1274, 1276, n. 1 (1983).

    That part of the order generally allowing normalization of tax expenses increased the Company's claimed expenses for cost of service by $2,777,949.00. That total increase breaks down as follows:

    Federal     State      Total
    CATEGORY I
    (1971-1979 Property)
    ADR with DDB over longer
    book lives with DDB        $1,074,293   $265,204   $1,339,497
    CATEGORY II
    (1980 Property)
    DDB over Straight-line
    Depreciation               $  941,286   $232,369   $1,173,655
    CATEGORY III
    (1981 Property)
    ACRS over Straight-line
    Depreciation               $  212,371   $ 52,426   $  264,797
    __________   ________   __________
    Total:        $2,227,950   $549,999   $2,777,949
    

    The difference between the tax expense allowable for rate purposes under straight line depreciation and the tax deduction allowable under accelerated depreciation is credited *504 to a reserve account for future taxes. However, the revenue required to meet this increased rate-making expense is much greater than the hypothetical tax increase itself. This is because the revenue used to establish the reserve fund is itself largely taxable. In this case, the revenue requirement approximates $5,000,000.00. Under currently applicable corporate tax rates about two dollars of revenue will be generally necessary for every one dollar added to the reserve tax account.[4]

    In effect, IRC Section 168(e)(3)(C) with its cross-reference to Section 167(l) makes rapid depreciation unavailable to utilities in computing their federal income tax if they are not allowed to charge the higher current rates necessary to fund the tax reserve account required by normalization. Because the loss of this benefit in computing federal income taxes would result in even higher rates than those resulting from normalization of that tax expense, the Consumer Advocate does not, on this appeal, contest normalization of federal income tax expense for depreciation on post-1980 property (Category III). The Consumer Advocate points out that ERTA requires normalization of federal tax expense when a utility uses ACRS. A utility's revenue requirement is higher if it does not take advantage of ACRS depreciation allowances than if it does not normalize. This result occurs because the ratepayer receives a rate base deduction under normalization which, in turn, reduces the *505 revenue requirement.[5] Thus, flow-through would, for post-1980 assets, increase the cost to the ratepayers. It is for this reason that the Consumer Advocate has not contested the action of the Commission in applying normalization to post-1980 assets.

    A reserve for taxes, like any reserve, is a source of current funds. In theory, those funds will be used to finance new capital expenditures until higher taxes fall due. The funds generated from the reserve are, however, available for any proper business purpose. By allowing normalization and assuming the Company will use its deferred tax account to finance its capital expenditures, the Commission is, in fact, requiring current ratepayers to make a forced investment in the Company.[6] The Commission has previously recognized this. In Pennsylvania PUC v. Duquesne Light Co., 43 PUR 4th 27 (1982), the Commission rejected normalization stating that it:

    1. forces ratepayers to become involuntary investors;
    2. forces the ratepayer to provide an additional dollar to pay the resulting increase in current federal and state taxes for every deferred tax dollar provided by ratepayers;
    3. produces an inequity since that ratepayer who pays in the first part of an asset life may not be the one who bears the benefit of underpaying in the latter part of the assets life, and
    4. threatens to deny ratepayers a chance to reap its (normalization's) benefit for a minimum payback period of twenty-two years, and possibly never if inflation surpasses 10 per cent.

    Id. at 69. The Commission then correctly observed:

    In developing an appropriate and reasonable level for income taxes, we are guided by the Internal Revenue *506 Code and the "actual taxes paid" doctrine as enunciated by the [C]ommonwealth [C]ourt. Where the code is silent on the treatment to be afforded a particular matter, we perceive our duty to approximate the actual taxes to be paid by the utility, and pass that cost and only that cost, on to ratepayers.

    Id. at 69 (footnote omitted). See also Pennsylvania PUC v. Pennsylvania Power and Light Co., 54 Pa. PUC 645, 670-71 (Jan. 1980 — Feb. 1981). The amount of the forced investment may be somewhat reduced because the Commission is permitted to remove from the rate base the amount in the tax reserve account. 26 Code of Fed.Reg. 1.167(1)-1(h)(6). Pittsburgh, Appellant v. Pennsylvania Public Utility Commission, 182 Pa.Superior Ct. 551, 128 A.2d 372 (1956) (Pittsburgh I); Pennsylvania PUC v. Duquesne Light Co., supra.[7] In other words, the Commission may deny the Company a return through its rates on the funds provided by consumers to the tax reserve account.

    III.

    In establishing rates, any regulatory body must determine whether the advantages derived from accelerated depreciation, i.e., a larger deduction for federal income tax expense, should accrue to a firm's shareholders or to its customers, the ratepayers. Regulators must choose between the "flow-through" and "normalization" methods. Under the flow-through approach, utilities using accelerated depreciation for tax purposes use the same depreciation method for computing the tax expense includable in their cost of service. See Federal Power Commission v. Memphis Light, Gas and Water Div., 411 U.S. 458, 460, 93 S.Ct. 1723, 1725, 36 L.Ed.2d 426 (1973). Flow-through comports with Pennsylvania's longstanding "actual taxes paid" principle. Under that principle a utility is only permitted to *507 charge rates which reflect its actual tax expense for any given year. Public Systems v. F.E.R.C., 709 F.2d 73, 76 (6th Cir.1983) (Public Systems II). Flow-through also benefits the utility's current customers since the entire advantage derived from the utility's tax deduction "flow through" to them. See Distrigas of Massachusetts Corp. v. F.E. R.C., 737 F.2d 1208, 1212 (1st Cir.1984).

    Normalization, in contrast, allows the utility to compute the income tax component of its cost of service as though it were using the straight-line depreciation method. Thus, during the first years of normalization, customers are charged for more taxes than the utility actually pays. This method is said to protect future ratepayers from unfairly high tax bills which may occur if current customers receive the entire tax benefit attributable to accelerated depreciation. See Public Systems II, supra. However, as the Commission has previously recognized,[8] normalization also benefits the utility itself since it retains use of the funds allocated for deferred taxes until the taxes actually fall due. Distrigas of Massachusetts Corp. v. F.E.R.C., supra.

    Advocates of normalization argue that if tax expenses are not normalized utilities may have to raise future rates to pay higher future taxes, thus shifting expenses incurred by current consumers to future taxpayers. Public Systems v. Federal Energy Regulatory Commission, 606 F.2d 973 (D.C.Cir.1979) (Public Systems I), later app. 709 F.2d 73 (D.C.Cir.1983) (Public Systems II).[9] The theory supporting normalization assumes that accelerated depreciation merely provides a deferment of taxes, not a true saving, and that proper income accounting requires the establishment of a reserve for deferred taxes. Under this theory, the annual credit to the reserve is considered a business cost and the *508 credit to the tax reserve is an allowable expense in figuring the cost of service component of the rate base.[10]

    The Consumer Advocate, on the other hand, argues that normalization, absent IRC § 167(l) and § 168(e)(3)(c) penalties, requires the consumer to pay for purely speculative future tax obligations. He suggests that normalization of all taxes violates the "actual taxes paid" doctrine[11] on this record because the Company failed to provide evidence that the sums reserved for taxes will in fact be paid to the taxing authorities in the reasonably foreseeable future.[12]

    Unless tax benefits are passed on to the taxpayer, the argument goes, the utility will reap a permanent tax saving, since so long as the expenses subject to normalization remain stable or increase, the benefits of postponing present tax liabilities will more than offset previously deferred taxes that currently fall due. This proposition is strengthened in periods of inflation, when current liabilities exceed previously deferred expenses.

    Public Systems II, supra (footnotes omitted). On this view, as long as total company assets grow at all, future taxes will not increase despite accelerated depreciation. See also Federal Power Commission v. Memphis Light, Gas and Water, 411 U.S. 458, 93 S.Ct. 1723, 36 L.Ed.2d 426 (1973), conformed to Memphis Light, Gas and Water Div. v. F.P.C., 500 F.2d 798 (D.C.Cir.1974); City of Los Angeles v. PUC, 15 Cal.3d 680, 125 Cal.Rptr. 779, 542 P.2d 1371 (1975) (accelerated depreciation produces permanent tax *509 savings at least in the case of expanding or stable plants). The Consumer Advocate's argument plainly implies that the Commission's approval of normalization of federal taxes for Categories I and II rate-making purposes is so unreasonable that it constitutes an error of law.

    A choice between these competing theories requires a detailed analysis of IRC § 167(l) and its history. With respect to federal taxes on certain property placed in service after 1969 and before December 31, 1980, the Tax Reform Act of 1969 and subsequent amendments permit a utility to use: (1) straight-line depreciation; (2) accelerated depreciation with normalization; or (3) accelerated depreciation with flow-through of its tax benefits to ratepayers, if the utility used such flow-through for its July 1969 accounts. 26 U.S.C. § 167(l)(2)(B) and 26 U.S.C. § 167(l)(2)(C). Prior to August, 1969, Pennsylvania Power Company used accelerated depreciation with flow-through for its post-1969 assets.

    Since the Tax Reform Act of 1969, IRC § 167(l) has had a further special category of post-1969 "expansion" property. On such expansion property a utility could elect, by June 29, 1970, to abandon accelerated depreciation with flow-through in favor of accelerated depreciation with normalization. Central Maine v. PUC, 405 A.2d 153 (1979), cert. denied, Maine Public Utilities v. Central Maine Power Co., 447 U.S. 911, 100 S.Ct. 2999, 64 L.Ed.2d 862 (1980). If it so elected, the utility was, in effect, prohibited from continuing to receive the tax benefits of accelerated depreciation unless the regulatory body with jurisdiction over its rates approved normalization.[13]

    *510 Despite a Public Utility Commission statement of policy dated June 24, 1970 which assured public utility companies that they would be permitted to normalize post-1969 expansion property if the utilities filed the "election" required by the IRC, Pennsylvania Power Company failed to make that election. Thus, the Commission is not now required to allow normalization as a pre-requisite to this utility's use of accelerated depreciation for its pre-1981 expansion property. Similarly, the IRC, as amended, does not require normalization of tax benefits as a prerequisite to the use of double-declining balance on 1980 property, even if the utility had made a Section 167(l)(4)(A) election to normalize. See F.P.C. v. Memphis Light Gas and Water, supra, 411 U.S. at 471-472, 93 S.Ct. at 1731. Since the rapid depreciation method this utility used for 1980 property involved only DDB[14] no current tax benefit is lost on that category of property.

    Finally, on the facts of this case, neither the Internal Revenue Code nor the Treasury Regulations require normalization of pre-1981 assets for rate purposes before this company can compute its federal taxes using accelerated depreciation. See IRC § 167(m); 26 CFR 1.167(a)-11(b)(6).[15] Thus, there is nothing in the Internal Revenue Code which would have jeopardized accelerated depreciation for any of this utility's property in Categories I and II, if the Commission *511 had continued to require flow-through of the federal tax benefits resulting from that rapid depreciation. The propriety of normalizing federal income tax expense for all of the property in those categories may, therefore, be considered together on the facts of this case.[16]

    IV.

    The Company's own experts did not testify that the assumed federal tax expenses appellant questions will ever become due under reasonable projections about the rate of inflation and the Company's growth. On such a record we cannot assume that the Commission considered and applied the "actual taxes paid" doctrine to federal taxes before it granted the Company's normalization request.[17]

    The filing of six separate complaints, after the PUC instituted a formal investigation of the lawfulness and reasonableness of the Company's proposed rates, places the burden of proving that the proposed rates are just and reasonable on the Company. That burden includes an obligation to prove that the tax expense included in the rates charged to the consumer bears a fair and substantial relationship to the actual tax expense or that the denial of *512 normalization accounting would in fact penalize ratepayers. Public Utility Code § 315(a), 66 Pa.C.S. § 315(a).[18]

    In Pittsburgh I, supra, Superior Court quoted with approval the statement: "[T]he utility [is] not entitled to an allowance on account of a tax which it had not been called on to pay and which there was no evidence that it would be required to pay," citing Chambersburg Gas Co. v. PUC, 116 Pa.Superior Ct. 196, 224, 176 A. 794, 805 (1935) (company filing consolidated return may not show as expense for rate-making purposes income taxes it would have paid had it filed a separate return).

    Two years later in Pittsburgh, Appellant v. Pennsylvania Public Utility Commission, 187 Pa.Superior Ct. 341, 144 A.2d 648 (1958) (Pittsburgh II),[19] Superior Court further explained the "actual taxes paid" doctrine, restating its earlier conclusion that a tax saving should not be retained by a utility or recovered from its ratepayers unless it is shown that the saving is merely deferred. If the Commission now allows a company to pass on costs or expenses in the absence of evidence it will actually incur them, it abuses its discretion.

    In an effort to bring itself within the doctrine so redefined, the utility company here has argued that accelerated depreciation merely defers taxes. It is true that accelerated depreciation of each asset taken as a single unit would result in a mere tax deferral if overall effective tax rates were static and if the mechanical application of that static rate to the annual depreciation deduction attributable to each individual asset resulted in a pro-tanto tax reduction. *513 Under such assumptions, the tax reduction on each asset would be correspondingly reduced in later years and the tax due would be increased by the sum of all such reductions in a simple linear relationship. In reality, however, the relation is dynamic; utilities replace depreciated assets. Their capital expenditures increase with growth and inflation and tax rates are not static. Indeed, the Commission's rate case handbook states:

    "[A] continuing construction program adding assets at a constant or increasing dollar rate may perpetuate indefinitely the fact that the income tax claim of a utility remains lower under flow-through than under normalization. This is true because, while for any given single asset tax depreciation eventually becomes lower than book depreciation, when you continuously purchase new assets and depreciate them, the total depreciation for all assets will remain higher than book depreciation indefinitely."

    J.H. Cawley and N.J. Kennard, supra, C3-C4 (emphasis in original).

    In Alabama-Tennessee Natural Gas Co. v. Federal Power Commission, 359 F.2d 318, (5th Cir.1966), cert. denied 385 U.S. 847, 87 S.Ct. 69, 17 L.Ed.2d 78 (1966), the Court of Appeals summarized the effect of normalization, stating that a regulated utility:

    "will never be required to pay higher income taxes because of its election to claim liberalized depreciation unless its gross plant declines in dollar value as a result of lower demand or lower plant construction cost. Normalization during a period of growth or stability would force ratepayers to provide funds for a hypothetical tax liability that might never become payable or, at the very least, to provide funds many years in advance of the time they are needed."

    Id. at 336 (emphasis in original).[20]See also Pennsylvania PUC v. Phila. Electric Co., 31 PUR 4th 15, 75 (1978).

    *514 In this case, the Commission has not told us whether it views this utility as a "going concern", subject to growth, inflation and the necessity of replacing retired physical plant, or whether it is instead treating the reduced tax expense from accelerated depreciation on an asset-by-asset basis. The distinct tax treatment the Internal Revenue Code now requires for pre and post-1981 assets, in the absence of overall normalization of federal tax expense, may require separate rate-making treatment of the earlier assets. In such case, the PUC must determine whether the reserve created by normalizing depreciation of post-1980 assets will be sufficient to offset the decreasing depreciation on the utility's pre-1981 property.[21] On this record we cannot tell whether the Commission examined the amounts, collected from current ratepayers and placed in the tax reserve account, in light of the Company's projected growth rate and projected replacement costs. It should do so to determine if a time will come in the reasonably near future when the Company's total annual tax due under an accelerated method of depreciating that property will be greater than under straight-line depreciation. Only if that total tax *515 is greater will there be a need to draw on the tax reserve account to meet it.

    As the expert body, the Commission must make these determinations as part of its duty to see that rates are "just and reasonable." 66 Pa.C.S. § 1301. See also Pennsylvania PUC v. Pennsylvania Gas and Water Co., 492 Pa. 326, 424 A.2d 1213 (1980), cert. denied 454 U.S. 824, 102 S.Ct. 112, 70 L.Ed.2d 97 (1981), later proceeding Pennsylvania Gas and Water Co. v. Pennsylvania PUC, 79 Pa. Commonwealth Ct. 416, 470 A.2d 1066 (1984). Those determinations must be supported by specific findings as provided in 66 Pa.C.S. § 703(e).[22]See also 2 Pa.C.S. § 507, Administrative Agency Law.

    In the absence of specific findings of fact and conclusions of law, this Court cannot infer that the Commission actually and properly determined that rates which include normalized federal taxes on Categories I and II property are "just and reasonable" as a matter of law. Further proceedings are therefore necessary on these issues relating to federal tax expense. Page's Department Store v. Velardi, 464 Pa. 276, 346 A.2d 556 (1975). National Fuel Gas Distribution Corp. v. PUC, 76 Pa. Commonwealth Ct. 102, 464 A.2d 546 (1983), later proceeding National Fuel Gas Distribution Corp. v. PUC, 81 Pa. Commonwealth Ct. 148, 473 A.2d 1109, 1123 (1984). See Noerr Motor Freight v. PUC, 180 Pa.Superior Ct. 62, 118 A.2d 248 (1955); UGI Corporation v. PUC, 49 Pa.Commonwealth Ct. 69, 410 A.2d 923 (1980). *516 Following those proceedings the Commission must make the determinations set forth at pp. 96 and 97 of this opinion.

    V.

    In the case before us the Commission, over the objections of its staff and the Administrative Law Judge, declined to consider our redefined "actual taxes paid" doctrine and permitted normalization on the ground that it would improve the company's downgraded bond rating. On this basis, without evidence or findings on the extent to which normalization's increase in service cost would be offset by a decrease in capital cost, the Commission concluded that normalization is in the "public interest" and that "it is unnecessary to review all arguments regarding the issue of normalization and flow-thru [sic]." Reproduced Record, Appendix B at p. 27. If there were evidence and findings to support it, this position would be consistent with some federal cases holding that normalization is permissible even though it results in perpetual tax deferrals, because it is a ratemaking device which promotes the financial integrity of utilities. See, e.g., Public Systems v. F.E.R.C., 709 F.2d 73, 83 (D.C.Cir.1983) (Public Systems II). This position has been rejected by Pennsylvania's Legislature and our courts in favor of the policy considerations underlying the "actual taxes paid" doctrine as redefined in Pittsburgh II, supra.

    Arguably tomorrow's consumers may be better served by requiring current ratepayers to infuse interest-free capital into our utilities through the collection of excess expenses from today's ratepayers. This may also reduce the need for additional borrowing in the capital market, improve the bond rating of utilities and lower the cost rate for common equity. However, if utility rates are to reflect such a capitalization requirement from current consumers to protect the financial integrity of the system for the future, that is a matter for the Legislature, not the courts or the Commission, to decide. In adding Section 1315 to the Public Utility Code, Act of December 30, 1982, P.L. 1473, effective immediately, the Legislature has recently reaffirmed the *517 policy that current ratepayers should shoulder only the actual expenses of providing current utility service. This section pertains to those costs incurred by electric utilities which may be included in their rate bases and provides:

    Except for such nonrevenue producing, nonexpense reducing investments as may be reasonably shown to be necessary to improve environmental conditions at existing facilities or improve safety at existing facilities or as may be required to convert facilities to the utilization of coal, the cost of construction or expansion of a facility undertaken by a public utility producing, generating, transmitting, distributing or furnishing electricity shall not be made a part of the rate base nor otherwise included in the rates charged by the electric utility until such time as the facility is used and useful in service to the public. Except as stated in this section, no electric utility property shall be deemed used and useful until it is presently providing actual utility service to the customers.

    66 Pa.C.S. § 1315.

    We believe we should take into account this declaration in a plainly related area in determining whether the Commission has the power to make a general change in policy with respect to normalization of tax expense to the benefit of investors as well as future ratepayers and to the detriment of current ratepayers.

    We simply cannot tell on this record whether the use of normalization for pre-1981 federal tax expenses, as opposed to flow-through, results in a rate which justly and reasonably allocates responsibility for those federal tax expenses among the utility and its current and future ratepayers. At a minimum, the PUC should have examined relevant factors such as the potential effect of separating utility assets into two distinct categories for purposes of federal tax deductions: (a) pre-1981 assets, which will forever decline in value, and (b) assets placed in service after 1980, which will be constantly augmented or replaced as they wear out or as demand increases. If the Commission's summary adoption of normalization of tax expense without a real discussion of *518 its reasons reflects a new policy generally favoring normalization, without regard to the particular circumstances surrounding a company's request, it cannot stand. We are uncertain whether the Commission has intended such an about face because it has previously rejected normalization of tax expense as a general policy, stating that:

    It may be argued that the matter of normalization versus flow-through could be handled better in a generic proceeding, since the underlying theories are applicable to all electric utilities who could provide input. A thorough review of this matter convinces us, however, that not only would it be appropriate — but even preferable — to approach at least certain aspects of the normalization flow-through problem on a case-by-case method. Normalization and flow-through, where not required for accelerated depreciation, do not lend themselves to an "all or none" treatment. While it may be true that a general commission policy on certain aspects of this problem may be better reached on a generic basis, on the other hand, other aspects may be better met on an ad hoc basis upon consideration of the varying situations with each utility.

    Pennsylvania PUC v. Philadelphia Electric Co., 31 PUR 4th 15, 75 (1980).

    VI.

    We turn now to the PUC's authorization of normalization of state income taxes in determining rates. Here there is no likelihood that the Company will lose federal tax benefits since both the IRC and the Treasury Regulations requiring normalization are expressly inapplicable to state income taxes. 26 CFR 1.167(1)-1(a). See Pennsylvania PUC v. Philadelphia Electric Co., 33 PUR 4th 319, 353-54 (1980). Whether accelerated depreciation is normalized in rate-making has no effect on the Pennsylvania net income taxes which corporate utilities pay. To the extent that Pennsylvania accepts as its tax base a public utility's federally taxable income, that utility receives the benefit of any accelerated depreciation federal tax law allows. We believe the discussion *519 in Part III of this opinion shows that the PUC will have insured the availability of accelerated depreciation on this utility's federal tax returns by permitting normalization of the federal tax expense attributable to IRC § 168 ACRS rapid depreciation on post-1980 assets.

    Normalization has no effect on the amount due for our state income taxes. The extent by which the Commission adjusts a public utility's tax expense by removing the current tax benefits of accelerated depreciation in determining cost of service has no effect on the state taxes a utility pays. The Pennsylvania corporate net income tax law makes no reference to normalization. Unlike the Internal Revenue Code, it nowhere requires or prohibits the use of normalization in connection with the allowance or disallowance of rapid depreciation, either as a general tax benefit or that kind of tax benefit specifically denominated a "tax preference." See Tax Reform Code of 1971, Act of March 4, 1971, P.L. 6, Art. IV, Section 401, as amended, 72 P.S. § 7401(3)(d)-(o), especially § 7401(3)(l) (Supp.1984-85).[23]

    On the other hand, normalization does affect the rates which a utility charges its customers. When the Commission uses a hypothetical state tax expense which reflects depreciation calculated by the straight line method the rates set will be higher than rates based on the actual taxes paid unless the cost of capital is lowered enough to produce a fully offsetting reduction in the "fair return" component of the rate. There is no evidence to support such a finding on this record.

    Finally, since our state's tax laws require no artificial separation of assets into categories depending on when they are placed in service, no impediment appears to using a going concern analysis in dealing with state taxes. That analysis, not the separate asset analysis underlying normalization, seems to us in accord with the dynamics of the real *520 world and we will require its use in expensing Pennsylvania taxes in the absence of contrary legislative intent. Under that going concern analysis the normalization of state taxes would on this record violate our longstanding and well-recognized doctrine of "actual taxes paid," as it has been redefined.

    We believe Commonwealth Court incorrectly read Pittsburgh I as holding only that the Commission did not abuse its discretion in denying the normalization request before it in that case. That reading of Pittsburgh I overlooks its ratio decidendi. Superior Court expressly determined the Commission did not abuse its discretion in denying normalization only because allowance of normalization in that case would have violated the rule that only "actual taxes paid" can be considered as a tax expense when rates are set. Pittsburgh I, supra, 182 Pa.Superior Ct. at 580, 128 A.2d at 385.

    Prior to its decision in the case now before us, the Commission had consistently followed Pittsburgh I and II. In these earlier cases, such as Pennsylvania PUC v. Duquesne Light Co., 43 PUR 4th 27, 71 (1982), the Commission also relied on the United States Supreme Court's decision in Federal Power Commission v. Memphis Light, Gas and Water Division, supra. In Memphis Light the Court noted that normalization of accelerated depreciation:

    [I]n practice resulted in permanent tax savings. Because most utilities had growing or at least stable plant investments, the depreciation allowances from additional and replacement equipment offset the declining depreciation allowance on existing property.

    Id. 411 U.S. at 460, 93 S.Ct. at 1725. We find the actual taxes paid doctrine, as redefined in Pittsburgh II, supra, to permit consideration of tax expense which is merely deferred, is in accordance with the policy our legislature has set in this area.

    Although we have held that the power to fix "just and reasonable rates" imports flexibility to a complicated *521 regulatory function by a specialized decision-making body, we do not believe that the appropriate ratemaking treatment of deferred taxes is a matter within the unbridled discretion of the Commission. See PUC v. Pennsylvania Gas and Water Co., 492 Pa. 326, 424 A.2d 1213 (1980). We have never interpreted that discretion to be absolute. We are unpersuaded by the cases cited in support of the proposition that the decision to permit normalization is necessarily and always a matter within the Commission's discretion. These decisions generally rest on a bald conclusion that normalization is not unlawful, is supported by reason and is not arbitrary and capricious. State ex rel. Utility Consumers Council, Inc. v. Public Service Commission, 606 S.W.2d 222 (1980), cert. denied sub nom. Utility Consumers Council, Inc. v. Public Service Commission, 450 U.S. 1042, 101 S.Ct. 1761, 68 L.Ed.2d 240 (1981).[24] If normalization violates the "actual taxes paid" doctrine, the rates approved by the Commission cannot be "just and reasonable" as required by law unless some extraneous penalty, such as that which Congress imposed with respect to post-1980 property, would leave the ratepayer relatively worse off if normalization is not allowed. An agency order imposing a rate which is not just and reasonable is unlawful. This Court will not affirm an unlawful agency decision.

    We believe that the Pennsylvania version of the "actual taxes paid" doctrine, as developed in Pittsburgh I and II and the Commission in its earlier cases, accurately interprets the statutory requirement that rates be "just and reasonable," found in 66 Pa.C.S. § 1301. That view is further supported by the Legislature's recent quoted mandate *522 in the directly related area set forth in 66 Pa.C.S. § 1315. See, supra, at 104.

    Reversed and remanded to the Public Utility Commission for further proceedings consistent with this opinion.

    JUDGMENT

    ON CONSIDERATION WHEREOF, it is now here ordered and adjudged by this Court that the judgment of the COMMONWEALTH COURT OF PENNSYLVANIA be, and the same is, hereby reversed and remanded to the Public Utility Commission for further proceedings consistent with this opinion.

    NOTES

    [1] "Normalization" is a rate-making concept designed to adjust a utility's operating expenses in its test year by eliminating abnormal, non-annual events which are known and certain to change in a regularly recurring manner. J.H. Cawley and N.J. Kennard, Rate Case Handbook, 150-153 (1983). As applied to income tax expense it permits a public utility to include in its current rate base an income tax expense higher than that which it has been required to pay on the assumption that the taxes saved by accelerated depreciation are merely deferred. A particular kind of normalization of tax expense utilizing reserve accounting for income taxes is required by United States Internal Revenue Code §§ 167(l) and 168(e) if the utility wishes to use rapid depreciation on certain assets in computing its federal tax.

    [2] The "Asset Depreciation Range" (ADR) sets the useful life of assets for tax depreciation by classifying them into broad types. The tax life is generally less than the period of actual use. Thus, ADR confers a current tax benefit.

    The "Double Declining Balance" (DDB) method of depreciation offers another current tax benefit. It permits tax depreciation in the first year an asset is used at twice the rate allowable under the straight line method. The same double rate is thereafter applied to the declining balance instead of the original cost. This results in lower deductions in later years.

    The "Accelerated Cost Recovery System" (ACRS) replaced all other statutory accelerated depreciation methods for property placed in service after December 31, 1980. Under ACRS, a taxpayer may choose between only two primary methods: (1) accelerated statutory rates or (2) a straight-line rate using one of three elective recovery methods.

    All these tax benefits are designed to encourage capital investment.

    [3] In this opinion we have retained the traditional terminology of "actual taxes paid" for this doctrine. Strictly speaking, however, our present version of this doctrine, as impliedly redefined by Superior Court in Pittsburgh, Appellant v. Pennsylvania Public Utility Commission, 187 Pa.Superior Ct. 341, 144 A.2d 648 (1958) (Pittsburgh II), would be more aptly titled "actual taxes payable," since it does allow consideration of deferred taxes. See, infra, at p. 512.

    [4] An example may help clarify this situation. Suppose a company with $1,000,000.00 of depreciable plant applies straight-line depreciation for the purpose of computing its federal income taxes, it gets a depreciation deduction of $30,000. If it uses one of the liberalized methods, its depreciation deduction will be $60,000. Assuming a fifty percent tax rate the company can thus reduce its current income tax liability by $15,000.00. Under IRC § 168 the company must credit the $15,000 to a reserve for deferred taxes. If "flow-through" were used, i.e., if the taxes are not "normalized" by creating the reserve, this $15,000 tax saving would be passed on to ratepayers, resulting in lower current rates. 1 Priest, Principles of Public Utility Regulations 124 (1969). The revenue required in our case has a ratio of less than two to one because the applicable tax rate is only 46%, not 50%. If the rate were 50% the company would have to charge its customers $30,000 to get the $15,000 it must reserve.

    [5] See note 7 at 506, infra.

    [6] Indeed, the policy argument for normalization of tax expense recognizes this. It holds that the capital formation incentives behind accelerated depreciation are contradicted if the rate-making process passes them on to consumers instead of reserving them for investment.

    [7] The Consumer Advocate states that elimination of the deferred tax reserve account from the rate base ultimately reduces the additional revenue required for normalization by 20%. The Company and the Commission have not disputed that assertion. If it is correct, the reserve required by the Internal Revenue Code offsets only one-fifth of the rate increase normalization entails.

    [8] See Pennsylvania PUC v. Duquesne Light Co., supra at 69, quoted, supra, at 98.

    [9] Accord E.C. Segar, in Thimble Theater, speaking through Wimpy, "I'll gladly pay you Tuesday for a hamburger today." Contra, Aesop, The Ant and the Grasshopper.

    [10] Thereafter, however, the utility's used and useful assets may be reduced by the amounts accumulated in the tax reserve to arrive at the capital component of the rate base on which the Commission figures the utility's "fair return." See 26 Code of Fed.Reg. 1.167(1)-1(h)(6).

    [11] See, infra, at 510-513.

    [12] The Consumer Advocate does not distinguish his arguments with respect to federal and state taxes. Such distinction is necessary to proper analysis for the reasons discussed below. However, the arguments for or against normalization of state taxes do not require separation of depreciable assets into the IRC categories since state law permits accelerated depreciation without regard to time of acquisition. See discussion in Part VI, infra, at 518.

    [13] In this respect the Tax Reform Act of 1969 provided that "if the taxpayer makes an election (to normalize) . . . before June 29, 1970. . . paragraph (2)(C) shall not apply with respect to any post-1969 public utility property to the extent that such property . . . increases the productive or operational capacity of the taxpayer. . . ." 26 U.S.C. § 167(l)(4)(A). The result of adding paragraph (A) to 26 U.S.C. § 167(l)(4) "is that even though utility commissions retain discretion to approve accounting methods, approval of flow-through accounting will cause . . . utilities . . . (having made the election under paragraph (4)(A) and those which did not use flow-through accounting in 1969) to lose their rights to claim accelerated depreciation tax deductions for post-1969 expansion property." "Recent Decisions on Accelerated Depreciation and Normalization", Public Utilities Fortnightly, Vol. 105, No. 10, at 49 (May 8, 1980). This resulted in a strong practical compulsion for regulatory approval of normalization. Prior to 1969, rates were generally based on flow-through. Today, normalization is the common method.

    [14] See, supra, at 503-504.

    [15] However, neither ADR nor DDB are permissible methods of depreciating assets put in service after 1980. For such assets the ACRS method of IRC § 168(e) is the only available method of accelerated depreciation. Section 168(e)(3)(C) prohibits utilities that are not allowed to normalize tax expense in figuring their cost of service from using ACRS. The increase in tax expense this prohibition entails is apparently so large that the Consumer Advocate concedes the necessity of normalizing federal taxes to account for the effects of ACRS depreciation on these assets.

    [16] The facts presented by other utilities may require separate analyses for Categories I and II as well as for replacement and expansion property within those categories.

    [17] The Commission's own expert utility tax consultant, William F. Doyle, testified against normalization of the Company's deferred state and federal taxes:

    First, . . . . The ratepayers are captive investors. . . . These ratepayers-investors are then not allowed to earn a proper return on their investment since a part of every dollar they are forced to contribute is not deducted from the utilities' rate base. For every one dollar required for normalization the ratepayer must provide approximately $2.00 of revenues. Approximately one dollar goes to the State and Federal government as income tax on the increased revenue received by the utility. . . . Second, with regard to State deferred income taxes, there is no State or Federal law requiring normalization of state income taxes. . . . . Fourth, flow-through of State deferred income taxes is an established policy with this Commission.

    Reproduced Record at 35A-37A. Mr. Doyle also testified that the Commission had denied normalization of the ADR benefit in Pennsylvania Power and Light's recent rate filing (R-80031114). Reproduced Record at 35A.

    [18] This section reads:

    In any proceeding upon motion of the Commission, involving any proposed or existing rate . . . or in any proceeding upon complaint involving any proposed increase in rates, the burden of proof to show the rate involved is just and reasonable shall be upon the public utility.

    [19] This case involved the question of whether the utility had to flow-through benefits of accelerated depreciation even though it elected to use straight-line depreciation instead of the accelerated method for tax purposes. The court held this issue was within the discretion of the Commission.

    [20] We recognize that later federal cases do not reflect this view. Thus, Federal Power Commission v. Memphis Light, Gas and Water Div., 411 U.S. 458, 93 S.Ct. 1723, 36 L.Ed.2d 426 (1973), held that the authority to approve normalization for pre-1970 property and post-1969 replacement property is within the discretion of the Federal Power Commission. The FPC's view of normalization seems to us to be based on economic assumptions and projections accepted by the Congress with respect to future energy requirements and who should pay for them. The Pennsylvania Legislature has not accepted those assumptions and projections. See 66 Pa.C.S. § 1315 and discussion, infra, at 516.

    [21] This issue was raised in Memphis Light, Gas and Water Div. v. Federal Power Commission, 500 F.2d 798 (D.C.Cir.1974). There, the D.C. Circuit Court recognized that the funds in the reserve account established by IRC § 167(l)(3)(G) are not statutorily earmarked for the specific property which generated the deferred tax expense. However, the Court ruled that Congress contemplated that the reserve fund would be used exclusively to offset future tax liability, that is, declining depreciation on the same post-1969 (now 1980) expansion property. Id. at 804-05. See H.R.Rep. No. 413, 91st Cong. 1st Sess., pt. 1 at 132 (1969). These authorities do not bind us. The statute contains no mechanism for effectuating that intent for utilities, such as Pennsylvania Power Company, which did not make a section 167(l) election to normalize.

    [22] Prior to its repeal in 1978, Section 1005 of the Act of May 28, 1937, P.L. 1053, 66 P.S. § 1395, required the Commission to make findings "in sufficient detail to enable the court on appeal to determine the controverted question and whether proper weight was given to the evidence."

    Section 704 of the Administrative Agency Law, 2 Pa.C.S. § 704 now governs our scope of review on appeals from the Commission's adjudications and is substantially identical to the repealed provision of the Public Utility Law.

    "[T]he court shall affirm the adjudication unless it is in violation of the constitutional rights of the appellant or is not in accordance with the law . . . or that any finding of fact made by the agency and necessary to support its adjudication is not supported by substantial evidence."

    [23] The Capital Stock Franchise tax imposed on utility corporations whose business is not confined to Pennsylvania is similarly neutral with respect to normalization. Act of March 4, 1971, P.L. 6, Art. VI, 72 P.S. §§ 7601-7606.

    [24] The battle for normalization in the courts having been largely lost, in 1981 Congress denied the use of ACRS rapid depreciation to those utilities which are not permitted to normalize. By limiting these utilities to straight-line depreciation the cost to the utility, and ultimately to its ratepayers, is sufficiently onerous to compel PUC approval of normalization of federal tax expense in connection with assets put in service thereafter. This federal law does not apply to state taxes. Although the investment policy set nationally by Congress through the income tax law appears in conflict with our redefined "actual taxes paid" doctrine, that policy is not binding on us.