Boston Gas Company, doing business as Key span Energy Delivery New England (company), made timely application to the board of assessors of Boston (assessors) for abatement of the tax imposed on its rate-regulated3 utility property in the city of Boston (city) in fiscal year 2004.4 At issue are separate assessments of the company’s personal property and real property. After being denied abatements by the assessors, the company appealed to the Appellate Tax Board (board). With respect to the personal property, the board determined that a valuation methodology according equal weight to the property’s net book value and its reproduction cost new less depreciation (RCNLD) provided a reliable estimate of the fair cash value of the property. As that value was in excess of the assessed value, the board denied the company relief. With respect to the real property, the board concluded that neither the company nor the assessors had provided a sufficient basis for valuing the property, and so left the assessed value undisturbed.
The plaintiff filed a notice of appeal from the decision of the board, and we granted direct appellate review. On appeal, the plaintiff claims (1) the board lacked substantial evidence in support of its determination that a valuation method other than net book value was permissible, and that the board therefore erred in according equal weight to net book value and RCNLD; (2) the board lacked substantial evidence to support the analysis of earnings before interest, taxes, depreciation, and amortization (EBITDA) in the income capitalization approach for valuing the personal property; (3) the board erred in failing to use a tax factor to account for property taxes in the income capitalization *717approach; (4) the board erred in its weighing of certain evidence and in its assessment of the credibility of a key witness; and (5) the board erred in concluding that there was insufficient evidence to determine the value of the real property. We remand to the board for further consideration of the use of a tax factor in the income capitalization approach, and of certain subsidiary conclusions related to the EBITDA analysis in that approach. On all other issues, we affirm the board’s decision.
1. Background, a. The legal framework. City assessors are charged with making a “fair cash valuation” of property that is subject to taxation. G. L. c. 59, § 38. We have determined “fair cash value” to mean “fair market value,” or “the price an owner willing but not under compulsion to sell ought to receive from one willing but not under compulsion to buy.” Boston Gas Co. v. Assessors of Boston, 334 Mass. 549, 566 (1956). When challenging an assessment before the board, the taxpayer bears the burden of establishing its right to an abatement of the assessed tax. See Schlaiker v. Assessors of Great Barrington, 365 Mass. 243, 245 (1974), quoting Judson Freight Forwarding Co. v. Commonwealth, 242 Mass. 47, 55 (1922). The assessment is valid unless the taxpayer sustains its burden of proving otherwise. See Schlaiker v. Assessors of Great Barrington, supra at 245.
Various methods are used to value taxable utility property. These include (1) a determination of the property’s net book value, (2) an income capitalization valuation, (3) a sales comparison valuation, and (4) a determination of RCNLD.5 See Tennessee Gas Pipeline Co. v. Assessors of Agawam, 428 Mass. 261, 263 (1998), citing Montaup Elec. Co. v. Assessors of Whitman, 390 Mass. 847, 850 (1984).
The Department of Public Utilities (DPU) regulates the rates that gas companies charge to consumers. See G. L. c. 164, § 94. The net book value of regulated utility property, also known as the “rate base” value, plays an important role in the DPU’s calculation of the revenue that a regulated gas utility is permitted *718to earn. See Tennessee Gas Pipeline Co. v. Assessors of Agawam, supra at 263. The DPU allows a utility to recover, through the rates charged to consumers, its reasonable operating expenses, taxes, depreciation and amortization, and other costs. Boston Gas Co. v. Department of Telecomm. & Energy, 436 Mass. 233, 234 (2002), quoting Theory and Implementation of Incentive Regulation, D.P.U. 94-158 at 3 (1995); Boston Gas Co., D.T.E. 03-40-B at 13-20 (2004) (breaking out the components of Boston Gas’s revenue requirements). A utility is also permitted to earn a reasonable return on investment, which is calculated as a percentage return on the utility’s rate base. See Boston Gas Co. v. Department of Telecomm. & Energy, supra at 234; Boston Gas Co., supra at 16 (calculating return on rate base for company). The cost of utility property may be included in the utility’s rate base if the property is “used and useful to customers” and if the costs were “prudently incurred.” See Hingham v. Department of Telecomm. & Energy, 433 Mass. 198, 202 (2001). For ratemaking purposes, the value of property included in the rate base is its net book value, which has been defined as “the original cost of the property at the time it was originally devoted to public use, less accrued depreciation.” Tennessee Gas Pipeline Co. v. Assessors of Agawam, supra at 263.
In the context of a sale of utility assets, the DPU has maintained a general policy of limiting the net book value of the assets in the hands of the buyer to the existing net book value in the hands of the seller. See id. In this way, any acquisition premium paid for the assets — that is, an amount paid above net book value6 — would be excluded from the buyer’s rate base, and the buyer would thus not earn the DPU-specified rate of return on the premium; as of 2003, the DPU stated that such exclusion remains the norm. See Boston Gas Co., D.T.E. 03-40 at 323 (2003). This policy has been referred to as the “carry-over rate base principle.” Montaup Elec. Co. vs. Assessors of Whitman, supra at 852-853.
As a result of this regulation, we have stated that the net *719book value of utility assets is the proper value for assessment purposes, absent “special circumstances” that would induce a buyer to pay more than net book value. Tennessee Gas Pipeline Co. v. Assessors of Agawam, supra at 263-264. Such circumstances may include (1) that “the utility company’s net earnings actually may exceed the rate of return approved by the regulatory agency”; (2) that “the profit available from this transaction may exceed that which an investment of comparable risk could bring in the open market”; (3) that “the applicable regulatory agency may change its policies and abandon the carry-over rate base principle, thereby making an investment in the company more attractive,”7 Montaup Elec. Co. v. Assessors of Whitman, supra at 852-853; or (4) “[t]he potential for growth in a utility’s business.” Boston Edison Co. v. Assessors of Watertown, 387 Mass. 298, 305-306 (1982). The special circumstances that could induce a buyer to pay more than net book value are not limited to the examples enumerated above. See id. at 306.
b. Facts. With this understanding of the legal framework for valuing regulated utility property, we turn to the facts of the present case. We recite the basic facts here, reserving a more detailed discussion of some of the facts for our subsequent analysis.
The personal property at issue consists primarily of the pipes, lines, and meters used to transport and monitor the distribution of natural gas to the company’s customers within the city. The pipes, or “mains,” represent about eighty percent of the value of the personal property. Approximately two-thirds of the mains are made of cast iron — a material used between 1850 and 1950 — and the remainder consists of steel, which was introduced in the 1930’s, and plastic, which was introduced by 1970.
The real property at issue is a parcel known as Commercial Point. The majority of the parcel is used as a liquid natural gas storage and distribution facility, and is improved with a 1.13 billion cubic foot storage tank, a cooling tower, a containment dike, and a monitoring and control building.
For fiscal year 2004, the assessors valued the personal property *720at $223.2 million and the real property at $28 million. The net book value of the personal property was approximately $159.2 million, and the net book value of the real property was approximately $1.8 million. The plaintiff paid the assessed taxes and, as stated, timely filed applications for abatement with the assessors for both the real and personal property. The assessors denied the applications, and the plaintiff appealed both matters to the board.
At the hearing before the board, the assessors’ expert, George Sansoucy, presented his appraisal report for the personal property. He derived a value for the property based on three valuation methodologies: an income capitalization approach,8 a sales comparison approach,9 and a RCNLD approach.10 The board ultimately concluded that an equal weighting of the RCNLD value (once adjusted for a discovered error) and the net book value provided a reliable estimate of the fair cash value of the property. That weighting yielded a value of $248 million, exceeding the assessed value of the personal property of $223.2 million. The board accordingly concluded that the plaintiff had not met its burden of showing that the property was assessed at a value above its fair cash value.
The board concluded that a number of factors justified the use of a valuation methodology other than net book value. Specifically, the board determined that (1) there has been a trend in Massachusetts regulatory policy away from a strict carry-over rate base valuation model; (2) the useful life of gas utility pipeline vastly exceeds its depreciable life,11 so that the *721personal property has residual value well in excess of net book value; and (3) sales activity in the marketplace indicates that, in practice, purchasers of utility property have paid substantially more than net book value.
Finally, with respect to the real property, the board determined that the appraisal evidence offered at the hearing was substantially flawed, and thus did not establish a sufficient basis on which to determine a value for the parcel that differed from the assessed value. As a result, the board concluded that the plaintiff had not met its burden of proof in challenging the assessment. As neither of the experts who testified on the matter provided a sufficient basis for valuing the parcel, the board did not address a dispute between the experts as to the size of the parcel or the portion of it that was “upland acreage.”
2. Discussion. A decision of the board will not be reversed or modified if it is based on substantial evidence and a correct application of the law. Koch v. Commissioner of Revenue, 416 Mass. 540, 555 (1993), citing Commissioner of Revenue v. Wells Yachts S., Inc., 406 Mass. 661, 663 (1990). Substantial evidence is “such evidence as a reasonable mind might accept as adequate to support a conclusion.” Tennessee Gas Pipeline Co. v. Assessors of Agawam, 428 Mass. 261, 262 (1998), quoting New Boston Garden Corp. v. Assessors of Boston, 383 Mass. 456, 466 (1981). Our consideration of the substantiality of the evidence “must take into account whatever in the record fairly detracts from its weight.” New Boston Garden Corp. v. Assessors of Boston, supra at 466, quoting Cohen v. Board of Registration in Pharmacy, 350 Mass. 246, 253 (1966). “Where there is substantial evidence to support the board’s decision, we defer to the board’s judgment as to what evidence to accept and which method or methods of valuation to rely on.” General Electric Co. v. Assessors of Lynn, 393 Mass. 591, 608 (1984), quoting Boston Edison Co. v. Assessors of Watertown, 387 Mass. 298, 302 (1982). However, we must set aside a finding of the board if “the evidence points to no felt or appreciable probability of *722the conclusion or points to an overwhelming probability of the contrary.” Tennessee Gas Pipeline Co. v. Assessors of Agawam, supra at 262, quoting New Boston Garden Corp. v. Assessors of Boston, supra at 466.
a. Valuation of the personal property, i. Use of a valuation methodology other than net book value. The company asserts that the board erred in using an equal weighting of RCNLD and net book value in valuing the personal property, arguing that the board should have been limited to using net book value. To support this contention, the company maintains that the board’s stated reasons for departing from strict adherence to net book value were flawed.12
First, the company claims that substantial evidence does not support the board’s findings that changes in the regulatory environment for utilities justified the use of a valuation method other than net book value. We do not agree.
In Boston Edison Co. v. Assessors of Boston, 402 Mass. 1, 13 (1988), we affirmed the board’s decision to value utility property by equally weighting net book value and depreciated reproduction cost. There, we held that the board reasonably saw, based on a prior decision of the DPU upheld by this court, “the possibility that the department might allow adjustments in a purchaser’s rate base to reflect a prudent purchase price above the plant’s net book cost.” Id. at 15, citing Attorney Gen. v. Department of Pub. Utils., 390 Mass. 208, 210-211 & n.3, 217 (1983) (affirming DPU decision to allow company to recover, through amortization expenses, its prudent investment in power plant reasonably abandoned before completion, and to receive carrying charge to compensate it for delay in recovery).
The DPU formalized a shift in its policy with respect to the carry-over rate base principle in a 1994 order regarding mergers and acquisitions of utilities. See Guidelines & Standards for Acquisitions & Mergers, D.P.U. 93-167-A (1994) (Mergers & Acquisitions). There, the DPU stated that it would “no longer follow the practice of denying acquisition premium recovery on a per se basis,” id. at 18, concluding that “[mjerger proposals *723that include an acquisition premium will henceforth be judged on a case-by-case basis.” Id. at 7. The DPU discussed proposals regarding the precise treatment of the premium, such as including the premium in rate base, but concluded only that “the [DPU] will consider [the] appropriate level of a recoverable acquisition premium on a case-by-[case ]basis.” Id. at 19. The ruling appeared to contemplate the possibility both of a return of the acquisition premium — for example, as a recoverable cost to the company — and a return on the acquisition premium by including it in the acquirer’s rate base.
This court acknowledged the DPU’s regulatory change in Stow Mun. Elec. Dep’t v. Department of Pub. Utils., 426 Mass. 341, 347 (1997). In that case, the town of Stow was purchasing the electricity distribution system of Hudson Light and Power Department, and Stow petitioned the DPU13 for a determination of a purchase price. See id. at 342-343. The DPU used an equal weighting of RCNLD and original cost less depreciation,14 and we upheld that determination. See id. at 343, 345. Although Stow argued that the DPU had “impermissibly speculated” in discussing possible regulatory change, we acknowledged the DPU’s shift from a mandatory carry-over rate base policy to a case-by-case approach, and concluded that the DPU had not erred in considering the effect of that change on value of the utility. Id. at 347.
Finally, in Attorney Gen. v. Department of Telecomm. & Energy, 438 Mass. 256, 258 (2002), we reviewed a decision by the DPU15 to approve a rate plan proposed by a group of utility companies. Those companies planned to merge, but made the merger contingent upon approval of the rate plan by the DPU. See id. at 258-259. The rate plan allowed the companies to recover the acquisition premium paid to consummate the merger. See id. at 259-260. We again acknowledged the regulatory policy *724change announced in the Mergers & Acquisitions order, and affirmed the DPU’s decision and order.16 See id. at 259, 263.
These cases and DPU orders amply demonstrate the type of regulatory change anticipated in Boston Edison Co. v. Assessors of Watertown, supra at 305-306, justifying the use of a valuation methodology other than net book value. The DPU has declared its abandonment of a strict carry-over rate base policy, this court has repeatedly and recently acknowledged that policy change, and the DPU has, in practice, allowed the recovery of a premium in a utility merger.17
The company contends that developments since the 1994 change in DPU policy in the Mergers & Acquisitions order have made it “less . . . likely” that a step-up in rate base would be allowed. The company acknowledges that the DPU has allowed recovery of an acquisition premium in a merger context but claims that the DPU has not allowed an increase in rate base value following an acquisition since the Mergers & Acquisitions order. Such an argument may speak to a diminished probability of a buyer earning a return on an acquisition premium, but factors bearing on valuation need not be certainties before the board may consider how they would manifest in a hypothetical sale. The company’s evidence does not rebut evidence of the DPU’s current practice of considering inclusion of premiums in the buyer’s rate base on a case-by-case basis. The evidence in the record warrants a finding that a potential buyer of the subject property could reasonably conclude that the DPU no longer fol*725lows a policy of “denying acquisition premium recovery on a per se basis,” and that the buyer may be able to earn a return on the premium. As the board cited substantial evidence for its conclusion that regulatory change had called into question the exclusive use of net book value, we affirm the board’s decision in this regard.
The company also disputes the board’s second reason for concluding that it was proper to rely on a valuation method other than net book value: evidence from prior transactions that utility assets in fact sold for more than net book value. The validity of the sales comparisons are relevant to the resolution of this case in two ways. First, as noted, they serve as a second reason cited by the board for departing from a strict adherence to net book value as the value of the subject property. Second, the assessors’ expert relied on such sales in each of his three valuation approaches, and the board ultimately credited the expert’s use of those sales in determining the value of the property. The company argues that the sales comparisons were flawed and, as a result, that they did not justify the use of a valuation method other than net book value. The company likewise maintains that the sales should not have been relied upon in estimating the value of the subject property.
At the hearing, the assessors’ expert, George Sansoucy, testified that he reviewed twenty-two sales of gas utility property in the United States over the preceding decade, and chose from among them six sales he thought most comparable to the utility property in the present case.18 Each of the six sales was part of an acquisition or merger transaction that included more than the regulated utility assets alone. As a result, for each sale, the expert endeavored to isolate the portion of the sale price that was paid for the regulated utility assets, and described his procedure for removing various value elements not related to that property. The expert’s analyses indicated that utility assets were sold for prices well in excess of net book value.
The company takes the position that the assessors’ expert had not in fact isolated the price paid for the utility assets. It alleges that in merely subtracting unrelated assets from the price paid for entire utility enterprises, the expert had attributed to the *726remaining tangible utility property a portion of the purchase price that had actually been paid for intangible sources of value, and that such intangible value is not taxable. See G. L. c. 59, §§ 2, 18 (authorizing taxation of tangible personal property). See also G. L. c. 59, § 5, Twenty-fourth (exempting intangible personal property from taxation). The company sought to support its view with evidence showing (1) that no buyer would pay more for the subject utility assets than their net book value, and (2) that sales of utility enterprises include intangible value that the assessors’ expert failed to deduct from the purchase price when analyzing prior transactions. The company asserts that, in overstating the price paid for tangible assets in its analysis of prior comparable transactions, the board in turn overstated the price that would be paid for the company’s tangible assets.
To that end, the company presented evidence and testimony from Susan Tierney, an economic and regulatory consultant and former DPU commissioner. She described how the value of a utility enterprise as a whole may include sources of value other than the regulated utility assets. Sources of value beyond the utility assets could include, inter alia, intellectual property, brand name, management acumen, customer base, business relationships, and economies of scale. She also opined that, because utilities earn a return only on their rate base, a buyer of regulated utility property would not pay more for the property than the expected net book value, and that the “norm” of the carry-over rate base principle would restrict the willingness of a buyer to pay more than the seller’s existing net book value. Thus, she reasoned, the portion of the price paid that was in excess of net book value was for sources of value in the enterprise other than the regulated utility assets.
The company’s other relevant evidence came primarily in the testimony of Joseph Bodanza, who was the chief accounting officer and head of regulatory affairs for Keyspan Corporation in 2003, and who had been an executive with the company prior to Keyspan Corporation’s acquisition of the company’s parent, Eastern Enterprises. He testified that, with respect to the Eastern Enterprises acquisition of utility company Colonial Gas Company (Colonial Gas) (one of the six transactions analyzed by the assessors’ expert), the premium paid was projected to be offset by synergies from elimination of redundant staff and facilities, and *727acquisition of management expertise.19 In testimony pre-filed with the DPU regarding the rate plan for Colonial Gas after the merger, Bodanza also projected a reduction in gas prices as a result of operational efficiencies and synergies in the merger, as well as savings on avoided technology investments. The company claims that those synergies and other intangibles are among the sources of value that account for the premium paid in prior transactions.
To counter the assertion by the company that any premium was necessarily paid for value other than the tangible assets, and thus to reinforce the analysis of its own expert, Sansoucy, the assessors presented evidence from an expert witness in public utilities. This expert discussed a prior series of utility transactions in Rhode Island, a State that he described as having a regulatory regime somewhat similar to that of Massachusetts.20 In the described transactions, a company had acquired two utility enterprises through mergers, and had paid substantial premiums in the transactions. Upon consummation of the transactions, the utility operations of the acquired companies became an unincorporated division of the acquiring company. Approximately six years later, the acquiring company sold nearly the same group of assets it had acquired in the mergers for approximately the same price and premium it had paid for the two utility enterprises. The assessors contended that such evidence demonstrates that the premiums paid in acquiring utility enterprises are for the underlying utility assets rather than for the other sources of value described by the company’s witnesses.
In its decision, the board credited Tierney’s distinction between the value of rate-regulated utility property and the value of a broader business enterprise. The board found unsubstantiated, however, her insistence that any amount paid in excess of net book value was necessarily for enterprise value other than the utility assets. With respect to the potential intangible sources of value that Tierney had raised, the board noted that she could not recall an instance of a utility company owning intellectual *728property; that brand name was not shown to hold discernable value in the present case, as evidenced by the change of the name on the company’s Commercial Point tank after changes in ownership; and, consistent with evidence introduced at the hearing, that there was no discernable value in the customer base of a gas company whose customers are essentially captive.
The board concluded that the company had not shown that the hypothetical sources of intangible value that Tierney had discussed were in fact present in the sales analyzed by the assessors’ expert, and noted also that the company did not present any expert testimony from a qualified appraiser of utility property who might have offered such evidence. With respect to Bodanza’s testimony, the board found that he had not broken out the components of value in the Eastern Enterprises acquisition, or the contribution of intangible value to the purchase price. The board also emphasized that Bodanza was neither presented nor qualified as an expert on the valuation of utility property, and that his testimony was accordingly of limited worth in determining the value of the subject assets.
By contrast, the board credited the assessors’ expert’s testimony as to the portions of prior sale premiums attributable to regulated utility assets, and credited the evidence from the Rhode Island transactions as to the minimal intangible value involved in regulated utility sales. The board accordingly found that a valuation method other than net book value was warranted in the present case, and that the methods of the assessors’ expert — including implicitly the sales comparisons upon which they rely — were sound.
In the present appeal, the company restates its contention that the analysis of prior sales failed to account for intangible assets, and argues that the board’s findings were thus unsupported and amounted to impermissible taxation of intangible value. As such, the company claims, the analysis of prior sales does not provide a legitimate justification for using a valuation methodology other than net book value. We disagree.
“Although the burden of establishing overvaluation is on the [company],. . . until there is some evidence offered by the assessors to show that, because of [special] circumstances, the relevance of rate base value is put in question,” the company is *729not required to show the absence of such circumstances. Montaup Elec. Co. v. Assessors of Whitman, 390 Mass. 847, 855 (1984), citing Foxboro Assocs. v. Assessors of Foxborough, 385 Mass. 679, 691 (1982). Thus, net book value remains the appropriate method until the assessors introduce evidence of such special circumstances. Evidence in the record supports the board’s conclusion that the assessors did so, including the assessors’ detailed evidence of prior utility sales that featured acquisition premiums, and the evidence introduced relating to the Rhode Island transactions.
Once the assessors put the exclusive use of net book value in question, the company could have prevented the use of methods other than net book value by rebutting the assessors’ evidence. See Tennessee Gas Pipeline Co. v. Assessors of Agawam, 428 Mass. 261, 264 (1998). As the company acknowledges, the board adopted the testimony of the company’s expert that the value of a utility enterprise may well include value beyond that of its regulated utility assets. The board found, however, that the company did not offer evidence that would allow the board to deduct the value of such intangibles from the prices paid for the utility enterprises examined in this case, if those intangibles were indeed present. While Bodanza’s pre-filed testimony in particular presented figures for expected savings based on his projection of synergies, the board provided cogent reasons for making little use of his testimony. We do not think Bodanza’s evidence so detracts from the substantiality of the evidence that there is no longer a “felt or appreciable probability” of the board’s conclusion. See id. at 262. In sum, we conclude that the board relied on sufficient evidence in determining that special circumstances warranted the use of a valuation method other than net book value.21
ii. Substantiality of the evidence supporting the EBITDA analysis in the income capitalization approach. Although the board rested its final valuation on an equal weighting of net *730book value and the RCNLD approach, the RCNLD approach relied on the outcome of the assessors’ expert’s separate income capitalization approach. Specifically, after estimating the cost of reproducing the property in the RCNLD approach, and after accounting for physical depreciation and functional obsolescence, the expert took the further step in the RCNLD analysis of accounting for the external, or economic, obsolescence of the property as hypothetically reproduced. External obsolescence is a type of depreciation that takes account of market factors external to the property that have an impact on its fair market value, such as an economic recession that decreases demand for the property. See Appraisal Institute, The Appraisal of Real Estate, 442 (13th ed. 2008) (Appraisal of Real Estate).
To account for external obsolescence in the RCNLD approach, the expert simply decreased the RCNLD value to the value derived from the income capitalization approach. He reasoned that the income approach accounts for the effect of regulation and other external factors on the value of the property in a way that the RCNLD approach, before an external obsolescence adjustment, does not. He opined that the difference between the higher RCNLD value (before accounting for external obsolescence) and the lower income approach value was thus itself a measure of external obsolescence. Subtracting this difference from the RCNLD approach, of course, renders the value from the RCNLD approach equivalent to that of the income approach. It was this final RCNLD value that the board weighted equally with net book value to reach its final valuation.22 In this regard, the company argues that the board’s valuation, despite purporting to rely on the RCNLD approach, effectively reduces to reliance on the income capitalization approach.
We accord deference to the expertise of the board in its choice of an appropriate methodology for valuing the subject property. Here, the board gave weight to the assessors’ expert’s income capitalization approach for the purpose of estimating economic *731obsolescence in the board’s RCNLD valuation. Given the importance of the income capitalization approach to the board’s final valuation, we conclude that the income approach must itself be sound. The company alleges that there are errors in the estimation and capitalization of earnings in the income capitalization analysis, and we partially remand these issues to the board for further consideration.
In his analysis, the assessors’ expert first estimated the annual EBITDA attributable to the company’s subject property. He then capitalized this figure, using an “EBITDA multiplier,” to arrive at a valuation of the property.23 The EBITDA multiplier was derived by looking at six comparable sales of regulated utility property, and calculating the ratio of sales price to annual EBITDA for each. The approximate average of these figures was the expert’s EBITDA multiplier. Multiplying the EBITDA of the company’s property by the EBITDA multiplier, the expert derived an estimated sales price of the company’s property.
Rather than capitalizing a single year of the company’s EBITDA from the subject property, the expert chose to “smooth” the EBITDA estimate by taking a seven-year sample of the company’s annual EBITDA figures. Those years were calendar years 1997 through 2003. The expert eliminated year 2000, the year for which there was the lowest EBITDA, for reasons that are not challenged by the company. He also eliminated year 2001, the year for which there was the second-lowest EBITDA, on account of the abnormal amounts of deferred income taxes and amortization expenses that were taken that year.
Because neither income taxes nor amortization expenses enter into EBITDA, the company alleges that the board provided no evidence upon which to exclude the 2001 figure. The board’s decision states only that the expert “removed 2001 because of discrepancies relating to depreciation and amortization.”24
*732Even according substantial deference to the board, we can discern no reason from the evidence in the record why abnormal depreciation and amortization expenses — or a Federal income tax issue discussed in the expert’s testimony — should result in the exclusion of the 2001 EBITDA figure, and the board’s statement on the issue does not clarify its decision in that regard. The document on which the expert relied shows the components of EBITDA, none of which are among the figures he identifies as anomalous. As this change would have the potential to decrease the assessed value, contingent upon the outcome of other issues remanded in this opinion, we remand this issue for further consideration by the board.
The company also objects to the inclusion of EBITDA from calendar year 2003, because those earnings were generated after the relevant assessment date of January 1, 2003. See G. L. c. 59, § 18 (establishing January 1 assessment date for personal property). The fact that the data arise after the relevant assessment date, standing alone, does not mean that they are per se excluded from the board’s consideration. Cf. Teele v. Boston, 165 Mass. 88, 91-92 (1896) (allowing consideration of comparable sale that took place after date of eminent domain taking); Roberts v. Boston, 149 Mass. 346, 354 (1889) (holding that “[t]he mere lapse of time after [a] taking did not render the evidence of. . . sales incompetent,” and that “the discretion of the court seems to have been rightly exercised in admitting the evidence”). Such evidence can be used, where its probative value is not outweighed by the risk of hindsight bias or other factors, for the limited purpose of determining the value of the property as of the assessment date. In so doing, the relevant inquiry is whether the evidence reflects information that would be knowable to a hypothetical buyer and seller of the subject property as of the assessment date.
In the present case, the 2003 figure was one of five data points used for the purpose of “smoothing” an annual EBITDA figure, so that evidence from before the assessment date formed the principal basis of valuation.25 The evidence from after the assessment date was limited to the immediately subsequent *733year, and the earnings were generated from property that was largely unchanged. We conclude that, in the circumstances, the board did not err in allowing the 2003 figures to be included in the averaged EBITDA value.
Finally, the company argues that the board erred in failing to adjust the expert’s EBITDA multiplier for an alleged error in his analysis. For the reasons outlined below, we remand this issue for consideration by the board.
As discussed, the EBITDA multiplier was calculated as an approximate average of the ratio of sale price to EBITDA in six prior sales. One of the six sales used was Keyspan Corporation’s acquisition of Eastern Enterprises in 2000. The expert’s report indicates that at the time of that sale, Colonial Gas was a subsidiary of Eastern Enterprises, and the expert testified that Colonial Gas had become a subsidiary of Eastern Enterprises in August of 1999. The company maintains that the expert erred in calculating the sales price to EBITDA ratio for that sale by using a sale price from 2000, which included the amount paid for Colonial Gas, while using an EBITDA from the end of 1998, which did not include Colonial Gas’s contribution to EBITDA. As a result, the company asserts that the expert’s sales price to EBITDA ratio was inflated.
The company cross-examined the expert in detail on the issue, introduced an exhibit showing the proposed recalculation, and noted the error in the post-hearing brief it submitted to the board. While the board recognized that the expert had made a similar error in his exclusion of Colonial Gas’s net book value from his ratio of sales to net book value — and made a correction for that error — the board made no findings with respect to the sales to EBITDA ratio, and indeed did not discuss the issue.
While the board is specifically exempted from the Massachusetts Administrative Procedure Act, G. L. c. 30A §§ 1, 11 (8), the board is nonetheless bound by “general principles affecting administrative decisions and judicial review of them.” Schlaiker v. Assessors of Great Barrington, 365 Mass. 243, 245 n.2 (1974), quoting Assessors of New Braintree v. Pioneer Valley Academy, Inc., 355 Mass. 610, 612 n.l (1969). Accordingly, we have held *734that “[t]he board’s decision must state adequate reasons in support of its decision so as to permit meaningful appellate review.” Blakeley v. Assessors of Boston, 391 Mass. 473, 476 (1984), citing New Boston Garden Corp. v. Assessors of Boston, 383 Mass. 456, 467 (1981). Because this issue was adequately raised before the board, and because we have no findings of the board to review, we remand this issue to the board for consideration.
iii. Use of a tax factor in income capitalization. The company next claims that the board erred in adopting an income capitalization approach that did not utilize a “tax factor.” We remand the matter to the board for further findings and rulings.
As noted, the assessors’ expert’s income capitalization method took an average annual EBITDA figure for the company’s utility property, and capitalized that figure using an EBITDA multiplier to arrive at a valuation of the property. In calculating the company’s EBITDA figure for each of the years that were included in the average, the expert deducted the property tax expense actually incurred by the company.
The company asserts that the proper method to account for property taxes is not to deduct the tax expense from EBITDA, but to include a tax factor in the capitalization rate —■ or, in the present case, in the EBITDA multiplier.26 The included tax factor would be the relevant fiscal year’s tax rate.27 “The purpose of a tax factor, in a formula for capitalizing earnings, is to reflect the tax [that] will be payable on the assessed valuation produced by the formula.” Assessors of Lynn v. Shop-Lease Co., 364 Mass. 569, 573 (1974), citing Assessors of Lynnfield v. New England Oyster House, Inc., 362 Mass. 696, 700 n.2 (1972). We have recognized that, in using a tax factor rather than the tax expense actually incurred, one avoids including the very tax assessment in dispute in the valuation of the property for the *735purpose of resolving that dispute.28 See Assessors of Lynnfield v. New England Oyster House, Inc., supra, quoting New Brunswick v. New Jersey Div. of Tax Appeals, 39 N.J. 537, 546 (1963) (“The expense of local taxation turns on the very point in dispute, the fair cash value of the property. Logically, therefore, income should be capitalized before taxes ‘with the capitalization rate increased to yield the return the investor expects plus the amount of local taxes payable’ ”).
We have discussed the use of a tax factor on a number of occasions. While we have never held that a tax factor is required in income capitalization analyses — and we do not so hold today — we have noted the board’s preference for the use of a tax factor in accounting for local real estate taxes, Assessors of Lynnfield v. New England Oyster House, Inc., supra at 700 n.2; we have discussed the logic underlying its use, id.; and we have addressed the appropriateness of the tax rate used in its application, Assessors of Lynn v. Shop-Lease Co., supra at 573. A tax factor is often used in income capitalization analyses before the board. See, e.g., Black Rock Golf Club, LLC vs. Assessors of Hingham, Mass. App. Tax Bd. Rep., Nos. F284357, F288545 (Mar. 1, 2010).
The company appropriately raised this issue, together with a proposed recalculation of the income capitalization valuation,29 in its reply brief before the board. The board did not discuss the issue in its decision. Given the board’s expressed preference for the use of a tax factor, the frequent use of a tax factor in income capitalization analyses, the consideration of its use by this court, and its potential importance in this case, we conclude that the board should have addressed this issue.
There may well be facts or methodological considerations in the present case that would justify the method used to account for property taxes in the income capitalization analysis. If so, *736we would accord our usual deference to the board’s findings. See General Elec. Co. v. Assessors of Lynn, 393 Mass. 591, 608 (1984). However, as we are unable to review meaningfully the board’s decision in the absence of findings or rulings on the matter, we remand this issue to the board for further consideration.
iv. Weight of the evidence and credibility of the witnesses. The company also asserts that the board made a number of errors in its weighing of the evidence introduced at trial. We conclude that these claims are without merit.
The company first claims that the board erred in finding that evidence of the assessed valuation of comparable utility property was not probative of the proper valuation of the company’s property. The evidence in question included the assessment of other gas utility property in the city at 1.07 times its net book value and evidence that the company’s utility property in other Massachusetts communities had only rarely been assessed at a value above its net book value. The board admitted this evidence at the hearing.
To support its claim, the company cites G. L. c. 5 8A, § 12B, which states that “[a]t any hearing relative to the assessed fair cash valuation or classification of property, evidence as to the fair cash valuation or classification of property at which assessors have assessed other property of a comparable nature or class shall be admissible.” As that statute speaks to the admissibility of evidence, rather than the board’s assessment of its weight, the statute is not a basis for relief here.
With respect to the evidence of assessments in other Massachusetts communities, the board described that evidence and noted in its decision that the testimony on that matter had been brief. Indeed, the testimony did not discuss any of the individual circumstances or other details of those assessments, or provide any other evidence in that regard. The board did not comment specifically on the evidence of Boston utility property valued at 1.07 times net book value, but indicated that the testimony was not helpful in valuation. We conclude that it was within the discretion of the board to determine the weight given to such evidence.
The company next argues that the board failed to address the impact on the income capitalization approach of a regulation mandating replacement of cast-iron mains. See 220 Code Mass. *737Regs. §§ 113.00 et seq. (1993). The company asserts that the costs to the company of such regulations should have been accounted for “through the adoption of a reserve or other means within the income approach.” In support of its contention, the company cites prior decisions of the board in which a reserve for replacements was adopted in an income approach. In those instances, the reserve was adopted in the form of an addition to expenses to account for the projected replacements. See Olympia & York State Street Co. v. Assessors of Boston, 23 Mass. App. Tax Bd. Rep. 96, 103 (1997), aff’d on other grounds, 428 Mass. 236, 237 (1998) (setting reserve for replacements of one per cent of effective gross income from subject property); Saunders v. Assessors of Boston, 15 Mass. App. Tax Bd. Rep. 1, 7 (1993) (adding sum to operating expenses per square foot per year). See also Appraisal of Real Estate, supra at 490 (discussing inclusion of replacement allowance in estimates of income).
In its findings of fact, the board recounted testimony regarding the regulatory requirement that cast-iron mains be replaced, and noted other expenses associated with the use of cast-iron mains. In its later assessment of excess operating expenses,30 the board found it appropriate to increase the company’s expenses from $2,100 to $3,600 per mile of pipe, because the assessors’ expert had inadvertently used the company’s system-wide expenses instead of those specific to Boston.31 In so doing, the board noted that the higher sum “accounted for the disproportionately high amount of cast iron pipe in Boston.” While the board did not break out the specific factors contributing to the higher expenses associated with such pipe in the city, or the value contribution of each,32 it is apparent that the board had considered the evidence related to the regulatory requirements, *738together with other costs inherent in a system that uses such pipe, and made an adjustment accordingly. In the circumstances, greater specificity was not required of the board. Cf. Jordan Marsh Co. v. Assessors of Malden, 359 Mass. 106, 110 (1971) (board is “not required to specify the exact manner in which [a valuation figure] was arrived at” where its decision “[made] reasonably clear what it was deciding as to each element of value”).
Finally, the company alleges that the board did not sufficiently take account of the fact that the expert had prepared a prior appraisal report that differed significantly in its valuation from the appraisal presented at the hearing. The specific redress requested by the company in this respect is somewhat unclear, but the company implies that the board should have found that the expert’s credibility had been undermined.
The board admitted the expert’s prior report as evidence, acknowledged the report in its decision, and ultimately decided that the appraisal presented by the expert at trial was credible. Where the board has reviewed such evidence in the record, we will not second-guess the board’s conclusions as to witness credibility. See Cummington School of the Arts, Inc. v. Assessors of Cummington, 373 Mass. 597, 605 (1977), citing Fisher School v. Assessors of Boston, 325 Mass. 529, 534 (1950) (“[t]he credibility of witnesses, the weight of the evidence, and inferences to be drawn from the evidence are matters for the board”).
b. Valuation of the real property. The company maintains that the board erred in ruling that there was insufficient evidence to determine the value of the real property. In this respect, the company cites several valuations proposed during the hearing on which the board could have drawn, and concludes that, given the board’s conclusions as to the personal property, consistency dictates that the real property “should be weighted at an appropriate ratio with [net book value].” We do not agree.
In an appeal before the board, the assessment is presumed valid until the taxpayer demonstrates its right to an abatement. Schlaiker v. Assessors of Great Barrington, 365 Mass. 243, 245 (1974). The board heard testimony and received evidence from several witnesses as to the proper method of estimating the value of the real property. Tierney posited, on the basis of *739detailed evidence in her report, that the property’s highest and best use33 was its current use as rate-regulated utility property. The board credited her testimony in that regard. Relying on her prior assertion that no buyer would pay more than net book value for rate-regulated utility property, she concluded that the value of the real property was its net book value. As noted previously, the board found unsubstantiated Tierney’s broad assertion that the value of rate-regulated utility property is its net book value, and accordingly concluded that her testimony did not provide credible evidence of the value of the property.
The board also heard testimony from two real estate appraisal experts — one offered by the company, the other by the assessors. The board found that the testimony of the company’s expert was of limited probative value because he appraised the land under the assumption that it was not rate-regulated utility property. He did so despite credited evidence from Tierney, and indeed despite his own determination, that the property’s highest and best use was its current use. The board similarly concluded that the assessors’ appraisal expert’s valuation “did not provide sufficient probative evidence” to establish the value of the real property. He valued the land on the assumption that it was vacant and available for development, an assumption that the board found at odds with the evidence that under no foreseeable circumstances would the property be used in any capacity other than its current use.
It was well within the discretion of the board to determine, on the basis of the evidence in the record, that the company had failed to demonstrate its right to an abatement. Having found no credible evidence on which to appraise the real property, the board also left undisturbed the parcel size used by the assessors, which had been a subject of disagreement between the parties’ experts. The board noted in this regard that neither of the parties had offered evidence from a registered land surveyor. This conclusion was likewise within the board’s discretion.
3. Conclusion. The board did not err in using a valuation method that equally weighted net book value and RCNLD. *740Further, the board did not err in taking account of evidence from after the assessment date, or in its weighing of evidence related to the assessors’ expert’s credibility, the impact of cast-iron replacement regulations, or the assessments of other utility properties in Massachusetts. With respect to the real property, we affirm the board’s decision to leave the assessed value undisturbed.
We remand the matter to the board for further consideration, consistent with this opinion, of (1) its decision not to use a tax factor to account for property taxes in the income capitalization analysis; (2) its exclusion of 2001 EBITDA from the average EBITDA generated by the company’s personal property; and (3) the assessors’ expert’s alleged failure to account for Colonial Gas’s EBITDA in Keyspan Corporation’s acquisition of Eastern Enterprises.
So ordered.
As will be discussed in further detail, the Department of Public Utilities (DPU) regulates the rates that gas utilities charge to their consumers. See G. L. c. 164, § 94.
Boston Gas Company, doing business as Keyspan Energy Delivery New England (company), also sought abatements with respect to fiscal years 2005-2009. By agreement between the parties and the Appellate Tax Board (board), the fiscal year 2004 appeals were tried as a “test year” for adjudicating the relevant issues. The adjudication will provide guidance for the disposition of the remaining appeals.
The latter three methods are common methods of appraisal used beyond the context of regulated utility property. See Correia v. New Bedford Redevelopment Auth., 375 Mass. 360, 362 (1978), quoting State v. Wilson, 6 Wash. App. 443, 447-448 (1972); see also Appraisal Institute, The Appraisal of Real Estate 141 (13th ed. 2008) (Appraisal of Real Estate).
The DPU has defined an acquisition premium in this context as “the difference between the purchase price paid by a utility to acquire plant that previously had been placed into service and the net depreciated cost of the acquired plant to the previous owner.” Guidelines & Standards for Acquisitions & Mergers, D.P.U. 93-167-A at 9 (1994).
The reasoning behind the regulatory change circumstance is that if a buyer were allowed to earn a return on the premium paid, it would be willing to pay such a premium.
The income capitalization approach measures the present value of the future benefits of the property. See Appraisal of Real Estate, supra at 142. The specific method of income capitalization used in this case by the expert for the board of assessors of Boston (assessors) in this case, called direct capitalization, uses one year of the property’s income and converts it into a valuation of the property using a market-derived capitalization rate or income multiplier. See id.
The sales comparison approach “produces a value indication by comparing the subject property with similar (i.e., comparable) properties.” Appraisal of Real Estate, supra at 141.
The reproduction cost new less depreciation (RCNLD) approach is a type of “cost approach” in which, as applied here, one estimates the current cost of constructing a reproduction of the subject matter and then “subtract[s] the amount of depreciation (i.e., deterioration and obsolescence) in the structures from all causes.” Appraisal of Real Estate, supra at 142.
The “useful life” is “[t]he period of time over which a structure or a *721component of a property may reasonably be expected to perform the function for which it was designed.” Appraisal of Real Estate, supra at 413. The depreciable life to which the board refers is the period of time over which the property is “depreciatfed] for rate purposes.”
The board’s finding that special circumstances argued against using net book value as the sole determinant of fair cash value applies to both the personal and real property appeals. Our review here of the board’s findings in that respect applies also to the real property.
The petition was authorized by G. L. c. 164, § 43. See Stow Mun. Elec. Dep’t v. Department of Pub. Utils., 426 Mass. 341, 343 (1997).
This measure is evidently very similar to net book value. The court did not discuss whether the measures are in fact identical.
At the time, and for the period between November, 1997, and April, 2007, the DPU was known as the Department of Telecommunications and Energy. See St. 1997, c. 164, § 186; St. 2007, c. 19, § 21.
In another DPU order, the department concluded that it would allow an acquiring company to seek recovery of its acquisition premium after a proposed merger. See Bay State Gas Co., D.T.E. 98-31 at 45 (1998). Recovery of the premium would be allowed if the company could show that the merger-related benefits were equal to or greater than the share of the premium proposed “to be included in base rates.” Id.
We also agree with the board that the DPU’s adoption of “performance-based rates” (PBR) could contribute to a buyer’s willingness to pay more than net book value for rate-regulated utility property. Under the PBR regime, a “cast off” rate of return is set in a given year, and the DPU sets a fixed annual upward inflation adjustment and a downward productivity adjustment to encourage more efficient operation. See Boston Gas Co. v. Department of Telecomm. & Energy, 436 Mass. 233, 235 (2002). A buyer who anticipates being able to perform more efficiently than is contemplated by the productivity adjustment could thus earn a higher return than otherwise would be available under existing rate regulation.
One of those sales was the company’s acquisition of Eastern Enterprises, the prior owner of the utility assets at issue in this case.
The assessors sought to undermine these assertions in part by evidence that few in management were in fact retained.
Rhode Island was characterized as being less liberal than Massachusetts in the allowance of recovery of acquisition premiums.
The company did not dispute specifically a third reason given by the board for concluding that a valuation method other than net book value was warranted, namely, that the useful life of gas utility pipeline exceeds its depreciable life, so that the property has residual value in excess of net book value. As the board’s decision is supported by the reasons discussed, we do not address this third justification.
A stated reason given by the board for not relying on the income capitalization approach more directly was that it “is not typically used to estimate the value of special purpose property.” The board concluded that while the approach was “generally reliable,” it was “better suited as support for the value derived under the [RCNLD] approach rather than as the primary valuation methodology."
In the direct capitalization method used by the assessors’ expert, one capitalizes the earnings estimate by either dividing it by an appropriate capitalization rate, or multiplying it by an income factor. Appraisal of Real Estate, supra at 499. The expert’s EBITDA multiplier approach is presented as a form of the latter.
The board also states that the 2001 EBITDA figure was $24,556,000, while the range of values of the other five years included in the average was $31,323,000 to $40,432,000. The board does not state that this is why the figure was excluded.
It is notable also that the average EBITDA value with 2003 excluded — approximately $35.78 million — is very similar to the approximate EBITDA *733in 2003 of $35.76 million. The values of the four pre-assessment years ranged from $31.3 million to $40.4 million.
Rates used in capitalization are the reciprocals of multipliers. Dividing annual income by a capitalization rate yields a valuation; multiplying annual income by an income multiplier achieves the same result. See Appraisal of Real Estate, supra at 516; id. at 499.
For example, a tax rate of $33.08 per $1,000 would first be expressed as a decimal (0.03308) and then added to the capitalization rate. The annual income would then be divided by this sum to determine the value of the property. See, e.g., Assessors of Brookline v. Buehler, 396 Mass. 520, 523-524 (1986). An equivalent adjustment can be made to an income multiplier.
This problem may become more acute when the 2004 “test year” is applied to more recent years. The average EBITDA calculated by the assessors’ expert considered the six years prior to the assessment date. If he were to look to the six years prior to fiscal year 2009, for example, nearly all of the tax expenses that would be deducted from EBITDA in those years would be a result of the assessments challenged in this case.
Should the board utilize a tax factor on remand, the question whether the company’s specific figures and calculations should be adopted is a matter for the board.
Part of the assessors’ expert’s RCNLD approach accounts for the fact that, if reconstructed new today, the system would be composed of plastic pipe rather than cast iron or steel pipe. The excess operating costs are those associated with the operation and maintenance of a system composed of these older materials.
The board also decided to average the final RCNLD figure with net book value, in part, to account for the “contemporaneous regulatory environment,” although the board’s primary focus in that respect appears to have been on changes to the carry-over rate base principle.
The board noted that the record did not reflect the precise excess operating costs incurred by the company in the city of Boston.
In an appraisal, the “highest and best use” of property is “the reasonably probable and legal use of vacant land or an improved property that is legally permissible, physically possible, appropriately supported, financially feasible, and that results in the highest value.” Appraisal of Real Estate, supra at 278.