Minnesota Energy Resources Corporation, Relator v. Commissioner of Revenue, Commissioner of Revenue, Relator v. Minnesota Energy Resources Corporation, A15-422

886 N.W.2d 786

Opinion

OPINION

STRAS, Justice.

In a proceeding before the Minnesota Tax Court, Minnesota Energy Resources Corporation (MERC) challenged the Commissioner of Revenue’s valuation of its natural-gas pipeline distribution system for the years 2008 through 2012. With the exception of 2012, the lone year in which it increased the assessed value of the pipeline distribution system, the tax court reduced the Commissioner’s valuation and ordered the Commissioner to recalculate MERC’s tax liability. Both parties appeal from the tax court’s order and raise a variety of challenges to the tax court’s findings and conclusions. For the reasons that follow, we affirm the tax court’s decision in part, reverse it in part, and remand to the tax court for further explanation of the beta factors it used to calculate MERC’s cost of equity and to reconsider whether external obsolescence impacted the pipeline distribution system’s market value.

I.

MERC, a wholly owned subsidiary of Integrys Energy Group, Inc. (Integrys), owns a natural-gas pipeline distribution system in Minnesota. During the tax years at issue, 2008 through 2012, MERC delivered natural gas over 3,600 miles of pipeline to approximately 205,000 customers in 50 Minnesota counties. As a regulated utility, MERC’s pipeline distribution system is taxable personal property under Minn.Stat. § 273.33 (2014).

MERC’s pipeline distribution system stretches south from Canada across several states, including Minnesota. Each year, the Commissioner of Revenue determines a market value for MERC’s pipeline distribution system, which includes distribution pipes, gas mains, gate stations, gas meters, distribution-regulation stations, gas valves, and odorizing equipment. See MinmStat. § 273.33, subd. 2. The Commissioner uses information provided by MERC to make her calculations. See Minn.Stat. § 273.371, subd. 1 (2014). After calculating the total market value of MERC’s pipeline distribution system within Minne sota, the Commissioner apportions the value among the taxing districts through which it passes. See Minn.Stat. § 278.33, subd. 2; see also Minn. R. 8100.0200 (2015) (“[B]y the process of apportionment, the portion allocated to Minnesota is distributed to the various taxing districts within the state.”). Each district then assesses MERC’s personal property based on the share allocated to it by the Commissioner. See Minn.Stat. § 273.062 (2014). MERC’s real property, by contrast, is assessed separately by the county or taxing district in which each parcel is located. See Minn. Stat. § 273.17, subd. 1 (2014).

Before the tax court, MERC challenged the Commissioner’s 2008 to 2012 valuation of the pipeline distribution system. In support of its position that the Commissioner’s valuation was excessive, MERC presented an expert report and testimony from Kevin Reilly of American Appraisal Associates, Inc. The Commissioner relied on the expert opinion of Brent Eyre, an independent accredited senior appraiser with a background in property-tax valuation, to support an even higher valuation of MERC’s pipeline distribution system than the amount originally calculated by the Commissioner.

Following a 4-day trial, the tax court issued findings of fact and conclusions of law, in which it found that Reilly’s report and testimony were sufficient to overcome the presumptive validity of the Commissioner’s valuation. See Minn.Stat. 272.06 (2014). It then conducted its own valuation of -MERC’s property; based on the relevant law and its consideration of the testimony of both experts. For each of the years from 2008 to 2011, the court determined that the market value of MERC’s property was lower than the Commissioner’s valuation. For 2012, it reached the contrary conclusion, deciding that the Commissioner had undervalued MERC’s pipeline distribution -system by approximately $13 million. In valuing MERC’s property, the court used a combination of two of the three approaches to valuing property, the cost and income approaches, and rejected the third approach, the market approach, after determining that it would not lead to an accurate assessment of market value. The court also deducted the value of nontaxable intangible assets and working capital on the basis that neither is taxable under Minnesota law, a point on which the parties disagree. The following table summarizes the Commissioner’s original valuation, the valuations proposed by both experts, and the market-value determination of the tax court, for each taxable year:

Commissioner’s Eyre’s Reilly’s Tax Court’s Taxable Apportionable Apportionable Apportionable Apportionable , Year_Value_Value_Value_Value_

2008_$118,247,871_$199,951,677_$51,461,168_$94,732,200

2009_$112,627,661_$231,954,372_$65,250,150_$102,981,800

2010_$144,628,839_$258,799,869_$99,360,276 $131,233,100

2011_$155,934,300_$271,870,280_$106,518,546_$144,747,800

2012_$161,525,900 $273,892,276 $120,510,785 $174,125,500

Both MERC and the Commissioner appeal from the tax court’s decision. MERC challenges four decisions made by the tax court: its failure to adopt a company-specific risk factor, its rejection of the buildup method, its lack of explanation of the beta factors it applied, and its adoption of the Eurofresh standard for proving external obsolescence. Eurofresh, Inc. v. Graham Cty., 218 Ariz. 382, 187 P.3d 530, 535, 538 (Ariz.Ct.App.2007). We will explain the background principles underlying each of-these challenges in more detail below.

The Commissioner, by contrast, challenges only two aspects of the tax court’s decision. She objects to the deductions for intangible assets and working capital and asserts that the tax court clearly erred by rejecting the market approach in its entirety without at least considering the price paid by Integrys when it purchased MERC in a 2006 arms-length sale. We consolidated the two appeals, designating MERC as the appellant for briefing and oral argument. We now address the challenges to the tax court’s decision, beginning with those raised by MERC and then ■turning to the Commissioner’s arguments.

II.

Our review of the. tax court’s decision is limited and deferential. Cont’l Retail, LLC v. Cty. of Hennepin, 801 N.W.2d 395, 398 (Minn.2011). Specifically, “[w]e review tax court decisions to determine whether the tax court lacked subject matter jurisdiction, whether the tax court’s decision is supported by the evidence in the record, and whether the tax court made an error of law.” Hohmann v. Comm'r of Revenue, 781 N.W.2d 156, 157 (Minn.2010). More generally, we review the tax court’s legal determinations de novo and its factual findings for clear error. Cont’l Retail, 801 N.W.2d at 398. With respect to the tax court’s valuation of the property, we defer to the tax court’s determination unless it clearly misvalued the property or failed to explain its reasoning. Id. at 399.

A.

MERC’s first challenge is to the tax court’s decision to reject the application of a company-specific risk factor to MERC’s cost of equity; The cost of equity is one of the components each expert used to calculate the value of MERC’s pipeline distribution system under the income approach. The court, as well as both experts, estimated value using direct capitalization, one of two methods of determining value under the income approach. The direct-capitalization approach “convert[s] a single year’s income expectancy into” an indication of market value by dividing the estimate of a single year’s net operating in,come by a capitalization rate. Appraisal Institute, The Appraisal of Real Estate 491 (14th ed.2013); see also Eden Prairie Mall, LLC v. Cty. of Hennepin, 797 N.W.2d 186, 195 (Minn.2011) (explaining the direct-capitalization approach).

The parties’ disagreement in this case focuses on the capitalization rates applied by the tax court, and in particular, the cost of equity it used to determine each year’s rate. The tax court calculated the capitalization rates by estimating both the cost of debt and the cost of equity for each taxable year, based on the straightforward principle that most businesses, including utilities, finance the purchase of property through a combination of debt and equity. See In re Minn. Power & Light Co., 435 N.W.2d 550, 559 (Minn.App.1989). Applying this principle, the tax court multiplied the percentage of equity by the cost of equity and the percentage of debt by the cost of debt using the figures submitted by Eyre, the Comiqissioner’s expert. It then added those two figures together to generate a capitalization rate, which the court then used in combination with the yearly estimates of MERC’s net operating income to calculate the value of MERC’s pipeline distribution system. Several of MERC’s challenges, including its argument about the allegedly erroneous exclusion of a company-specific risk factor, suggest that a single component of the calculation, the cost of equity, is unrealistically low in light of the risks involved in MERC’s business.

Specifically, MERC argues that the tax court erred when it failed to apply a company-specific risk factor to account for the increased risk of a utility business that operates largely within a single state— Minnesota—and distributes only a single product—natural gas—to its customers. This circumscribed portfolio of business, according to MERC, raises the risk for equity investors and necessarily creates a higher cost of equity. MERC argues that an additional, company-specific risk factor in the cost-of-equity formula would account for this risk.

Both experts used the standard capital-asset-pricing model to estimate the cost of equity:

cost of equity = risk-free rate + (beta*market-risk premium) + additional risk factor.

However, only MERC’s expert, Reilly, applied an additional risk factor of 3 percent. Eyre, by contrast, used the same model and formula but simply concluded that an additional company-specific risk factor was unwarranted. The tax court’s decision to exclude the risk factor had a substantial impact on MERC’s cost of equity: with the 3 percent additional risk factor proposed by Reilly, the cost of equity would have increased from about 8 percent to about 11 percent, which in turn would have substantially decreased the estimated market value of MERC’s pipeline distribution system under the income approach.

Before addressing the merits of the parties’ competing positions, we first address the applicable standard'of review. In defending the tax court’s decision, the Commissioner argues that the court’s refusal to apply a company-specific risk factor was a.factual determination subject to clear-error review because the tax court’s decision reflected its acceptance of the testimony of one party’s expert (Eyre) and the rejection of the testimony of the other expert (Reilly). In contrast, MERC suggests that a de novo standard governs because, in its view, the court adopted two bright-line legal rules, neither of which allegedly finds support in the record: (1) regulated entities cannot have their cost of equity adjusted based on company-specific risk; and (2) mathematical precision is required before applying a company-specific risk factor. We agree with the Commissioner that the tax court’s rejection of the additional risk factor was a factual determination subject to clear-error review. See Kohl’s Dep’t Stores, Inc. v. Cty. of Washington, 834 N.W.2d 731, 735 (Minn.2013) (applying the clear-error standard of review to the tax court’s capitalization-rate calculations).

The tax court excluded a company-specific risk factor from its calculation of MERC’s cost of equity based on a lack of evidéntiary support in the record for the proposition that MERC’s business was riskier than the market, not because it determined, as a matter of law, that a regulated entity’s cost of equity can never be augmented to account for additional risk. As the tax court found, there, was “no creditable evidence in the record to support [Reilly’s] claim that MERC, as a rate-regulated entity, experiences more risk (even if considered a standalone company) than Reilly’s diversified guideline companies.” The tax court further found that “Reilly provided no support” for choosing 3 percent' as the applicable risk factor rather than some other figure. Accordingly, because the tax court’s decision to exclude a company-specific risk factor from its calculation of the cost of equity was a factual determination, not a legal conclusion, we review it for clear error.

Faced with conflicting testimony by competing experts, the tax court agreed with Eyre’s position that “[property specific risk or nonsystematic risk should not be accounted for in the cost of equity, but rather should be accounted in the forecast of expected cash flows.” Eyre also testified about a study that showed that “there is no conclusive empirical evidence to support the general practice of adding a small firm risk premium to the discount rate when valuing small firms.” Another witness, Jon Van Nurden, an employee of the Department of Revenue, largely agreed with Eyre, stating that he had not seen support for the application of an additional risk factor in any source other than the one relied upon by Reilly.

To be sure, MERC accurately observes that Reilly’s report, relying on a multi-volume publication on business valuation, supports the addition of a company-specific risk factor to the cost of equity for small, undiversified firms. See 1 Jay E. Fishman et al., Guide to Business Valuations 5-21 (5th ed.1995). Even so, the tax court, as the finder of fact, was entitled to resolve the conflicts in the record and determine how much weight to give each expert report. See City of New Prague v. Hendricks, 286 N.W.2d 696, 702 (Minn.1979). The tax court, in other words, was in the best position to weigh the conflicting opinions, and based on our review of the record, we cannot say that the tax court clearly erred when it decided to adopt the expert opinions of Eyre and Van Nurden over Reilly’s opinion. 1

B.

MERC’s second- challenge is to the tax, court’s explanation of the beta factors it used to calculate MERC’s cost of equity, which again requires us to examine the capital-asset-pricing model. Once the court decided to exclude the company-specific risk factor from the cost-of-equity calculation, it condensed the calculation to the following formula:

cost of equity = risk-free rate + (beta*market-risk premium).

The risk-free rate adopted by the court was the yield on long-term United States Treasury securities, long viewed by investors as a safe investment. See Appraisal Institute, supra, at 145; Fishman et al., supra, at 5-15; see also United States v. Doud, 869 F.2d 1144, 1145 (8th Cir.1989) (discussing “a ‘riskless’ rate, usually commensurate with the interest paid on government issue bonds and bills”). The market-risk premium, which represents the additional return required by investors to own an equity security rather than a bond, was the difference between the return on large-company stocks and long-term government bonds. See Steiner Corp. v. Benninghoff, 5 F.Supp.2d 1117, 1134 (D.Nev.1998) (“The market risk premium is a measure of the additional return needed, on average, to convince investors to invest in the stock market rather than in risk-free Treasury Notes.”). Finally, the beta factor, a concept unique to corporate finance, accounts for the relative volatility of a specific investment compared to the volatility of the market as a whole. See P.R. Tel. Co. v. Telecomm. Regulatory Bd., 665 F.3d 309, 327 (1st Cir.2011) (explaining the “ ‘beta’ factor, which estimates the risk of investing in a particular company by measuring the volatility of its stock as compared to the market as a whole”); Steiner Corp., 5 F.Supp.2d at 1134 (“[A] beta is the covariance of a company’s rate of return against the market rate.”). The tax court’s adoption of a beta factor that was less than 1 for each taxable year indicated that MERC’s volatility, likely due to its status as a highly regulated entity, was lower than the market’s risk, 2 which resulted in a lower cost of equity for MERC than a similar investment with a higher beta factor.

Other than stating that the beta factor was less than 1 for each of the years in question, the tax court’s order does not specify the value of the beta factors it used for each year, much less explain how or why it selected them. MERC argues that this omission by the tax court requires us to remand the case for further explanation so that we can meaningfully review the tax court’s decision. The Commissioner’s position is that, even though the tax court’s order does not reveal the beta factors it used to calculate MERC’s cost of equity, we can reverse engineer the beta factors by plugging each of the other figures identified in the tax court’s order into the cost-of-equity formula. Moreover, the Commissioner says that she independently calculated each of the beta factors and that they are similar, though not identical, to the beta factors used in Reilly’s report. The Commissioner’s assertion finds support in the tax court’s statement that its beta factors were “adapted [from] Reilly’s [b]eta figures.”

Even though the Commissioner is correct that we can isolate the beta factor in the formula and determine each year’s beta factor arithmetically, the Commissioner’s suggestion does not resolve the problem with the tax court’s lack of explanation. Calculating the beta factor for each year would still leave us in the dark about why the tax court selected any specific beta factor for a particular year, which itself necessitates a remand. See Westling v. Cty. of Mille Lacs, 512 N.W.2d 863, 866 (Minn.1994). Therefore, because we simply cannot determine whether the record supports the beta factors adopted by the tax court, we reverse the portion of the tax court’s order discussing MERC’s cost of equity and remand to the tax court for further explanation.

C.

MERC’s third challenge is to the tax court’s decision to reject the build-up method as an alternative technique for calculating MERC’s cost of equity. The tax court, as well as both expert witnesses, used the capital-asset-pricing model to calculate MERC’s cost of equity. Reilly supplemented his ■ valuation of MERC’s pipeline distribution system with another technique, the build-up method, to determine MERC’s average cost of equity. The b'uild-up method, much like the capital-asset-pricing model, begins with a risk-free rate, but rather than using a beta factor, it adds a market premium for equity investments and a size premium to reflect the higher expected rate of return for investments in smaller companies. See Shannon P. Pratt & Roger J. Grabow-ski, Cost of Capital 177 (5th ed.2012). Reilly’s report deviated slightly from the typical methodology of the build-up method by substituting' the company-specific risk factor for the size premium. He then used MERC’s average cost of equity, which he derived from averaging the cost of equities calculated from both the buildup method and the capital-asset-pricing model, to estimate the market value of MERC’s pipeline distribution system using the direct-capitalization technique.

The tax court declined to incorporate the build-up method into its calculations, but the parties disagree on the reason for the court’s decision. The Commissioner suggests that the court’s decision simply reflects its choice between competing expert opinions, with the court adopting Eyre’s position that the build-up method would not accurately estimate MERC’s cost of equity. MERC once again asserts that the court adopted a bright-line rule that categorically prohibits using the buildup method in the valuation of regulated entities such as utilities. We agree with the Commissioner’s reading of the court’s decision.

At trial, Eyre identified problems with Reilly’s use of the build-up method. First, Eyre explained that the build-up method is traditionally used in situations in which there are “no comparables out there from which to derive betas,” which was not the case with MERC. Second, Eyre testified that, neither Minnesota Administrative Rule 8100 nor “corporate finance tax” theory requires application of the build-up method. Third, Eyre criticized Reilly’s decision to use a corporate-bond rate as the risk-free rate, because corporate bonds are a riskier investment than United States Treasury securities. The tax court echoed Eyre’s final point, stating in its order that “a government bond rate would represent a less risky investment than a corporate bond rate.” The court was also “troubled by Reilly’s failure to use a specific risk premium related to the gas distribution industry” in calculating MERC’s cost of equity under the build-up method.

MERC responds with a general assertion that each of these reasons for rejecting the build-up method is unpersuasive. Yet MERC references little testimony, evidence, or authority to support its position. Rather, one of the few sources it cites says only that the build-up method is “commonly used,” not that it must be used to value business property or that it would be particularly useful under these circumstances. See Robert F. Reilly & Robert P. Schweihs, Guide to Property Tax Valuation 166 (2008). Indeed, nothing relied on by MERC contradicts Eyre’s testimony that the build-up method is commonly used only when there are no comparable firms from which to draw reliable beta factors.

The tax court’s decision to reject the build-up method is similar to the situation it faced with respect to the company-specific risk factor: a disagreement between two competing expert opinions. Reilly and Eyre had differing opinions about the value of the build-up method, with Eyre opining that its use in this case was inappropriate and Reilly advocating for its inclusion. The tax court was in the best position to judge the credibility of these two experts and to assign weight to their testimony. See City of New Prague, 286 N.W.2d at 702. We defer to its decision to adopt Eyre’s view. We therefore conclude that the tax court did not clearly err when it declined to incorporate the build-up method into its own calculation of MERC’s cost of equity.

III.

Each of MERC’s first three challenges addressed the tax court’s calculation of the cost of equity, which was one of the variables it used to estimate the market value of MERC’s pipeline distribution system under the income approach. MERC’s final challenge, which involves the court’s determination that MERC’s pipe line distribution system did not suffer from external obsolescence during the years in question, affects the court’s estimate of market value under the cost approach.

Thé cost approach is “founded on the proposition that an informed buyer would pay no more for the property than the cost of constructing new property having the same utility as the subject property.” Equitable Life Assurance Soc’y of the U.S. v. Cty. of Ramsey, 530 N.W.2d 544, 552 (Minn.1995). As we have recognized, the cost approach, which estimates market value based on the current cost to construct new or substitute property, is particularly useful when trying to determine the market value of “special purpose property” such as pipelines and specialized equipment. S. Minn. Beet Sugar Coop v. Cty. of Renville, 737 N.W.2d 545, 555-56 (Minn.2007).

It is clear that MERC’s pipeline distribution system qualifies as “special purpose property” because “it was designed and built for a special purpose.” Fed. Reserve Bank of Minneapolis v. State, 313 N.W.2d 619, 621 (Minn.1981). It is also clear that “[e]xternal obsolescence often relates to the business enterprise that operates at the special-purpose property, such that a change in industry conditions could cause the taxpayer to incur a reduction in revenue, profit margin, or return on investment metrics.” Guardian Energy, LLC v. Cty. of Waseca, 868 N.W.2d 253, 263 (Minn.2015) (citation omitted) (internal quotation marks omitted). The tax court correctly defined external obsolescence, but then found that MERC’s evidence insufficiently demonstrated that MERC’s pipeline distribution system was externally obsolete.

External obsolescence “is a loss in value caused by negative externalities” that is “almost always incurable.” Appraisal Institute, supra, at 632. It is one of three forms of depreciation—functional obsolescence and physical depreciation being the others—that can decrease the market value of a property under the cost approach. Id. at 576, 633; see Guardian Energy, 868 N.W.2d at 262. Yet rather than treating external obsolescence in the same manner as these other forms of depreciation, the tax .court has adopted a special standard for evaluating taxpayer claims of external obsolescence. This standard, borrowed from the Arizona Court of Appeals, requires “a taxpayer claiming external obsolescence [to] offer probative evidence of the cause of the claimed obsolescence, the quantity of such obsolescence, and that the asserted cause of the obsolescence actually affects the subject property.” Eurofresh, Inc. v. Graham Cty., 218 Ariz. 382, 187 P.3d 530, 538 (Ariz.Ct.App.2007) (emphasis added). In addition to this case, the tax court has applied the Eurofresh standard in at least two other cases since 2009. See, e.g., Guardian Energy, LLC v. Cty. of Waseca, Nos. 81-CV-10-365, 81-CV-11-348, 81-CV-11-741, 2014 WL 7476215, at *41 (Minn,T.C. Dec. 9, 2014); Am. Crystal Sugar Co. v. Cty. of Polk, Nos. C1-05-574, C4-06-367, 2009 WL . 2431376, at *25 (Minn,T.C. Aug. 5, 2009).

The tax court examined MERC’s claim that it was entitled to receive a reduction in market value for external obsolescence under the Ewrofresh standard.' In an attempt to satisfy Eurofresh, MERC presented the' testimony of five witnesses, including Reilly. The witnesses pointed to “regulation and rate lags, mild weather, the' economic crisis in 2008, and [an] increase [in] use of energy efficient appliances” as contributing to MERC’s decreased revenues and profit ' margins. Reilly, in particular, attempted to demonstrate the existence of external obsolescence by comparing MERC’s lower return on equity to nine other utility companies that had collectively performed better than MERC.

Even so, the tax court concluded that MERC’s evidence was insufficient to warrant an adjustment to the market value of MERC’s property under the cost approach. The court instead evaluated each of MERC’s explanations individually, rather than viewing the evidence as a whole, to determine whether MERC’s pipeline distribution system suffered from external obsolescence during the years in question. In the court’s view, MERC “had failed to demonstrate that any of these factors affected the subject property,” even though the court acknowledged “the difference between MERC’s return on equity and the average return on equity for the gas distribution industry could indicate that the subject property suffered from external obsolescence.” Clearly, the causal-nexus requirement from Eurofresh played a decisive role in the court’s decision to reject MERC’s external-obsolescence evidence. See 187 P.3d at 538.

As MERC observes, we have never adopted the Eurofresh standard as the appropriate framework for evaluating taxpayer claims of external obsolescence. See Guardian Energy, 868 N.W.2d at 264 (“Because neither party argues that the tax court applied the incorrect analytical framework in this case, we assume, without deciding, that [Eurofresh] is the appropriate analytical framework.”). For two reasons, we decline to adopt the Euro-fresh standard now.

First, we have never required taxpayers to make the heightened showing required by Ewrofresh. In Northwest Racquet Swim & Health Clubs, Inc. v. County of Dakota, for example, we affirmed a decision by the tax court to accept a taxpayer’s claim of external obsolescence based on an expert opinion that was no more specific than the evidence presented in this case. 557 N.W.2d 582, 586, 588 (Minn.1997). There, the taxpayer’s expert opined that, because a health club received an actual rate of return that was half as large as its expected rate of return, the property in question suffered from external obsolescence of 50 percent. Id. at 586. The County’s expert, by contrast, estimated the health club’s obsolescence based on its location to be 15 percent. Id. Rather than requiring the taxpayer to show a specific causal nexus between the asserted cause of the obsolescence and the subject property, the tax court simply “chose a compromise value” of 25 percent, which was between the estimates of the two experts. Id. We affirmed the tax court’s decision, reasoning that “a court confronted with conflicting appraisals may conclude that a compromise in valuation is required, provided it has evidentiary support and is not unreasonable or clearly erroneous.” Id. at 588. Nothing in Northwest Racquet supports the imposition of heightened requirements for proving external obsolescence. See also Am. Express Fin. Advisors, Inc. v. Cty. of Carver, 573 N.W.2d 651, 660 (Minn.1998) (stating that the “the financial losses of [a conference facility]” was probative evidence of external obsolescence).

Second, the fact that the taxpayer cannot identify the specific causes of external obsolescence and precisely calculate the contribution of each to decreased revenues or profit margins does not mean that a property does not suffer from external obsolescence. As Northwest Racquet suggests, external obsolescence can exist but the cause of it can be difficult to quantify, resulting in variation among experts in their estimation of the impact of external factors on the fair-market value of certain properties. 557 N.W.2d at 588 (affirming the tax court’s use of a compromise figure for obsolescence because of the lack of supporting data to ascertain a reliable fig ure). In fact, the precision required by the Eurofresh standard is inconsistent with what The Appraisal of Real Estate calls the “most persuasive measurement of the effect of negative externalities on value”: “[djirect comparison of similar properties with and without external obsolescence.” Appraisal Institute, supra, at 684. This was precisely the type of evidence that MERC introduced in this case through Reilly’s testimony, which compared MERC’s returns on equity to those of nine other similar properties.

We do not suggest that the tax court, on remand, is required to find the existence of external obsolescence or accept the testimony of MERC’s witnesses. Rather, we hold that it evaluated MERC’s evidence of external obsolescence under the wrong legal standard by relying on Eurofresh, and that MERC’s evidence was at least sufficient to make out a prima facie case of external obsolescence. It will be the tax court’s task on remand to consider all of the evidence presented to determine whether the evidence of external obsolescence is sufficient to support a downward adjustment to the estimated market value of MERC’s property under the cost approach.'

IV.

We now turn to the Commissioner’s cross-appeal, beginning with her argument that the tax court erred as a matter of law when it deducted the value of intangible property and working capital from its valuation of MERC’s pipeline distribution system under the income approach. The Commissioner’s argument requires us to determine whether these two categories of property are exempt from taxation under Minnesota law.

A.

All real and personal property in Minnesota is taxable, unless exempt by law. Minn.Stat. § 272.01, subd. 1 (2014). Minnesota Statutes § 272.03, subd. 2(5) (2014), specifies that taxable “personal property” includes “[a]ll gas, electric, and water mains, pipes, conduits, subways, poles, and wires of gas, electric light, water, heat, or power companies.” Minnesota Statutes § 273.33 specifically addresses “pipeline companies” and grants authority to the Commissioner of Revenue to assess ad-valorem taxes on “pipeline system[s],” which include “mains, pipes, and equipment attached thereto.’’ See also Minn. R. 8100.0100, subp. 11 (2015) (defining a utility’s “operating property” as “any tangible property that is owned or leased, except land, which is directly associated with the generation, transmission, or distribution of electricity, natural gas, gasoline, petroleum products, or crude oil”). The statute specifically exempts the “products transported through the pipelines” from ’ taxation,, as well as pipelines that transport “petroleum products” exclusively for the pipeline owner’s consumption. See Minn.Stat. § 273.33, subd. 2. Based on these statutes, the tax court concluded that MERC’s tangible personal property was taxable, but that its intangible property, including intangible assets and working capital, was not. Our review of the tax court’s conclusion presents a question of statutory interpretation that we review de novo. See Cont’l Retail, LLC v. Cty. of Hennepin, 801 N.W.2d 395, 398 (Minn.2011).

The Commissioner disagrees with the tax court’s interpretation of these statutes, stating that a calculation of the market value of MERC’s pipeline distribution system must necessarily include the value of all intangible assets that affect its value. Those intangible assets include, according to the Commissioner, MERC’s working capital and other intangible assets, which reflect the going-concern value of MERC’s property. We disagree.

The Commissioner’s argument fails to distinguish between tangible and intangible property, only the former of which is taxable under the statutes and rules.that apply to pipeline companies. By focusing on “mains, pipes, and equipment attached thereto,” Minn.Stat. § 273.33, subd. 2, allows the Commissioner to tax only a pipeline company’s tangible property. The administrative rules make this distinction even clearer by granting authority to the Commissioner to assess taxes on a utility’s “operating property,” which one of the rules defines as “any tangible property that is-owned or leased, except land, which is directly associated with the generation, transmission, - or distribution of electricity, natural gas, gasoline, petroleum products, or crude oil.” Minn. R. 8100.0100, subp. 11 (emphasis added); see also id., subps, 14, 16 ■ (2015) (defining “system plant” and “unit value,” both of which are used in calculating the taxes a utility owes on its personal property). “Nonoperating property,” by contrast, is not taxed according to “the formula provided ... for the valuation of utility property.”- Minn. R. 8100.0500, subp. 2 (2015).

The tax court’s analysis followed the statutes and rules that treat a pipeline company’s intangible property as nontaxable. Critically, there is no indication that the tax court deducted the going-concern vaiue attributable to MERC’s tangible assets, but rather limited the deduction to the income provided by those intangible assets that are exempt from taxation under the relevant statutes and rules. In fact, Reilly’s analysis, upon .which the tax court relied, noted that the deduction for intangible assets did not reduce MERC’s going-concern value. By deducting only the income provided by MERC’s intangible assets, which are nontaxable, the tax court’s analysis is fully consistent with the plain language of Minn.Stat. §§ 272.03 (2014), 273.33, and Minn. R. 8100.0100-.0700 (2015).

B.

In addition to its purely legal argument, the Commissioner challenges the tax court’s specific deductions of 5 percent for working capital and 5 percent for intangible assets, both of which the tax court adopted from Reilly’s report. The tax court stated that its deduction for intangible assets was for an “assembled and trained management team and workforce, computer software, and operating manuals and procedures.” The deduction for working capital, on the other hand, was based on the “observed levels of working capital presented by the guideline companies as of the valuation date as well as historically and consideration of the actual levels of working capital observed by MERC.” The Commissioner argues that, even if the relevant statutes and rules authorize these deductions, the tax court made them at the wrong stage of the analysis under the administrative rules.

The Commissioner relies on Rule 8100.0500, which addresses how to make deductions from the unit value of taxable property for nonoperating and tax-exempt property. Rule 8100.0500, subpart 1, requires the deductions for nontaxable property to be made after the calculation of the property’s unit value, which is “the value of the entire system plant of a utility company taken as a whole without any regard to the value of its component parts.” Minn. R. 8100.0100, subp. 16. Calculating unit value requires weighing the three indicators of value, which the rule refers to as the market, income, and cost indicators of value. Minn. R. 8100.0300. Once the unit value is determined using some combination of these three indicators, the unit value must then be allocated based on “the portion of value [that] is attributable ■ to Minnesota.” Minn. R. 8100.0400. Only at the final stage, after calculating unit value and allocating value, does Rule 8100.0500 allow the tax court to.make deductions for the value of nonoperating and tax-exempt property.

The tax court deviated from the requirements of Rule 8100.0500 by deducting the working capital and intangible assets from the income indicator of value, rather than making the deduction at the end of the process, after each of .the indicators of yalue has been considered and weighed in calculating the property’s unit value. Relying on our decision in Northwest Airlines, Inc. v. Commissioner of Revenue, 265 N.W.2d 825, 830 (Minn.1978), the tax court concluded that it was not bound by the process set forth in the Commissioner’s administrative rules. Northwest Airlines does not provide a basis fQr ignoring the requirements of binding administrative rules.

Northioest Airlines involved an appeal from the tax court’s valuation of Northwest’s “airflight property,” including its aircraft. Id. at 827. More specifically, the case involved the validity of the Commissioner of Revenue’s adoption of a cost-less-depreciation formula to value Northwest’s property. Id. at 828, In reviewing the decision, we stated that the tax court, in an appeal from an assessment by the Commissioner, was not required to use the Commissioner’s formula if it would result in an inaccurate valuation of the property. Id. at 830. However, Northwest Airlines, unlike this case, did not involve any administrative rules. In fact, the Commissioner unilaterally adopted the formula after negotiations with Northwest Airlines had collapsed prior to the filing of the appeal. Id. at 828.

To be sure, Northwest Airlines reflects the general principle that the tax court has an independent obligation to determine the market value of property when the taxpayer challenges the Commissioner’s assessment of taxes. Id. at 830. But it' does not stand for the additional proposition, contrary to the tax court’s analysis, that courts are free to ignore administrative rules when they believe those rules will result in an inaccurate valuation. Administrative rules have the force and effect of law, Minn.Stat. § 270C.06 (2014); U.S. W. Material Res., Inc. v. Comm’r of Revenue, 511 N.W.2d 17, 20 n. 2 (Minn.1994), and courts are free to reject them only when they conflict’ with the statutes they implement, see Billion v. Comm’r of Revenue, 827 N.W.2d 773, 781 (Minn.2013). Mere disagreement with an administrative rule is not -a valid reason to disregard it.

Even though the administrative rules are binding on the tax court, the rules themselves recognize that the Commissioner—and by extension, the tax court—can “exercise discretion” to depart from the .valuation formula “whenever the circumstances of a valuation estimate dictate the need for it.” Minn. R. 8100.0200. “Discretion may be used,” among other reasons, “to ensure a balance between a prescriptive rule and sound appraisal judgment; .to ensure that all relevant data pertaining to value is considered; [and] to ensure that a reasonable estimate of market value is derived.” Id. The administrative rules, in other words, allow for the exercise of discretion when deviating from the formula will lead to a more accurate valuation. Accordingly, .even though the tax court was wrong to rely on Northwest Airlines to support its decision to deviate from the Commissioner’s formula, the administrative rules recognize that it nevertheless had the authority to do so “to ensure that a reasonable estimate of market value [was] derived.” Minn. R. 8100.0200.

C.

The Commissioner also raises a purely factual objection to the size of the deductions for working capital and intangible assets. She says that MERC failed to satisfy its burden of proof to show that it was entitled to separate 5 percent deductions for both categories of intangible property. See Minn. R. 8100.0500, subp. 5 (“The utility company has the burden of proof to establish that the value of any property should be excluded from the Minnesota portion of the unit value.”). Reilly deducted 5 percent of the previous year’s income for intangible assets “based on his experience in valuing energy properties” and. stated that intangible assets for these types of properties generally fall between 5 percent and 20 percent of the “total business enterprise.” The tax court adopted Reilly’s 5 percent figure in part because it was on the low end of the range for intangible assets for comparable businesses. Although Eyre did not directly challenge the deduction for working capital, the tax court explained that the 5 percent deduction was based on “observed levels of working capital” presented by comparable companies as well as “the actual levels of working capital observed by MERC.”

Here again, the tax court was faced with conflicting expert opinions, with Reilly opining that the 5 percent deductions for intangible assets and working capital were supported by comparable energy companies and Eyre challenging the deductions altogether. As with MERC’s arguments on the addition of a company-specific risk factor and the use of the build-up method to calculate MERC’s cost of equity, the tax court was in the best position to evaluate the credibility of each expért and to weigh the conflicting opinions. See City of New Prague v. Hendricks, 286 N.W.2d 696, 702 (Minn.1979). We therefore conclude that the tax court did not clearly err when it adopted Reilly’s opinion on the necessity of deducting the value of intangible assets and working capital from the estimated value of MERC’s pipeline distribution system under the income approach.

V.

The final question presented by this case is whether the tax court erred when it declined to incorporate the 2006 sale of MERC into its calculation of the market value of MERC’s pipeline distribution system. The tax court is required to consider all relevant evidence when determining the market value of property. See Indep. Sch. Dist. No. 99 v. Comm’r of Taxation, 297 Minn. 378, 384, 211 N.W.2d 886, 890 (1973). We have previously acknowledged that prior sales of the subject property can be important evidence when valuing real property, even though they are “not conclusive.” Archway Mktg. Servs. v. Cty. of Hennepin, 882 N.W.2d 890, 896 (Minn.2016). Relying on these principles, the Commissioner argues that MERC’s 2006 sale, which occurred just 2 years before the first of the assessments that MERC challenges, is reliable evidence of the market value of MERC’s pipeline distribution system during the years in question, from 2008 to 2012. For three reasons, we are not persuaded that the tax court erred.

First, MERC’s sale included all of its property, not just the pipeline distribution system that the Commissioner assessed and the tax court valued. MERC’s purchase price in 2006 captured the overall value of the entire enterprise—including MERC’s tangible assets, goodwill, investments, and working capital, some of which we have already determined is nontaxable. It also included the value of ServiceChoice, MERC’s appliance-repair business, which is completely separate from MERC’s pipeline distribution system. Thus, although MERC’s 2006 purchase price includes the tangible personal property at issue in this case, it also reflects the value of other property that is either nontaxable or not part of the valuation undertaken by the tax court.

For this reason, the cases relied on by the Commissioner are unpersuasive. The Commissioner cites various cases for the proposition that a recent sale of the subject property is relevant evidence that the tax court must consider. See, e.g., Minn. Entm’t Enters., Inc. v. State, 306 Minn. 184, 188, 235 N.W.2d 390, 393 (1975); Schleiff v. Cty. of Freeborn, 231 Minn. 389, 394-95, 43 N.W.2d 265, 268-69 (1950). But these cases involved evidence of a recent sale involving the subject property itself, not a sale involving a combination of different types of property, some of which are not even taxable. Minn. Entm’t Enters., 306 Minn, at 184, 235 N.W.2d at 391; Schleiff, 231 Minn, at 392, 43. N.W.2d at 267. Moreover, these cases involved real-property valuation, which can require a different approach than the valuation of the tangible personal property of a utility company. In fact, the Commissioner’s administrative regulations, which the tax court was bound to follow, reflect the differences between valuing the tangible personal property of utilities and other types of property. Minn. R. 8100.0300, subp. 1 (“Because of the unique character of public utility companies, the traditional approaches to valuation estimates of property (cost, capitalized income, and market) must be modified when utility property is valued.”).

Second, the Commissioner cannot point to any statute- or administrative rule that required the tax court to consider the 2006 sale in its valuation. Instead, Rule 8100.0300, subpart 1, requires only that the tax court must consider each of the three approaches (cost, market, and income), but it may not use any approach that is “not demonstrated to be reliable.” (Emphasis added.) Here, the tax court considered and rejected application of the market approach, but only after concluding that it was unreliable and unhelpful.

Third, the experts did not rely on the market approach or MERC’s 2006 sale in their analyses. In fact, the Commissioner’s expert,.Eyre, “placed no weight on the [market] approach” because “sales [of utility property, of this type] rarely occur.” Reilly reached a similar conclusion, stating that it is difficult to identify “how much of the purchase price [of an energy-distribution company] was paid for intangible assets,” which precludes “an effective [market] comparison.”

Recognizing that neither expert relied on the market approach, the Commissioner suggests that the tax court should have used MERC’s 2006 sale as a “benchmark” showing “that its valuations [of MERC’s property],were far too low.” But without evidence allowing the court to disaggre-gate MERC’s purchase price into the component parts of MERC’s business, the Commissioner does not show how the court could have derived a reliable value for MERC’s pipeline distribution system from only the 2006 purchase price of MERC’s entire business. 3 We accordingly conclude that the tax court did not err when it declined to incorporate MERC’s 2006 purchase price into its calculation of the estimated value of MERC’s pipeline distribution system under the market approach.

VI.

For the foregoing reasons, we affirm the tax court’s decision in part, reverse in part; and remand to the tax court for further proceedings consistent with this opinion.

Affirmed in part, reversed in part, and remanded.

CHUTICH, J., not having been a member of this court at the time of submission, took no part in the consideration or decision of this case. McKEIG, J., not having, been a member of this court at the time of submission, took no part in the consideration or decision of this case.

1 . Because the tax court did not clearly err when it excluded the additional risk factor from the cost-of-equity calculation, we need not resolve whether a 3 percent risk factor would have been appropriate to account for MERC’s company-specific risk.

2 . An example illustrates ■ beta’s operation. Suppose that a security has a beta factor of 2 and then the market’s overall value increases by 5 percent. The individual security with a beta factor of 2 can be expected, on average, to increase in value by 2*.05, or 10 percent, because the individual security moves in the same direction as the market but with twice the volatility, based on its beta. See Richard A. Brealey et al., Fundamentals of Corporate Finance 346-48 (7th ed.2012).

3 . The Commissioner criticizes the' tax court for failing to consider MERC's internal documents, including the annual goodwill-impairment studies conducted by Ernst & Young and the purchase-price allocation done at the time of the sale, in its valuation analysis. However, the Commissioner does not show how MERC’s 2006 púrchase price' necessarily reflected the value of its tangible assets more than 2 years later in 2008 or, for that matter, in any subsequent year. Nor does the Commissioner indicate how the tax court could have used these documents to calculate the value of MERC’s tangible personal property under any of the three approaches mentioned in the administrative rules in light of the failure of the parties’ experts to do so.