This is an ad valorem tax case in which petitioners, Southern Railway Company and Norfolk Southern Railway Company,1 *179hereinafter “Railroads,” challenge the Property Tax Commission’s, hereinafter “Commission,” appraisal of their companies’ market value. The Court of Appeals affirmed the Commission’s decision. We conclude the Commission erred in ruling that the Railroads failed to rebut the presumption of correctness accorded the appraisals of the Department of Revenue, hereinafter “Department.” We also conclude the Commission erred in adopting certain appraisal methods used by the Department. We, therefore, reverse the decision of the Court of Appeals and remand to that court with instructions to remand the matter to the Commission for a re-determination of the Railroads’ market value in a manner consistent with this opinion.
I.
Subchapter II of Chapter 105 of our General Statutes, hereinafter “Machinery Act” or “Act,”2 provides for the listing, appraisal, and assessment of property for ad valorem tax purposes and for the collection of the tax. Under the Act, § 333(14), the Railroads are “public service companies” subject to ad valorem taxation. Public service companies are appraised initially by the Department, § 335, which also apportions the values subject to North Carolina taxation, § 337, and allocates the values among local taxing units, § 338. Pursuant to § 342 of the Act, the Department duly notified the Railroads of its tentative appraisals of their systems for the 1980 tax year; the Railroads objected to the appraisals and requested a hearing before the Commission.
At this hearing the Railroads supported their challenges to the Department’s appraisal methods by the testimony of Dr. Arthur A. Schoenwald, a nationally recognized expert in appraisal of railroads and utilities, and by Dr. Thomas Keller, Dean of the Fuqua Business School at Duke University and R. J. Reynolds Industries Professor of Business Administration. The Department offered the testimony of one of its employees, Mr. William R. Underhill, an experienced appraiser of public service companies. The Department appraised Southern Railway at $1,025,000,000 and Norfolk Southern at $59,500,000. The Railroads’ witnesses appraised Southern Railway at $690,166,000. Although the Rail*180roads’ witnesses made no formal, independent appraisal of Norfolk Southern, the testimony of Dr. Schoenwald demonstrated that if the methods he advocated had been used by the Department, the Department’s own appraisal of Norfolk Southern would have been $46,156,000. As our opinion will show, the Commission on this record should have adopted Dr. Schoenwald’s methods.
The Commission issued its final decision on 19 May 1981 in which it adopted the tentative appraisals made by the Department and rejected entirely the appraisal methods urged by the Railroads. The Court of Appeals affirmed. We allowed the Railroads’ petition for futher review on 11 January 1983.
II.
Railroads contend that the Commission erroneously concluded that the Railroads failed to rebut the presumption of correctness inasmuch as this conclusion was based only upon a review of the Department’s evidence and is unsupported by the evidence of record. Railroads also argue that it is clear from the language used by the Commission in its second conclusion that the Commission “misconstrued its role to be that of an appellate agency.” These arguments have merit.
Under § 342(b) of the Act, Department appraisals of public service companies are “deemed tentative” since they are made without notice or opportunity for hearing. The Department is required to give the public service company notice of its tentative appraisal, after which the company may, by timely request, secure a hearing before the Commission. This is the first and only evidentiary hearing to which the public service company is entitled. This hearing presents the first opportunity for a public service company to challenge the Department’s appraisal methods. At this hearing the Commission does not sit as an appellate tribunal. Its function under § 342(d) is to hear all the evidence offered by the taxpayer and the Department and from this evidence to make findings of fact, from the findings to make conclusions of law, and from the conclusions to issue its decision. The Commission’s function is “to determine the weight and sufficiency of the evidence and the credibility of the witnesses, to draw inferences from the facts, and to appraise conflicting and circumstantial evidence.” In re McElwee, 304 N.C. 68, 87, 283 S.E. 2d 115, 126-27 (1981).
*181It is true that the Department’s appraisal as it stands before the Commission is presumed to be correct. In re Appeal of AMP, Inc., 287 N.C. 547, 215 S.E. 2d 752 (1975). The presumption, however, “is only one of fact and is therefore rebuttable.” Id. at 563, 215 S.E. 2d at 762. The presumption is rebutted when the taxpayer’s evidence before the Commission shows that the Department used either an arbitrary or an illegal method of valuation and that the method used resulted in “a substantially higher valuation than one which would have been reached” under a proper valuation method. In re McElwee, 304 N.C. at 86, 283 S.E. 2d at 120; accord, In re Appeal of AMP, Inc., 287 N.C. at 563, 215 S.E. 2d at 762. An illegal appraisal method is one which will not result in “true value” as that term is used in § 283 and, for public services companies, in § 335. In re Appeal of AMP, Inc., 287 N.C. at 563-65, 215 S.E. 2d at 762 (tax assessor’s method of using book value of inventory to arrive at “true value” was illegal); In re McElwee, 304 N.C. at 88-91, 283 S.E. 2d at 127-29 (where statutory appraisal standard was “present use value” of agricultural land, tax assessor’s use of comparable sales held an illegal method when the “comparable” land was not shown to be used for same purpose as land being valued).
Here, Railroads offered testimony which demonstrated that the appraisal methods used by the Department would not result in ascertainment of “true value” of the Railroads. Further, the Railroads’ evidence showed that the Department’s methods resulted in substantially higher valuations than those which would have been reached had proper methods been followed. The Railroads’ evidence showed that the methods were not, in this case, simply matters of appraisal judgment. Rather, it showed that the Department’s methods would inevitably and always produce substantially higher valuations than the “true value” of the companies called for in the appraisal statutes.3
Despite this evidence and notwithstanding the Commission’s duty to consider the case as a trial tribunal of original jurisdic*182tion, the Commission in support of its decision concluded (1) the Railroads did not overcome the presumption of correctness given to the Department’s appraisals and (2) the Department’s appraisals were “supported by substantial competent evidence of record.” The first conclusion is legally erroneous and the second indicates that the Commission’s decision may have been based on an erroneous view of the Commission’s duty vis-a-vis the evidence.
When the Railroads offered evidence that the appraisal methods used by the Department would not produce true values for the Railroads and that the values actually produced by these methods were substantially in excess of true value, they rebutted the presumption of correctness. The burden of going forward with evidence and of persuasion that its methods would in fact produce true values then rested with the Department. And it became the Commission’s duty to hear the evidence of both sides, to determine its weight and sufficiency and the credibility of witnesses, to draw inferences, and to appraise conflicting and circumstantial evidence, all in order to determine whether the Department met its burden. In re McElwee, 304 N.C. at 86-87, 283 S.E. 2d at 126-27.
The Court of Appeals found no error in the Commission’s ruling that the Railroads did not rebut the presumption of correctness of the Department’s appraisals. The Commission, as we have shown, did err in this ruling. But the Department argues that, even if error, the ruling did not affect the outcome because the Commission’s order, viewed as a whole, demonstrates that the Commission did not rely on the presumption of correctness but carefully weighed the conflicting testimony in reaching its decision.
The Department also argues that notwithstanding the language in the Commission’s order, when the order is considered as a whole, it is clear that the Commission did consider both the Railroads’ and the Department’s evidence and determined in the manner of a trial tribunal of original jurisdiction which evidence it thought more worthy of acceptance. There are aspects of the Commission’s order which indicate that it might have done this. /in, its findings the Commission does refer to the evidence of both the Railroads and the Department. The Commission also gives ex*183planations as to why it adopted the Department’s methods rather than those supported by the Railroads’ experts. The Court of Appeals concluded that because of these aspects of the order the Commission did consider the evidence as a trial tribunal of original jurisdiction.
It is difficult, if not impossible, for an appellate court to divine the decision making process of an administrative agency unless the agency clearly sets it out in its order. We cannot say on the basis of its order before us that the Commission’s decision was not affected by its erroneous conclusion on whether the presumption of correctness was rebutted or the erroneous statement of how it should view the evidence, or both.
We need not, however, dwell further on whether the Commission’s decision rested to any degree on these errors. Assuming arguendo that it did not, we find, nevertheless, as we shall now demonstrate, more fundamental errors in the Commission’s decision.
We conclude that even if the Commission had considered the evidence as a trial tribunal of original jurisdiction, its decision would not have been supported by the evidence before it when the whole record test is, as it must be, applied to that evidence. There is, therefore, no reason to remand this case for reconsideration by the Commission because of the possibility that it looked at the case as an appellate, rather than a trial, tribunal. Neither should we vacate these entire proceedings and remand the case for an entirely new hearing on new evidence in the hope that the Department could produce evidence which might sustain its position. This disposition of the appeal would be not only novel; it would not be authorized by the statutes governing these appeals. See §§ 345-346. Only § 345.1 expressly authorizes the appellate division to direct the Commission to take new evidence. This section reads, in pertinent part:
[I]f any party shall satisfy the court that evidence has been discovered since the hearing before the Property Tax Commission that could not have been obtained for use at that hearing by the exercise of reasonable diligence, and will materially affect the merits of the case, the court may, in its discretion, remand the record and proceedings to the Com*184mission with directions to take such subsequently discovered evidence. . . .
Otherwise the statutes contemplate that decisions of the appellate division will be based on the record as the parties have chosen to make it. There is no suggestion in this case that the Department has evidence which meets the test of § 345.1. This Court in In re McElwee, 304 N.C. 68, 283 S.E. 2d 115, determined that the Commission’s decision was not supported by the evidence in light of the whole record. We reversed its decision and remanded the matter to the Commission for determination of values consistent with the Court’s opinion. We did not vacate the proceedings and order new proceedings in order to give the taxing authorities in McElwee a second opportunity to bolster its position with new evidence, although such evidence might have been available. We concluded that the property owners in that case were entitled to a decision on the record before the Commission and before the Court. There is no reason grounded in legal principle not to accord the Railroads here the same treatment we accorded the property owner in McElwee.
III.
Before the Commission the Railroads challenged various appraisal methods used by the Department. Three of these challenges have been brought forward to this Court.
The Machinery Act, § 336, requires that public service companies, which include railroads, be appraised by determining the company’s “true value ... as a system.” True value means “market value, that is, the price estimated in terms of money at which the property would change hands between a willing and financially able buyer and a willing seller, neither being under any compulsion to buy or to sell. ...” § 283. Thus, in appraising a railroad for ad valorem tax purposes, the appraisers seek to determine the fair market value of the railroad’s system properties, ie., that amount which a willing and financially able buyer would pay and a willing seller would accept, neither being under any compulsion to buy or to sell. The entire operating system, without geographical or functional division, is appraised and a portion of the appraised value is allocated to North Carolina by various statutory formulae. § 337.
*185Railroads, like other public service companies, are not often sold as going concerns, or operating systems; therefore evidence of comparable sales to prove fair market value is generally not available. For this reason, appraisals of such systems rely on a combination of methods which, in North Carolina, are prescribed in the Act. These are commonly referred to as (1) the “cost approach,” (2) the “income approach,” and (3) the “stock and debt” approach to value.4
The income approach to value is based on the principle that something is worth what it will earn. Fair market value of a railway system, using the income approach, is determined by capitalizing at a specified rate of return the net railway operating income (NROI), that is, the income from system property after depreciation and taxes are deducted but before distribution to the railway’s debt and security holders. The rate of return which an investor would demand as an inducement to commit capital to the purchase of the system which generates the NROI determines the rate at which this income is to be capitalized. Because appraisers assume that a purchaser will commit both debt and equity capital to the purchase, the overall capitalization rate is derived by calculating a weighted average of the cost of debt and equity capital (sometimes called the “band of investment”). The capitalized value of a given income stream varies directly with the amount of income and inversely with the capitalization rate. Value equals income divided by rate. Slight variations in the capitalization rate can result in large variations in value.
The stock and debt approach to value is based upon the premise that the aggregate market value of a public service company’s outstanding stocks and bonds reflects the market value of the company’s assets. Normally companies being appraised consist of system and nonsystem property. Section 335 requires that the “influence” of nonsystem property (which is taxed by other authorities) must be removed from the company’s stock and debt. The “income influence method” is an accepted means for elim*186inating the influence of nonsystem property. Using this method, the appraiser multiplies the total stock and debt value by the “income influence percentage,” which is derived by dividing the annual income from nonsystem property by the combined annual income from both system and nonsystem property. The product, representing the value of nonsystem property, is deducted from the total value of the company’s stock and debt.
With regard to the income approach to value, Railroads contend the Commission erred in its adoption of the Department’s capitalization rate. In determining this rate the Department used the Railroads’ historic interest rates applicable to its debt, sometimes referred to as “the embedded cost of debt,” in figuring the debt component of the rate rather than the current cost of borrowed money. Railroads urge that current cost of borrowed money must be used in determining the debt component of the capitalization rate.5 Second, Railroads urge that the Commission erred in adopting the Department’s method of adding the five-year average of Southern’s deferred income taxes of $15,524,000 to Southern’s NROI to arrive at the income to be capitalized. Finally, Railroads urge that the Commission erred in adopting the Department’s method of eliminating undistributed earnings of Southern’s nonsystem subsidiary companies and in adding back deferred taxes to system income in determining the “income influence percentage” under the stock and debt approach to value.
In determining whether the Commission erroneously adopted the challenged methods of the Department, we do not “substitute our judgment for that of the [Commission] when the evidence is conflicting.” In re McElwee, 304 N.C. at 87, 283 S.E. 2d at 127. The standard for our review is the “whole record” test. Id. “The ‘whole record’ test is not a tool of judicial intrusion; instead, it merely gives a reviewing court the capability to determine whether an administrative decision has a rational basis in the evi*187dence.” In re Rogers, 297 N.C. 48, 65, 253 S.E. 2d 912, 922 (1979). Under this test the reviewing court is permitted to take into account whatever evidence in the record fairly detracts from the evidence relied on by the Commission. Thompson v. Board of Education, 292 N.C. 406, 233 S.E. 2d 538 (1977). “Under the whole evidence rule, the court may not consider the evidence which in and of itself justifies the [agency’s] result, without taking into account the contradictory evidence or evidence from which conflicting inferences could'be drawn.” Id. at 410, 233 S.E. 2d at 541. See In re McElwee, 304 N.C. at 87-88, 283 S.E. 2d at 126-27, for an application of the foregoing principles to a decision of the Property Tax Commission.
When the whole record test is applied to the challenged decisions of the Commission, it is clear that these decisions have no rational basis in the evidence and that the Court of Appeals erred in affirming them.
A.
Department’s Erroneous Approach to Market Value
As we have shown, the Machinery Act requires that public service companies, including railroads, be appraised at “market value, i.e., the price at which the company “would change hands between a willing and financially able buyer and a willing seller, neither being under any compulsion to buy or to sell. ...” § 283. The testimony of the Department’s expert, Mr. Underhill, is seriously flawed because of his repeated insistence that he did not attempt to appraise the Railroads from the standpoint of their value to a hypothetical purchaser. His methods were designed to arrive at the value of the Railroads simply from the standpoint of the seller-owner. Mr. Underhill stated: *188Later, during cross-examination, Mr. Underhill reiterated: “I find the true value of the railroad system property by determining the value to the owner of the property. I explained that I do not consider value to a willing buyer because railroad sales are few and those few sales are abnormal and don’t represent fair market.” (Emphasis added.) Mr. Underhill continued:
*187In my opinion —even though the laws in the State of North Carolina and most states require the exact willing buyer, willing seller —if I can appraise Southern Railway Company as a value to the present owner . . ., I will determine a value that will not be greater than Southern Railway Company would require as a seller. And in my opinion any time you attack or discuss an appraisal from a purchaser point of view then you are getting the absolute lowest indication of that.
*188My appraisal of fair market value is determined based on my opinion of the appraisal of the fair market value of the railroad to the present owner and in light of the fact that everyone seems to agree there is no willing buyer or seller, and that satisfies the criteria. I think that if you put it in the perspective of a willing purchaser under a hypothetical restructuring of the capital and everything, also, you come out with a value that is not realistic to market value. So, for that reason I confine my approach to the value of this property to the owner.
Throughout the appraisal there are areas where you say prospective purchaser, or this is what would happen. And any appraiser would do that; but in light of the fact that it’s not going to be sold, I think that the value of the present owner represents a reasonable market value.
Mr. Underhill’s appraisals of the Railroads from the perspective of the present owner to the exclusion of the willing buyer were in clear violation of the statutory “market value” standard.
Mr. Underhill was the Department’s only witness. Arrayed against his testimony was the testimony of the Railroads’ experts, Dr. Schoenwald and Dean Keller. Both Railroads’ witnesses approached their appraisals from the standpoint of the willing and able buyer and the willing seller as the Act requires. Mr. Under-hill’s failure to follow the statutory standard by approaching his appraisals solely from the seller-owner’s standpoint so detracts from the usefulness of his methods that, on the whole record test, we must conclude it was error for the Commission to adopt them and to fail to adopt the methods urged by the Railroads’ experts.
*189B.
The Commission Erred on this Record in Using Embedded, Historical Cost of Debt Rather than Current Cost in Arriving at the Proper Capitalization Rate
To determine the capitalization rate, the Department calculated a weighted average of the cost of equity and the cost of debt, the “band of investment,” as follows: The cost of equity component was based upon current cost of money. However, the Department’s figure for the debt component reflected not the current cost of money, but rather the actual historic (and lower) interest rates applicable to the Railroads’ existing debt (the “embedded cost of debt”). The Department justified its use of embedded cost of debt in terms of value to the seller. In terms of value to both seller and buyer, the Railroads’ expert, Dr. Schoenwald, explained:
Now, if you use this embedded cost theory, this five percent, the issue rate some years ago is the basis for your capitalization rate, you are saying those railroad assets, the one we are appraising, are still worth one hundred million dollars and investors would pay one hundred million dollars for those assets. I submit that that is not true.
If in today’s market given that risk in this company, whatever it is, the appropriate market rate of interest is ten percent, and that reflects everything that’s happened in the interim from their issues five years ago to the present. It reflects the tight money policies in existence. That asset can only be worth in today’s market fifty million dollars because it can only generate five million dollars annual income. No investor would come in and pay for that asset any more than fifty million dollars, because he can go in the competitive market and buy some equally risky asset for the fifty million and generate his five million a year and obtain the required, current ten percent. As a consequence, the bonds related to that industry could only be worth fifty million dollars. And while it may be in a vault some place and have a price marked on it from some years ago of a hundred million dollars, at today’s rates it’s only worth fifty million. And the related asset is only worth fifty million.
*190In approving the embedded cost of debt rate, the Commission’s conclusion on this issue reflects the Department’s emphasis on value to the owner. The Commission concluded in accord with the Department’s evidence that “it is more reasonable to expect a purchaser to assume the debt and pay it off as provided in the actual existing instruments than it would be to expect him to refinance the transaction at current interest rates.” In response, the Railroads correctly point out, as their evidence shows, that this rationale, ie., that an assumable, existing debt at a low interest rate enhances the value of the property, “confuses valuation with methods of financing. Once value has been determined, the means of payment are a matter of further negotiation between the parties.” An existing debt with a low interest rate simply does not affect the fair market value of the property subject to the debt. See Crockett v. Savings & Loan Assoc., 289 N.C. 620, 224 S.E. 2d 580 (1976) (seller receives “fair market value” even when buyer, because of due-on-sale clause in mortgage, cannot assume a mortgage at lower than market rates).
Under the income approach to value, fair market value must be determined by current market conditions, not existing contractual obligations with reference to the asset being valued. This Court has held that valuations of real property for ad valorem tax purposes using the income approach must be based on fair market rents, not actual contract rents. In re Pine Raleigh Corp., 258 N.C. 398, 128 S.E. 2d 855 (1963) (contract rents produced less than market rents); In re Valuation, 282 N.C. 71, 191 S.E. 2d 692 (1972) (contract rents produced more than market rents). Just as use of actual, contract rents not reflective of market rents is illegal in making market value appraisals under the income approach, it is likewise illegal to use “actual cost of debt,” not reflective of market cost of debt in making such appraisals. Market value appraisals for ad valorem tax purposes must be based on market data.
We further agree with the Railroads that the Commission placed undue significance on the past use of the embedded cost of debt by the ICC and other regulatory bodies in determining an adequate rate of return for rate-making purposes. The Railroads properly note that “the purpose of those proceedings was to develop an adequate return to the current owners on their present investment, and perhaps it is in this sense that the Department *191referred to ‘the value of the property to the seller.’ ” Significantly, the ICC has recently determined that even for rate-making purposes, the current cost of debt must be used:
The minimum rate of return that will allow railroads to obtain investment funds is the cost of capital. The cost of capital is, by definition, the rate at which the market values investment funds. As we have said, investments earning less than the cost of capital will, in general, not maintain existing funding nor obtain new funding because investors will have sufficient opportunities to invest their funds elsewhere at a higher rate of return. It is extremely important to add, however, that this is true of funds generated internally as well. Railroad management has little incentive to reinvest funds generated by ratepayers in continued rail use if greater returns are available elsewhere. Railroads are private companies whose stockholders would not permit such reinvestment. Thus, even retained earnings will not be invested in the company if they cannot earn a rate of return equal to the cost of capital.
Ex parte No. 393, Standards for Railroad Revenue Adequacy, 364 I.C.C. 803, 810 (1981), affirmed sub nom., Bessemer and Lake Erie Rr. Co. v. Interstate Commerce Comm’n, 691 F. 2d 1104, 1111 (3rd Cir. 1982) (emphasis added).
The only reported tax case involving this issue holds that current cost rather than embedded cost of debt must be used in valuing railroads for ad valorem tax purposes. Soo Line R. Co. v. Wis. Dept. of Rev., 89 Wis. 2d 331, 278 N.W. 2d 487 (Wis. App. 1979), aff’d, 292 N.W. 2d 869 (Wis. 1980); see also County of Washtenaw v. State Tax Commission, 126 Mich. App. 535, 337 N.W. 2d 565 (1983) (a real property ad valorem tax case).
In summary, we hold that under our statutory definition of market value, which focuses on both a willing seller and a willing buyer, the value of a railroad’s assets to the prospective investor must be measured in terms of current market cost of both equity and debt. It is only by doing so that the prospective prudent investor is able to measure the value of the railroad against the value of other potential investments. The Department’s reliance on “value to present owners” not only ignores the statutorily mandated buyer component of the market value definition, but *192results in an excessively high value under the income approach which, realistically, would be rejected by any prospective investor seeking a reasonable rate of return under present market conditions.
C.
The Commission Erred on this Record in Refusing to Deduct Deferred Income Tax Expense from Income to he Capitalized
Railroads compute depreciation expenses for book and financial reporting purposes under the straight-line method using asset lives prescribed by the ICC. Under accelerated depreciation provisions of the Internal Revenue Code, the Railroads, for income tax purposes, show depreciation expense greater during the early life of an asset than will be shown during the asset’s later life. This means that depreciation expense for income tax purposes will, during the later life of the asset, be less than the expense shown on the books measured by the asset’s life. This result is required by the proposition that total depreciation deductions over the life of an asset, even for income tax reporting purposes, cannot exceed the cost of the asset. In the later years of an asset’s life the taxpayer will not be able to deduct for income tax purposes the full depreciation shown on the books under the straight-line method. General accounting principles suggest that an expense item, denominated deferred income tax expense, be offset against income during periods when accelerated depreciation is being used so as not to overstate actual after-tax income earned during these years. Even the Department’s witness, Mr. Under-hill, conceded that this was a generally accepted accounting principle but, in his opinion, the deferred income tax expense should not be deducted from the income stream for purposes of ascertaining value.
The Department initially approached the deferred income tax expense issue two ways. Under one calculation it added back to Southern’s NROI the sum of $15,524,000, representing the average deferred expense over the last five years as shown on Southern’s books. In the second calculation the Department treated deferred taxes as an expense in determining the income to be capitalized but also treated accumulated deferred taxes on the liability side of the balance sheet as a cost-free source of capital, ie., an interest-free loan from the government, at a zero percent *193interest rate in the band of investment. This latter treatment reduced income to be capitalized, but also decreased the overall capitalization rate. Before the Commission the Department’s witness testified:
In my opinion, deferred income tax is a reality or it is not a reality. It is a legitimate expense or it is not a legitimate expense. If it is or should be considered a legitimate expense on the income statement to reduce income since it is absolutely not paid, it is then a liability. That must be paid in the future, and it constitutes an interest free loan from the federal government. That is the calculation covered in Method B. If we ignore current deferred income tax, since it is not paid, then it would be improper to consider the accumulated deferred as an interest free loan. In other words, ignore it or use it. But do not use it in one place and ignore it in the other.
On cross-examination the Department’s witness said that he preferred disallowing the deferred tax expense as an offset to the income stream to be capitalized. He conceded, however:
I recognized this morning that whatever value deferred taxes have to the railroad, they are not as valuable as a dollar of earnings because they have to be paid back. I said that by using the five-year average of current deferred that you add back to income, you are really kind of discounting them by thirty percent from what this year’s deferred tax was. As to whether I feel that’s proper because since they will have to be paid back and they are just not as valuable as income itself, well, there is considerable question about being paid back. I think that if they do have to be paid back that they are an interest-free loan.
I do acknowledge that on a single equipment purchase such as a boxcar, the tax is deferred by use of accelerated depreciation and shorter asset lives or, life, in that case. It will eventually be paid back on that particular boxcar. [Emphasis added.]
The Commission determined not to treat the deferred tax expense as an interest-free source of capital; instead the Commission simply adopted the Department’s alternative treatment and *194added back the five-year average for deferred taxes ($15,524,000) to Southern’s NROI in arriving at income to be capitalized. Again, applying the whole-record test to this issue, we are satisfied the Commission erred.
The basis for the Commission’s determination was that deferred income taxes are not a presently outstanding indebtedness but a mere contingency which may never be paid. The Commission relied on Realty Corp. v. Coble, Sec. of Revenue, 291 N.C. 608, 231 S.E. 2d 656 (1977). In Coble this Court held that an installment method taxpayer may not deduct from its franchise tax base deferred federal and state income tax liabilities. The decision in Coble was based on the specific language of the franchise tax statute which permitted certain deductions from the franchise tax base only “for definite and accrued legal liabilities” and for “taxes accrued.” The Court held that deferred income taxes carried as an expense on the taxpayer’s books were not “definite and accrued legal liabilities” or “taxes accrued” within the meaning of the franchise tax statute.
The question before the Commission was whether the deferred income tax expense is properly deductible from the Railroads’ NROI in arriving at income to be capitalized under the income approach to value. On this issue the Railroads’ evidence, much of which was not challenged by the Department’s evidence, is not only clear and cogent, it is overwhelmingly convincing.
This testimony demonstrated, and the Department’s witness did not contravene it, that in order for deferred income taxes to be perpetually immune from payment, the Railroads would have to maintain increasingly greater levels of investment necessary to obtain new depreciation in amounts sufficient to offset the reduced depreciation attributable to older assets. Further, the accelerated depreciation provisions of the income tax laws would have to remain in place. Railroads’ evidence demonstrated that potential buyers and sellers would not appraise the railroad system on the assumption that these kinds of investments would continue to be made, or that accelerated depreciation provisions would be forever with us. This is true notwithstanding the fact that Southern’s capital acquisitions over the last several years *195have been so large that it has continued to accumulate deferral tax expenses and, in fact, has paid no income tax.6
Normally prospective buyers and sellers evaluate the income of a company by eliminating from it expenses associated with it, even though the expense may not be payable until later. The expense is recognized as a cost of earning the income and should be accounted for accordingly. Dean Keller testified:
Tax expense is essentially related to the earning of income. So, in any particular period in which income is earned, the tax expense should be recognized as having been incurred in that particular period.
Current deferred taxes are not payable currently; however, in my opinion they should be recognized as a current expense. The tax expense is, as I said, related to the earning of income; so in the period in which the income is earned you would recognize the expense.
If I were consulted by a willing seller or a willing buyer to give advice as to the proper income stream to be capitalized in valuing a business, I would not consider it reasonable to add back any portion of the current deferred tax expense of the business to the income stream. None at all.
NROI, the income stream to be capitalized, is, after all, income after depreciation and taxes are deducted. To fail to deduct a tax expense which would have to be paid but for accelerated depreciation schedules, from the standpoint of a prospective buyer, over*196states the NROI for the period in which both the income and the taxes attributable to it occur.
Both the Railroads and the Department seem to agree that deferral of federal income taxes is beneficial to the owner-seller largely because it provides cash for asset acquisition which would otherwise be paid to the government. The Railroads’ evidence showed, however, that the accumulated deferred tax expense has no value to a prospective buyer. It cannot be transferred from seller to buyer. If deferred taxes can be viewed as value to the buyer, that value is more properly reflected in the value of income generated by assets purchased with funds attributable to the deferral. The Department’s expert, Mr. Underhill, admitted that the value of assets purchased with deferred taxes would appear in the capitalization of earnings from those assets. But to include in income to be capitalized both the deferred tax expense and the income earned from the use of that money, in the testimony of Dr. Schoenwald, “double counts the benefit,” resulting in an overstatement of value which, rather than attracting a prudent investor, would discourage investment.
Our holding on this issue finds support in the only two reported cases considering the question in a similar context. See Southern Railway Co. v. State Board of Equalization, 682 S.W. 2d 196 (Tenn. 1984); Pac. Power & Light Co. v. Dept. of Revenue, 596 P. 2d 912 (Or. 1979).
D.
The Commission Erred on this Record in Determining the “Income Influence Percentage” in the Stock and Debt Approach
The Court of Appeals correctly summarized the stock and debt approach to value as follows:
This appraisal technique operates on the premise that the true property value of a company equals the total market value of all its outstanding debt and equity securities. However, all non-system property must be eliminated to arrive at the true value of the system operation. Under the ‘income influence approach,’ the appraiser determines the ratio of non-system income to total income before fixed charges (i.e., the income available to both bondholder and stockholder), and then multiplies that ratio by the total value of the *197company’s stock and debt. The resulting figure is the ‘income influence’ of the non-system property. This figure is deducted from the total stock and debt value. The final figure represents the true stock and debt value for the Railroads’ system property.
59 N.C. App. at 131, 296 S.E. 2d at 471.
In arriving at the total income upon which the income influence ratio is figured, the Department in one calculation added back the deferred income tax expense. As we have already held, the Commission erred on this record in adopting this method of arriving at income. Further, the Department excluded undistributed earnings ($20,660,000 in Southern’s case) of subsidiaries from both the nonsystem and the total income. This resulted in reduced nonsystem income, a smaller “income influence” for the nonsystem assets, and a larger system value. The Railroads argue that undistributed earnings from nonsystem subsidiaries should be included. In support of their position, the Railroads offered the testimony of Dr. Schoenwald who explained:
If you have more coverage for a debt security or the interest expense on a debt security by virtue of additional income from nonsystem sources, this increases the market value of that security. The greater the coverage, the greater the safety; therefore, the higher the price of that security. In other words, investors would be willing to take a lower interest rate from that type of company than from a company which has marginal coverage and has no nonsystem income to support that coverage.
Earlier, Dr. Schoenwald stated:
The elimination of those earnings from the nonoperating income reduces the nonoperating income influence percentage; therefore the deduction for nonsystem property under the stock and debt approach is inadequate, and the additional value flows through into the estimate of railroad value made by the Department of Revenue. The market value of Southern Railway stock definitely reflects all of the undistributed earnings of all the companies. You have got to deduct influence of that equity in order to get the proper value of Southern Railway Company.
*198Mr. Underhill’s reason for excluding undistributed earnings of nonsystem subsidiaries was that they offset the exclusion of the long-term debt of the nonsystem subsidiaries in determining the system’s stock and debt value. Mr. Underhill conceded that to consider the nonsystem subsidiaries’ long-term debt in determining the nonsystem’s influence on the system’s stock and debt values was a “new concept” in railroad appraising but one which he personally thought ought to be used. Traditionally, only stock and debt values of the parent company, i.e., the system, offset by the influence of nonsystem values, are considered in the stock and debt approach.
In its determination of this issue the Commission merely adopted the Department’s figures without reference to the inclusion or exclusion of undistributed earnings. The Court of Appeals affirmed by simply concluding that “[rjetained earnings of a subsidiary have little or no effect on the value of Southern’s common stock.”
We hold that this record does not support exclusion of non-system subsidiaries’ undistributed earnings in determining their income influence on the system’s stock and debt values. We disagree with the Court of Appeals that the subsidiaries’ retained earnings have little or no effect on the value of the parent Southern’s common stock. The only evidence of record supports a contrary conclusion.
IV.
In its appraisal of Norfolk Southern, the. Department relied almost entirely on the income approach to value. In determining Norfolk Southern’s NROI to be capitalized, it added back the deferred income tax expense. It then capitalized the income at a rate based in part on Norfolk Southern’s embedded cost of debt rather than current market cost. It arrived at a value of $59,624,725 with this method. Had the Department not added back the deferred income tax expense to arrive at NROI and had it used current market cost of debt in determining the capitalization rate, it would have arrived at a value under the income approach of $46,156,000.
We have held that the Commission on this record erred in adopting the Department’s methods of adding back deferred in*199come tax expense to determine NROI and using embedded rather than market cost of debt. We conclude, therefore, that the Commission erred in adopting the Department’s appraisal of Norfolk Southern based in large part on these methods.
V.
We wish to emphasize that this is an ad valorem tax evaluation case. Our resolution of the questions presented would not necessarily be the same were we addressing the proper methods of valuation for rate making purposes.
We reverse the Court of Appeals and remand the case to that court with instructions that it remand to the North Carolina Property Tax Commission to determine the system valuation of Railroads’ property in a manner consistent with this opinion.
Reversed and remanded.
Justice VAUGHN did not participate in the consideration or decision of this case.. After these proceedings were begun, Norfolk Southern’s name was changed to Carolina and Northwestern Railway Company. To be consistent with the record, we shall refer to this company as Norfolk Southern.
. Since all references to statutes herein are contained in Subchapter II of Chapter 105, we shall refer only to section numbers of the chapter.
. The Railroads’ evidence tended to show, for example, that the Department’s appraisal methods challenged on this appeal resulted alone in an income approach to value approximately 27 percent higher in the case of Southern Railway and 29 percent higher in the case of Norfolk Southern than would have been the case if the methods advocated by the Railroads had been used.
. Under the “cost approach” the true value of a system is presumed to be equivalent to its original or “book value” cost “less a reasonable allowance for depreciation.” See § 336(a)(2). In this case the Railroads and the Department, because of the nature of the Railroads’ assets, agree that the cost approach should be accorded little or no weight as a method of appraisal.
. The Department used a 15 percent return on common equity and Southern’s actual 7.2 percent embedded debt cost for the debt component to arrive at a weighted capitalization rate of 12 percent for Southern. The Railroads’ expert used an 18 percent rate for the equity component and a 10.5 percent rate for the debt component (based on current cost of borrowed money) to arrive at a weighted capitalization rate of 15.25 percent for Southern. Railroads are contesting only the Department’s use of the 7.2 percent debt component rate based on Southern’s actual embedded cost.
. We acknowledge, too, several scholarly articles cited in the Department’s Brief which tend to support the Department’s position on this question. See Fn. Davidson, “Accelerated Depreciation and the Allocation of Income Taxes,” 33 Accounting Rev. 173 (1958); Warren, “Tax Accounting in Regulated Industries, Limitations on Rate Base Exclusions,” 31 Rutgers L. Rev. 187 (1979); Davidson, Kirsch and Palast, “Utilities, Accelerated Depreciation and Income Tax Allocation: An Empirical Study,” Public Utilities Fortnightly (July 2, 1981); “Is Generally Accepted Accounting for Income Taxes Possibly Misleading Investors,” Price Waterhouse and Co., 1967. It suffices to say that when we apply the whole record test, as we are bound to do, to determine whether the Commission erred, we are limited to a consideration of the record as it existed before the Commission. There is no indication in the record or briefs that these articles were made a part of the record below or considered by the Commission in reaching its decision.