Opinion
TC 5411
05-24-2023
PACIFICORP, Plaintiff, v. DEPARTMENT OF REVENUE, State of Oregon, Defendant.
OPINION (VALUATION)
Robert T. Manicke, Judge
I. INTRODUCTION
This case concerns the real market value of Plaintiff's Oregon property as of January 1, 2020. Plaintiff is a rate-regulated electric utility with operating property in Oregon, as well as California, Idaho, Utah, Washington, and Wyoming. (See Ptf's Pretrial Br at 1-2; Def's Ex A at 8.) Because Plaintiff is in the electricity business, its property is subject to central assessment by Defendant. See ORS 308.515(1)(k) ; see generally DISH Network Corp. v. Dept. of Rev., 364 Or. 254, 257-58, 434 P.3d 379 (2019) (explaining central assessment). One consequence of central assessment is that Defendant may determine the value of Plaintiff's Oregon property by first valuing all of Plaintiff's property used in the same business, within and without Oregon, as a unit and then allocating a share of that "system" value to the Oregon property based on a formula. See ORS 308.555 (authorizing unit valuation); 308.550(2) (authorizing allocation). Another consequence is that intangible property (subject to exceptions not relevant here) is included in the definition of Plaintiff's "property." See ORS 308.505(14).
Plaintiff is, indirectly, wholly owned by Berkshire Hathaway Energy Company, which in turn is a consolidated subsidiary of Berkshire Hathaway Inc., the stock of which is publicly traded. (See Ptf's Ex 1 at 15; Def's Ex A at 9.) Plaintiff's financial results are not reported separately; they are combined with those of several other energy-related affiliates. (See Ptf's Ex 1 at 16.)
Unless otherwise noted, the court's references to the Oregon Revised Statutes (ORS) are to the 2019 edition.
As is customary, the court refers to "system" value to mean the value of all property, within and without Oregon, used or held for future use in a business listed in ORS 308.515(1) and considered as a unit, as allowed by ORS 308.555. See, e.g., Alaska Airlines, Inc. v. Dept. of Rev., 307 Or. 406, 410, 769 P.2d 193 (1989) (describing formulary allocation of taxpayer's "total system value").
On May 22, 2020, Defendant issued a Notice of Proposed Assessment asserting that the real market value of Plaintiff's Oregon centrally assessed property was $3,180,000,000 on the January 1, 2020, assessment date. (Ptf's Ex 6.) On June 18, 2020, Plaintiff requested a director's conference, as allowed by ORS 308.584; the conference was held July 23, 2020; and on July 31, 2020, Defendant issued an opinion and order (O&O) sustaining the real market value of $3,180,000,000. The O&O showed that the assessment was based on a value of $19,500,000,000 for the entire system. (See Ptf's Compl, Ex 2.)
Plaintiff timely appealed to the Magistrate Division from the O&O under
ORS 308.584(5) and ORS 305.280(1), initially claiming only that Defendant had overvalued its property, including by allocating excessive value to Oregon. (See Ptf's Compl at 2-5.) On the motion of both parties, the court specially designated the case to be heard in the Regular Division. Before trial, Plaintiff moved for leave to amend its complaint to add two claims based on a legal issue already before the court in Delta Airlines, Inc. v. Dept. of Rev., TC 5409, namely, that taxation of Plaintiff's intangible property violates the Oregon Uniformity Clauses. (See Ptf's 1st Amend Compl at 4-7 (citing Or Const, Art I, § 32; Or Const Art IX, § 1).) The court allowed the amendment but stayed proceedings on the Uniformity Clause claims until after trial on the valuation issues. (See PacifiCorp v. Dept. of Rev., TC 5411 (Nov 8, 2021).) The court then held a trial (Nov. 15-19, 2021) and accepted extensive post-trial briefing on the valuation issues. This opinion addresses only the valuation issues; the court will issue its order on the Uniformity Clause claims separately.
At trial, both parties presented expert reports and testimony complying with the requirement to "consider[ ] three different approaches to valuation: the cost approach, the comparable sales approach, and the income approach." Powell Street I v. Multnomah County Assessor, 365 Or. 245, 249, 445 P.3d 297 (2019) ("The appraiser is not required to use all three approaches, but the appraiser must consider them.") (Emphasis in original.).
II. LEGAL ISSUES
Before determining the real market value of Plaintiff's property, the court must resolve two issues of law. First, the parties have different views about the extent to which the court must defer to two of Defendant's administrative rules relating to the determination of value. Second, the parties differ as to which party bears the burden of proving in this court a value above or below the value recorded on the central assessment roll.
A. Court's Deference to Defendant's Administrative Rules
The parties disagree over the level of deference the court should afford to OAR 150-3080690 and OAR 150-308-0590. The court addresses these two rules separately.
1. OAR 150-308-0690: WSATA Handbook as Authority for Valuation Under ORS 308.205
OAR 150-308-0690 states in its entirety:
"The 2009 Western States Association of Tax Administrators Appraisal Handbook: Unit Valuation of Centrally Assessed Properties is adopted as the official valuation guide for property assessed by the Oregon Department of Revenue under ORS 308.505 to 308.665 for ad valorem tax purposes."
The "WSATA Handbook" consists of approximately 350 pages of guidance, in varying degrees of depth, about methods and tools for unit valuation and allocation of utilities and other centrally assessed property. It "serves as the primary textbook * * * to train state government appraisers of centrally assessed properties." WSATA Handbook at vi. As such it "incorporates WSATA's opinion of the current state of academic and appraisal theory" as applicable to centrally assessed property. Id. at I-1. As will be seen, the authors of the WSATA Handbook identify strengths and weaknesses in various valuation methods and theories; they rarely mandate or entirely condemn the use of a particular method, but they express reasons to prefer some techniques and to reject others. Cf., e.g., Minn. R. 8100.0300 (prescribing specific procedures for determining and weighting cost, income and other indicators of value for "public utility companies"); Minn Stat Ann § 270C.06 (rules "have the force of law."); Utah Admin. Code r R884-24p-62(4)(b)(iii) (setting forth "preferred valuation methods" as "rebuttable presumptions"; requiring "any party challenging a preferred valuation method [to] demonstrate, by a preponderance of evidence, that the proposed alternative establishes a more accurate estimate of fair market value").
Plaintiff argues that the WSATA Handbook does not "restrict the discretion of this Court or any other party (including the Department) in a de novo proceeding." (Ptf's Post-Trial Br at 56.) Elsewhere, Plaintiff argues that "this Court is not bound by the WSATA Handbook." (Ptf's Response at 5.) Defendant, on the other hand, argues that, because of the WSATA Handbook's status as an administrative rule, "this court need not engage in debate over the theoretical propriety of various of the issues raised in this case because the administrative rule answers the questions." (Def's Opening Post-Trial Br at 30.) Defendant argues further: "Contrary to PacifiCorp's assertions, a de novo appeal to the Tax Court does not give the taxpayer license to disregard the approaches called for by law. In this centrally assessed property matter, that means following the WSATA Handbook." (Def's Response at 3.) Defendant concludes: "Thus, to determine the real market value of PacifiCorp's centrally assessed property, the court must apply the 'methods and procedures' found in the WSATA Handbook." (Def's Reply at 1-2.)
a. Analysis under published sources of statutory authority for OAR 150-308-0690: ORS 305.100 and ORS 308.655.
Any analysis of the level of deference to afford an administrative rule must begin by identifying the statute or statutes that are the source of the agency's authority to adopt the rule. See Trebesch v. Employment Division, 300 Or. 264, 267, 710 P.2d 136 (1985) ("We seek to derive the legislature's intent from an analysis of the statutes by which a particular agency operates."). The Administrative Procedures Act generally requires agencies to specify that source when adopting or amending a rule. See ORS 183.335(2)(b)(A), (B). Although neither party refers to this requirement, it is clear that courts may look to the statute or statutes referred to in the published rule when analyzing what level of deference to apply. See, e.g., Clackamas Cty Assessor v. Village at Main Street Phase II, 349 Or. 330, 336 n 8, 245 P.3d 81 (2010) (referring to published statutory authority (ORS 305.100) in determining degree of deference to afford to Defendant's rule). In this case, a notation included with the officially published OAR 150-3080690 states: "Statutory/Other Authority: ORS 305.100 &308.655; Statutes/Other Implemented: ORS 308.655." Therefore, the court begins its analysis with those two statutes.
The first statute referred to in the published rule states in relevant part: "The Department of Revenue shall: (1) Make such rules and regulations it deems proper to regulate its own procedure and to effectually carry out the purposes for which it is constituted." ORS 305.100(1). The second statute states in full: "The Department of Revenue may prescribe directions, rules and regulations to be followed in answering any requirement of ORS 308.505 to 308.674." ORS 308.655. Both provisions originate in the 1909 act that created Defendant's predecessor (the Board of State Tax Commissioners) to supervise the statewide tax system and to centrally assess the property of utilities and other listed businesses. See Or Laws 1909, ch 218, § 4(1)("It shall be the duty of the said Board of State Tax Commissioners * * * [T]o make such rules and regulations as the board shall deem proper, effectually to carry out the purposes for which the board is constituted, and to regulate its own procedure."); id. § 7(14) ("The board is hereby given the power to prescribe directions, rules and regulations to be followed in answering any of the requirements of this act.").
(1) ORS 305.100 does not require court to defer to OAR 150-308-0690.
As to the first statute, the Oregon Supreme Court repeatedly has characterized ORS 305.100 as "granting to department interpretative, but not substantive, rulemaking authority," such that, "[t]o the extent that [a rule adopted under the authority of that statute] is inconsistent with legislative intent, as determined by this court, the rule is invalid." Avis Rent A Car System, Inc. v. Dept. of Rev., 330 Or. 35, 41, 995 P.2d 1163 (2000); see also Village at Main Street, 349 Or at 336 n 8 (concluding that Defendant's rule 150-311.216 was "interpretative, not substantive," and that court's role was to "interpret the [governing] statute independently" to decide "whether the department's rule is consistent with [that] statute"). Based on that case law, this court concludes that it need not defer to OAR 150-308-0690 to the extent that the source of the rule's authority is ORS 305.100.
(2) ORS 308.655: Application of Springfield principles to determine level of deference.
As to the second statute, this court has found no reported decisions analyzing the degree of deference or weight that applies to rules adopted under ORS 308.655. The court turns to the analytical framework that the Oregon Supreme Court has prescribed for judicial review of administrative agency actions. Under that framework an administrative rule, or the agency's application of the rule or of underlying law, is entitled to deference if the legislature has delegated to the agency the task of completing the legislature's policymaking role, and the rule or action is within the range of discretion that the delegation confers. See Springfield Education Assn. v. School Dist., 290 Or. 217, 223-230, 621 P.2d 547 (1980) (outlining standard of review for agency implementation of "exact," "inexact," and "delegative" statutory terms in both adjudication and rulemaking). More recent decisions have made clear that judicial deference is required only if the legislature's intention was "delegative." Schleiss v. SAIF, 354 Or. 637, 642, 317 P.3d 244 (2013) (no deference for agency interpretation of "inexact" terms). And to determine whether the legislature's intention was delegative, the court must analyze the text and context of the statute, and the court may consider any helpful legislative history, following the steps laid out in State v. Gaines, 346 Or. 160, 171-72, 206 P.3d 1042 (2009). See OR-OSHA v. CBI Services, Inc., 356 Or. 577, 584-85, 588, 341 P.3d 701 (2014). As an aid in deciding whether a statutory term is delegative under the Gaines framework, the Supreme Court has identified the following considerations:
"(1) whether the court has concluded that the term, or one like it, is delegative in another context; (2) whether the term is defined by statute or, on the other hand, susceptible to many different interpretations; (3) whether the term is one that invites a value or policy judgment; and (4) whether other, related provisions suggest a legislative intent that the term be considered a delegation."Penn v. Board of Parole, 365 Or. 607, 628, 451 P.3d 589 (2019) (summarizing OR-OSHA discussion).
This analytical framework has developed primarily under Oregon's Administrative Procedures Act, portions of which do not apply to Defendant or to this court's review of Defendant's actions. See, e.g., ORS 183.315(1). However, the court has found no provisions of the Act that would make the Springfield framework inapplicable to Defendant's rules. Furthermore, the Supreme Court has applied Springfield principles in at least one appeal from this court. See YMCA v. Dept. of Rev., 308 Or. 644, 784 P.2d 1086 (1989). There, citing Springfield, the taxpayer argued that the term "charitable" in Oregon's property tax exemption statutes was a delegative term that required rulemaking to complete the legislature's policymaking role. Because Defendant had not promulgated such a rule before the taxpayer applied for exemption, the taxpayer claimed that Defendant could not deny the claim. See id. at 649. The court rejected the taxpayer's argument on the grounds that the term had been part of the statute since before statehood and had been construed by the court numerous times. See id. at 649-52. The court also implied that, even if "charitable" was a delegative term, the statute authorized Defendant to fulfill its obligation to complete the legislature's general policy decision by applying it adjudicatively to various individual fact situations, rather than by adopting a definition by rule. The court stated: "Even if there were no history of construction and use, we would not agree that the ORS 307.130(1)(a) is an example of delegative legislation requiring an agency to adopt a 'legislative' rule to complete its meaning before it may be applied to an individual case decision." Id. at 652 (footnote omitted; citing Springfield, 290 Or at 228-30). This court has not found a decision where the Supreme Court has suggested that a framework other than the principles of Springfield apply to tax cases.
As in Village at Main Street and Avis, the court in YMCA referred to "legislative" and "interpretive" rules. Id. & n 2. Although those terms have specific meaning under federal administrative law, this court does not understand the Oregon Supreme Court to have adopted the federal standard of deference to Treasury regulations as the standard for review of an Oregon administrative rule. See generally Boris I. Bittker & Lawrence Lokken, Federal Taxation of Income, Estates & Gifts ¶ 110.5 (explaining development of standards for deference to Treasury regulations); Health Net Life Ins. Co. v. Dept. of Rev., 24 OTR 514, 520-21, 537-38 (2021) (applying federal principles of deference to regulations of federal Department of Health and Human Services); see also West House, Inc. v. State Tax Com., 228 Or. 167, 172, 364 P.2d 598 (1961) (rejecting Defendant's argument that its regulations adopted pursuant to 1953 version of ORS 308.205 "became a legislative rule which must be given judicial recognition unless it can be said that the regulation is unreasonable").
Following the Penn considerations, and applying the Gaines framework, the court starts by examining the single sentence comprising the text of ORS 308.655 for any "delegative" terms. Specifically, the court focuses on whether the text indicates an intention to delegate to Defendant the authority to adopt a set of processes that must be followed to determine the real market value of property assessed under ORS 308.505 to 308.674. The court observes first that the text says nothing specifically about determining real market value. The first part of the sentence ("The Department of Revenue may prescribe directions, rules and regulations") is a simple declarative phrase that briefly authorizes Defendant to take a broad set of actions. The second part is a passive phrase with an unidentified actor ("to be followed"). The last part is a phrase specifying the circumstances in which the directions are to be followed ("in answering any requirement" of the central assessment statutes). Overall, the court interprets the text to mean that, to the extent the central assessment statutes impose any requirement, Defendant can direct taxpayers (and its own employees) how to comply. The central assessment statutes say relatively little about how real market value is to be determined, apart from authorizing Defendant to inspect property, consider taxpayer's annual reports and other evidence, and use unit valuation and formulary allocation to determine the value of Oregon-situs property. See ORS 308.545 to 308.555. Based on the text, the court finds it doubtful that the legislature intended ORS 308.655 to delegate authority to Defendant to prescribe the specifics of valuation procedures.
The statutory context, including other central assessment statutes, provides numerous examples of "requirements" as to which guidance from Defendant is obviously helpful. Most requirements are imposed on Defendant itself, such as the procedures for apportioning Oregon value among the counties, and preparing, finalizing and correcting the annual roll. See ORS 308.560 to 308.636. But, in contrast to the laws governing most locally assessed property, ORS 308.524 requires each company subject to central assessment to file an annual tax return, and ORS 308.525 sets forth 16 items that must be disclosed, the last of which is "[a]ny other facts or information the Department of Revenue requires in the form of return prescribed by it." ORS 308.525(16). This context tends to confirm the court's impression that ORS 308.655 is not delegative, but rather is an all-purpose authorization of guidance, serving a function for centrally assessed property that is similar to the function of ORS 305.100 for all state taxes.
Finally, the court looks to the report commissioned by the Oregon legislature in 1905 and delivered in 1906 (1906 Oregon Report), which examined Oregon's property tax system and proposed constitutional amendments and a statutory scheme. See Frederick W. Mulkey, E.B. Seabrook, Wm. J. Lachner, Report of the Board of Commissioners Appointed Under the Provisions of Chapter 90, Law of 1905, for the Purpose of Examining and Reporting on Matters of Assessment and Taxation, Etc. (June 30, 1906) discussed in Jarvill v. City of Eugene, 289 Or. 157, 175-78, 613 P.2d 1 (1980); Level 3 Communications LLC III v. Dept. of Rev., 23 OTR 440, 469-71 (2019); aff'd 368 Or. 303, 490 P.3d 149 (2021). The 1906 Oregon Report includes a proposed statutory provision that reads:
"The board is hereby given the power to prescribe such directions, rules, and regulations to be followed in answering any of the requirements of this section, or as herein authorized, as in its judgment shall be best calculated to insure accuracy and uniformity in reporting the facts."1906 Report at 121. The writers included this sentence in an unnumbered portion at the end of section 8 of their proposed statute, which contains provisions materially similar to the tax return requirements now found in ORS 308.524 and 308.525. Read in that context, the authorization contained in the provision in the 1906 Oregon Report was limited to guidance regarding tax returns ("this section") and to the information to be reported by taxpayers. The 1909 act enacted the first portion of the 1906 provision but broadened the scope from "this section" to "this act," which if read literally would have caused it to apply to all state tax laws, coextensively with what is now ORS 305.100. See Or Laws, 1909, ch 218, § 7(14) ("The board is hereby given the power to prescribe directions, rules and regulations to be followed in answering any of the requirements of this act."). However, in the 1953 codification that created the Oregon Revised Statutes, the legislature narrowed the scope of the provision from all tax laws to the centrally assessed property laws, enacting ORS 308.655 to read: "The State Tax Commission may prescribe directions, rules and regulations to be followed in answering any requirement of ORS 308 505 to 308 660." Former ORS 308.655 (1953); cf. Or Laws 1951, ch 586, § 4(14) ("The commission hereby is given the power to prescribe directions, rules and regulations to be followed in answering any requirement of this Act."). From this statutory development, the court cannot discern an intention to delegate rulemaking authority regarding the application of valuation methods. The court concludes that it need not defer to OAR 150-308-0690 to the extent that the source of the rule's authority is ORS 308.655.
b. Analysis under ORS 308.205
Rather than look to the published sources of statutory authority, both parties frame their arguments on the assumption that the statutory authority for OAR 150-308-0690 is ORS 308.205, which provides:
"(1) Real market value of all property, real and personal, means the amount in cash that could reasonably be expected to be paid by an informed buyer to an informed seller, each acting without compulsion in an arm's-length transaction occurring as of the assessment date for the tax year."
"(2) Real market value in all cases shall be determined by methods and procedures in accordance with rules adopted by the Department of Revenue and in accordance with the following:
"(a) The amount a typical seller would accept or the amount a typical buyer would offer that could reasonably be expected by a seller of property.
"(b) An amount in cash shall be considered the equivalent of a financing method that is typical for a property.
"(c) If the property has no immediate market value, its real market value is the amount of money that would justly compensate the owner for loss of the property.
"(d) If the property is subject to governmental restriction as to use on the assessment date under applicable law or regulation, real market value shall not be based upon sales that reflect for the property a value that the property would have if the use of the property were not subject to the restriction unless adjustments in value are made reflecting the effect of the restrictions."(Emphasis added.) The court assumes, without deciding, that Defendant may rely on the emphasized portion of ORS 308.205 as the source of authority for OAR 150-308-0690 even though Defendant did not refer to ORS 308.205 when adopting the rule. Cf. OAR 150-3080240(2) (specifying "Methods and Procedures for Determining Real Market Value" and citing ORS 308.205 as statutory authority). However, because the relevant case law spans many decades of statutory change, the court briefly sets forth its understanding of how the parts of the statute originated and operate together.
Plaintiff has not challenged the validity of OAR 150-308-0690 on procedural grounds, and the court expresses no view as to whether any procedural error occurred.
Subsection (1) of ORS 308.205 stems primarily from a 1991 act comprising 180 pages; most provisions of the act were devoted to implementing constitutional changes pursuant to Measure 5, a property tax limitations measure enacted by citizen initiative in the prior year. See Or Laws, 1991, ch 459, § 88; Or Const, Art XI, § 11b. Among other things, Measure 5 introduced the term "real market value" in lieu of the prior term "true market value." See Or Const, Art XI, § 11b(2)(a) ("'Real market value' is the minimum amount in cash which could reasonably be expected by an informed seller acting without compulsion, from an informed buyer acting without compulsion, in an 'arm's-length' transaction during the period for which the property is taxed."). The 1991 act essentially replicated the new constitutional term and its definition in subsection (1) of ORS 308.205, replacing the former short definition that defined "true market value" as "market value." See Or Laws, 1991, ch 459, § 88.
The 1997 referendum known as Measure 50 revised the definition of real market value in ways not relevant to this case, retaining the arm's-length standard. See Or Const, Art XI, § 11(11)(a)(A) ("The real market value of property shall be the amount in cash that could reasonably be expected to be paid by an informed buyer to an informed seller, each acting without compulsion in an arm's length transaction occurring as of the assessment date for the tax year, as established by law."). The 1997 legislature amended subsection (1) of ORS 308.205 to conform to the Measure 50 definition, creating the text applicable to this case. See Or Laws 1997, ch 541, § 152 ("Real market value of all property, real and personal, means the amount in cash that could reasonably be expected to be paid by an informed buyer to an informed seller, each acting without compulsion in an arm's length transaction occurring as of the assessment date for the tax year.") (Omitting text deleted by amendment.)
Subsection (2) of ORS 308.205 contains a blend of text added in 1991 and older provisions:
• The introductory flush language, relating to rulemaking by the Department of Revenue, dates substantially to 1953 and 1955. See Or Laws 1953, ch 701, § 1 ("True cash value of all property, whether real or personal, shall be held and taken to mean the amount such property would sell for at a voluntary sale made in the ordinary course of business, under normal conditions, in accordance with rules and regulations promulgated by the State Tax Commission.") (Emphasis added.); Or Laws 1955, ch 691, § 1 ("True cash value of all property, real and personal, means market value as of the assessment date. True cash value in all cases shall be determined by methods and procedures in accordance with rules and regulations promulgated by the State Tax Commission."), § 2 (prescribing effective date of January 1, 1961) (emphasis added).
• Paragraph (a) relates to a "typical" buyer and seller, and paragraph (b) refers to a "financing method." This text was first added by the 1991 act but does not appear to correspond to or implement a specific provision in Measure 5. Paragraph (c) consists substantially of text first added to ORS 308.205 in 1955. See Or Laws, 1955, ch 691, § 1 (adding "With respect to property which has no immediate market value, its true cash value shall be the amount of money that would justly compensate the owner for loss of the property."). Paragraph (d) consists substantially of text first added to ORS 308.205 in 1977. See Or Laws, 1977, ch 423, § 2 (adding: "With respect to property that is subject to governmental restriction as to use on the assessment date under applicable law or regulation, true cash value shall not be based upon sales that reflect for the property a market value that the property would have if the use of the property
were not subject to the restriction unless adjustments in value are made reflecting the effect of the restrictions.").
Before the 1953 act, the statute defining "true cash value" did not refer to rulemaking. See Or Laws 1941, ch 440, § 4 ("True cash value of all property, whether real or personal, shall be held and taken to mean the amount such property would sell for at a voluntary sale made in the ordinary course of business, taking into consideration its earning power and usefulness under normal conditions.").
(1) Portland Canning and Other Cases Addressing Rulemaking Under ORS 308.205 In 1965, four years after the text now substantially found in subsection (2) had become effective, the court construed the statutory reference to rulemaking as follows:
"Clearly the dominant note of the legislation is that, if possible, value is to be ascertained in accordance with market value. While the commission has been given power to make regulations setting forth procedures as to how this may be done, it cannot vary the mandate of the law under this guise."Portland Canning Co. v. Tax Com., 241 Or. 109, 113, 404 P.2d 236 (1965) (emphasis added). Defendant's rule in effect for the tax year at issue in Portland Canning allowed market value to be determined by using "[a]ny one of" the three standard methods (the cost approach, the income approach, or the market approach). Id. at 113 (quoting former Reg Art 8205.1). The court found that a market existed for the majority of the component parts of the taxpayer's food canning plants. See id. at 113-14. Without expressly invalidating the rule's allowance of a single valuation approach, the court concluded that defendant erred by ignoring the existence of a market for the canning equipment in favor of relying solely on the cost approach. See id. at 113 ("The commission has no power to permit the evaluation of the property by the exclusive means of the cost approach to determine the value to the owner when a market in fact exists.").
Defendant later added to the rule the requirement that "[r]eal property shall be valued through the market data approach, cost approach and income approach," retaining, however, the provision that "[a]ny one of the three approaches to value, or all of them, or a combination of approaches, may finally be used by the appraiser in making an estimate of market value, depending upon the circumstances." R308.205-(B) Personal Property Former Rule 308.205-(A) (1969) (emphasis added); cf. OAR 150-308-0240(2)(a) (2018) ("For the valuation of real property all three approaches, sales comparison approach, cost approach, and income approach, must be considered. For a particular property, it may be that not all three approaches are applicable. However, each approach must be investigated for its merit in each appraisal.").
Some 20 years later, the Oregon Supreme Court reached a similar conclusion in Alsea Veneer, Inc. v. Dept. of Rev., 297 Or. 512, 516, 687 P.2d 137 (1984). At that time, the rule provided in part:
"'A "unit of property" is the item, structure, plant or integrated complex as it physically exists on the assessment date as real or personal property. The market value of a unit of property is not ascertained from the market price of its component materials, such as wood, glass, concrete, pipe, wire, furnaces, elevators, etc., each priced separately as an item of personal property, without regard to its being integrated into the total unit. Similarly, in the appraisal of industrial properties the fixed machinery and equipment comprise an integral part of a manufacturing plant and as such is usually real property. The market value test of such a plant is predicated upon the sales of comparable plants. No market test of unit property exists when there are no sales of comparable property at times and places which are reasonably relevant to the appraisal date and subject property under the existing circumstances.'"Former OAR 150-308.205-(A) (emphases added). In Alsea Veneer, Defendant argued that the "unit of property" provision in its rule precluded consideration of the market value of component parts of the taxpayer's veneer plant. The court rejected that argument, pointing out that Defendant had "used the component parts method" in its own valuation of the plant. Id. at 517. The court implicitly criticized Defendant for incorrectly relying on the "unit of property" provision in its rule to limit the field of comparable sales, concluding that the point of the rule provision was instead to "assist in the determination of whether property is personal or real * * *." Id. As to the portion of the rule that required sales of "comparable plants" as a predicate for using the market test, the court stated: "We cannot agree with the Department that this rule prevents the Tax Court from considering plaintiff's method of evaluating the machinery and equipment * * * based on an appraisal of the component parts," which included auction sales of comparable equipment. Id. However, as in Portland Canning, the court stopped short of expressly declaring portions of the rule invalid or explaining what level of deference, if any, the court was required to apply to it.
A Supreme Court opinion from 1986 commented in passing on Defendant's rulemaking authority under ORS 308.205, stating: "The department may prescribe a method of valuation by rules, ORS 306.120, ORS 308.205, but in the absence of a single prescribed method, the proper method is itself a matter to be resolved by evidence, if it is disputed." Lewis v. Dept. of Rev., 302 Or. 289, 292, 728 P.2d 1378 (1986). This court does not read the quoted passage as implying any required level of judicial deference to Defendant's rules. Rather, this court reads the Supreme Court as refuting the taxpayer's assertion that "'[a]ll three [methods] must be used'" to determine value. Id. at 292. Read together with the rule in effect at that time, the court's comment appears to summarize the rule's requirement that all three approaches be considered, while allowing for the possibility that only one or two approaches may be applied, depending on the facts. See id. at 293 ("What the Department of Revenue's rule states is not that the 'market data approach, cost approach, and income approach' must all be employed, but that '[a]ny one of the three approaches to value may finally be used * * *.'" (quoting former OAR 150-308.205-(A)(2) (1981)).
In 1999, the Oregon Supreme Court relied heavily on Defendant's administrative rules under ORS 308.205 in determining the real market value of centrally assessed airline property. Delta Air Lines, Inc. v. Dept. of Rev., 328 Or. 596, 984 P.2d 836 (1999). By that time, Defendant had promulgated a rule that adopted by reference an earlier version of the WSATA Handbook. See id. at 609-10 (reprinting former OAR 150-308.205(B) (1999)). The same rule also contained text specifically addressing one of the issues in the case--whether and how to adjust the value of airline property to account for the fact that taxpayer was merely the lessee of some of the aircraft it used. See id. at 609-11. The Oregon Supreme Court applied both parts of the rule, stating first that the rule "answer[ed] the question" whether to make a leased-equipment adjustment by providing that "'[t]he reconciled unit value estimate shall be adjusted to include taxable property not included in the unit, i.e., full value of lessors' interest in equipment leased from others or to exclude nontaxable property included in the unit.'" Id. at 609-10 (quoting former OAR 150-308.205-(B)(7)(b)). As to several other issues, the court described the WSATA Handbook as "pivotal." Id. at 611. Citing these references, Defendant in this case urges this court to simply apply the WSATA Handbook. However, the court in Delta did not purport to articulate any particular standard of deference with respect to the WSATA Handbook. The court did not need to do so, because "Delta d[id] not challenge the validity of the administrative rule * * *." Id. at 610 n 8 (emphasis added); cf. Hewlett-Packard Co. v. Benton County Assessor, 357 Or. 598, 600-03, 356 P.3d 70 (2015) (taxpayer challenged application, but not validity, of rules under ORS 308.205(2); stating that an "appraiser must follow" "complex scheme of administrative rules outlining the methods and procedures that the department requires").
The court apparently did not interpret Delta's challenge to the making of any leased-equipment adjustment as a challenge to the validity of the rule that declared that such an adjustment was required.
(2) Application of Springfield and Portland Canning assuming authority under ORS 308.205
In summary, none of the decisions the court has found under ORS 308.205 declares a standard of review of rules under that section. The court returns to the analytical framework under Springfield Education Association. Neither party sought to apply the framework at trial or as part of extensive post-trial briefing; it therefore falls to the court to decide which terms to analyze. The court starts with the flush language in subsection (2), which implicitly directs Defendant to adopt rules containing "methods and procedures" to "determine[ ]" "[r]eal market value." The court finds that these terms are not in dispute; rather, the parties disagree about the requirement that real market value be determined "in accordance with" Defendant's rules. If the court were writing on a blank slate, its task under Springfield would be to determine whether the phrase "in accordance with" is delegative, in the sense of conferring a "range of discretion" upon Defendant, and if so, whether Defendant's adoption of the WSATA Handbook, to the exclusion of any other method of determining real market value, exceeded that range. As to this point, however, the court views the decision in Portland Canning as binding. The court again reprints the key language holding that Defendant's procedures are subordinate to the concept of "market value" (now real market value):
"Clearly the dominant note of the legislation is that, if possible, value is to be ascertained in accordance with market value. While [Defendant] has been given power to make regulations setting forth procedures as to how this may be done, it cannot vary the mandate of the law under this guise."241 Or at 113. No subsequent Supreme Court opinion overturns the holding in Portland Canning.
The court concludes that, whatever discretion might be delegated to Defendant to prescribe valuation methods and procedures for centrally assessed property, under Portland Canning that discretion does not include the authority to compel the use of methods and procedures that fail to result in real market value in a particular case. Stated positively, a taxpayer remains free to argue that a value determined in accordance with Defendant's rules is inconsistent with the definition of real market value in subsection (1) of ORS 308.205, and the court is not bound to accept a value determined under Defendant's rules if the court finds that another value, determined under methods or procedures not in accordance with Defendant's rules, is correct. The court concludes that it need not defer to OAR 150-308-0690 with respect to methods and procedures of determining the real market value of Plaintiff's system, to the extent that the source of the rule's authority is ORS 308.205(2).
This conclusion does not render the WSATA Handbook, or any other rules, meaningless. Importantly for the property tax system as a whole, Defendant's rules under ORS 308.205 bind not only Defendant, but also the assessors in all counties, who are charged, under Defendant's supervision, with determining value for all real and personal property other than centrally assessed or large industrial property. See ORS 308.215(1)(a)(F), (b)(B) (county assessor to record real market value on roll); ORS 306.115 (1) (Defendant to exercise "general supervision and control" over property tax system); ORS 306.120 (1) (Defendant to issue regulations to assessors "as to the methods best calculated to secure uniformity according to law"). Therefore, a significant result of the flush language of subsection (2) is to enable Defendant to ensure compliance with the constitutional mandate that "[a]ll taxes shall be levied and collected under general laws operating uniformly throughout the State." Or Const, Art IX, § 1; see also Or Const, Art I, § 32; see generally C. W. Macy, Some Legal and Administrative Aspects of the Property Tax in Oregon, 33 Or L Rev 179, 182 (1954) (discussing legislative development of Or Laws 1953, ch 701 and recounting "gross inequalities in the appraisal of property for taxation purposes" within and among counties, caused by "lack of a clear-cut legal method of determining values for taxation purposes.").
2. OAR 150-308-0590: Defendant's Allocation Formula for Electricity Companies
The parties also dispute the degree of deference the court should afford to Defendant's separate administrative rule that prescribes a formula to allocate the system value used in an electricity business to Oregon. Defendant argues that this rule, OAR 150-308-0590, "has the full force of law." (Def's Response at 23.) Plaintiff asks the court to apply a different formula, arguing that the rule is inconsistent with the allocation statute, ORS 308.550. As with OAR 150308-0690, neither party has presented the court with an analysis under the Springfield/Penn framework. And once again, three statutes are in contention as authority for the rule. The published rule refers to ORS 305.100 and ORS 308.550. It also is conceivable that OAR 150308-0590 could be considered to prescribe "methods and procedures" for determining real market value under ORS 308.205(2), and thus might be covered by the Portland Canning rule discussed above. Ultimately, the court finds it unnecessary to reach the level of deference to this rule. As discussed near the end of this opinion, even if Plaintiff were to show that the rule is entitled to no deference as a matter of administrative law, the court concludes that Plaintiff has not shown that Defendant's allocation results in a "gross distortion" as required under case law.
The court also notes that the formula in OAR 150-308-0590 is materially identical to a formula for electricity business property set forth in the WSATA Handbook at page XIV-23, but Defendant apparently does not view the WSATA Handbook's version as incorporated by reference under OAR 150-308-0690. This may be because the formula is not part of the main text of the WSATA Handbook but is appended as part of a 1960 committee report in which the committee "recommend[s]" that WSATA member states adopt and use allocation formulas that the committee developed for various businesses. See WSATA Handbook at XIV-6. The body of the WSATA Handbook does not specifically endorse any of the formulas in the appendix, but instead states that "[r]ecommended formulas will be contained in a separate publication." WSATA Handbook at VIII-1. The parties have not referred the court to any such separate publication, assuming it exists.
B. Burden of Proof
Each party argued at trial for an Oregon real market value different from the value shown on the central assessment roll prepared under ORS 308.560 and certified under ORS 308.610. Cf., e.g., Level 3 Communications LLC III v. Dept. of Rev, 23 OTR 440, 442-43 (2019) (describing parties' positions relative to roll values). The parties present differing views as to the burden of proof, which is set forth in ORS 305.427 (2021):
"In all proceedings before the judge or a magistrate of the tax court and upon appeal therefrom, a preponderance of the evidence shall suffice to sustain the burden of proof. The burden of proof shall fall upon the party seeking affirmative relief and the burden of going forward with the evidence shall shift as in other civil litigation."
Plaintiff acknowledges that it "seek[s] affirmative relief" in the form of a reduction of the real market value shown on the roll, and that it bears the burden of proving any such reduction. (See Ptf's Post-Trial Br at 4.) Defendant agrees with that point but disagrees with Plaintiff's next assertion, that Defendant bears the burden of proving any "'increase compared to the roll RMVs.'" (Id. n 3 (quoting Level 3, 23 OTR at 443 (2019).) Defendant relies on the following passage from a 1989 central assessment opinion by the Oregon Supreme Court:
"Pacific next argues that the Tax Court erred in not requiring the Department to carry the burden of proving that its appraisal, which suggested a value significantly greater than that found by the Director of the Department of Revenue at an earlier, administrative step in this case, should be adopted. Pacific cites no binding authority for this proposition, which contradicts the basic idea that the burden in an appeal by a taxpayer to the Tax Court is on the taxpayer.ORS 305.427." PP&L v. Dept. of Rev., 308 Or. 49, 54-55, 775 P.2d 303 (1989). (See Def's Opening Post-Trial Br at 1-2.) In quoting this passage, Defendant omitted the remaining sentences in the paragraph:
"The Tax Court did not err. To the extent that Pacific is also asking that this court impose such a burden on the Department in this court, the request is denied. The burden is the same 'upon appeal' from the Tax Court. ORS 305.427."PP&L, 308 Or at 55. On the other hand, Plaintiff cites the following passage from another central assessment opinion decided three years later:
"With respect to the case as it is presented in this court, we note only that, as to the claims that UP makes that would require a modification of the decision of the Tax Court, UP has the burden of proof by a preponderance of the evidence. We
have required UP to meet that burden throughout our review of its evidence. As to the matters concerning which the Department asks this court to change the decision of the Tax Court or as to which it cross-assigns error, it is the Department that shoulders the burden of proof. See PP&L v. Dept. of Rev., supra, 308 Or at 54-55, (discussing burden of proof)."Union Pacific Railroad v. Dept. of Rev., 315 Or. 11, 17, 843 P.2d 864 (1992) (referring also to the same passage quoted above in PP&L).
This court seeks to reconcile the above statements in PP&L and Union Pacific. Under PP&L, the Supreme Court concluded that Defendant did not bear the burden of proof before this court as to an increase above the value Defendant's own director had set and from which the taxpayer appealed. But under Union Pacific, Defendant did bear the burden of proof, on appeal to the Supreme Court, as to any value higher than the value determined by this court. Yet, as the omitted sentence in PP&L declares, the same law assigning the burden applied in both courts, as it does today. See ORS 305.427 (1989) (referring to "all proceedings before the tax court and on appeal therefrom") (emphasis added); cf. ORS 305.427 (2021) (referring to "all proceedings before the judge or a magistrate of the tax court and on appeal therefrom) (emphasis added).
Prior to 1995 the standard of review before the Supreme Court was de novo; the Supreme Court's task was to sit as a trier of fact and apply all relevant law, including ORS 305.427, to the facts based on the record before this court. See Delta Air Lines, Inc. v. Dept. of Rev., 328 Or. 596, 600-03, 984 P.2d 836 (1999) (discussing former standard); Or Laws 1995, ch 650, § 25 (limiting supreme court's scope of review to "errors or questions of law or lack of substantial evidence in the record to support the tax court's decision or order").
Because both PP&L and Union Pacific predate the Supreme Court's announcement of the current statutory interpretation framework in State v. Gaines, the court now applies that framework to ORS 305.427 in an effort to discern principles that would harmonize the two opinions for purposes of this case. The statute, enacted in 1965, places the burden on the "party seeking affirmative relief." Or Laws 1965, ch 6, § 5 (SB 4). Because the statute relates to legal proceedings, the court turns initially to a contemporaneous edition of Black's Law Dictionary for the following technical legal definition of Affirmative Relief:
"Relief, benefit, or compensation which may be due and granted to defendant. Garner v. Hannah, 6 Duer (N. Y.) 262. Relief for which defendant might maintain an action independently of plaintiff's claim and on which he might proceed to recovery, although plaintiff abandoned his cause of action or failed to establish it. Southwestern Surety Ins. Co. v. Walser, 77 Okl 240, 188 P 335, 336."Black's Law Dictionary 75 (4th ed 1951). Comcast Corp. v. Dept. of Rev., 356 Or. 282, 296, 337 P.3d 768 (2014) ("[W]hen a term is a legal one, we look to its 'established legal meaning' as revealed by * * * legal dictionaries."). This definition indicates that a party proceeding as a defendant could be in a position of seeking affirmative relief, at least if that party made a counterclaim.
Although the definition describes the party seeking affirmative relief as the "defendant," the cases cited do not indicate that affirmative relief is unavailable to a plaintiff. Garner v. Hannah, 13 NY Super Ct 262, 273, 6 Duer 262 (1857) (Defining "affirmative relief" in terms of the defendant due to statute case was brought under); Southwestern Surety Ins. Co. v. Board of Com'rs of Coal County, 77 Okla 137, 187 P 467 (1920) (discussing whether defendant's reply constituted "affirmative relief").
The court proceeds to consider statutory context, starting with other procedural statutes in place in 1965. The only contemporaneous use of the full phrase "affirmative relief" in the Oregon Revised Statutes provided:
"If a counterclaim established at the trial exceeds the plaintiff's demand so established, judgment for the defendant shall be given for the excess; or if it appears that the defendant is entitled to any other affirmative relief, judgment shall be given accordingly."Former ORS 18.100 (1963) (emphasis added), repealed by Or Laws 1981, ch 48, § 1. This statute can be read to mean that affirmative relief included "relief, benefit or compensation" available to a defendant regardless whether the defendant made a counterclaim.
Indeed, case law at the time indicated that, depending on the circumstances, plaintiffs and defendants could seek affirmative relief, and whichever party did so had the burden of proof. See, e.g., Chance v. Carter, 81 Or. 229, 240, 158 P 947 (1916) (concluding that statute enabled "the Defendant, upon proving his claim, not only to defeat the action of the plaintiff, but also to secure affirmative relief") (emphasis added); Hanna v. Hope, 86 Or. 303, 310, 168 P 618 (1917) ("The counterclaim upon which a defendant may have affirmative relief in an equity suit must contain matters of equitable cognizance."); Comegys v. Hendricks, 55 Or. 533, 535, 106 P 1016 (1910) ("The plaintiff did not offer any evidence in support of the controverted question, and thereby failed to substantiate her right to any affirmative relief.").
A statute perhaps more closely on point was ORS 41.210 (1963), which provided:
The legislature repealed ORS 41.210 (1963) in 1981, essentially moving the phrase to ORS 40.115 (OEC 307). See Or Laws 1981, ch 892, § 98 (repealing ORS 41.210); ORS 40.115 ("The burden of producing evidence as to a particular issue is on the party against whom a finding on the issue would be required in the absence of further evidence.").
"The party having the affirmative of the issue shall produce the evidence to prove it. Therefore, the burden of proof lies on the party who would be defeated if no evidence were given on either side."
Some three months before ORS 305.427 was enacted, this court applied ORS 41.210 (1963) in its first suit filed by a county assessor, who sought to increase the value of locally assessed property. Strawn v. Commission, 1 OTR 98, 120, 149-50 (1963). Following the procedure in place at that time, the taxpayer initially contested the assessor's valuation in an administrative appeal before Defendant's predecessor, the State Tax Commission. The Commission reduced the valuation "very substantially." Id. at 121. The assessor appealed to this court. The taxpayer counterclaimed for even lower values than determined by the Commission. The court devoted some six pages to discussing the burden of proof, concluding that the assessor had "the affirmative" as to whether the assessed value should be raised, and the taxpayer had the affirmative as to any lowering of the assessment. Id. at 155 ("The burden of proof remains fixed with the party, or in this case, parties, seeking affirmative relief[.]"
With this context in mind, the court returns to the text of ORS 305.427. Notably, the text does not say that the burden of proof shall fall upon the "plaintiff," much less upon the "taxpayer." By using instead the less specific phrase "party seeking affirmative relief," the statute seems to contemplate the possibility that either party might seek affirmative relief, in line with the discussion in Strawn and the longstanding non-tax cases cited above.
The court has found nothing in the legislative history of SB 4 that sheds light on the burden of proof under ORS 305.427. SB 4 was requested by the Interim Committee on Taxation and by Senator Ben Musa, a proponent of the Oregon Tax Court Act enacted four years earlier. See House and Senate Journal, Regular Session, 1965, S-4 (listing Interim Committee on Taxation and Senator Musa as requesting bill); Or Laws 1961, ch 533 (creating Oregon Tax Court). SB 4 amended a half-dozen provisions of the Oregon Tax Court Act and made conforming amendments referring to this court in other tax and procedural statutes. At several points, legislators or witnesses referred to the text that became ORS 305.427 as assigning the burden of proof to "the taxpayer." For example, the minutes of a Senate Revenue Committee hearing state that Senator Anthony Yturri supported the bill's placement of the "burden on the taxpayer," consistent with "civil cases[, in which] it is the plaintiff, the one who asserts the claim, who has the burden." Minutes, Senate Committee on Taxation, Jan 20, 1965 at 2. Carlisle Roberts, then-counsel to the Commission, "discussed the importance of the word 'preponderance' * * *, which he said puts more of the burden of proof on the taxpayer." Minutes, Senate Committee on Taxation, Jan 18, 1965 at 2; see also id. ("[Eugene Feltz, of the Oregon State Bar] thought the burden of proof is already with the taxpayer in case of an appeal."); Minutes, Senate Committee on Taxation, Jan 20, 1965 at 1 ("[Judge Edward Howell, Oregon Tax Court] agreed '100 percent' that the burden of proof should be on the taxpayer, as he is seeking affirmative relief."). This court considers these statements as shorthand, reflecting the fact that taxpayers are plaintiffs in nearly all cases, and that a taxing authority's assertion of a value higher than shown on the roll is a nuance that does not commonly arise or at least is not commonly discussed in reported cases. The minutes indicate that the burden of proof provision was not controversial, and that the discussion quickly moved to the next passage in the bill, a contested provision that would have imposed a presumption in favor of the initially assessed value; the Senate Revenue Committee ultimately voted to reject the statutory presumption.
Subsection (2) of Section 4 of SB 4 as introduced provided:
"The tax court shall recognize a rebuttable presumption of validity and correctness to have attached to the final determination of tax or assessed value from which and appeal or other proceeding may be taken to an authority independent of the authority making such first determination. The presumption shall remain with such first determination until the tax liability at issue is finally determined."This court continued to apply a judicial presumption of assessment validity until the Oregon Supreme Court abolished it in 1972. See J. R. Widmer, Inc. v. Dept. of Rev., 261 Or. 371, 378, 494 P.2d 854 (1972) ("The Tax Court (ORS 305.425(1)) and this court (ORS 305.445) review de novo without any presumption as to the correctness of the assessor's valuation.").
Against this backdrop of the text and context of ORS 305.427, this court now returns to the apparently conflicting passages in PP&L and Union Pacific. The passage in PP&L describes the value from which the taxpayer appealed to this court as a value "found by the Director of the Department of Revenue at an earlier, administrative step in this case." 308 Or at 54-55. As of the 1989 PP&L opinion, the setting of the central assessment roll did not have the finality that exists under current law. This court recently concluded that, before 2007, Defendant had discretionary authority to correct a valuation error on a central assessment roll for up to three years. See ORS 306.115 (2005); D.E. Shaw Renewable Investments, LLC v. Dept. of Rev., ___ OTR ___ (Apr 25, 2022) (slip op at 14-15, 30) (appeal pending) (concluding that ORS 306.115 grants discretionary authority for department to correct valuation errors, but is limited by ORS 308.624, enacted in 2007). The 2007 legislature curtailed that authority with respect to centrally assessed property by enacting what is now ORS 308.624(4): "For purposes of this section, the director may not correct an error in valuation judgment [on the central assessment roll] that is an error in the department's opinion of the value of property." Consistent with the breadth of Defendant's authority in 1989 to change roll values unilaterally, this court reads the statement in PP&L to mean that Defendant was not seeking affirmative relief in this court, even when Defendant sought an increase above the roll value. From there, it follows that the plaintiff taxpayer in PP&L was the only party bearing the burden of proof in this court.
By contrast, even before 1989, a local county assessor generally was prohibited from correcting a valuation error on a prior roll. See ORS 311.205 (1) (1987) ("The officer may not correct an error in valuation judgment. Such errors are those where the assessor would arrive at a different opinion of value after the roll has been returned to the assessor by the board.").
Under post-2007 law, however, Defendant may correct a valuation error on the central assessment roll only if ordered by this court or by the Supreme Court on appeal. See ORS 308.624(3)(b). Reading PP&L in light of changes in the procedural law, this court concludes that the rule articulated in Union Pacific applies today in valuation disputes before this court: As to changes to the roll that a taxpayer requests for its benefit, including a valuation lower than that shown on the roll, the taxpayer seeks affirmative relief and therefore bears the burden of proof. As to a valuation higher than that shown on the roll, Defendant seeks affirmative relief and bears the burden of proof.
III. REAL MARKET VALUE ANALYSIS
The court now seeks to apply the foregoing understanding of applicable law to determine the real market value of Plaintiff's property, based on the evidence presented at trial. As discussed above, the court first determines the real market value of the "system" comprising all of Plaintiff's property used in its electricity business, without regard to location, before determining the percentage of that value to allocate to Oregon. To a great extent, the issue is fact intensive and involves "competing testimony of appraisal experts, including the credibility and persuasiveness of those experts." Hewlett-Packard, 357 Or at 609.
A. Cost Approach
Each party's appraiser gave approximately 20 percent weight to his value indicator derived from a cost approach. (Ptf's Pretrial Br at 4.) Plaintiff's appraiser, Thomas K. Tegarden, determined a cost indicator of $15,838,880,747. Defendant's appraiser, Brent Eyre, determined a cost indicator of $20,735,886,213. (See Ptf's Ex 1 at 29; Def's Ex A at 23.) Tegarden and Eyre each started with the "net book value" Plaintiff reported on its annual "FERC 1," a form Plaintiff filed with the Federal Energy Regulatory Commission (FERC). (E.g., Ptf's Ex 2 at 1983 (FERC 1 for Q4 2019).) Net book value for FERC purposes is not identical with the locally determined "rate base" on which Plaintiff is allowed to earn a reasonable return, but local regulators refer to net book value when setting the rate base. See WSATA Handbook at II-8 to II-9. Net book value for FERC purposes is based on a variation of the cost approach to valuation known as "historical cost less depreciation" (HCLD), which Defendant uses only for centrally assessed property. See OAR 150-308-0240(2)(f) (for other property, cost approach must use "reproduction, replacement or used equipment technique"); WSATA Handbook at I-4 (HCLD may be a "meaningful" indicator if company is rate base regulated). The parties agree that net book value is not likely to be equal to real market value, but they disagree about the extent of any adjustments that must be made in order to make the HCLD approach useful as an indicator of real market value. Each party started with the historical cost of $28,843,430,112 taken from Plaintiff's FERC 1 form for 2019, and each deducted $10,870,776,722 in depreciation, an entry also taken from the same FERC 1 form. (See Ptf's Ex 1 at 29; Def's Ex A at 23; cf. Ptf's Ex 2 at 2065 (FERC 1).)
Plaintiff started with a net book value of $28,661,071,569, citing a different page and line number of the same FERC 1 cited by Defendant. (Compare Ptf's Ex 1 at 29 and Ptf's Ex 2 at 2065 & 4807 with Def's Ex A at 23.) The court considers the difference immaterial for purposes of this decision.
1. Plaintiff's Method: Additional Deduction for Economic Obsolescence
Nearly all of the difference between the parties' cost method indicators arises because Tegarden took an additional deduction for "obsolescence" in the amount of $4,428,401,079, while Eyre did not. (Ptf's Ex 1 at 23; see Ptf's's Post-Trial Br at 59.) Plaintiff argues that an approach that starts with historical cost is valid only if it allows deduction of all three of the widely recognized forms of depreciation: physical depreciation, functional obsolescence, and economic obsolescence. (See Ptf's's Post-Trial Br at 52; Ptf's Ex 8 at 1 (citing WSATA Handbook at II-6.).) Tegarden testified that the deduction for depreciation on Plaintiff's FERC filings is too low because it does not adequately account for a type of economic obsolescence peculiar to regulated utilities, namely, the utility's inability to earn a return on property acquired or paid for with the following:
The parties refer to depreciation allowed for FERC filing purposes as "accounting" or "regulatory" depreciation.
(1) Deferred Income Taxes (DIT).
Broadly speaking, DIT is the difference between (1) the amount that ratemaking authorities consider Plaintiff to owe in income taxes for a particular year and (2) the amount Plaintiff actually owes for that year. The amount Plaintiff actually owes typically is less because Congress has repeatedly increased the annual amount otherwise allowed as a deduction from gross income to depreciate or amortize the cost of acquired property over the property's useful life. See generally Bittker &Lokken, ¶¶ 23.1-9 (discussing "accelerated" depreciation, "bonus" depreciation, and "expensing" of purchases). By allowing higher
depreciation deductions right away, Congress reduces taxable income, and thus income tax liability, with the goal of encouraging businesses to use the additional cash to buy machinery, equipment, and other property, thereby stimulating the economy. Accelerated depreciation provides additional cash flow in the early years after property is acquired, but the tax liability is, in theory, merely "deferred" and is recouped in later years when the property is fully depreciated and the taxpayer must cease to claim deductions sooner than it otherwise would have. Regulatory authorities do not allow Plaintiff to earn a return on property purchased with the cash saved through DIT, but the acquired property used in Plaintiff's electricity business contributes to the system value, an apportioned share of which is deemed to be the real market value of Plaintiff's property in Oregon.
(2) Income Tax Credits (ITC).
As with DIT, property attributable to certain income tax credits is excluded from the rate base but contributes to real market value for property tax purposes to the extent used in Plaintiff's electricity business.
(3) Contributions in Aid of Construction (CIAC).
Money or property contributed by particular utility customers (such as large industrial plants) to fund extensions of electricity infrastructure that primarily benefit those customers is likewise outside the rate base. However, the value of the property contributes to the real market value of Plaintiff's system to the extent that Plaintiff uses the property.(See Transcript at 1293-94; Ptf's Ex 21 at 32; Ptf's Post-Trial Br at 58-59.) See generally PP&L v. Dept. of Rev., 308 Or. 49, 52-53, 775 P.2d 303 (1989) (describing above categories of property). Tegarden's sources included an email exchange with an economist employed by FERC, who stated that "accounting depreciation is a cost allocation process, not a valuation process. Consequently, the net book value of a plant asset may differ significantly from its market value." (Ptf's Ex 10 at 1.)
2. Does Plaintiff's method fail as a matter of law?
Defendant acknowledges the general point that net book value equals real market value "'only by coincidence.'" (Def's Opening Post-Trial Br at 4 (quoting WSATA Handbook at II-10).) However, Defendant argues that an additional deduction is unavailable both as a matter of law and as a factual matter. As to the first point, Defendant argues that a deduction for economic obsolescence would "violat[e] the WSATA Handbook guidelines * * *." (Def's Opening PostTrial Br at 7.) However, based on the court's reasoning above, violation of the WSATA Handbook guidelines is not a basis to reject an additional deduction under the cost approach if the taxpayer can persuade the court that a value determined pursuant to the WSATA Handbook is not the real market value of the property. Plaintiff's proffered method does not fail as a matter of law.
The WSATA Handbook states:
"HCLD cost indicators are generally not adjusted further to account for appreciation or depreciation. A deduction from HCLD for obsolescence is just as inconsistent as adding value to HCLD because some of the utility's property has increased in value since it was acquired, or because the utility's earnings are extraordinarily high for some reason (e.g., lax regulatory oversight). The practice of not adjusting HCLD for perceived obsolescence does not mean that obsolescence has not been considered and measured, since as noted previously, regulatory depreciation should, in theory, reflect all forms of obsolescence. The degree to which regulatory depreciation reflects an accurate estimate of market depreciation for a particular property is taken into account when reconciling the value indicators."WSATA Handbook at II-12.
3. Does Plaintiff's "income shortfall" method have adequate factual support?
The court thus turns to the parties' factual arguments. Notwithstanding Defendant's arguments that regulatory depreciation adequately accounts for all obsolescence, the court finds it clear that the historical cost of property included on the FERC 1 form, less the FERC subtraction for accounting depreciation, may not accurately reflect the real market value used in a centrally assessed business. The question, then, is whether Tegarden's proposed adjustment accurately translates net book value into real market value. Tegarden offered an "income shortfall" computation that is based on capitalizing the "income * * * loss attributable to the negative influence," in this case, "governmental regulation * * * [that] affects what you can do with the property, how much you can earn on it * * *." (Transcript at 240-43.) Tegarden computed the shortfall by comparing Plaintiff to similarly situated companies that are not restricted from earning income on DIT, ITC and CIAC, concluding that Plaintiff earned a rate of return 1.65 percent below the return investors required as of the assessment date. (See Ptf's Ex 1 at 31.)
As an example of the potential mismatch between historical cost less regulatory depreciation, on the one hand, and the real market value of property (whether or not paid for with cash savings from DIT, ITC and CIAC) on the other hand, Plaintiff's tax director Norman Ross testified that regulatory depreciation of a coal-fired power plant in Utah continued into the tax year at issue (2020-21) even though the plant was decommissioned, dismantled, and reclaimed in 2017. (See Ptf's Ex 9; Transcript at 47.)
By "governmental regulation," Tegarden primarily meant DIT and ITC. (See Ptf's Ex 1 at 33 ("The primary cause for the lower-than-adequate rate of return on all property is the regulatory commissions' specific exclusion of certain properties from the rate base, i.e. those properties financed by funds provided by the deferral of federal income taxes. Simply stated, the regulatory agencies will not allow PacifiCorp to earn a return on those properties which it has purchased using funds provided by deferring the payment of federal and state income taxes.").)
At one point in briefing, Plaintiff seems to ask the court to simply subtract the CIAC amount shown on Eyre's cost approach calculations ($366,356,390, shown on Def's Ex A at 23) from the cost indicator. (See Ptf's Response at 41 ("CIAC amounts should be valued at zero.") Yet Plaintiff also purports to account for CIAC in its income shortfall calculation. (See Ptf's's Post-Trial Br at 63.) To do both would amount to "double dipping" for Plaintiff. The court assumes for purposes of analyzing the cost approach that Plaintiff advocates for accounting for CIAC solely in its income shortfall calculation.
By Tegarden's own description, an income shortfall computation requires (1) proof of the existence of a "negative influence" and (2) proof of a causal relationship between that influence and diminution of Plaintiff's income. In this case, Defendant does not contest that Plaintiff had some amount of DIT, ITC and CIAC, and some of Plaintiff's financial documents admitted into evidence refer to various items of DIT, ITC or CIAC. (See, e.g., Ptf's Ex 2 (Rocky Mountain Power report to Wyoming Public Service Commission) at 3, 15, 26-28).) However, none of Plaintiff's witnesses attempted to identify or quantify the relevant amount of any of those items as of a relevant time; nor did the computation establish the amount of DIT, ITC or CIAC of the comparator companies. (But see Transcript at 666 (Ptf's cross-examination of Brian Conway) (noting "accumulated deferred income taxes 4 billion"); (Ptf's Ex 5 at 11 (excerpt from Ptf's Annual Report to Def, entitled "Results of Operations Summary * * * December 2018 2017 Protocol").) It is true, as the Oregon Supreme Court has found, that a mere statement of these amounts, without more, would not likely suffice to support an adjustment to net book value.On the other hand, without an understanding of those amounts, in relation to Plaintiff's assets and in comparison with the data of the comparator companies, the court finds that Plaintiff's income shortfall computation has inadequate evidentiary foundation and becomes an abstract exercise that lacks important context. The court finds that the preponderance of the evidence does not support Plaintiff's proffered economic obsolescence adjustment to net book value.
Although not precedential for its valuation methodology (see Delta Air Lines, Inc. v. Dept. of Rev., 328 Or. 596, 615 n 11, 984 P.2d 836 (1999)), the Oregon Supreme Court's rejection in PP&L of Plaintiff's attempt to simply subtract DIT and ITC amounts from its cost indicator of value is persuasive, at least on the ground that simply subtracting those amounts would amount to "a portion of the tangible assets of Pacific inappropriately being exempted from taxation." PP&L, 308 Or at 57. (To the extent that the court in PP&L may have determined that a subtraction for DIT and ITC is unwarranted because the regulatory restrictions on earning disappear in the hands of a new buyer, this court notes that the authors of Defendant's study, Ben Johnson Associates, Inc., appear to disagree. See id. ("it seems clear to us that a willing buyer of the plant and equipment would be agreeable to paying a figure close to HCLD, because the buyer could then earn off all that expenditure") (emphasis added). (See Def's Ex D at 14.))
Defendant criticizes Plaintiff's income shortfall approach for other reasons: the WSATA Handbook discourages its use, and the Oregon Supreme Court rejected it as duplicative of an income approach in Delta Air Lines, Inc. v. Dept. of Rev., 328 Or. 596, 618-619, 984 P.2d 836 (1999), and United Telephone Co., Inc. v. Dept. of Rev., 307 Or. 428, 433, 770 P.2d 43 (1989). (See, e.g., Def's Opening Post-Trial Br at 7.) Plaintiff argues that its income shortfall computation is distinguishable and does not suffer from the infirmities identified in prior cases. This court declines to reject all variations of an income shortfall approach per se. The court reiterates that the Supreme Court in Delta and United Telephone took pains to emphasize that the opinions are not precedential as to value methodology. The court concludes only that it cannot evaluate the viability of Plaintiff's income shortfall approach without the foundation described above.
4. Defendant's Cost Analysis
The court turns to Defendant's analysis applying a cost approach. Eyre essentially applied the HCLD method described in the WSATA Handbook, which applies depreciation and other data from the FERC 1 form and resulted in a value indicator very close to Plaintiff's FERC net book value. (See WSATA Handbook at II-8; Def's Ex A at 22-24; Ptf's Ex 2 at 2065.) He was satisfied that the result was not too low because, under a "market-to-book ratio" study summarized in his appraisal report, the historical sales prices of other regulated companies always exceeded the company's net book value. (See Def's Ex A at 25; see also Def's Ex D at 35-78 (Ben Johnson Associates, Inc.'s report examining 50 transactions occurring from 2001 through 2020).) The court finds Defendant's HCLD analysis unhelpful in determining real market value. The WSATA Handbook describes the HCLD method in equivocal terms, on the one hand characterizing the method as "a valid indicator of value" for a rate base-regulated utility such as Plaintiff, but also stating that "the market's perception of value is likely to be at odds with rate base at any given time." WSATA Handbook at II-9 to II-10. A significant reason for the acknowledged mismatch between net book value and real market value is that mandatory depreciation methods for regulatory purposes may result in a net book value that is "vastly different" from the market's perception of the value of property. Id. at II-10.
The WSATA Handbook states further that "HCLD cannot be considered either a lower or upper limit of value," but "rate regulation provides a vague limit on how far market value will stray from HCLD" because regulators, if attentive, are supposed to allow utilities an opportunity to earn a "fair return" on "prudently invested" funds and to recover "prudent costs." Id. at II-10. These terms stop far short of assuring the court that the result of HCLD, by itself, is a meaningful indicator of real market value.
As for DIT specifically, the WSATA Handbook states that "[a] utility's inability to earn an accounting 'return on' investment acquired with DIT is offset by its ability to collect revenue for a tax it does not yet owe." Id. at II-11. While a conceptual "offset" may justify the regulatory policy of including DIT, ITC and CIAC in net book value, the WSATA handbook does not state that the included amount accurately corresponds to the value of property a utility may acquire with the extra cash from its tax savings, or from CIAC. Just as Plaintiff has been unable to show that its income shortfall calculation accurately matches the effect of DIT, ITC or CIAC on net book value, Defendant's ignoring of those items is a flaw that renders Defendant's HCLD approach unreliable as an indicator of real market value. And even if an exhaustive analysis of Defendant's historical market-to-book ratio studies were to reveal that HCLD was below real market value in the subject transactions, those historical studies of sales of other companies do little to help the court decide the real market value of the specific property at issue in this case as of January 1, 2020.
5. Conclusion Under Cost Approach
The WSATA Handbook's solution to the tenuous connection between net book value and real market value is to assign less weight to the HCLD cost indicator than to other indicators in the reconciliation process. See, e.g., WSATA Handbook at II-12 ("The degree to which regulatory depreciation reflects an accurate estimate of market depreciation for a particular property is taken into account when reconciling the value indicators."). In this case, however, the court finds no factual basis to assign any weight to either party's cost indicator of value.
In light of the court's conclusions as to each party's analysis, the court finds it unnecessary to decide cost approach issues other than those specified. (E.g., Ptf's Post-Trial Br at 65 (arguing Defendant erred by not deducting "Asset Retirement Obligations" from HCLD but acknowledging amount at issue is de minimis).)
B. Income Approach
The parties agree that the income approach to valuation (1) determines the future expected cash flow from the property, and (2) divides that cash flow by a rate. The result is the present value of the future expected cash flow, which is an indicator of the value of the property. (See Ptf's Post-Trial Br at 7, 10-11; WSATA Handbook at III-3 (cash flow as proper measure of future "income"); id. at III-17 to III-20 (rates).) The simple formula can be expressed as:
Value = Cash Flow ÷ Rate
The two main methods under the income approach are direct capitalization (which uses a single year's projected cash flow) and yield capitalization (which uses a projection of cash flow over future years). See WSATA Handbook at III-8; Appraisal Institute, Appraisal of Real Estate at 36 (15th ed 2020). The court begins by examining the parties' analysis under the yield capitalization method.
Only Defendant also used the direct capitalization method, which the court discusses at the end of this section.
1. Yield Capitalization: Constant Growth Model
Both parties applied a variation of yield capitalization known as the constant growth yield capitalization model. The constant growth model assumes that the company's future cash flows will grow at a constant rate into perpetuity and determines value by applying the following formula:
Value = Next year's cash flow ÷ (weighted average cost of capital - growth rate) (See Def's Ex A at 28; Ptf's Ex 1 at 45.) The court now determines each component of the formula based on the parties' evidence.
2. Yield Capitalization: Cash Flow and Growth Rate for Constant Growth Model
Each party determined an estimated cash flow amount for purposes of its yield capitalization method, with markedly different results. Tegarden determined an estimated cash flow of $1.1 billion, Eyre $800 million. (Ptf's Ex 1 at 47; Def's Ex A at 30.) Consistent with the WSATA Handbook, each appraiser used as his starting point the net operating income that Plaintiff reported on its FERC 1 form for 2019 ($1,038,196,981). (See Ptf's Ex 2 at 2022.) See WSATA Handbook at III-3 ("accounting income, such as net operating income * * * should be transformed into cash flow prior to its use"). The parties agree that, to determine cash flow, depreciation must be added back to net operating income because the amount that was deducted for depreciation is an accounting construct rather than an actual cash expenditure. The parties also agree that capital expenditures must be subtracted in determining cash flow. (See Ptf's Ex 1 at 44 (Tegarden estimate is "net of CAPX &Depr."); Def's Ex A at 30 (Eyre estimate adds depreciation, subtracts capital expenditures); WSATA Handbook at III-4.)
Tegarden's estimated cash flow of $1,100,000,000 is only slightly greater than Plaintiff's actual 2019 net operating income, reflecting his "assum[ption] that the amount of capital reinvestment is equal to the depreciation expense." (Ptf's Ex 1 at 41.) Tegarden took into account net operating income in the prior three to five years and various statistical measures of those amounts, including a comparison against net plant investment. (See Ptf's Ex 1 at 44; Ptf's Ex 2 at 2022; Transcript at 255-60 (Tegarden).) As for assumptions about the future relevant to growth, Tegarden considered the effect of anticipated positive and negative rate changes in different service territories, which he estimated would result in a modest net reduction ($1,957,538), as well as construction work in progress expected to be placed in service, which he estimated would cause a slight immediate increase. (See Ptf's Ex 1 at 45-46.)
Eyre determined an estimated cash flow of $800 million. (See Def's Ex A at 28-31.) Eyre first increased 2019 net operating income slightly, from $1,038,196,981 to $1,059,000,000 by applying a two percent growth factor. He created a substantial net decrease ($270,000,000) when he added back estimated depreciation and amortization of $1,030,000,000 while subtracting estimated capital expenditures of $1,300,000,000. Eyre also subtracted $43,000,000 to account for estimated changes in regulatory assets and liabilities, and made other, smaller adjustments, resulting in his rounded amount of $800,000,000. (See id. at 30.)
The smaller adjustments were to subtract $25,000,000 in DIT and $8,600,000 in working capital. (See id.)
The $300 million difference separating the parties derives almost entirely from Eyre's large net subtraction ($270 million) that resulted when Eyre added $1,030,000,000 for depreciation and amortization while deducting $1,300,000,000 for capital expenditures, as opposed to Tegarden's assumption that capital expenditures will equal depreciation expense. The $300 million difference does not cut the way one might expect. Counterintuitively, under the simple formula for value shown above, if one assumes the same rate, Eyre's lower cash flow amount in the numerator would result in a lower value, and Tegarden's higher cash flow amount would result in a higher value.
In arguing for their respective cash flow estimates, the parties refer frequently to their differing expectations of future growth. For that reason, the court will consider the issues of cash flow and the growth factor together.
3. Plaintiff's Method
Tegarden's method generally fits within what the WSATA Handbook describes, in critical terms, as a "No Growth Model." See WSATA Handbook at III-15 to III-16. According to the WSATA Handbook, this model "assume[s] that the company has no 'effective' future growth potential which would contribute to value." WSATA Handbook at III-15. At this point, the court pauses to consider the meaning of "growth" as the parties (and the WSATA Handbook) use that term. Both parties distinguish between "real growth, i.e., earning more than the cost of capital" and "nominal growth, meaning that the overall size of the plant increased, and the nominal cash flows increased due to the new size, but the company was still earning its cost of capital." (Ptf's Response at 13; see Def's Ex A at 48 ("Real growth occurs when a company's rate of return is greater than its cost of capital.").) See also WSATA Handbook at III-16 (an assumption of no future growth in cash flow means that "future capital expenditures are only expected to earn their cost of capital.").
Without implying any precedential effect, the court notes that the Oregon Supreme Court has similarly explained: "Growth in net cash flows will arise either because the present asset base generates additional income or because additions to the asset base produce income beyond the cost of the additions themselves, including the costs of operating them." Union Pacific Railroad v. Dept. of Rev., 315 Or. 11, 22, 843 P.2d 864 (1992) (analyzing parties' arguments for and against growth factor in case involving centrally assessed railroad).
Tegarden's approach to cash flow ignores any additional capital expenditures in excess of depreciation, while Defendant points to such excess expenditures in 2018 and 2019, as well as future projections, as a source of "real" growth in the value of Plaintiff's property. Plaintiff presented evidence that, while its investment in its plant has grown over time, its growth in net operating income has not kept pace. Plaintiff points to a graph showing that, at least since 2001, the rate of annual increases (or losses) in net operating income corresponds closely to the rate of annual increases in net plant. (See Ptf's Ex 11 at 1.) A more detailed table shows that, over the years leading up to and including 2019, the historical rate of change in net plant generally has exceeded the rate of change in net operating income by, on average, 0.9 percent. (See Ptf's Ex 11 at 2.) The court finds that this historical evidence supports, at least by correlation, Plaintiff's position that increases in net operating income require increases in capital expenditure, rather than happening as a result of increased usage of existing property or other efficiencies, as Defendant argues.
The table shows all date ranges that start with 2019 and end in a prior year, through 1989. For 24 ranges over those 30 years, the rate of change in net operating income was less than the rate of change in net plant. For six ranges, the rate of change in net operating income was greater than the rate of change in net plant; however, the excess was modest, ranging from 0.2% to 0.9%. (See Ptf's Ex 11 at 2.)
Plaintiff also presented officer testimony and an exhibit showing that, for the assessment year ended December 31, 2019, Plaintiff earned a regulatory rate of return of 7.373 percent, which was less than the allowed regulatory rate of return (7.60%). (See Ptf's Ex 5 at 11, line 63 (7.373% for 2018); Transcript at 129.) In addition, Plaintiff's managing director of revenue requirement, Steven McDougal, testified that Plaintiff earned less than its allowed return on equity for each of the approximately ten years preceding 2020. (See Transcript at 130.) The WSATA Handbook recognizes this evidence as a potential predictor of future increases in net operating income. See WSATA Handbook at III-5 ("Another technique for predicting net operating income is to evaluate a utility's historical performance as compared to the regulatory body's allowed rate of return. An analysis of performance ratios can be used to predict where current earnings on net plant will fall in relationship to allowed earnings. For example, if it can be shown statistically that a centrally assessed company's historical earnings have been, for example, 95 percent of its allowed rate of return on rate base, it may be reasonable to assume a similar relationship in the future."). The court finds that this evidence further supports Plaintiff's "no growth" position and enhances the reliability of net operating income as a proxy for cash flow for yield capitalization purposes.
4. Defendant's Method
Rather than use a proxy, Eyre attempted to build an actual cash flow estimate based on prior-year and projected data. The quality of his estimate depends, therefore, on the reliability of the data and the logic behind his assumptions, which the court now explores.
As to prior-year data, Eyre testified that he "normalized" depreciation and capital expenditure data from 2016 through 2019 in arriving at his projections for 2020. (See Def's Ex A at 28-30.) Eyre's normalization seems to have consisted mainly of simple arithmetic processes applied without investigation, as opposed to adjustments designed to account for identified anomalies, transactions, or other events. Eyre treated the four-year increase in Plaintiff's depreciation in 2016 through 2018 as a "trend" and determined his 2020 depreciation amount of $1,030,000,000 by multiplying the 2019 amount by the four-year average percentage increase. (See Def's Ex A at 30; Transcript at 561-62.) On cross-examination, however, he acknowledged that some of the increase in the two latter years may have been caused by changes in federal tax law, rather than by the addition of new property, and that he did not know whether those law changes had an effect beyond 2020. (See id. at 562.) Also during 2018 and 2019, capital expenditures increased dramatically, jumping more than 60 percent in 2018 and more than 70 percent in 2019. There is no evidence that Eyre investigated the business reasons for these increases or the likelihood that they would continue to affect future cash flow. The court finds this troubling, particularly since capital expenditures declined approximately 14 percent from 2016 to 2017. Eyre simply averaged the amounts for all four years without explanation.
Specifically, the 2017 federal Tax Cuts and Jobs Act (TCJA), Pub L 115-97, 131 Stat 2054 (2017).
Eyre derived his estimates from the following prior-year data:
2020 Eyre Est.
2019
2018
2017
2016
Depreciation (Eyre uses 8% growth rate)
$1,030,000,000
$953,983,000
$979,350,000
$796,220,000
$770,251,000
Capital Expenditures (Eyre uses four-year average of these amounts)
$1,300,000,000
$2,247,610,000
$1,291,567,000
$797,524,000
$930,851,000
In briefing, Defendant argues that the fact that capital expenditures in 2018 and 2019 exceeded the sum of depreciation and DIT is evidence of "real growth." (See Def's Opening Post-Trial Br at 15.) However, Defendant presents no evidence linking Plaintiff's capital expenditures to growth in Plaintiff's potential for generating new, additional cash flow, as opposed to, for example, expenditures needed to replace existing property that can be expected to generate the same cash flow. Documents in the record, such as Plaintiff's 2019 FERC-1 forms and the discussions of Plaintiff's activities in the 2019 10-Ks filed by Plaintiff's ultimate parent company, do not aid the court on this point. The filings contain brief discussions of a wide range of anticipated capital expenditures, but the court has found no narrative or quantitative summary that would allow the court to conclude that Plaintiff expected "real" growth as opposed to "nominal" growth. (See, e.g., Ptf's Ex 2 at 2010 (2019 FERC-1) (discussing Plaintiff's 2019 "integrated resource plan" filed with state regulators that calls for investment in additional "renewable capacity," energy storage capacity and a new transmission line); id. at 4335 (2019 K-1) (discussing "repowering" of all of Plaintiff's wind-powered generating facilities "by the end of 2020" to requalify them for income tax credits, "extend the lives" of the facilities, and "increase[e] the anticipated electrical generation * * * by approximately 26%").) Without testimony or other evidence tying together the data, the court cannot accept evidence of increased capital expenditures as proof of real growth.
The following table is drawn from Eyre's appraisal report. (See Def's Ex A at 30.)
2019
2018
2017
2016
Depreciation
$953,983,000
$979,350,000
$796,220,000
$770,251,000
Deferred Income Tax
- $125,091,000
-$202,299,000
$75,165,000
$145,070,000
Total
$828,892,000
$777,051,000
$871,385,000
$915,321,000
Capital Expenditures
$2,247,610,000
$1,291,567,000
$797,524,000
$930,851,000
As to other prior-year data, for the item "net changes in regulatory assets and liabilities," Eyre selected $43 million as his estimate for 2020, even though the amount for 2019 was below zero (negative $55,014,000), the amount declined for the two prior years, and there was an overall decline from 2016. (See Def's Ex A at 30.) He acknowledged on cross-examination that he "ha[d]n't looked into" why the 2019 amount declined below zero. (Transcript at 553.) This computation further undermines the court's confidence in Eyre's estimate of cash flow.
Eyre's estimate of the net change in regulatory assets and liabilities for 2020 was $43 million. He determined this amount by averaging the following:
2020 Eyre Est.
2019
2018
2017
2016
$43,000,000
-$55,014,000
$87,483,000
$18,492,000
$122,115,000
Turning to Eyre's data of future projections, Eyre relied heavily on a "2020 Plan" generated by Plaintiff. (See Def's Ex A at 50; Def's Ex F.) The 2020 Plan is a 19-page document, consisting of tables and narrative, that bears the heading "Version 2 - 11/08/2019." (Def's Ex F at 2.) The tables are labeled "2020 Plan Summary," "Income Statement," "Balance Sheet," and "Cash Flow"; each table contains columns for the years 2020 through 2029, and some tables also show one or two prior years. Eyre determined that the 2020 Plan forecasts a compound annual growth rate in net income of 8.6 percent for the first three years and 5.6 percent over the entire ten-year period. (See Def's Ex A at 48 (finding 8.6% rate for "the first five years"); Transcript at 597 (correcting to "just the first three years" based on an acknowledged typographical error).) Eyre compared these rates to a forecasted compound annual growth rate in property, plant and equipment of only 2.63 percent and concluded that Plaintiff was itself predicting "real growth." (Def's Ex A at 48.) Based on this information, and on his observation that inflation for 2009 through 2019 had an annual average rate of 1.61 percent, Eyre selected a rate of growth of two percent, which affected two components of the constant growth model formula set forth above: increasing the 2019 cash flow by two percent and, separately, setting the growth rate to subtract from WACC. (See id. at 48-49.)
The court finds little basis to rely on the 2020 Plan. Eyre understood the 2020 Plan to be Plaintiff's "business plan closest to the end of the year." (Transcript at 581.) Eyre did not speak to anyone who had prepared it; he surmised, without evidence, that Plaintiff has "a large group of people that do this on an annual basis." (Id. at 579.) He did not ask the purpose for which the 2020 Plan was prepared. He acknowledged that the forecasted amounts become "more speculative * * * the farther you get out from the present days." (See id. at 582.) Eyre was unaware whether the 2020 Plan contained any mathematical errors or other errors. (See id.) Defendant presented no evidence correlating amounts shown in prior such plans with actual historical results. Nothing informs the court whether the 2020 Plan represented a loose corporate aspiration, as opposed to, for example, a thoroughly developed metric that might be used to set officer compensation. Plaintiff's tax director Norman Ross testified that the 2020 Plan was not released to the public and that Plaintiff did not intend that it be relied on by people outside the company. (See Transcript at 1411.) Overall, the court finds little basis to treat the 2020 Plan as a reliable forecast of future data.
Unexplained differences compared to the amounts used in the parties' appraisal computations also undermine the persuasive value of the 2020 Plan. First, the "income statement" to which Eyre refers in the 2020 Plan does not show amounts of "net operating income" (the starting point for Eyre's cash flow computation); instead, the 2020 Plan refers to "net income" and to "operating income." Using 2018 (the only completed year shown on the 2020 Plan) as a reference, the court finds that the amount for "operating income" ($1,050,700,000) is similar to, but not identical with, "net operating income" ($1,041,866,000), but the manner of computing the two amounts is sufficiently different that the court is not confident that "net operating income" is comparable to "net income." (Compare Def's Ex F at 6 with Def's Ex A at 30 and Ptf's Ex 2 at 1499-1982 (FERC 1).) The court also notes that the heading of the relevant page of the 2020 Plan is "Income Statement (PPW Holdings)." Plaintiff's Form 10-K for 2019 describes PPW Holdings LLC as Plaintiff's parent company. (See Ptf's Ex 2 at 4415.) The court finds no evidence in the record indicating whether PPW Holdings LLC has income or operations, apart from its ownership of Plaintiff's stock, that might have been combined with projections for Plaintiff.
Eyre concluded that the 2020 Plan was a forecast of "real growth" because the plan predicts net income to grow at a compound annual growth rate of 5.6 percent (averaged over the ten years), while the plan predicts that property, plant and equipment (PP&E) will grow at a compound annual growth rate of only 2.63 percent over the same period. (See Def's Ex A at 48.) However, Ross testified that Eyre's computations contain a mismatch that overstates the difference between the two rates. In determining the compound annual growth rate in PP&E, Eyre included assets in each year that constituted "construction work in progress" (CWIP).
Although Ross agreed that PP&E together with CWIP grew at 2.63 percent, Ross determined that PP&E without CWIP grew at 4.8 percent. (See Transcript at 1373-75; Ptf's Ex 38.) Ross concluded that CWIP must be subtracted from PP&E because Plaintiff's regulators do not allow Plaintiff to earn a return on CWIP, so CWIP does not contribute to net income. The court finds no argument from Defendant in response to this point. The court agrees with Plaintiff: Where the goal is to test for "real growth" by comparing the rate of growth in Plaintiff's net income to the rate of growth in PP&E, the dollar amount of PP&E must be limited to property capable of generating net income. Ross's calculations show that the effect of this correction is that the ten-year compound annual growth rate in PP&E rises from the rate Eyre derived (2.63%) to 4.8 percent, which is much closer to the 2020 Plan's projection of 5.6 percent growth in net income. (See Ptf's Ex 38.)
Ross also testified that the 2020 Plan contained an error in the projection of income amounts for 2028 and 2029; for each of those years, income was overstated by $100 million. (See Transcript at 1373-74.) According to Ross, correcting that error reduced the compound rate of growth in net income from 5.6% to 4.8%, which matches the post-CWIP compound rate of growth in PP&E. (See Ptf's Ex 38.) Defendant objects to Ross's attempted correction as untimely and unsubstantiated; Plaintiff argues that the error is sufficiently obvious that Eyre should have spotted it. (See Def's Opening Post-Trial Br at 49; Ptf's Response at 30.) The court need not resolve this point because the court finds the 2020 Plan unreliable for the reasons stated, especially as to the later years.
The narrative in the 2020 Plan does not, by itself, support or refute Eyre's conclusion of real growth. Like Plaintiff's public filings on Forms FERC-1 and 10-K, the narrative in the 2020 Plan mentions many future projects and prospects, including new renewable energy and storage projects as well as coal plant retirements. (See, e.g., Def's Ex F at 2.) However, the 2020 Plan does not appear to have the purpose of distinguishing between real growth and nominal growth, and its data are not organized in a way that informs the court on that point. The court finds no basis in the 2020 Plan to conclude that Plaintiff expected real growth.
The court finds no other evidence of future real growth. In post-trial briefing, Defendant argues that "PacifiCorp's customer base has been increasing and is expected to continue to increase." (Def's Response at 11.) Defendant cites no evidence for this assertion. Eyre testified, without reference to any documentary or other evidence: "A growing company like PacifiCorp, they're growing. They're adding new customers every day. So they're required to add new plant as they--as they continue to grow, they'll probably on a normalized basis have more capital expenditure and depreciation because they're growing." (Transcript at 1016.) Even if these circular assertions had a basis in evidence, they do not help the court because they fail to distinguish between real growth and nominal growth.
5. Conclusion as to Cash Flow and Growth Rate
The court is compelled to choose between two imperfect measures of cash flow. Plaintiff offers net operating income as a proxy based on its firmly held view that the fact that it is a regulated utility means, per se, that it will never experience real growth because regulators will prevent it from earning more than its cost of capital. To some extent, Plaintiff relies on this theoretical position instead of addressing Defendant's points, notably the fact that capital expenditures grew substantially in excess of depreciation and DIT in 2018 and 2019. On the other hand, Defendant insists just as strongly that all regulated utilities experience real growth that is not reflected in net operating income, building up a store of property value ignored by regulators, with the result that utilities always sell for more than their net book value. The court need not, and does not, resolve the parties' competing theories to determine cash flow in this case. The court finds too many logical leaps and unsupported assumptions in Defendant's built-up determination of cash flow. Plaintiff's use of net operating income as a proxy, despite flaws, is superior because it is better supported by historical data. On balance, the court finds that the preponderance of the evidence supports Plaintiff's estimated cash flow of $1,100,000,000. For the same reasons, the court finds that the growth rate in the yield capitalization formula is zero. In adopting this cash flow and the growth rate of zero in this case, the court neither validates nor rejects a generalized theory that the net operating income of a regulated utility is necessarily an appropriate proxy for cash flow consisting of net operating income plus depreciation and DIT, less capital expenditures.
Consistent with the use of net operating income as a proxy, in lieu of a computed cash flow, the court also accepts Plaintiff's position that no adjustment for inflation is appropriate. While the court agrees with
Defendant also criticizes Tegarden's no-growth assumption as inconsistent with his assumptions of positive growth in computing the weighted average cost of capital. (See Def's Opening Post-Trial Br at 9.) However, only the parties' Dividend Growth Model (DGM) estimates relied on an assumed rate of future growth. Because of flaws unrelated to the projection of Plaintiff's future cash flow stream, the court gives no weight to either party's DGM analysis, and the court therefore finds any such inconsistency irrelevant.
6. Yield Capitalization: Rate (Weighted Average Cost of Capital)
In a yield capitalization method, the rate by which cash flow is divided often is referred to as the "weighted average cost of capital" (WACC). See, e.g., WSATA Handbook at III-19 to III-20. The rate is composed of two components: the cost of debt, and the cost of equity. The WACC is the average of the two, weighted according to the company's capital structure, i.e., its relative proportions of equity and debt. See id. at III-26 to III-30. The court reviews each party's proffered capital structure, cost of debt, and cost of equity.
7. Capital Structure
In this case, the parties' determinations of Plaintiff's capital structure are very similar, and each is supported by substantial evidence. Tegarden determined a capital structure of 35 percent debt and 65 percent equity. (See Ptf's Ex 1 at 51.) Eyre determined a capital structure of 37 percent debt and 63 percent equity. (See Def's Ex A at 34.) The court determines a capital structure of 36 percent debt and 64 percent equity.
Defendant that regulators may provide for inflation, either in a rate case under ORS 757.259 or through an automatic adjustment clause under ORS 757.210, the adjustment would affect the numerator and the denominator of the constant growth formula; inflation would not represent real growth. (See Ptf's Post-Trial Br at 46; Ptf's Response at 24; Def's Opening Post-Trial Br at 39; Def's Response at 10-12.)
8. Capitalization Rate: Cost of Debt
Tegarden determined a cost of debt of 4.25 percent; Eyre determined a cost of debt of 3.73 percent. (Ptf's Post-Trial Br at 18; Def's Response at 13.) Both Tegarden and Eyre based their cost of debt on the actual yield to maturity of bonds issued by utilities they considered comparable. They agreed on one point: they each looked to bonds that were assigned the bond rating of BBB+. As discussed below, the parties disagree about whether to use effective rates for newly issued debt, as opposed to the yield to maturity for existing debt. They also disagree about the selection of comparator companies. Finally, Defendant argues that Plaintiff failed to select yields as of the relevant date.
BBB+ is the Standard & Poor's rating. The parties sometimes referred to the equivalent Moody's rating of Baa. (See, e.g., Transcript at 444.) Tegarden arrived at the BBB+ rating because it was the median bond rating "for the entire electric industry no matter how it is divided." (Ptf's Ex 1 at 53.) Eyre reached the same rating by selecting 10 guideline companies that he considered "reasonably similar" members of the "Electric Utility segment." (Def's Ex A at 31-33; Transcript at 444.) The average rating of those 10 companies was Baa (Mergent). (Def's Ex A at 46.)
Tegarden started with "the yields to maturity for all rated electric utility bonds traded at year end 2019," as reported in Mergent Bond Record in January 2020. (Ptf's Ex 1 at 53 (narrative), 172-76 (list), 176 (publication date).) From this list of approximately 300 bonds (which the court refers to as Tegarden's Master List), Tegarden selected those rated at BBB+. There were 37 such bonds; the court refers to this subset as Tegarden's BBB+ List. (See id. at 172-76.) The average yield to maturity of the bonds on Tegarden's BBB+ List as of January 1, 2020, was 4.40 percent. (See id.) Tegarden also considered the average yield of all utility bonds with the same BBB+ rating (3.73%), as reported by Mergent's Bond Record, as well as other data points. (Ptf's Ex 1 at 54-55.) He concluded that 4.25 percent was appropriate as his cost of debt.
Eyre reviewed one of the data sources Tegarden considered, namely, the Mergent data for all utility bonds. Eyre described this source as "the Mergent Bond Record for January 2020," showing "the average yield to maturity for a public utility bond of [the BBB+] rating issued on 1/1/20 * * *." Eyre selected that percentage as his cost of debt: 3.73 percent. (Def's Ex A at 46.)
a. Use of Effective Rates for Newly Issued Debt vs. Yield to Maturity For Existing Debt
Defendant criticizes Tegarden's approach on the grounds that it does not rely on the yield to maturity for debt issued on or around the assessment date, January 1, 2020. Defendant asserts: "Mr. Eyre's 3.73% cost of debt was as of year-end 2019 but Mr. Tegarden's 4.25% debt rate was not." (Def's Response at 13.) Defendant then qualifies this statement somewhat, stating that "certain of" Tegarden's yields "are not year-end rates." (Id. n 3.) Eyre distinguished between (1) the effective rates for bonds actually issued on or about the assessment date, and (2) the effective rates for bonds issued months or years before the assessment date but traded on or about the assessment date. (See Def's Ex E at 24; see also Def's Ex A at 46.) At trial, he criticized Tegarden because "Mr. Tegarden went and used the yields of bonds that have already been issued. Some of them had been issued years before the [assessment] date." (Transcript at 444.) In briefing, Defendant takes the position that "[t]his practice of using existing, or embedded debt, is inconsistent with the determination of real market value on the appraisal date * * *." (Def's Opening Post-Trial Br at 16.) The apparent premise of this position is that only the effective rates for actual issuances on or about a particular date are reliable data for the yield to maturity of comparable bonds as of that date. The court finds Defendant's premise unreliable and its position overstated.
The court makes no finding as to whether Eyre relied solely on yields on bonds issued at or around January 1, 2020. Eyre's exhibits include no source document that identifies whether the subject bonds are newly issued or pre-existing. On cross-examination, Eyre was asked whether Mergent's, the source of his data, was "looking at all sorts of seasoned bonds and * * * determining what that yield is based on the current interest rate that is being provided." (Transcript at 622-23 (emphasis added).) Eyre responded: "I don't know what procedure they go through to determine that yield. But it is * * * reported as a yield to maturity for this rated bond." (Id. at 623.)
The court discusses Defendant's use of the term "embedded debt" below.
The WSATA Handbook's discussion of debt rates includes the following statement:
"An excellent source of debt money costs is the effective rate at which new debt issues are sold. The effective rate may vary from the coupon rate depending upon the market at the time. Wide publicity is given to the sale of new debt issues and these issues are very frequently put out to competitive bidding."
WSATA Handbook at III-28. Thus, the WSATA Handbook implies that there is an evidentiary reason to prefer rates for newly issued debt, i.e., that those rates are more likely to be reliable because they are more likely to be based on bids from a larger number of potential buyers. The court questions whether this rationale, first stated more than thirty years ago, has a basis in fact in today's bond market. Even if it does, the court finds Defendant's reliance on the quoted passage misplaced because the WSATA Handbook elsewhere emphasizes that debt rates are "relatively easy to determine" and are "usually obtainable from current loan rates and current ratios of yield to price of bonds." WSATA Handbook at III-28 (emphasis added). At another point, the WSATA Handbook states that "debt with a low nominal interest rate will be discounted in the marketplace at an effective rate equivalent to the current cost of debt." Id. at III-29. Reading the WSATA Handbook's discussion of debt rate as a whole, the court is left with the impression that the marketplace can efficiently and accurately determine the yield to maturity of debt as of an assessment date, even if the debt was issued earlier. The court has found nothing in the Department's other exhibits that discusses the relative reliability of effective rates of current vs. prior issuances. The court finds no evidence that Tegarden's debt rate is flawed for failure to rely solely on debt issuances on or around the January 1, 2020, assessment date. Nor does the court find any evidence that Eyre's debt rate is rendered more reliable because it purports to be based solely on debt issuances on or around the January 1, 2020, assessment date.
The quoted sentences originally appeared in the 1989 edition of the WSATA Handbook and were reprinted verbatim in the most recent edition, which was published in 2009. Western States Association of Tax Administrators, Appraisal Handbook: Valuation of Utility & Railroad Property at 50-51 (1989). In particular, the court is uncertain whether the reference to "[w]ide publicity" for new debt issues remains valid in light of substantial changes in the delivery of information since the development of the internet. The WSATA Handbook offers no further detail, nor does anything in the record in this case substantiate the sentences.
This impression is confirmed by another of Defendant's administrative rules, entitled: "Derivation of Capital Structure and Discount Rates for Valuing Industrial Properties and Department-Assessed Properties." OAR 150-308-0250(2)(b). That rule states in relevant part:
"The cost of debt is the current market rate for new securities. The embedded rate on securities previously issued is not a proper measure. In order to determine the cost of debt the appraiser should:
"(A) Refer to the rates for seasoned bond issues from Moody's Utility, Industrial, and Transportation weekly news reports or other rating services for at least two months immediately prior to the appraisal date. This should be done by bond rating (Aa, A, Baa, etc.) and industry type.
"(B) Obtain information on new bond issues by industry type and bond rating from Moody's Bond Survey or other publications for at least two months immediately prior to the appraisal date.
"(C) Consider recommendations on debt rates submitted by industry.
"(D) Select rates for each industry group by bond rating after analyzing the data in the steps above."In other words, the rule directs appraisers to both refer to "rates for seasoned bond issues" (paragraph (A)) and to "[o]btain information on new bond issues" (paragraph (B)). Nothing directs an appraiser to prefer, or to rely solely on, rates for new issuances.
Eyre's testimony also suggests that he conflated the use of current yields on seasoned debt (which, as discussed, is permitted by the WSATA Handbook and by OAR 150-308-0250(2)(b)) with the use of "embedded" debt rates. (See Transcript at 445 ("You do not use embedded costs of debt or -- or the yields on bonds that have already been issued sometimes many years prior to the appraisal date. That is not indicative of market value today.").) The WSATA Handbook cautions:
"Some appraisers advocate using the actual coupon rates on existing debt (embedded debt). The logic is that during times of rising interest rates a prospective purchaser would most likely assume the existing debt rather than refinance. This position lacks merit because, even in the case of an assumption, debt with a low nominal interest rate will be discounted in the marketplace at an effective rate equivalent to the current cost of debt. The use of embedded debt rates in estimating the current cost of capital results in a capitalized earnings indicator which reflects high or low interest debt instruments at their face value rather than at their market value. Regardless of the regulatory practice of using embedded debt rates, their use is contrary to the market value concept."WSATA Handbook at III-29. (See Def's Opening Post-Trial Br at 16-17.) All evidence indicates that Tegarden relied on market data for yields to maturity, rather than coupon rates, when estimating the cost of debt. (See Ptf's Response at 17.) Therefore, the court sees no factual basis for Defendant's criticism that Tegarden relied on "embedded debt," as defined in the WSATA Handbook.
b. Use of Trades Occurring Near January 1, 2020; Appraisal Date vs. Earlier Dates
Defendant's second significant criticism is that, even though the Tegarden BBB+ List may have relied on yields to maturity for all rated electric utility bonds "traded at year-end 2019," those data nevertheless were inaccurate because they were not yields to maturity "as of" December 31, 2019. (Def's Reply at 15.) Defendant's expert Dr. Bradford Cornell testified that Tegarden may well have assembled a list of bonds that were traded at the end of 2019, but rather than use the yields attributable to trades of those bonds occurring at the end of 2019, he instead selected yields attributable to trades occurring throughout the year. (See Transcript at 919-24; Def's Ex J at 5-20.) Defendant presented evidence that yields on corporate and utility bonds generally declined significantly during 2019, such that relying on average rates during 2019 would be inaccurate. (See Def's Ex I at 20-21.) In a rebuttal exhibit, Cornell reprinted the entire Tegarden Master List, then selected five of the bonds shown on that list--bonds that had been issued by Plaintiff itself. (Transcript at 921-22; Def's Ex J at 8-12 (corresponding to Ptf's Ex 1 at 172-76).) Cornell presented a series of five screen shots from Bloomberg, which he described as "the traders' internet site that produces actual market results." (Transcript at 921; Def's Ex J at 15-19.) For each of the five bonds, the screen shot showed seven to 12 dates in December 2019 or January 2020 on which the bond was traded, the volume and number of trades on each date, and a last yield on each date. On each screen shot the last yield reported was between approximately one percent and two percent lower than the yield shown on the Tegarden Master List. (Compare, e.g., Def's Ex J at 15 (for bond at coupon rate of 5.9% and maturing August 15, 2034, apparently showing eight trades from December 4, 2019, through January 16, 2020, at yields ranging from 3.343% to 3.497% with Def's Ex J at 10 (PacifiCorp bond with coupon rate of 5.9%, showing 5.12% yield).) Cornell testified: "I can't square anything like Mr. Tegarden's number with the actual trading data for PacifiCorp's bonds." (Transcript at 923.) Defendant argues that this sample of Plaintiff's own bond rates undermines the accuracy of the Tegarden Master List as a data source. (See, e.g., Def's Response at 13.) In its response brief, Plaintiff reiterates Tegarden's statement in his appraisal report that his Master List showed "the yields to maturity for all rated electric utility bonds traded at year end 2019" and Tegarden's testimony on cross-examination that his Master List "represent[s] the cost of debt at year-end December 31st, 2019." (Ptf's Response at 17-18; Ptf's Ex 1 at 53; Transcript at 354.) In rebuttal, Tegarden testified that he did not know why there might be a difference between the yields he used and those derived from Cornell's Bloomberg screen shot approach, noting that Cornell did not explain how the Bloomberg yields were calculated. (Transcript at 1241-42.) Tegarden also testified that his BBB+ List was "basically the same data as Mr. Eyre, I believe, used[:] a summary of BBB rated debt." (Id. at 1242.)
The court finds that the comparison of the Tegarden Lists with Bloomberg screen shots showing yields of actual trades near the assessment date raises questions but falls short as a reason to reject Tegarden's debt rate. The trades on Cornell's screen shots are of five bonds that Tegarden did not use because they had a rating different from the rating on which both appraisers agree. The list Tegarden actually used--his BBB+ List--includes only 37 bonds, but Cornell made no effort to provide comparable screen shots showing trades for those bonds. Cornell also made no effort to show that Tegarden used the wrong set of published data and that Tegarden should have used a more accurate set of published data. Rather, Cornell relied on unpublished data, apparently obtaining the screen shots by contacting "a trader [who] was nice enough to give [the screen shots] to me from Bloomberg * * *." (Transcript at 923.) Cornell's evidence implies that an appraiser seeking data on actual trades of seasoned bonds during a particular window of time must resort to a favor from a friend in the industry because the data are unavailable from publishing services such as Mergent. The court finds this implication implausible. The court finds the Department's criticism of Tegarden's debt rate inadequately substantiated by the evidence.
c. Use of Electric-Only Utilities as Guideline Issuers vs. Inclusion of Other Kinds of Utilities
Plaintiff criticizes Eyre's cost of debt analysis because Eyre relied exclusively on a data point that includes not only electric utilities but also water and gas distribution utilities. (Ptf's Post-Trial Br at 18-19.) Tegarden stated that data for electric-only utilities allowed a more precise measurement of Plaintiff's cost of debt. (See Ptf's Ex 1 at 53.) Eyre did not refute this point, except to testify that "[e]lectric utilities dominate the utility bond market * * * and they would be the major component leading up to that typical market yield." (Transcript at 445.) Although the evidence on both sides is thin, the court finds it plausible that other electric utilities are more comparable to Plaintiff than a mix of electric, water and gas companies. See WSATA Handbook at III-28 (recommending segregating data based on industry type); III-9 to III-12 (discussing use of "reasonably similar" companies as comparators for capitalization analysis).
Plaintiff claimed in briefing that water and gas utilities have risks and capital structures different from those of electric utilities, but the court has found nothing in the record that supports this assertion. (See Ptf's PostTrial Br at 18.)
d. Conclusion as to Cost of Debt
The court finds that the preponderance of the evidence supports Plaintiff's determination that the cost of debt was 4.25 percent.
9. Capitalization Rate: Cost of Equity
Tegarden determined a cost of equity of nine percent, excluding "flotation costs," which are discussed below. (See Ptf's Ex 1 at 57; Ptf's Post-Trial Br at 20.) Eyre determined a cost of equity of 6.7 percent. (See Def's Ex A at 43.) Each expert used three methods. Both Tegarden and Eyre used the Dividend Growth Model (DGM), which the court will discuss first. Both also used the historical Capital Asset Pricing Model (CAPM), as well as variations; the court will discuss the CAPM second. Finally, Tegarden used the Build-Up (Risk Premium) Method, which the court discusses third.
Plaintiff argues for the cost of equity that Tegarden determined. (See Ptf's Post-Trial Br at 20.) However, Plaintiff also engaged additional expert Dr. Roger A. Morin, who prepared a separate estimate solely of the cost of equity, which was 9.6%. (See Ptf's Ex 3 at 52.)
Other names for the DGM include "discounted cash flow" and the "discounted growth model." (See Transcript at 160 (Morin testimony).)
a. Cost of equity: DGM
Using the DGM, Tegarden determined a cost of equity of 8.35 percent. (See Ptf's Ex 1 at 59.) Plaintiff's expert, Dr. Roger A. Morin, determined a cost of equity of 8.8 percent to 9.4 percent. (See Ptf's Ex 3 at 24.) Eyre determined a cost of equity of 7.60 percent. (See Def's Ex A at 42.) The parties agree on the formula for the DGM and agree that the only material difference is one component of the formula: the rate of growth. (See Ptf's Post-Trial Br at 2223; Def's Opening Post-Trial Br at 18.) Under the DGM formula, the cost of equity estimate is (1) the expected dividend from the company at the end of the first future year divided by the company's current stock price, expressed as a percentage, plus (2) the expected rate of growth in dividends, also expressed as a percentage:
A finance textbook that both parties cite frequently refers to the cost of equity estimate variously as the "discount rate," the "market capitalization rate for the firm's common stock," or the "expected rate of return on other securities of comparable risk." Richard A. Brealey, et al., Principles of Corporate Finance 87 (13th ed 2020) (Brealey).
Cost of equity = (Dividend in Year 1 ÷ Stock Price in Year 0) + (Growth Rate)
The relatively uncontroversial fraction in clause (1) of the foregoing sentence sometimes is referred to as the current yield, or the dividend yield. (See, e.g., Ptf's Ex 1 at 63; Ptf's Ex 3 at 12; Def's Ex A at 40.) For publicly traded companies, analyst firms regularly publish the current stock price, as well as estimated dividends and estimated growth rates. Because Plaintiff's stock as a wholly-owned subsidiary is not publicly traded, the parties' experts used published data for a group of investment-grade, dividend-paying electric utilities they selected. (See Ptf's Ex 1 at 53 (Tegarden); Ptf's Ex 3 at 11, 18 (Morin); Def's Ex A at 42 (Eyre).) Morin concluded a dividend yield of 3.15 percent, and Eyre concluded a dividend yield of 3.10 percent. Ptf's Ex 3 at 22 (Morin; 3.15%); Def's Ex A at 42 (Eyre; 3.10%).) Tegarden's appraisal report does not specify his conclusion as to a dividend yield, but it lists companies with average or mean dividend yield percentages ranging from 2.82 percent to 3.24 percent, and Plaintiff states in briefing: "The only material difference between Mr. Tegarden's and Mr. Eyre's DGM estimates is their growth estimates." (Ptf's Post-Trial Br at 23; see Ptf's Ex 1 at 100-102.) Accordingly, the court finds the dividend yield of 3.10 percent, consistent with Eyre's conclusion and Plaintiff's position.
Turning to the growth rate, because the DGM formula adds the growth rate to the dividend yield, the higher the growth rate, the higher the cost of equity. The parties' experts differed over which companies' growth rate estimates they selected and whether they modified those growth rate estimates. Tegarden and Morin used average estimated growth rates of selected companies, as published by Value Line, Standard &Poor's, and Zacks, without modification, resulting in average growth estimates ranging from 4.75 percent to 5.71 percent (Tegarden) and 5.45 percent to 6.05 percent (Morin). (See Ptf's Post-Trial Br at 23.) According to Defendant, to rely solely on the published growth rates is flawed because those rates are "short term growth rates," while the object of the DGM analysis is to estimate "sustainable growth into perpetuity." (Def's Opening Post-Trial Br at 18.) Eyre attempted to correct for this deficiency in the published data by using a multi-stage model that reduces the estimated growth rate over time, taking into account the growth rate for the economy as a whole, for which he used a federal government estimate of 3.70 percent. (See Def's Ex A at 42.) Plaintiff does not appear to contest that the published growth rates are for the short term, and Plaintiff acknowledges that an approach that assumes lower growth rates after a period of years could be reasonable. (See Ptf's Response at 19-20.) However, Plaintiff cautions that a downward adjustment to the growth rate must be coupled with an upward adjustment to the assumed dividend--which Eyre's model lacks. (See Def's Ex A at 34, 42 (Eyre).) Plaintiff's expert Morin testified:
"Another way of explaining this is as follows. If growth is going to decrease, less [of] the earnings will be required to finance the capital required by growth. So these liberated earnings, these freed up earnings, they have to go somewhere. Where do they go? Dividends. They can't go anywhere else. And the big problem here with * * * Mr. Eyre's multistage model is that he failed to take into account that if you lowered the growth you've got to beef up the dividend."(Transcript at 1209 (testimony of Morin).) Plaintiff's tax director, Norman Ross, testified similarly:
"Mr. Eyre has used a three-stage version of the dividend growth model and that can, in fact, be a valid model, but it * * * can only be a valid model when the inputs are correctly applied, and Mr. Eyre has not correctly applied those. He's left the dividend yield, and thus, the payout ratio constant over time while he's shrunk the growth rate and that seriously understates the cost of equity estimate."(Transcript at 1318 (testimony of Ross).)
The court finds that each party states valid criticisms about the other's DGM, but neither party offers a DGM that incorporates corrections of the flaws identified by the other party. As to Plaintiff's use of analyst-published growth rates to make a long-term forecast, the court finds Defendant's criticism warranted. Neither party has submitted evidence from the publications themselves as to the intended duration of the published growth rates they used. However, Eyre testified that the published data are "a short-term growth rate" for "the next three to five years," and Plaintiff has not refuted that evidence. (Transcript at 437.) Brealey seems to corroborate Defendant's position, stating: "Analysts are rarely prepared to stick their necks out by forecasting dividends to kingdom come, but they often forecast growth rates over the next five years, and these estimates may provide an indication of the expected long-run growth path." Brealey at 88. Cornell testified for Defendant that a growth rate that outpaces the annual rate of growth of the economy for an extended period is invalid. (Transcript at 925; see Def's Ex J at 23 (Cornell quoting Warren Buffett).) On cross-examination, Plaintiff's witness Morin agreed with this general point but insisted that analyst forecasts nevertheless correlate with the stock price more than any other variable. (See Transcript at 1231.)
Morin acknowledged the general validity of a passage in Brealey cautioning entry-level finance students to "resist the temptation to apply the [simple constant-growth DCF] formula to firms having high current rates of growth. Such growth can rarely be sustained indefinitely, but the constant-growth DCF formula assumes it can." Brealey at 89. On the other hand, Brealey's hypothetical high-growth rate is 20%, far above the rate that Plaintiff assumes. See id.
Defendant, meanwhile, proffers a multi-stage DGM that reduces the rate of growth over time. Brealey at 90-91. The court finds warranted Plaintiff's criticism that Eyre failed to increase the expected dividend amount (the numerator in clause (1) of the formula). That some increase is likely, is supported by the testimony quoted above, as well as by a comment from a nationally recognized expert whose work is cited by both parties. (See Ptf's Ex 32 (email from Dr. Aswath Damodoran, New York University) ("A model that just changes growth and leaves discount rates and payout ratios unchanged is fundamentally flawed and that is one reason I would reject models that have 3 stages of growth and hold all else constant or the H-model, another over used and rigid model, when it comes to payout ratios.").) Cornell seemed to intend to address this point in rebuttal testimony regarding "disappearing earnings" (Transcript at 1424.); however, the testimony that followed merely reiterated that Plaintiff's growth rates were too high, and the court cannot discern a refutation of Plaintiff's criticism. (Id. at 1424-30.)
Each party has identified significant, credible shortcomings in the other's DGM analysis. The parties have not attempted to present estimates that adjust for these issues, and the evidence does not leave the court equipped to determine any such adjustment. The court concludes that neither party's DGM estimate is entitled to any weight.
b. Cost of equity: CAPM (historical, ex post formula)
The parties agree that the CAPM is a widely accepted and widely used method to estimate the cost of equity. (See Ptf's Post-Trial Br at 31; Def's Ex A at 35.) See WSATA Handbook at III-20 (CAPM one of "two most widely used and recognized methods for estimating the cost of equity"). Under the standard CAPM formula (also referred to as the "historical" or "ex post" formula), the cost of equity estimate is (1) the "risk-free rate" of return on investment plus (2) the product of (a) the "beta" factor (a number representing the company's risk in relation to that of the average publicly traded company), and (b) the "market risk premium." (See Ptf's Post-Trial Br at 31; Def's Ex A at 35.) The market risk premium is the expected return in excess of the risk-free rate that an investor would demand; in other words, the sum of the risk-free rate and the market risk premium is the rate of return that a hypothetical investor could expect to derive by investing in a well-diversified portfolio of publicly traded stocks. (See id.) See also Brealey at 186, 205.
Cost of equity = Risk-free rate + (Beta * (market risk premium))
Market risk premium = market rate of return - risk-free rate
Each party derived estimates using this basic formula, as well as variations discussed below.
The parties do not disagree significantly regarding the beta factor to be applied to Plaintiff. Tegarden determined an estimated beta of 0.58; Morin and Eyre estimated 0.60. (See Ptf's Ex 1 at 62; Ptf's Ex 3 at 35; Def's Ex A at 35.) Both estimates are supported by substantial evidence, and the court finds that the beta of Plaintiff was 0.60 based on the findings of at least one expert for each party.
Tegarden and Eyre each presented one historical CAPM estimate that used nearly identical inputs, with very similar results. For that estimate, Tegarden and Eyre selected, as the risk-free rate, the long-term Treasury bond rate of 2.25 percent as of the assessment date. They then used a historical market risk premium of approximately 7.20 percent, based on readily available published figures. (See Ptf's Ex 1 at 62-64 (Tegarden 7.20%); Def's Ex A at 37 (Eyre 7.15%).) Not surprisingly, the resulting estimates of the cost of equity are nearly the same: Tegarden's estimated cost of equity based on these inputs is 6.43 percent; Eyre's is 6.54 percent. (See Ptf's Ex 1 at 64; Def's Ex A at 38.) There the agreement ends, however. While Eyre placed one-third reliance on this result, Tegarden essentially disavowed it.
In selecting his final cost of equity of 6.70%, Eyre gave equal weight to (1) his DGM estimate of 7.60%, (2) his historical CAPM estimate of 6.54%, and (3) his supply side CAPM estimate of 5.95%. (See Def's Ex A at 43.)
Tegarden gave this result "a whole lot less weight" when reconciling it with other indicators. (Transcript at 309.)
c. CAPM: Selection of risk-free rate
One reason Tegarden assigned little to no weight to his historical CAPM estimate was a concern that the federal government, as of the assessment date, had long been holding the Treasury bond rate below the rate that would have been set by the market. (See Transcript at 308.) Specifically, Tegarden testified that the federal government's practice of quantitative easing, starting around 2008, which included purchasing great quantities of its own bonds, "manipulated" the market by driving up the price, thereby reducing the yield to an artificial level. (See Transcript at 309, 1090; Ptf's Opening Post-Trial Br at 33.) Thus, although Tegarden's CAPM estimate applied the 2.25 percent long-term Treasury bond rate, he essentially rejected it. Morin shared Tegarden's view but addressed the issue by using a risk-free rate of 3.9 percent derived from forecasts of future long-term Treasury bond yields. (See Ptf's Ex 3 at 30-34, 51.) Morin gave about 16.7 percent weight to his CAPM estimate based on this higher risk-free rate (and based as well as on a "forward-looking" market premium discussed below). (Ptf's PostTrial Br at 35.)
Morin derived a cost of equity estimate of 8.9% using CAPM and a cost of equity estimate of 9.6% using a variation he referred to as an empirical CAPM. (See Ptf's Ex 3 at 51.) He assigned equal weight to six estimates, including the two foregoing CAPM estimates, two DGM estimates, and two estimates using electric utility risk premium data and the yield on long-term U.S. Treasury bonds. (See id. at 51-52.) Morin's final cost of equity estimate of 9.6% is the average of those six estimates.
Defendant did not refute Plaintiff's evidence that government actions resulted in lower actual long-term Treasury bond rates than otherwise might have occurred. However, Cornell disagreed that the government actions invalidated current actual yields as a proxy for a risk-free investment. Cornell declared in his rebuttal report: "The current yield on long-term Treasury bonds as of December 31, 2019, reflects the investors' required return on long-term Treasury bonds and investor expectations regarding treasury yields for the future as of that date." (Def's Ex I at 11; see also Transcript at 959 (referring to long-term Treasury bond rate; "it's a forwardlooking rate. The 2.25 is what investors expect over the next 20 years. So it's an average of short-term expectations rolled over for 20 years.").) Cornell testified: "just because you have a big buyer, it's still a market rate * * * the market is super deep. You can buy it. You can short it. You can trade derivatives. If I thought that price was wrong, I'd be in the market." (Transcript at 959.) Cornell, as well as Defendant's witness, Dr. Steven Kihm, also testified that a forecast of future Treasury bond rates is inherently less accurate than the market-generated actual yield as of the assessment date. (See Def's Ex I at 11.) Answering this criticism, Morin testified that there should be consistency; if other components of the valuation model use forecasted rates, the risk-free rate should be derived by forecasting as well. (See Transcript at 1206.)
The court notes that Cornell undermined this position to some extent by referring approvingly in his rebuttal report to analysis by Duff &Phelps that relied on a risk-free rate derived from forecasts of long-term Treasury note yields rather than current yields. (See Transcript at 1216-17 (Morin; asserting that Duff &Phelps used 3.9% risk-free rate), 1432 (Cornell)) (referring to Def's Ex I at 14-15 &n 11); see also Ptf's Ex 12 at 9 (Def's 2020 Electric Capitalization Rate Study; using Duff &Phelps "normalized" risk-free rate of 3.5%).) In rebuttal testimony, Cornell acknowledged the apparent inconsistency but pointed out that, "in conjunction, [Duff &Phelps] also used a lower equity risk premium of * * * 5 percent," which, together with Plaintiff's beta of 0.6, would result in an overall cost of equity of 6 percent--lower even than Eyre's estimated cost of equity. (Transcript at 1432.) The court does not understand that response. Cornell did not explain--and the evidence presented does not inform the court--why or how a risk-free rate that is invalid as a matter of principle because it is clouded by potential forecasting error can become acceptable by adjustment of another factor in the formula. Accordingly, nothing in this opinion should be read as a finding that a forecasted rate is per se invalid as a proxy for a risk-free rate under the CAPM.
Kihm agreed, writing that "the market knows everything investment professionals know about the economy, quantitative easing by the Federal Reserve, threats of inflation, the impact of the COVID-19 pandemic, tax law revision possibilities, and everything else that might affect interest rates. While no one can forecast the future direction of interest rates with a high degree of accuracy, the evidence is compelling that the market will weigh those factors much more appropriately than investment professionals." (Def's Ex N at 18.)
The court finds Defendant's position as to the risk-free rate more persuasive. Plaintiff did not dispute that the market for long-term Treasury bonds remained liquid despite quantitative easing, nor did Plaintiff contend that the federal government's participation as a buyer in the bond market changed the character of federal bonds as essentially risk free. Rather, Plaintiff argued that the risk-free rate would have been higher if the government had not acted in the bond market. However, based on two scholars frequently cited by the parties, the court understands that the rationale for using a risk-free rate in the CAPM is to identify a point of relative certainty premised on market confidence in the United States government's promise to repay the borrowed sum plus the amount of interest stated on the face of the bond. See Brealey at 177 n 18 ("After all, the U.S. government can always print money to pay off its debts."). (See Def's Ex E at 35 (quoting Dr. Aswath Damodaran) ("The only securities that have a chance of being risk free are government securities, not because governments are better run than corporations, but because they control the printing of the currency.").) Adjusting this actual rate based on forecasts of future yields introduces uncertainty that will vary in degree based on the forecaster's ability to predict market conditions and future government policies, which is contrary to the purpose of the risk-free rate.
The court concludes that the most appropriate risk-free rate is the rate of interest on longterm Treasury bonds as of the assessment date. The court finds that the federal government's practice of purchasing large volumes of bonds may have lowered the rate but did not make the market illiquid or the resulting rate other than risk free. Morin's proffered forward-looking rate relies on forecaster judgment and thus adds uncertainty to the rate set by the market. Therefore, the court finds that the risk-free rate for purposes of a CAPM analysis in this case was 2.25 percent.
d. CAPM: Tendency of historical, ex post method to understate cost of equity for low-beta stocks
Plaintiff's expert Morin testified that cost of equity estimates using the historical CAPM (6.43% by Tegarden (Ptf's Ex 1 at 62); 6.54% by Eyre (Def's Ex A at 38)) are too low because that method (which he referred to as the "plain vanilla CAPM") has been shown to systematically understate the cost of equity for low-beta businesses. (See Transcript at 1217 ("[T]he plain vanilla CAPM understates the return on low beta securities and overstates the return on high beta securities."); Ptf's Ex 29 at 6 (Shannon P. Pratt &Roger J. Grabowski, Cost of Capital: Applications and Examples 164 (3d ed 2008) ("CAPM cost of equity estimates for high-beta stocks are too high and estimates for low-beta stocks are too low relative to historical returns.").) Defendant's expert Cornell seemed to agree with this general criticism of the historical CAPM. (See Transcript at 1435 ("[T]he fact that the security market line is flatter than the pure theory projects, tends to be right.").) The court finds the evidence compelling on this point. There is no dispute that Plaintiff, with a beta of .60, is a "low-beta" business. (See, e.g., Ptf's Ex 29 at 4 (Brealey graphic illustrating range of betas from 0.5 to 1.5).) Accordingly, the court finds that the historical CAPM estimates of 6.43 percent to 6.54 percent are likely below Plaintiff's actual cost of capital.
For this reason, the court also rejects Eyre's alternative CAPM analysis referred to variously as a "supply side" or "SBBI Yearbook" analysis, which he employed out of concern that the historical CAPM formula "overstates the market risk premium," which necessarily would overstate the cost of equity. (Transcript at 436 (emphasis added); see also id. at 632.) In his supply-side variation on CAPM, Eyre used a market risk premium of 6.17% published in the SBBI Yearbook, in lieu of the historical market risk premium of 7.15% to 7.20% on which he and Tegarden largely agreed. Changing solely this component of the CAPM formula, Eyre determined a cost of equity of 5.95% in lieu of 6.54%. (See Def's Ex A at 39-40; Transcript at 432-36.) Cornell and Kihm proffered estimated forward-looking market risk premiums of 5.06% to 5.20%, and 5.8%, respectively. (See Def's Ex C at 18 (Cornell; citing Damodaran reports); Def's Ex H at 12 (Cornell, citing Damodaran, Duff & Phelps, Fernandez/Apellaniz/Acin); Def's Ex N at 32 (Kihm).)
e. Morin's proffered correction ("empirical" CAPM) Morin sought to correct for the tendency of CAPM to understate value by applying a variation known as the "empirical CAPM" or "ECAPM." In oral testimony he summed up this variation as follows: "[Y]ou take the risk-free rate and increase it by 25 percent of the market risk premium and you eliminate the beta, the slope of the market risk premium by 0.75 * * *." (Transcript at 190-91; see Ptf's Ex 3 at 40 (full formula).) Appling the ECAPM formula, while otherwise using the same inputs, Morin's ECAPM resulted in an increase in his estimated cost of capital by approximately 0.76 percent, compared to his CAPM (ECAPM 9.40% - CAPM 8.64%). (See Ptf's Ex 3 at 37-40, 85, 89-104.) Cornell testified that empirical CAPM is "no longer * * * considered as a candidate" to improve the original CAPM:
Defendant's witness Kihm described ECAPM as the CAPM formula, with some of the factors weighted. (See Transcript at 1050.)
Morin's estimated cost of equity using the CAPM (8.64%) is significantly higher than those of Tegarden (6.43%) and Eyre (6.54%) because, as discussed elsewhere, Morin used a risk-free rate of 3.9% (vs. 2.25% for Tegarden and Eyre) and a market risk premium of 7.90% (vs. 7.20% for Tegarden and 7.15% for Eyre).
"[M]ost courts use the standard model * * * So the idea that there's an ECAPM out there that's used is wrong, and the fact that the security market line is flatter than the pure theory projects, tends to be right. But * * * if you were to try to use that, you would have to update all the analysis using current data, which I don't think has been done."(Transcript at 1434-35.) In response to this criticism, Morin testified that empirical CAPM continues to be discussed as an alternative to CAPM, referring to excerpts from several treatises. (Transcript at 1217-21.) However, the court finds that the treatises merely support the fact that economists continue to criticize the "plain vanilla" CAPM formula as resulting in cost of equity estimates that are too low for low-beta businesses; the cited treatises do not refer to ECAPM as a way to correct that shortcoming. (See Ptf's Ex 29.) Even assuming that ECAPM has a track record of accurately correcting the disparity between the predicted results under CAPM and actual observed results over time, the court is not persuaded that ECAPM's insertion of seemingly static weightings into a CAPM formula that otherwise consists of dynamic components (the risk-free rate, beta, and the market risk premium) will result in a reliable estimate of the cost of equity. Based on a lack of evidence that ECAPM is widely accepted, the court will assign no weight to Morin's ECAPM estimates of the cost of equity.
Kihm does not criticize the use of ECAPM per se, but he rejected Morin's use of a forecasted risk-free rate and Morin's 7.90% market risk premium in both CAPM and ECAPM. (See Def's Ex N at 11.)
f. CAPM: Plaintiff's forward-looking, ex ante, method
Tegarden and Morin each presented cost of equity analyses that used an "ex ante" (forward-looking) market risk premium in the CAPM formula. Tegarden used a forwardlooking market risk premium of 10.77 percent, derived from the "market-weighted average of the cost of equity capital" (13.02%) less the current long-term Treasury bond rate of 2.25 percent. (Ptf's Ex 1 at 63.) For his average of the cost of equity capital, Tegarden used as a proxy the cost of equity for each company in the S&P 500 Index. (See Id.) Weighting the results by market value, he determined a weighted average cost of equity for the S&P 500 companies of 13.02 percent, subtracted the current long-term Treasury bond rate of 2.25 percent, and thus derived his market risk premium of 10.77 percent. (See id.) Applying the beta, he determined an 8.50 percent cost of equity estimate. (See Ptf's Ex 1 at 62.) In terms of relative usefulness, Tegarden assigned the most weight to his DGM cost of equity result (8.35%), less to his ex ante CAPM result (8.50%) and minimal weight to his historical CAPM result (6.43%). (See Transcript at 313.)
Eyre rejected Tegarden's and Morin's analyses and used a different forward-looking method, discussed further below. (See Def's Ex E at 32-33.)
Morin, too, used a forward-looking market risk premium, together with his forwardlooking risk-free rate discussed above. Morin's market risk premium was 7.9 percent, which is the average of the historical market risk premium (7.20%) and a prospective market risk premium derived from the S&P 500 Index (8.6%). (See Ptf's Ex 3 at 38; Transcript at 184-85, 1207; Ptf's Post-Trial Br at 34-35.) Morin derived his prospective market risk premium by determining an expected market return on aggregate S&P 500 equities of 12.5 percent (lower than Tegarden's 13.02%) and subtracting his risk-free rate of 3.9 percent (higher than Tegarden's 2.25%). (See Ptf' Ex 3 at 38.) Morin did not provide a rationale for his weighting of his "raw" forward-looking market-risk premium by averaging it with the historical market-risk premium. With these inputs, Morin determined a cost of equity of 8.7 percent, similar to Tegarden's 8.5 percent estimate. (See Ptf's Ex 3 at 39.)
Defendant's witness Cornell criticized both Tegarden and Morin for using short-term (3-to 5-year) forecasts of the return on equity of publicly traded, dividend-paying stocks as their market rate of return. These rates, Cornell testified, are "mathematically impossible" to sustain when compared to the long-term expected growth rate for the U.S. economy. (See Def's Ex I at 25.; Def's Ex J at 23; Def's Response at 14.) Cornell proffered as a more appropriate guidepost an estimated market risk premium of 5.20 percent contained in a publication by Professor Damodaran. (See, e.g., Def's Ex H at 12; Transcript at 1433.) Cornell also used the U.S. Gross Domestic Product (approximately 4%) as a benchmark. (See Def's Ex I at 7; see also Def's Ex N at 9 (Kihm); see also Def's Ex N at 30-31 (Kihm; using multi-stage estimate of S&P 500 growth that adjusts downward after five years).)
The same exhibit includes "Estimates of ERP as of December 31, 2019" by Duff & Phelps (5.00%) and Fernandez/Appellaniz/Acin (5.60%). (Def's Ex H at 12.)
Like Cornell, Kihm criticized as unsustainable the use of three-year to five-year forecasted returns on stocks in the S&P 500 index. (Def's Ex N at 9.) Kihm proposed that a "three-stage model * * * fixes the growth rate problem." (Transcript at 1086.) Under this model, Kihm blended (1) the same published growth rates on which Morin relied, (2) the "limiting long-run growth rate" (current GDP, sometimes averaged with the current inflation rate), and (3) the average of rates (1) and (2). (See Def's Ex N at 11, 31-32.) Kihm offered no explanation of important details, such as how he decided weights and averages among the various components of his formula. Nor did he suggest that this method is accepted by others. He referred to a Table 6 that purportedly summarizes his computations, but his written report omits that table. (See id. at 32.) The court assigns no weight to Kihm's proffered correction.
g. Effect of Brealey range
Both Morin and Cornell referred to a range of market risk premium rates described in the Brealey treatise. Morin endorsed the range stated in the 2006 edition of Brealey:
"In their authoritative corporate finance textbook, Professors Brealey, Myers, and Allen conclude from their review of the fertile literature on the MRP that a range of 5% to 8% is reasonable for the MRP in the United States. My own survey of the MRP literature, which appears in Chapter 5 of my latest textbook, The New Regulatory Finance, is also quite consistent with this range."(Ptf's Ex 3 at 38 (footnote omitted).) Cornell acknowledged that the Brealey treatise is "a very respected finance textbook"; he criticized two details relating to Morin's reliance on the Brealey range. (Def's Ex I at 13-14.) First, Cornell claimed that the range should be updated to no more than 7.0 percent or 7.1 percent, as stated in the 2011 edition of the treatise. The court rejects this criticism because the 2020 edition of Brealey continues to declare as "reasonable" a range of five percent to eight percent. The 2020 edition states:
"Many financial economists rely on the evidence of history and therefore work with a risk premium of about 7%. The remainder generally use a somewhat lower figure. Brealey, Myers, and Allen have no official position on the issue, but we believe that a range of 5% to 8% is reasonable for the risk premium in the United States."Brealey at 174; see also id. at 171 (using 7.7% average historical risk premium).
On the other hand, the court finds Cornell's second criticism well taken: that Morin failed to account for the fact that the range for market risk premium in Brealey is calculated as a percentage above the current rate for short-term Treasury bills rather than the rate for long-term Treasury bonds (2.25%) to which all experts in this case have referred in their CAPM analyses. (See Def's Ex I at 14.) See Brealey at 169 (calculating hypothetical 7.7% risk premium on common stocks as 11.5% rate of return on common stocks minus 3.8% rate of return on Treasury bills). Plaintiff did not refute this criticism. Cornell testified that "[a]s of December 31, 2019, the yield on 3-month Treasury bills was 1.52 percent * * *." (Def's Ex I at 14.) By this logic, the court finds that an apples-to-apples comparison requires that the range expressed in Brealey must be adjusted to refer to the same risk-free rate. When Brealey's range is adjusted to use the long-term bond rate (2.25%) as the risk-free rate, it becomes 4.27 percent to 7.27 percent (e.g., 5% - (2.25% - 1.52%) = 4.27%)). (See Def's Ex N at 32 (Kihm; discussing similar adjustment to Brealey market risk premium range).)
Morin's determined market risk premium of 7.9 percent exceeds Brealey's adjusted top rate of 7.27 percent, as does Tegarden's estimate of 10.77 percent. Cornell's recommended rate (5.2%) is near the bottom of the adjusted Brealey range; the Gross Domestic Product benchmark is below the bottom of the Brealey range. The court finds that the Brealey range, as adjusted, is not in dispute, and the court adopts it as the reasonable range for the market risk premium in this case.
h. CAPM: Findings and conclusions
The court now summarizes its findings regarding Plaintiff's estimated cost of equity under the CAPM:
(1) The appropriate risk-free rate is 2.25 percent, which is the yield on long-term Treasury bonds as of the assessment date.
(2) Plaintiff's beta is 0.60, as determined by Morin and Eyre. Plaintiff's beta is "low" because it only slightly exceeds the midpoint between a risk-free investment and the level of risk of an average company.
(3) The cost of equity indicated by a historical CAPM is within the range of 6.43 percent to 6.54 percent, as determined by Tegarden and Eyre, respectively. However, historical CAPM likely understates the cost of equity for low-beta companies. Therefore, the court will select a cost of equity higher than 6.54 percent if supported by the evidence.
(4) The market risk premium is not more than 7.27 percent, derived as follows:
a. Consistent with the testimony of Morin, and all of Defendant's experts, the range of reasonable rates of market risk premium is as stated in Brealey: five percent to eight percent, using short-term Treasury bills as the baseline.
b. Consistent with Cornell's and Kihm's unrefuted testimony, the court finds that Brealey's published range must be adjusted by subtracting the difference between the rate of return on long-term Treasury bonds and the rate on short-term Treasury bills.
c. Consistent with Cornell's unrefuted testimony, the court finds that the rate on short-term Treasury bills as of the assessment date was 1.52 percent.
d. The court computes the difference between the rate of return on long-term Treasury notes and the rate on short-term Treasury bills as 0.73 percent (2.25% - 1.52%), and the resulting adjusted range of reasonable rates of market risk premium as 4.27 percent to 7.27 percent.
(5) Within this range, the court determines the actual market risk premium as follows:
a. The court first determines whether the low-beta understatement effect in finding (3) above requires any narrowing of the market risk
premium range. Applying the inputs of a risk-free rate of 2.25 percent, a beta of 0.60, and the range of potential market risk premiums of 4.27 percent to 7.27 percent results in a range of the cost of equity from 4.81 percent to 6.61 percent. However, finding (3) above requires the lower end of this range to be raised to at least 6.54 percent. In turn, the range of reasonable market risk premiums must be raised to at least 7.15 percent (2.25% + (0.6 * 7.15%) = 6.54%). The court thus finds that the range of reasonable estimates of the market risk premium is 7.15 percent to 7.27 percent.
b. Plaintiff presented evidence supporting a forward-looking market risk premium of 7.9 percent, based on returns on publicly traded stocks, which would result in a cost of equity of 6.99 percent. Defendant criticized this evidence as relying on excessively high returns on publicly traded stocks. However, the alternatives Defendant proffered to the court (market risk premiums of four% or 5.2%) would result in respective cost of equity estimates of 4.65 percent and 5.37 percent, well below the too-low historical CAPM estimate (6.54%) that is at the bottom of the court's narrowed range. Plaintiff's evidence results in a cost of equity much closer to the limited reasonable range of 6.54 percent to 6.61 percent and is therefore more persuasive than Defendant's evidence. The court finds that the market risk premium was 7.27 percent.
(6) Plaintiff's cost of equity under the CAPM was 6.61 percent. (6.61% = 2.25% + (0.6 * 7.27%)
i. Cost of equity: "Build-up" or "risk premium" method
Only Plaintiff's experts Tegarden and Morin used a "build-up" method, also called a "risk premium" method, determining a cost of equity of 9.40 percent (Tegarden) and 10.2 percent (Morin). Under the build-up method, the cost of equity is equal to the sum of a risk-free rate (which includes an inflation premium) and a risk premium. (See Ptf's Ex 1 at 60.)
j. Build-up: Tegarden's method
Tegarden determined a risk-free rate of 4.30 percent. He testified that he derived this from the historical average yields to maturity of Plaintiff's own traded bonds (4.30%), after comparing two other categories of corporate bonds: (1) a "generic" set of corporate bonds issued by electric companies rated at BBB+ (3.81%), and (2) a second set of corporate bonds issued by electric companies (4.14%). (See Transcript at 294-302; Ptf's Ex 1 at 60-61, 171.) He determined a risk premium of 5.70 percent by identifying the published historical arithmetic mean return since 1926 for large-company stocks (12.1%) and subtracting the long-term corporate bond total return since 1926 (6.4%), which is a difference of 5.70 percent. (See Transcript at 293-95; Ptf's Ex 1 at 171.) Although the sum of 4.30 percent and 5.70 percent is 10 percent, Tegarden testified that he made a downward adjustment to 9.40 percent in the exercise of conservative appraisal judgment. (See Transcript at 302.)
k. Build-up: Morin's method
For his risk-free rate, Morin used the same forward-looking risk-free rate of 3.9 percent as under his CAPM analysis, derived from forecasts of future long-term Treasury bond yields rather than the current yield on long-term Treasury bonds as of December 31, 2019 (2.25%). (Ptf's Ex 3 at 25-28.) He determined a risk premium of 6.3 percent using two methods. First, he computed the actual realized return on equity capital for the S&P Utility Index for each year, using the actual stock prices and dividends of the index, and then subtracting the long-term Treasury bond return (the "income component of bond yields") for that year. (See Ptf's Ex 3 at 42-43.) The result was 6.3 percent. Second, he examined the historical risk premiums implied in the cost of equity decisions rendered by regulatory utility commissions over the 1986-2019 period for which data were available, relative to the contemporaneous level of the long-term Treasury bond yield. Although the average spread was 5.59 percent, he observed an escalating trend of the risk premium and determined a 6.3 percent risk premium by statistical analysis of that trend. (See Ptf's Ex 3 at 44-45.)
l. Build-up: Defendant's criticisms and court's analysis
As to the risk-free rate component, the court sees no basis to depart from the 2.25 percent risk-free rate adopted as part of the CAPM discussed above. Tegarden offered no explanation for substituting Plaintiff's own historical bond rate (4.30%), and Morin's forward-looking rate of 3.90 percent suffers from the same uncertainty that the court already has found inconsistent with the purpose of a risk-free rate.
As to the risk premium, the court finds Morin's use of actual realized return on equity for the S&P Utility Index a reasonable method. Because Morin used utility data only, his result is more persuasive than Tegarden's approach using all large-company stocks and bonds. Defendant's only criticism of Morin's S&P Utility Index method is its reliance on historical data. (See Def's Ex J at 48; Def's Ex N at 9, 16.) Kihm in particular testified that historical earnings on utility stocks neither predict future earnings nor inform as to whether actual past earnings matched investors' expectations at the time. (See Def's Ex N at 16.) Although this criticism has some merit, the court disagrees that Morin's approach is entitled to "no weight." (Def's Ex N at 8.) Both parties in this case refer to historical data in various contexts, recognizing that past and present data are the only data actually available. In applying the CAPM, historical market risk premiums were essential to Eyre's analysis. And as discussed above, Defendant's experts, including Khim, referred repeatedly to an entirely backward-looking market-to-book analysis to rebut Plaintiff's assertion that regulated utilities should receive an additional obsolescence deduction and to urge the court to find evidence of "real growth."
m. Conclusion as to build-up method
Taking into account Defendant's criticisms, the court adopts a modification of Morin's determination. The court finds that the build-up method results in an indicator of the cost of equity of 8.55 percent, consisting of a risk-free rate of 2.25 percent and a risk premium of 6.3 percent.
The court need not address Defendant's extensive criticism of Morin's use of regulatory cost of equity data, as that method resulted in the same risk premium as his use of S&P Utility Index data (6.30%).
n. Yield capitalization: Conclusion as to cost of equity
The court finds that its analysis and conclusions discussed above under the CAPM method and the build-up method are comparable in the quality of supporting data and logic. The court therefore assigns equal weight to the indicators under each method. The court finds that the cost of equity is 7.58 percent, consisting of the average of 6.61 percent (CAPM) and 8.55 percent (build-up).
o. Yield capitalization: Flotation costs
Tegarden's estimate of WACC includes an upward adjustment of approximately 0.2 percent for "flotation costs," defined as legal and underwriting charges a buyer would incur to actually acquire all of Plaintiff's property. (See Ptf's Ex 1 at 49, 66-70.) Eyre's appraisal includes no adjustment for flotation costs. The parties do not dispute that these charges exist, and the parties agree that, as a matter of arithmetic, they can be accounted for either as an increase in WACC or as a reduction in cash flow. Defendant, citing the WSATA Handbook, urges reducing cash flow. (Def's Opening Post-Trial Br at 22-23; see WSATA Handbook at III-30 to III-31.) As to this point, Plaintiff argues that it is "generally accepted in the appraisal and finance communities to make a flotation cost adjustment to the cost of capital * * *." (Ptf's Response at 22.) The court's own research, however, suggests that courts have not favored an adjustment to the cost of capital in property tax cases. The Minnesota Tax Court has repeatedly refused to increase the cost of capital for flotation costs. See, e.g., CenterPoint Energy Res. Corp. v. Comm'r of Revenue, No. 9252-R, 2022 WL 995851, at *14 (Minn Tax Mar. 16, 2022) ("We * * * do not adjust the cost of capital for flotation costs."); see also Colorado Interstate Gas Co. v. Property Tax Administrator Huddleston, 28 P.3d 958, 962 (Colo App 2000) (board of assessment appeals did not abuse its discretion by excluding flotation costs as "hypothetical"). In this case, the court declines to adjust the WACC for flotation costs. And having adopted Plaintiff's cash flow estimate as the better of two imperfect alternatives, the court declines to further adjust cash flow for flotation costs.
For an overview of regulatory tribunal decisions, see Kansas City Power & Light Co. v. State Corp. Com'n of State, 52 Kan.App 2d 514, 540-42, 371 P.3d 923 (2016).
p. Yield capitalization: Conclusion as to WACC
The court applies a weighting of 63 percent to the cost of equity (7.58%), and the court applies the remaining 37 percent weight to the cost of debt (4.25%). This results in a weighted average cost of capital of 6.35 percent.
The court computes WACC as follows:
Weighted Average Cost of Capital = Cost of Equity weighted at 63% + Cost of Debt weighted at 37%
Weighted Average Cost of Capital = (7.58% * 63%) + (4.25% * 37%)
Weighted Average Cost of Capital = (4.78%) + (1.57%)
Weighted Average Cost of Capital = 6.35%
q. Yield capitalization: Constant growth indicator of value
The court applies the constant growth formula set forth above, which results in an indicator of a real market value for Plaintiff's system of $17,328,565,352:
System Value
Cash Flow
WACC
Growth Rate
$17,328,565,352
= $1,100,000,000
x
(6.35% - 0%)
r. Yield capitalization: Discounted cash flow model
The parties recognize the discounted cash flow (DCF) method of yield capitalization under the income approach, which attempts to convert the sum of future years' cash flows into present value:
Value = Cash Flow Year 1 + (1+ Rate) + CF Y2 + (1+ Rate) + CF Y3 + (1+ Rate), etc. The length of the forecast period is a matter of appraiser judgment; a terminal value may remain after the end of the forecast period. See WSATA Handbook at III-14 to III-15. (See Ptf's Ex 1 at 39-40 (general discussion).) Eyre performed a discounted cash flow analysis, concluding an indicator of value of $23,500,000,000. (See Def's Ex A at 49-53.) Tegarden did not perform a discounted cash flow analysis; his discussion suggests that he viewed it as unnecessary given his assumption of zero real growth. (See Ptf's Ex 1 at 40.)
Eyre's DCF analysis relies entirely on the 2020 Plan as a forecast of future cash flow. (See Def's Ex A at 52-53.) His rate is the same WACC discussed above (5.60%). (See id.) In the analysis above, the court has found the 2020 Plan unreliable as a measure of cash flow, and Eyre offers no new evidence or information in the context of his DCF. The court also has reached its own conclusion of WACC at 6.35 percent. For those reasons, the court gives no weight to Eyre's DCF analysis.
Plaintiff alleges other errors specific to Eyre's DCF analysis, which the court need not consider. (See Ptf's Post-Trial Br at 49 (arithmetic error in terminal value); Ptf's Response at 27-29 (treatment of interest and other non-operating income).)
10. Direct Capitalization (EBITDA)
Only Eyre used the direct capitalization method, which determines a single year's cash flow, then applies a multiplier, resulting in value:
Confusingly, Eyre seems to use the terms "multipliers" and "multiples" interchangeably. (Def's Ex A at 55-57.) The court understands the product of the equation ("value") to be the only "multiple," and the court refers hereafter to "multipliers" to mean Eyre's numbers between 10 and 16.3 by which he multiplies one year's cash flow to arrive at his value estimate.
Value = one year's cash flow * multiplier
Eyre selected $2,100,000,000 as his cash flow estimate. (Def's Ex A at 56.) This amount is Plaintiff's earnings before interest, taxes, depreciation and amortization (EBITDA) as reported on Plaintiff's FERC 1 form, which Eyre "normalized" over the four years 2016 through 2019 and multiplied by the same two percent growth factor discussed above. (Def's Ex A at 30.) Eyre started with EBITDA rather than reported net income out of concern that federal tax reductions enacted during the four-year period as part of the TCJA made reported net income (which is computed after tax) less comparable as between the earlier and later years of the period. (See Def's Ex A at 53-54.)
Eyre's written narrative under the direct capitalization method states that he normalized EBITDA over five years, but the exhibit to which he refers shows only the four years 2016 through 2019. (Compare Def's Ex A at 55 with id. at 30.)
Eyre selected 11.0 as his multiplier. He derived this by first determining the enterprise value of ten guideline companies, using the market value of their debt and equity securities, less cash accounts. He then estimated each company's normalized EBITDA based on EBITDA as reported on each company's Form 10-K for the year ending December 31, 2019. Next, Eyre multiplied normalized EBITDA by a percentage published by ValueLine that represents an estimated growth rate for a cash flow that is not EBITDA, but which Eyre considers to be very close to EBITDA.
Having determined these base amounts, Eyre divided each guideline company's enterprise value by its normalized and growth-adjusted EBITDA. The result was a multiplier; for example, the lowest multiplier was that of Portland General Electric (10.0), and the highest multiplier was that of PNM Resources (16.3). The average multiplier was 13.3, and the median was 13.0. (Def's Ex A at 56.) Eyre selected a lower multiplier of 11.0 "to eliminate the effects of any non-regulated income that may be affecting the value of the guideline companies." (Id. at 57.) The result of Eyre's cash flow of $2,100,000,000 multiplied by 11 is an indicator of value of $23,100,000,000.
Eyre's written narrative erroneously states that the highest multiplier was 15.
Plaintiff responds with numerous arguments, first, that using EBITDA as a measure of cash flow unrealistically ignores the importance of after-tax cash flow to investors. (See Ptf's Response at 45.) Although Eyre's intention seems to have been to enhance comparability of cash flows of any one company as to pre- and post-TCJA years, Plaintiff argues that the use of EBITDA reduces the comparability among the companies Eyre included in his guideline group. On cross-examination, Eyre acknowledged that corporations look to after-tax income to determine the amount of dividends they will pay. (See Transcript at 610-11.) Tegarden presented a chart listing the effective tax rates of Eyre's guideline companies (based on 10-K data), which showed that those rates ranged from -37.80 percent to 13.80 percent. (See Transcript at 1350; Ptf's Ex 21 at 20.) (Plaintiff's rate was 7.0%.) The court finds that Eyre's use of EBITDA as a measure for comparing cash flow within a group of guideline companies, without any attempt to adjust for the consequence of differing effective tax rates, undermines the reliability of his result.
Plaintiff also argues that Eyre proceeded incorrectly by using the "market value of the publicly traded debt" when determining the enterprise value of the guideline companies. (Def's Ex A at 56; see Ptf's Post-Trial Br at 75; Ptf's Ex 21 at 21.) Defendant describes the point of Eyre's determination of enterprise value as "determining the market value of the company to the typical purchaser." (Def's Opening Post-Trial Br at 53.) According to Defendant, "[t]his requires that the debt be valued at market value * * *." (Id.) However, Ross testified that a purchaser acquiring all the debt and stock of Plaintiff would not buy the debt from bondholders at market price and retire it, but instead would assume the debt. (See Ptf's Ex 21 at 19.) This is because "most of the debt held by companies[ ] such as PacifiCorp[ ] is first mortgage bonds. And * * * each of those first mortgage bonds are subject to early redemption provisions called 'make [whole] calls.' It's financially punitive to retire the debt early." (Transcript at 1353.) When assuming the debt, a buyer would become subject to the debt at its book or par value. (See Ptf's Ex 21 at 21.) Tegarden presented a table showing that the book value of the debt of each of Eyre's guideline companies was below the market value used by Eyre, and that Eyre's use of the market value of debt caused the collective enterprise value of the 10 guideline companies to be overstated by $5,454,000,000, which in turn caused Eyre's EBITDA multipliers to be overstated. (See id.) Defendant argued that the use of book value of the debt "would be a mismatch," but Defendant presented no evidence or reasoning to refute Tegarden's evidence that an actual acquirer of a utility such as Plaintiff or any of the guideline companies would assume the debt at book value. (See Def's Reply at 17 n 12.) The court finds that Eyre's use of the market value of debt materially undermines the reliability of his result.
Plaintiff asserts that Eyre miscalculated EBITDA for five of the ten guideline companies; Tegarden testified that Eyre incorrectly started with "operating income" and that the correct approach would have been to start with "net income." (Transcript at 1357; see Ptf's Post-Trial Br at 75-76; Ptf's Ex 24; Transcript at 612-17.) According to Tegarden, this error caused Eyre to "not capture the full level of income as the starting point." (Transcript at 1357.) Tegarden presented a table showing that this error understated EBITDA by between 10 percent and 28 percent for each of the five comparison companies. (See Ptf's Ex 21 at 22.) Defendant argues that Eyre's method was an attempt to ensure that the EBITDA he determined contained only the regulated operating income of the guideline companies. (Def's Opening Post-Trial Br at 53.) But Plaintiff responds that the result is a mismatch between total company equity value in the denominator and something less than total company EBITDA in the numerator. (See Ptf's Response at 48.) Plaintiff also points out that Eyre selected 11 as his multiplier (a lower number than the 13.3 average multiplier of his guideline companies), without providing any computations showing why he chose 11, rather than 13.3 or some other number. (Ptf's PostTrial Br at 72.) Eyre's stated reason for that selection was "to eliminate the effects of any nonregulated income that may be affecting the value of the guideline companies." (Def's Ex A at 57.) The court finds that Eyre either tried to remove nonregulated amounts twice or simply did not act with transparency or care in computing EBITDA; this error undermines the reliability of Eyre's result.
Tegarden's Exhibit 24 is an excerpt from a Form 10-K for one of Eyre's comparator companies, ALLETE, Inc. The consolidated income statement shows that "operating income" is computed before "net income." (See Ptf's Ex 24 at 2.) "Net income" includes items not found in "operating income," such as (on this sample) "interest expense," "equity earnings," "gain on sale of U.S. Water Services," and "Other." (Id.)
Plaintiff levels other criticisms, some of which overlap with arguments in connection with other methods and computations. (See Ptf's Post-Trial Br at 72-76; Ptf's Response at 4449.) The court finds it unnecessary to address those points here because, for the reasons stated above, the court finds Eyre's direct capitalization/EBITDA approach substantially unreliable for the reasons stated above and will assign it no weight.
11. Conclusion of System Value Under Income Approach
Having rejected other proffered methods under the income approach, the court adopts its constant growth indicator of value, determined using the CAPM and build-up (risk premium) methods of determining the cost of equity. As set forth above, that indicator is a real market value of $17,328,565,352 for Plaintiff's system.
C. Stock and Debt Approach
The stock and debt approach is a proxy for a sales comparison approach to value that can be useful when comparable sales are not available. See Level 3 Communications v. Dept. of Rev., 368 Or. 303, 320, 490 P.3d 149 (2021). Consistent with OAR 150-308-0240(2)(a), Eyre performed a stock and debt analysis, concluding an indicator of value in the amount of $25,863,750,000. (See Def's Ex A at 57-60.) However, because Plaintiff's stock is not publicly traded, Eyre took an indirect approach that involved comparing the price and earnings of stock of other utilities. After that step, he also was required to make adjustments to remove the value of property not used in Plaintiff's electric business. Eyre acknowledged "the subjective nature of the adjustment process and the equity valuation * * *." (Id. at 62.) For that reason, he assigned no weight to his result. Tegarden stated that he, too, considered the sales comparison approach and specifically the stock and debt approach, but he found neither approach useful because Plaintiff's stock is not publicly traded. (See Ptf's Ex 1 at 74-75.) He therefore did not undertake a formal analysis.
The court sees no basis to analyze at length an approach upon which neither party relies. Accordingly, the court assigns no weight to Eyre's indicator under the stock and debt approach.
D. Conclusion as to Value of Plaintiff's System
Having considered the parties' positions under the cost, income and sales comparison (stock and debt) approaches, the court concludes that neither party has carried the burden of proving the real market value for which it advocates. However, the evidence in total provides ample support for a conclusion of real market value. The court exercises its authority under ORS 305.412, which states:
"When the determination of real market value * * * is an issue before the tax court, the court has jurisdiction to determine the real market value or correct valuation on the basis of the evidence before the court, without regard to the values pleaded by the parties."
As discussed above, in this case the court relies entirely on the income approach. The court determines that the real market value of Plaintiff's system was $17,328,565,352 as of January 1, 2020.
IV. ALLOCATION TO OREGON
The court now considers Plaintiff's assertion, as part of its valuation claim, that Defendant erroneously and improperly used an allocation formula that overestimates the real market value of Plaintiff's property sitused in Oregon. (See Ptf's 1st Amend Compl at 4.) The formula in Defendant's rule consists of three factors: a "production" factor, a "generation" factor, and an "other plant" factor. Each factor relies in large part on the "original cost" of a portion of the system. The rule does not take into account any measure of depreciation. The first two factors, however, give some weight to components other than the cost of property, namely, the property's capacity to generate electricity and its actual generation in the prior year (weightings of 10% and 15%, respectively, in the production factor) and sales and revenues (weightings of 10% and 40%, respectively, in the distribution factor). Applying this formula, Eyre determined an Oregon allocation percentage of 16.4995 percent. (See Def's Ex A at 2; Transcript at 392; see also Def's Ex K (single-page sheet applying rule formula) and Transcript at 903-07.) If Eyre's allocation percentage of 16.4995 were applied to the court's concluded real market value of Plaintiff's system ($17,328,565,352), the result would be a real market value of $2,859,126,640.
The production factor is a percentage based on the sum of (a) the ratio the Oregon portion of the original cost of the production plant bears to the entire system's production plant, times 75%; (b) the ratio the Oregon portion of the system's electricity generation capacity (in kilowatts) bears to the entire system's generation capacity, times 10%; and (c) the ratio the Oregon portion of the system's energy generation (in megawatt hours) during the prior year bears to the entire system's generation, times 15%. See OAR 150-308-0590(1). The distribution factor is a percentage based on the sum of (a) the ratio the Oregon portion of the original cost of the distribution plant bears to the entire system's distribution plant, times 50%; (b) the ratio the Oregon portion of the system's energy production sold in the prior year (in kilowatt hours) bears to the system's total such revenue, times 10%; and (c) the ratio the Oregon portion of the system's revenue from the sale of energy bears to the system's total energy sale revenue, times 40%. See OAR 150-308-0590(2). The other plant factor is based on the ratio of original cost of the Oregon portion of the remaining plant to the total original cost of the remaining plant. See OAR 150-308-0590(3).
Plaintiff asserts that, in total, the original cost of property accounts for approximately 85% of the ultimate allocation percentage, while generating capacity, production, and megawatts distributed and sold account for the remaining 15%. (Ptf's Post-Trial Br at 78 n 53; see Def's Ex K.)
Eyre's percentage is close to the percentage Defendant apparently applied in its O&O: $3,180,000,000 Oregon real market value ÷ $19,500,000,000 system value = 16.3077%. (See Ptf's Compl, Ex 2.)
Plaintiff presented uncontested evidence that its property in Oregon is, on average, older and more heavily depreciated than Plaintiff's property system-wide. Tegarden testified that Plaintiff had made a relatively greater share of new investment in states other than Oregon in recent years. (See Transcript at 73.) He presented a table showing that, in the overall system, net book value was approximately 65 percent of overall cost, while the net book value of Oregon property was approximately 52 percent of original cost. (See Ptf's Ex 15.)
Plaintiff contends that basing the allocation percentage heavily on original cost causes Oregon to "import value" from other states for taxation. (Ptf's Post-Trial Br at 78.) The basis for this argument is that, "as a regulated utility, [Plaintiff] is only allowed to earn a return on the net book cost of its assets." (Id. at 79.) Therefore, according to Plaintiff, the Oregon property contributes relatively less to earnings than the proportion of value allocated to it. Plaintiff argues that this violates statutory requirements that the manner of allocation be "fair" and "just," and that the allocation determine a "fair proportion" of Plaintiff's overall property. (See id. at 79-82 (citing ORS 308.555, 308.550).)
Plaintiff quantifies its position by arguing that application of Defendant's formula "results in a 28 percent premium being added to the value of unitary property located [in] Oregon." (Ptf's Reply at 15-16.) Plaintiff's expert Tegarden derives this "premium" by dividing the value Eyre allocated to Oregon under Defendant's rule ($3,712,400,000) by the net book value of Plaintiff's operating property in Oregon ($2,773,846,275), which yields $1.34 of value allocated to Oregon for every $1 of net book value of property in Oregon. (See Ptf's Ex 21 at 26-27.) Tegarden then compares the $1.34 amount to the ratio that Eyre's concluded value of the property in states other than Oregon ($18,787,600,000) bears to the net book value of the property in states other than Oregon ($17,962,039,938), which yields $1.05 of value allocated to states other than Oregon for every $1 of net book value of property outside Oregon. The premium is the difference between the two results ($1.34 - $1.05 = $0.29).
Tegarden also found that the premium becomes $0.25 if the comparison is between the $1.34 Oregon-only value per $1 of net book value vs. the ratio that Eyre's concluded value for all states ($22,500,000,000) bears to the net book value of operating property in all states ($20,735,886,213). (See Ptf's Ex 21 at 27.)
Plaintiff asks the court to apply an alternative allocation formula that consists of the average of two factors: (1) "gross investment" in Oregon, which consists entirely of the original cost of property in Oregon ($5,221,374,481) divided by gross investment in the entire system ($31,138,058,548), yielding a ratio of 16.7685 percent; and (2) "net investment" (net book value) in Oregon ($2,692,080,863) divided by net investment in the entire system ($20,267,281,826), a ratio of 13.2829 percent. The average of the two factors is Plaintiff's proposed allocation percentage of 15.0257 percent. (See Ptf's Ex 1 at 3; Transcript at 331-35.) Compared to Eyre's rule-based percentage of 16.4995 percent, Plaintiff thus asks the court to reduce the allocation percentage by 1.4738 percent. If Plaintiff's allocation percentage of 15.0257 were applied to the court's concluded real market value of Plaintiff's system ($17,328,565,352), the result would be a real market value of $2,603,738,244, a reduction of $255,388,396 in the real market value of Oregon-situs property, or 8.93 percent.
The principal Oregon Supreme Court case construing Oregon's statutes governing allocation of value for central assessment purposes is Southern Pacific Trans. Co. v. Dept. of Rev., 302 Or. 582, 732 P.2d 18 (1987) (SoPac II); see also Alaska Airlines, Inc. v. Dept. of Rev., 307 Or. 406, 769 P.2d 193 (1989) (upholding airline allocation formula that included "flyover" time over Oregon in numerator of time-based factor). The property at issue was a "bridge" railroad owned by a corporation known as Cottonbelt that was a 99.7 percent subsidiary of the plaintiff Southern Pacific Transportation Company. See Southern Pacific Transp. Co. v. Dept. of Rev., 295 Or. 47, 49, 664 P.2d 401 (1983). As a bridge railroad, Cottonbelt connected traffic that originated with other railroads, which tends to be a more profitable business than the terminal and switching activities conducted by Southern Pacific on its directly held lines. See id. at 50. The plaintiff argued that Cottonbelt's property should be excluded from the system. This court agreed and went on to determine an allocation percentage for the system excluding Cottonbelt, based on relative property, "line haul," and "terminal activity." See Southern Pacific Transportation v. Dept. of Rev., 9 OTR 481, 498-99 (1982). Defendant appealed, and the Supreme Court reversed, concluding that Cottonbelt's property should not be excluded from the plaintiff's system. Rather than determine the value of the plaintiff's Oregon-situs property outright, however, the Supreme Court remanded the case to this court to reexamine the allocation percentage, given the large amount of value added to the system pursuant to the appellate decision. See SoPac II, 302 Or at 590 n 9. On remand, this court declined to change the allocation percentage, finding as a factual matter that the property of Cottonbelt was physically similar to that of Southern Pacific, with no "expensive terminals or other properties * * * 'dissimilar' from the rest of the unit." Southern Pacific Trans. Co. v. Dept. of Rev., 10 OTR 80, 84 (1985).
The plaintiff appealed to the Oregon Supreme Court, seeking an adjustment to the allocation formula on the grounds that Cottonbelt, as a relatively more profitable bridge railroad, made a "disproportionate contribution to the value of the system." SoPac II, 302 Or at 587. The Supreme Court affirmed this court's adherence to the allocation percentage determined at trial. The court found it a "fundamental flaw" to seek to determine the profitability of Cottonbelt as a standalone enterprise because Cottonbelt's property was used in a single, unitary business together with the property of Southern Pacific itself. Id. at 590-91. Apart from that major issue, the court found that the plaintiff had failed to prove that this court's apportionment formula had produced "an unfair result," pointing out that the factors in the formula reflected "investment in, and usage of, the various parts of the railroad system," including "[l]ine-haul tonnage and terminal activity, which Southern Pacific argues are the primary determinants of profitability * * *." Id. at 592.
Plaintiff's argument in this case has some similarity to that of Southern Pacific, as Plaintiff argues that its allocated Oregon real market value is disproportionate to the Oregon net book value on which it is allowed to earn a return. (See Ptf's Post-Trial Br at 79-80; Ptf's Reply at 15-16.) However, in contrast to Southern Pacific's facts, Plaintiff's asserted unfairness does have a basis in the physical characteristics of Plaintiff's Oregon operating property: it is, on average, older than the property elsewhere. Without deciding whether this case is factually distinguishable from Southern Pacific, the court turns to the standards for apportionment stated in that case and in the United States Supreme Court decision on which the Oregon Supreme Court primarily relied, Norfolk &Western R. Co. v. Missouri Tax Com., 390 U.S. 317, 88 S.Ct. 995, 19 L.Ed.2d 1201 (1968). In Southern Pacific, the court equated the Oregon statutory requirements of a "reasonable method," and a "proper" or "fair proportion," with the applicable requirements under the United States Constitution. SoPac II, 302 Or at 588 ("these rather vague [statutory] directives * * * are essentially a codification of the limits imposed by the due process and commerce clauses of the United States Constitution[, which themselves] are much alike."). The court turns, therefore to the United States Supreme Court's analysis in Norfolk & Western.
The issue in Norfolk &Western was the validity of a property tax allocation formula. The plaintiff acquired all of the rolling stock of Wabash Railroad Company in 1964, pursuant to a lease that required the plaintiff to pay applicable property taxes. See 390 U.S. at 320. Although there was no material change in the rolling stock or in the extent of its use in Missouri for the year after the acquisition, the state taxing authority determined an assessed value for the rolling stock of $19,981,000, more than double the assessed value for the year before the acquisition ($9,177,683). See id. at 320-22. The state arrived at this amount by "follow[ing] the literal command of the statute," multiplying the determined system value by 8.2824 percent, based on miles of track within the state vs. systemwide. The Court reaffirmed that the United States Constitution permits a state to determine the value of in-state property based on its "'enhanced value * * * through its organic relation to the system.'" Id. at 323 (quoting Pullman Co. v. Richardson, 261 U.S. 330, 338, 43 S.Ct. 366, 67 L.Ed. 682). However, the Court described the standards set by prior cases as nevertheless requiring any allocation formula to "bear a rational relationship, both on its face and in its application, to property values connected with the taxing State." Id. at 325; see also SoPac II, 302 Or at 589. The Court elaborated on this standard with the following points:
• "Considerable latitude." Because states may consider "going-concern value," which is "not susceptible of exact measurement," the states "have been permitted considerable latitude" in devising allocation formulas. 390 U.S. at 324; see also SoPac II, 302 Or at 589 ("wide latitude").
• Value need not be "precise." "A number of such formulas have been sustained by the Court, even though it could not be demonstrated that the results they yielded were precise evaluations of assets located within the taxing State." 390 U.S. at 324. The cases "do not require any close correspondence" between the formulary result "and the value of property actually located in the State * * *." Id. at 327. "We repeat that it is not necessary that a State demonstrate that its use of the mileage formula has resulted in an exact measure of value." Id. at 329. See also SoPac II, 302 Or at 591 ("Value assessment of the parts of a unit is necessarily a somewhat arbitrary process that must rely upon factors that are
quantifiable: e.g., miles of track, terminal activity, track usage and property investment).
• "Gross distortion" is not permissible. Missouri's "rigid application of the mileage formula led to a grossly distorted result." 390 U.S. at 326. "[W]hen a taxpayer comes forward with strong evidence tending to prove that the mileage formula will yield a grossly distorted result in its particular case, the State is obliged to counter that evidence or to make the accommodations necessary to assure that its taxing power is confined to its constitutional limits." Id. at 329; see generally Jerome R. Hellerstein & Walter Hellerstein, I State Taxation ¶ 8.16[5] (explaining standard for determining percentage of distortion).
• Convenience and wide applicability of a formula do not justify distortion. "[N]either does the Constitution tolerate any result, however distorted, just because it is the product of a convenient mathematical formula which, in most situations, may produce a tolerable product." Id. at 327.
Applying these standards, the Court concluded that the plaintiff must prevail, primarily because the increase in value was large and the evidence supporting it was lacking. See id. at 328 ("The difference between the assessed value and the actual value as shown by the evidence to which we have referred is too great to be explained by the mere assertion, without more, that it is due to an assumed and nonparticularized increase in intangible value.").
This court now applies the standards set by Norfolk & Western and incorporated by Southern Pacific II. The facts regarding the age difference as between Plaintiff's property within and without Oregon are uncontested; however, the court concludes that, as a matter of law, Plaintiff has not shown a "gross distortion" under Defendant's formula. The difference in allocated value in this case does not approach the 100 percent increase in Norfolk & Western. See 390 U.S. at 326. As shown, if the court in this case seeks to isolate the effect of Plaintiff's requested allocation formula change, that change translates to a reduction in the tax base (real market value of Oregon-situs property) of approximately nine percent. Cf., e.g., Hans Rees' Sons v. No. Carolina, 283 U.S. 123, 135, 51 S.Ct. 385, 75 L.Ed. 879 (1931) (striking down, in an income tax case, state's formula as "out of all proportion" to in-state business; finding more than a mere "negligible criticism[ ]" where state determined an apportioned tax base 269% higher than amount reported on taxpayer's return under separate accounting by type of income); Moorman Mfg. Co. v. Bair, 437 U.S. 267, 271 n 4, 98 S.Ct. 2340, 57 L.Ed.2d 197 (1978) (upholding, in an income tax case, state's apportionment formula based solely on in-state sales despite resulting increase in tax base of approximately 48%); Container Corp. v. Franchise Tax Bd, 463 U.S. 159, 174 n 11, 175 n 12, 184, 103 S.Ct. 2933, 77 L.Ed.2d 545 (1983) (upholding, in an income tax case focusing on foreign commerce, state's broad-based formula as "avoid[ing] the sorts of distortions that were present in Hans Rees' Sons, Inc."; taxpayer did not argue distortion, where increase in tax base averaged approximately 14%); see also General Dynamics Corp. v. Sharp, 919 S.W.2d 861, 869 (1996) ("The 28% difference [of tax base for Texas' franchise tax base] in this appeal is simply insufficient to show a grossly distorted result that is out of all proportion with the percentage actually apportioned."); but see Mississippi State Tax Comm'n v. ANR Pipeline Co., 806 So.2d 1081 (Mississippi 2001) (striking down, in a property tax case, state's allocation formula based entirely on original cost, allowing plaintiff's alternative formula of 50-50 original cost and depreciated cost, where result reduced tax base by 16.53%).
Plaintiff at times appears to argue for a different standard based on the emphasized language in the following passage in Southern Pacific:
"Of course, this court makes a more searching review of the factual bases of an apportionment formula than does the U.S. Supreme Court. We exercise de novo review. ORS 19.125(3), 305.445. To prevail, Southern Pacific need only prove to us by a preponderance of the evidence, ORS 305.427, that the assessment of its Oregon property was 'unfair,' i.e., that Oregon has taxed extraterritorial value. In practice, however, Southern Pacific has a difficult burden of proof because of the difficulty of isolating the value of its Oregon property from the value of its property in other states. While it need not prove that the assessed value of its Oregon property bears no 'rational relationship' to its 'true' value, the very difficulty of ascertaining that 'true' value makes our analysis in practice not so very different from that that the U.S. Supreme Court would employ."SoPac II, 302 Or at 590. (See Ptf's Post-Trial Br at 5, 77, 82 n 57.) This court interprets the Oregon Supreme Court's reference to taxing "extraterritorial value" as encapsulating all of the standards in Norfolk & Western, which this court outlines and applies above. See id. at 589 (citing Norfolk & Western) ("The constitutional requirement of fair apportionment in property taxation means that a state may not tax extraterritorial value.").
The Oregon Supreme Court in Southern Pacific also identified other weaknesses with the plaintiff's criticism of Defendant's formula. See SoPac II, 302 Or at 592. The court concluded that Defendant's formula incorporated components (in that case, "line-haul tonnage" and "terminal activity") that sought to measure some of the same determinants of profitability on which the plaintiff based its argument. Id. Similarly, in this case, Defendant's formula incorporates sales and revenue, in addition to giving modest weight to generation capacity and actual generation. Plaintiff has not persuaded the court that its proposed reliance on net book value would produce an allocation fairer than Defendant's. This is particularly the case because Plaintiff's proposed "net investment" factor is based on "accounting" depreciation, which Plaintiff itself has criticized as driven by conventions and inconsistent with the facts on the ground. (See Ptf's Ex 9 (illustrating "depreciation continuing well past the end of the Carbon plant's service life").
The court finds that Defendant's allocation percentage of 16.4995 percent applies in this case.
V. CONCLUSION
Based on the foregoing, applying the allocation percentage of 16.4995 percent to Plaintiffs system value of $17,328,565,352, the court determines that the real market value of Plaintiffs property sitused in Oregon, as of January 1, 2020, was $2,859,213,283. Now, therefore, IT IS THE OPINION OF THIS COURT that the real market value of Plaintiff s property sitused in Oregon, as of January 1, 2020, was $2,859,213,283.