These are consolidated appeals from an order of the public utilities commission issued in its docket DF 84-200 and reported in Re Public Service Co. of New Hampshire, 66 PUR4th 349 (N.H.P.U.C. 1985), authorizing Public Service Company of New Hampshire to issue and sell deferred interest bonds or tax exempt pollution control revenue bonds in amounts up to a total of $525 million. We affirm.
I. Facts and Procedural History
The object of this proposed financing is the provision of funds to allow the company to participate in the completion of construction of Unit I and “common facilities” at the Seabrook Nuclear Power Plant. Unit I is “the first unit of a [planned] two-unit nuclear facility” at Seabrook, New Hampshire, Re PSNH, supra at 398, and common facilities are “those facilities which are necessary to the operation of [each unit, such as] the portion of the plant devoted to the storage of nuclear waste.” Id. at 398 n.35. Public Service Company of New Hampshire is not the sole owner of the Seabrook project and thus does not exert complete and independent control over construction expenditure decisions; however, its 35.56942% share of the project is the largest of any of the Joint Owners. Id. at 360.
This is the third financing proposed in accordance with a three-step financing plan, devised in the spring of 1984, in response to the company’s financial problems. In Appeal of Seacoast Anti-Pollution League, 126 N.H. 789, 497 A.2d 847 (1985), we considered an appeal from the commission’s approval of the first step of that plan, the sale of $90 million of short-term securities. We considered the second step of the financing plan, the issuance of securities in the amount of *612$425 million, both in Appeal of Seacoast Anti-Pollution League, 125 N.H. 465, 482 A.2d 509 (1984), and in Appeal of Seaeoast Anti-Pollution League, 125 N.H. 708, 490 A.2d 1329 (1984). Those opinions provide further background on the issues before the court in the instant appeals.
The commission has reviewed the proposed financing under its docket DF 84-200, opened on August 2, 1984, to determine whether the company’s participation in the completion of the construction of Unit I would be in the public good, RSA 369:1, :4, a determination required by Appeal of Easton, 125 N.H. 205, 480 A.2d 88 (1984). Because of the precarious financial position of the company, the commission delayed the Easton inquiry until it reached consideration of this third step in the financing plan. We approved this delay in both 1984 Appeal of Seaeoast Anti-Pollution League cases.
In prehearing procedural rulings in docket DF 84-200, the commission delineated the scope of this Easton inquiry. Among the issues that the commission found to be subsumed within the determination of whether the completion of Unit I and the necessary financing would be in the public good are the following: the incremental cost to complete Unit I; the cost of alternative energy sources and a comparison of those costs to the incremental cost to complete Unit I; the reasonableness of the company’s rates that would result from completion of Unit I; the effect of those rates on demand for electricity; and the effects that would follow from a possible company bankruptcy.
Hearings in this docket began on December 3, 1984, and stretched over 38 hearing days, concluding on February 22, 1985. Transcripts of the testimony cover over 7,500 pages, and the record includes almost 180 exhibits. Although the company originally requested approval for third-stage financing of up to $730 million in securities to be issued to Newbrook Corporation, on the day before the hearings ended the company amended its financing request by eliminating Newbrook Corporation and requesting that the company itself be given authority to issue and sell up to $525 million in securities, the amount eventually approved by the commission.
On April 18, 1985, a majority of the commission issued a 211-page report, along with an order approving the requested financing subject to two conditions. First, before the securities could be issued and sold, the company would be required to persuade the commission either that all Unit I joint owners had received regulatory authorization to finance their respective ownership shares or that reasonable assurance existed that each participant could finance its share. Second, pending further order, the commission continued the lim*613itation, imposed upon authorization of the second stage of the financing plan, that the company could not contribute more to the cost of new construction at Seabrook than its proportional share of $5 million per week.
Commissioner Aeschliman filed an 82-page separate report. She also would have approved the financing request but only upon further conditions which in her view would protect ratepayers from full cost recovery through rates and from further risks if Unit I could not be completed.
When the 4]3Í>eliants’ timely motions for rehearing were denied, these appeals followed. While the appeals have been pending, we have issued three separate orders. First, on August 8, 1985, we heard oral argument on the appellants’ petition for a writ of prohibition to forbid the commission from granting the company’s motions to remove or amend the conditions imposed in the commission’s order of April 18, 1985, which was before this court on appeal. On August 13, 1985, we unanimously denied the petition for writ of prohibition but remanded the cases to the commission under RSA 541:14 for additional hearings, on proper notice, with respect to the company’s request for removal of the conditions that the commission had previously imposed. We specifically stayed the effect of any orders that might be issued by the commission upon remand, pending the completion of these appeals or further order of this court.
On September 13, 1985, the commission majority filed with the court a 51-page report resulting from the remand, proposing an order to remove both conditions. The company immediately moved this court to modify the stay, so that the company might contribute more to the cost of construction of Unit I than its share of spending at the $5-million-per-week level. We heard oral argument on this motion on September 17, 1985, and the next day issued an order granting the company’s motion, with the limitation that total expenditures in excess of the company’s share of $5 million per week could not exceed $32.9 million for the period between September 1, 1985, and December 31, 1985. The members of the court dissenting today also dissented from that order.
We issued our third and last prior order during the pendency of these appeals after oral argument on the merits on October 29, 1985. On the following day, we unanimously remanded the case to the commission for the issuance of a supplemental report, based on the present record, containing “specific findings, expressed in dollars and as percentages of the existing rates, of the reasonably probable range within which the actual customer rates will be set if Unit I is completed as authorized by the commission^]” In addition, we asked *614the commission if such findings would have any effect on the validity of its conclusions in its Report and Order of April 18, 1985. We based this remand on our view that the commission did not present this court with findings of fact sufficient for genuine appellate review on the issue of future rates, as required by Appeal of SAPL, 125 N.H. at 718, 484 A.2d at 1203. The commission majority responded on November 8, 1985, with a 45-page report and a supplemental order containing rate findings and indicating that those findings had no effect on the validity of the conclusions stated in its original order. Commissioner Aeschliman again filed a separate supplemental opinion. Thereafter, we allowed the parties to file supplemental briefs.
II. Scope of Subject Matter and Standards of Review
The scope of the issues before us in this appeal is determined by the law governing the commission's responsibility when considering a utility’s financing request and by the law governing the scope of this court’s review of an administrative agency’s findings and order. We will therefore preface our consideration of the merits of the appeal by examining each of these significant bodies of law.
The scope of the commission’s responsibility rests upon the mandate of RSA 369:1 and :4, which require the commission’s approval for the issuance of a utility’s securities and which condition the granting of that approval on a finding that the amount and objects of the proposed financing will be in the “public good,” id., as being “reasonable taking all interests into consideration.” Grafton Etc. Co. v. State, 77 N.H. 539, 542, 94 A. 193, 195 (1915). Thus in Appeal of Easton, 125 N.H. 205, 480 A.2d 88, we followed longstanding law in holding that a financing in the public good must be one “reasonably to be permitted under all the circumstances of the case.” Id. at 212, 480 A.2d at 91 (quoting Grafton, supra at 540, 94 A. at 194). Accordingly, we emphasized that the express statutory concern for the public good comprises more than the terms and conditions of the financing itself, and we held that the commission was obligated to determine whether the object of the financing was reasonably required for use in discharging a utility company’s obligation, which is to provide safe and reliable service, id. at 211, 480 A.2d at 90. Moreover, we specifically decided that the commission was obliged to determine whether the company’s plans to accomplish that object were economically justified when measured against any adequate alternatives; and whether the capitalization resulting from the utility company’s plans would be supportable. Id. at 212-13, 480 A.2d at 91.
The commission responded to the mandate of the Easton opinion *615by opening the present docket, and in two cases arising thereafter we further defined the scope of the required inquiry. In Appeal of SAPL, 125 N.H. 465, 482 A.2d 509, we referred to the issue of alternative sources of power by quoting with implicit approval from the commission’s order, which opened this docket to consider “‘the long term alternatives to completion of Seabrook Unit I in the context of the . . . incremental cost [of completion] and the assumptions found by the commission to be reasonable. .. .’” Id. at 473, 482 A.2d at 515. In the later, identically captioned, Appeal of SAPL, 125 N.H. 708, 490 A.2d 1329, we referred to the issue of capitalization by emphasizing that the Easton hearing must apply the standard expressed in Petition of the New Hampshire Gas & Electric Co., 88 N.H. 50, 184 A. 602 (1936), that “the primary public interest may be found to be affected injuriously” “if it appears, upon all the evidence, that the capitalization sought is so high that the utility, because of [its] inability to earn operating costs, depreciation and other charges, will not be able to give its consumers at reasonable rates the service to which they are entitled. . . .” Appeal of SAPL, 125 N.H. at 718, 490 A.2d at 1337 (quoting N.H. Gas & Elec., supra at 57, 184 A. at 607) .
Those cases thus held that the commission could not approve the present financing request except on the basis of findings that the company would have a need for its share of power to be generated by Unit I, that the company’s participation in the completion of Unit I would be preferable to any alternatives for obtaining that power, and that the company could support the resulting capitalization by the “reasonable . . . rates” mandated by RSA 378:27, :28. Although this appeal raises questions concerning both the commission’s methodology in considering these issues and the sufficiency of the evidentiary record to sustain the commission’s conclusions, we may say at this point that with the issuance of its supplemental report on November 8, 1985, the commission addressed each of the issues implicated by the statutory concern for the public good.
The scope of this court’s authority to review the commission’s methodology and its conclusions drawn from the evidence is limited by RSA 541:13, see RSA 365:21, and by the body of decisional law limiting the scope of judicial review of administrative decisions. RSA 541:13 governs appeals from agencies generally by placing the burden on appellants to demonstrate that an agency order is clearly unreasonable or unlawful. The statute provides that an agency’s findings of fact must be deemed prima facie lawful and reasonable and precludes appellate relief except for legal error or demonstration “by a clear preponderance of the evidence” that the agency action is unjust or unreasonable. RSA 541:13.
*616“We have frequently enunciated our recognition of the narrowness of our scope of review of commission orders. The ultimate issue before this court on appeal is whether the party seeking to set aside the decision of the commission has demonstrated by a clear preponderance of the evidence that such order is contrary to law, unjust, or unreasonable.” LUCC v. Public Serv. Co. of N.H., 119 N.H. 332, 340, 402 A.2d 626, 632 (1979) (citations omitted).
When, therefore, we are reviewing agency orders which seek to balance competing economic interests, or which anticipate such an administrative resolution, our “responsibility is not to supplant the Commission’s balance of . . . interests with one more nearly to [our] liking. . . .” Permian Basin Area Rate Cases, 390 U.S. 747, 792 (1968). The statutory presumption, and the corresponding obligation of judicial deference, are the more acute when we recognize that discretionary choices of policy necessarily affect such decisions, and that the legislature has entrusted such policy to “the informed judgment of the Commission, and not to the preferences of reviewing courts.” Id. at 767; see LUCC, supra at 340, 402 A.2d at 631; State v. New Hampshire Gas & Electric Co., 86 N.H. 16, 29, 163 A. 724, 731 (1932). Simply stated, as an appellate court we do not sit as a public utilities commission.
Although these principles limit our authority to disturb the commission’s resolution of factual and judgmental issues, we nonetheless have a broad responsibility to review the evidentiary record. While a reviewing court must “approach the task of examining some of the complex scientific issues presented in cases of this sort with some diffidence[,]” Lead Industries Ass’n v. Environmental Protection, 647 F.2d 1130, 1146 (D.C. Cir. 1980) (footnote omitted), we are obliged to study the record carefully in order “to assure [ourselves] that the [commission has given reasoned consideration to each of the pertinent factors” upon which the responsible derivation of policy and resolution of opposing interests must rest. Permian Basin Area Rate Cases, supra at 792. That is our responsibility no less than it is our obligation to refrain from arrogating to ourselves the role of a public utilities commission.
Subject to these limitations and responsibilities, we may turn now to the merits of the appellants’ issues. The appellants do not challenge the terms and conditions of the financing, but devote this appeal to the commission’s findings that it is in the public good for the company to participate in the completion of Unit I.
It would be appropriate in the abstract to deal first with the issues concerning the need for power, then with those dealing with the cost of Unit I as a means to supply that power, comparing it to other *617possible alternatives, and finally to review the commission’s findings about the rate effect incident to the completion of Unit I. The posture of the case, however, leads us to begin our review with the issues relating to the cost to complete Unit I. This is appropriate not only for the reason that the timing of this particular review allows resolution of the cost issues with a greater degree of certainty than the others, but also because of the interrelationship of all the issues, the conclusion on any one of which is necessarily dependent on the resolution of the others. Thus, the calculation of the Unit I completion cost not only provides the standard for measuring any feasible alternatives and determines maximum rate exposure when added to investment to date; it also affects the calculation of demand and need for power upon which any investment must rest, because demand is in part a function of cost as translated into rates. Hence if the commission’s conclusions about costs are found to be legally sufficient, they will furnish a comparatively firm point from which to review all the other economic issues in this appeal.
III. Completion Cost and Resulting Total Cost Attributable to Unit I
Taking August 1, 1984, as its point from which to measure, the commission found a $1 billion completion or “to go” cost for Unit I “reasonable for financing purposes.” Re PSNH, 66 PUR4th at 402. It also found that a $4.6-4.7 billion total cost was a reasonable planning estimate. Id. at 438. “[T]he $1 billion ‘to go’ cost. . . translates into a cost of approximately $870 per installed kilowatt of Seabrook capacity. . . . PSNH’s base case total cost figure [of] $4.7 billion . . . translates into a cost of approximately $4,087 per installed kilowatt of Seabrook capacity.” Id. at 394. These estimates were based on substantial evidence, including the schedule to completion proposed by the present construction manager, William Derrickson. See infra. The commission based its estimate on a finding that a December 1986 commercial operation date (COD) was achievable, but explicitly noted “a possibility of schedule slippage.” Re PSNH, 66 PUR4th at 402. The commission stated that the $1 billion figure provided for such slippage, given the company’s original estimate of the to go cost at only $882 million. Thus, the commission noted that “the difference in the $1 billion cost to go and the $832 [million] company estimate provides sufficient financial flexibility so that the company will be able to meet its construction costs even if it fails to meet the December 1986 [COD] by several months.” Id. at 402 (footnote omitted). We note in addition that the company has revised its estimate of total cash cost to go downward to $600 million and has targeted October 31,1986 as its COD. Commission Report and Fourteenth Supplemental Order No. 17,861 at 32-33 (September 13, *6181985). Given a $600 million cash cost to go, the company’s own completion cost is estimated to be $213,416,520. Commission Report and Fifteenth Supplemental Order No. 17,939 at 8 (November 8, 1985). Thus, the company may require even less financing than that originally approved by the commission.
Although the appellant Consumer Advocate challenges the commission’s finding as to COD, we do not find its determination clearly unreasonable. The commission bifurcated its analysis into two distinct time periods: the period preceding the fuel loading, and the period from fuel loading to COD. While citing the historical inaccuracy of the company’s construction time estimates, the commission noted that certain new elements boosted its confidence in the company’s current estimate of time to fuel loading; e.g., Mr. Derrickson’s prior experience at St. Lucie II, the excellence of his management staff, the nearness to completion of the plant itself, and the built-in flexibility that allows for recovery for time slippage. Id. at 399-401. The commission had evidence that Mr. Derrickson’s record in reaching relevant construction milestones had been exemplary, both before and since his employment at Seabrook. His success in meeting projected deadlines since coming to Seabrook indicated to the commission that its increased confidence in the company’s schedule was justified. In addition, Mr. Derrickson testified that the need for rework, which previously reached levels as high as 60%, had dropped enormously, in certain cases down as low as 10%. Thus, given management’s high degree of control over the pacing of construction and testing, the commission could reasonably prefer a Seabrook-specific estimate of time to fuel loading, as against generalized statistical evidence provided by other witnesses.
The commission, however, refused to accept the company’s fuel loading to COD estimate of 4 months. Nor did it accept appellant witness Paul Chernick’s 13.5-month estimate for planning purposes. The commission properly noted that several variables outside of the company’s control affect the fuel loading to COD time period; e.g., the NRC licensing process, which involves the approval of an evacuation plan. Given this uncertainty, the commission appropriately turned to generalized evidence from the nuclear industry. While noting that the Management Analysis Company employed by the company set a 7-to-10 month range, the commission considered that it had previously set a 6-to-11.5 month range and decided that, while optimistic, a 6-month fuel loading to COD interval was obtainable. Re PSNH, 66 PUR4th at 401-02. Although the commission’s conclusions and analysis may be optimistic, we do not find that they are unreasonable.
*619Turning now to the estimate of total cost, certain of the appellants in their joint reply brief maintain that the actual total cost of Unit I will be approximately $5.4 billion, using the company’s 1984 Annual Report as a basis for their calculation (Exhibit 173). However, the commission explicitly noted in its Report and Tenth Supplemental Order No. 17,601 at 18, that the calculation which resulted in this figure was incorrect, in that the use of the company’s share as a basis to calculate total cost fails to account for the fact that the company’s financing costs are abnormally high as compared to those of the other Seabrook Joint Owners.
We thus find that the commission had sustainable bases for its findings of completion and total costs. We turn now to review its conclusions about the relative desirability of possible sources of power alternative to Unit I power.
IV. Relative Economic Desirability and Technical Feasibility of Proposed Alternatives to Unit I
A. Methodology of Comparing Economic Desirability of Completing Unit I to Proposed Alternatives
The appellants’ first challenge to the commission’s determination of the economic desirability of completing Unit I, as compared to proposed alternatives, concerns the use of “incremental cost analysis.” An incremental cost analysis “ignores those costs which have already been spent on the project (sunk costs) and looks only at the costs which will be required to be spent from this day until completion.” Re PSNH, 66 PUR4th at 394. A total cost analysis evaluates both sunk costs and incremental costs of the project as compared to the alternatives.
An incremental cost analysis is the common standard used to assess the desirability of alternatives to partially completed utility plants. See Pierce, The Regulatory Treatment of Mistakes in Retrospect: Canceled Plants and Excess Capacity, 132 U. Pa. L. Rev. 497 (1984). The justification for this methodology is the recognition that sunk costs remain costs, which will be borne by either ratepayers or investors whether or not a plant is completed. Re Commonwealth Edison Co., 50 PUR4th 221, 258 (111. Commerce Comm’n 1982). “[S]unk costs are irrelevant to the cancellation decision. In deciding whether a partially completed plant should be finished or abandoned, the only costs that should be considered are the costs of completing the plant, the operating costs of the plant, and the opportunity costs associated with the inability to sell those relatively few components that have a net positive value upon removal from the plant.” Pierce, supra at 510.
*620The appellants argue, however, that an incremental cost analysis is inappropriate in New- Hampshire in light of RSA 378:30-a, the “anti-CWIP” statute, which prohibits utilities from charging customers for the costs of unfinished plants. See Appeal of Public Serv. Co. of N.H., 125 N.H. 46, 53-54, 480 A.2d 20, 25 (1984). Appellants note that “‘the primary concern of the commission in ascertaining the public interest for purposes of capitalization is the protection of the consuming public[,]’” Appeal of Easton, 125 N.H. at 210, 480 A.2d at 90 (quoting N.H. Gas & Elec., 88 N.H. at 57, 184 A. at 607), and therefore assert that the commission may not analyze comparative costs as if sunk costs were recoverable from ratepayers whether or not Unit I is completed. They argue that ratepayers will be charged either the total of prudently incurred costs if the plant is completed, or no costs if the plant is cancelled.
This argument, however, ignores the basic fact that sunk costs are economic costs, whether allocated to ratepayers or to investors. Whether Unit I is completed or cancelled, therefore, sunk costs cannot be eliminated. The commission thus declared that “in terms of the cost to society and the efficient allocation of economic resources, sunk costs exist equally whether the alternative of cancellation or completion is selected[,]” Re PSNH, 66 PUR4th at 396, and all three commissioners agreed that an incremental cost analysis is appropriate for the purpose of assessing alternatives to Unit I. We accept this reasoning as sound, and we certainly cannot find that the use of an incremental cost analysis was clearly unreasonable.
The Consumer Advocate raised a complementary analytical issue by contending that the value of incremental AFUDC (allowance for funds used during construction, see Appeal of Public Serv. Co. of N.H., supra at 50, 480 A.2d at 22) was incorrectly added to the cost of alternatives. He argues that AFUDC costs should be included only in the cost of completion because AFUDC costs would not be borne by the ratepayers, under the terms of the “anti-CWIP” statute (RSA 378:30-a), in the event Unit I were cancelled. AFUDC costs, however, reflect finance charges on the sunk investment, which will still exist whether or not the plant is completed. Whether or not these costs would be borne by the ratepayer in the event of cancellation, they are nevertheless economic costs that would be borne by someone. AFUDC remains a cost in an economic sense for both cancellation and completion, and as a future cost the commission properly included it in both the completion and the cancellation scenarios.
*621B. Commission’s Findings Concerning Unit I Relevant for Purposes of Comparison
The appellants next challenge the commission’s cost and other findings used to compare Unit I completion to the competing possibilities. In section III supra we have already examined challenges to the calculation of the Unit I completion cost. Here, we note findings as to other Unit I costs that are relevant to the comparison of Unit I to other sources of power: estimated cost of capital additions of $15 million in 1984 dollars, escalating at a nominal rate of 7.5% per year; a 35-year plant life; a 15% consumer discount rate, which is used to discount future dollars to present dollars to reflect the revenue requirements in terms of present value, Re PSNH, 66 PUR4th at 408-09; and a capacity factor of 60% for Unit I.
The appellant Conservation Law Foundation (CLF) has challenged the finding of capacity factor. Capacity factor may be defined as “the percentage of kilowatt-hours [kwh] that are actually generated at the plant as compared to the number of [kwh] which would be generated if the plant were generating at 100% of capacity for every hour of the year.” Re PSNH, 66 PUR4th at 403 n.44. Findings as to capacity factor are crucial because of their potential effect on the cost of electricity; i.e., the lower the capacity factor, the higher the unit cost of electricity, and vice versa.
The commission found a 60% capacity factor to be a reasonable assumption for use in comparative analyses. CLF maintains that this was error because the evidence in the record supports a capacity factor of 55% more readily than a 60% factor.
We note in passing that a 60% capacity factor was accepted by all three commissioners, and was also accepted as reasonable by the commission in a prior docket. More importantly, however, the commission had substantial evidence before it in the present docket from which it could determine a 60% capacity factor to be a reasonable assumption. While noting that capacity factors can be random to a degree, the commission stated that the National Economic Research Associates’ average projection of capacity factors was 65.8% and that witness Mr. Chernick’s regression analysis resulted in a projected capacity factor of 50-60%. Re PSNH, 66 PUR4th at 405. Given these projections, and Unit I’s freedom from the quality assurance troubles that have plagued other reactors, the commission concluded that a mature Unit I capacity factor would fall within a range of 52.5% (witness Dr. Richard Rosen’s projection) to 72% (the company estimate), and that a general 60% capacity factor was reasonable. After a review of the commission’s reasoning from the evi*622dence, we conclude that the commission’s finding is not unreasonable.
C. Comparison of Unit I Completion with Alternatives
Once the commission defined the assumptions associated with completion of Unit I, it compared the cost of completion to the costs of various cancellation alternatives. The commission defined the cancellation alternatives as conventional thermal generation, cogeneration, and aggressive conservation, and it also considered the significance of hydropower to be imported from Quebec.
In its evaluation of conventional thermal generation as an alternative, the commission considered a study submitted by the company and developed by its system planning engineer, Joseph Staszowski. The study presented two generation expansion plans in the case of cancellation, and analyzed 65 sets of scenarios comparing completion of Unit I to cancellation alternatives. In addition, the commission considered testimony of Campaign for Ratepayers’ Rights’ (CRR) witness Dr. Rosen, senior research scientist at Energy Systems Research Group, Inc., who presented a study that analyzed the economic costs and benefits of completing Unit I from the perspective of New Hampshire ratepayers. The study calculated revenue requirements for both completion and cancellation, deriving its assumptions from a statistical analysis of other nuclear reactors and applying them to the particular characteristics of Unit I. Dr. Rosen recommended cancellation.
The commission chose to use Mr. Staszowski’s plan for the purposes of its analysis, on the basis of the uncertainty surrounding the costs of Dr. Rosen’s alternatives and the lead time necessary to construct the large thermal units that his alternatives involved. Re PSNH, 66 PUR4th at 410. Under all of Mr. Staszowski’s scenarios except one, the cost of completion of Unit I was less than the cost of cancellation alternatives. In the one scenario where the cancellation case cost less, all the pessimistic assumptions regarding Unit I were combined. The commission stated that it was not reasonable to assume that all pessimistic assumptions, including the 100 percent loss of UNITIL load, would prove true. Id. at 411. The commission, therefore, found that completion is the preferred alternative to the cancellation scenarios presented by Mr. Staszowski and Dr. Rosen. The record presents sufficient evidence to sustain this finding, and we cannot say that it was clearly unreasonable.
The commission dealt with the possibility of relying on Canadian energy in finding that Hydro-Quebec Phase I, from which “PSNH [could] receive 7.6 per cent of 690 megawatts or 52 meg*623awatts of hydroelectric power . . . [,]” had not been shown to be a reliable source of power in terms of planning and peak demand. Re PSNH, 66 PUR4th at 390. In Phase II, which would expand hydro capacity from 690 megawatts to 2000 megawatts, the company could obtain 152 megawatts of hydropower. The commission nonetheless concluded that Phase II is not a “dependable source of capacity up to 152 megawatts by the 1990 time frame[,]” due, inter alia, to risks of construction and regulatory delays outside the company’s control. Id. The commission thus accepted Hydro-Quebec power not as a substitute for Unit I, but as “a supplemental source of power when available.” Id. We cannot say that this is an unreasonable conclusion.
The commission analyzed the cogeneration alternative by reviewing the testimony and exhibits of John Hilberg, President of Calcogen, a developer of cogeneration facilities. Mr. Hilberg proposed that an aggressive cogeneration program would cost less than completion of Unit I. The commission rejected this conclusion, however, largely because Mr. Hilberg’s analysis was based upon a total cost standard, rather than an incremental cost standard. Re PSNH, 66 PUR4th at 412. There is no support in the record for the conclusion that under an incremental cost analysis, the cogeneration alternative would cost less than completion of Unit I.
We should note here that CLF claims that the commission failed to apply the same standards evenhandedly to the alternatives and to Unit I. CLF cites commission language to the effect that “[f]orecasts of small power and cogeneration are undependable in the following respects:
1. Expenses may escalate beyond rate support for the project.
2. Operating characteristics may not be as favorable as the design of the small power project may predict.
3. Operating and maintenance costs may exceed estimates.
4. Design lives may not endure as planned.”
Re PSNH, 66 PUR4th at 389-90.
This, asserts CLF, sounds like a description of Unit I itself, and thus indicates a lack of regulatory evenhandedness on the part of the commission in that “[i]f the [c]ommission finds Seabrook a ‘reliable’ method of meeting demand, yet uses these criteria to minimize the possible contribution of small power producers and co-generators, it is clearly not engaging in even handed decision-making.” If the commission had merely listed these reliability problems with no elaboration, we might be more inclined to agree with CLF. How*624ever, the commission went on to note that small power producers can stop producing power when it is no longer economically advantageous for them to do so, in support of its conclusion that “the amount and reliability of future capacity from these sources are not adequate to compensate for loss of Seabrook I.” Re PSNH, 66 PUR4th at 390. Finally, the construction of Unit I is so near to completion that the remaining amounts of construction expense are far more certain than the costs involved in planning and building one or more small power plants. We do not find the commission’s conclusion unreasonable.
The commission next discussed the aggressive conservation alternative. The Consumer Advocate argued that this alternative was economically preferable to completion of Unit I. Amory Lovins, Director of Research at the Rocky Mountain Institute in Old Snow-mass, Colorado, testified that aggressive conservation resulting from technological improvements is the least cost alternative to meeting the company’s power needs. The commission concluded, however, that this proposal is not preferable to completion because (1) Mr. Lovins’ conclusion was based upon new technologies not yet proven in the marketplace; (2) capacity from conservation will not be developed sufficiently to meet the company’s needs in the early 1990’s; and (3) the projected savings from the conservation alternative were only theoretical, but not probable. Re PSNH, 66 PUR4th at 413. Commissioner Aeschliman, in her separate opinion, agreed that the conservation alternative does not provide a sufficient basis to reduce demand below the company’s estimate. Id. at 467. In support of these conclusions, there was evidence that much of the technology described by Mr. Lovins has yet to find general use in this country. We conclude that the evidence reasonably supports the commission’s conclusion that the potential of aggressive conservation is not a preferable alternative to completion of Unit I.
D. Comparison of Unit I Completion to Bankruptcy of the Company
From the time that the commission opened docket DF 84-200 to conduct the Easton inquiry it planned to assess the effects of the possible bankruptcy of the company in the event that it did not approve the request for financing to complete Unit I. In a strict sense, the commission would have been required to reach the bankruptcy issue only in two instances. First, once the need for power was established, if the commission had found that an alternative source would be more cost effective than completion of Unit I and that the consequent cancellation of Unit I would force the company into bankruptcy proceedings, the commission would have been *625required to determine whether the desirability of the cost saving outweighed the negative effects of bankruptcy. Second, if the commission had found that upon completion of Unit I the capitalization of the company could be supported only by full dollar recovery of Unit I investment, with bankruptcy the only alternative, the commission would have been required to consider the legal and factual questions of whether customer rates necessary to support full cost recovery could be “reasonable” as the only alternatives to the possible effects of corporate bankruptcy. (See section VI infra.) Since the commission did not find any alternatives to be less costly and did not find full dollar recovery to be necessary to support the proposed capitalization, and since bankruptcy is not an alternative that in itself generates power, the commission was not required to evaluate bankruptcy as an alternative to completion of Unit I. The so-called bankruptcy alternative upon which our dissenting brothers place such emphasis is therefore not even a relevant issue in this case. Because of the dissenters’ emphasis, however, and because of widespread concern about the significance of a possible corporate bankruptcy, we will address the appellants’ challenges to the commission’s conclusions. We find nothing that would warrant reversal even if an assessment of the effects of bankruptcy were required.
All three commissioners found that a denial of the proposed financing, and cancellation of Unit I without recovery of sunk costs, would compel the company to seek reorganization under Chapter 11 of the Federal Bankruptcy Code. Re PSNH, 66 PUR4th at 424--25, 442. Substantial evidence supports this conclusion.
Donald Trawicki, a partner of Touche Ross & Co., testified that in the event this financing request was denied, the company probably would be unable to secure the financing necessary for its continued participation in Unit I and would be likely to seek protection from its creditors under the Bankruptcy Code or be forced into reorganization proceedings. The commission noted that the “anti-CWIP” statute (RSA 378:3Q-a) would prevent recovery of the sunk investment if the plant was not completed. See Appeal of Public Serv. Co. of N.H., 125 N.H. 46, 480 A.2d 20 (1984). Total inability to recover investment funded by debt obligations would force the company into default on those obligations and into a bankruptcy reorganization. Re PSNH, 66 PUR4th at 425.
The commission accordingly undertook an investigation of the probable effects of bankruptcy, requesting the assistance of the attorney general. The attorney general retained the law firm of Devine, Millimet, Stahl & Branch for advice on this issue, and submitted a report prepared by the firm. Mark Vaughn, Esq., one of the authors of the report, testified at the Easton hearing, and the *626commission heard testimony from various other witnesses on the bankruptcy issue.
The commission concluded that the company’s bankruptcy would not serve the public interest. The commission emphasized the independence of this conclusion from its findings that the proposed financing would be in the public good as the source of funds for the most cost-effective way to meet the projected power need.
Although the commission found that consequences of bankruptcy would in many respects be uncertain, see Re PSNH, 66 PUR4th at 426, it concluded that reorganization would probably frustrate the company’s ability to meet future power needs, and would be more costly to ratepayers than completion of Unit I. Id. at 430, 434. The appellants contend that the commission’s conclusion that reorganization could frustrate the company’s ability to meet future power needs is not supported by a clear preponderance of the evidence. We cannot accept this contention. To begin with, the appellants mistakenly seek to place upon the commission the burden that the appellants themselves must sustain in order to overturn the commission’s findings; i.e., demonstration of error by a clear preponderance of the evidence. See RSA 541:13. It is a burden that the appellants cannot carry in the face of testimony from a number of witnesses that the availability and cost of capital would be uncertain following bankruptcy. The commission thus had an evidentiary basis to find that the company’s uncertain ability to raise capital would hinder the financing of those generation expansion plans that were put forward as alternatives to the completion of Unit I.
The appellants further assert that the commission unreasonably found that “bankruptcy reorganization will be more costly to ratepayers regarding reliable electric services at reasonable rates over the long term than the proposed financing to put Seabrook on line.” Re PSNH, 66 PUR4th at 437. The commission found that “[t]he estimated legal and accounting cost[s] of a bankruptcy are a minimum of [twenty million dollars].” Id. at 428. Far more significantly, the commission found, on the basis of testimony from its own financial consultant, that the rate effects of bankruptcy would be volatile and that a bankruptcy court could subordinate ratepayers’ interests to the interests of creditors. The commission consequently could not find that the “risks and uncertainties of a bankruptcy of PSNH would be resolved in a manner that best balances ratepayer and investor interests.” Id. at 427 n.66. Moreover, the commission recognized that the legal prohibition against recovery of investment in a cancelled plant, and the difficulty of regaining credibility in the investment community after reorganization, would raise the cost of capital, thus requiring ratepayers to pay a penalty for bankruptcy.
*627CLF raises the further argument that the commission did not specifically contrast the economic risks of Unit I completion with the risks inherent in a corporate reorganization. We find no merit in this argument. Although the commission did not literally organize its comparison in parallel columns, it specifically described the risks associated with the bankruptcy reorganization, as we have already described, just as it specifically confronted the risks inherent in the Unit I completion projections. It thus addressed the issues necessary to conclude that the risks of reorganization would substantially outweigh the risks of completion.
Not only does substantial evidence support the commission’s conclusion that a bankruptcy reorganization would not be in the public interest, but it is noteworthy that no witness explicitly recommended reorganization in bankruptcy as a preferable alternative to the completion of Unit I. Witnesses went no further than to recommend more study of the advantages and disadvantages of reorganization. But the commission responded that “[f]urther study of this issue would be fruitless . .. ,” Re PSNH, 66 PUR4th at 436, bearing in mind that no major electrical generating company has ever been reorganized under Chapter 11, and that no recent precedents of utility bankruptcy exist to guide the commission. Whether or not further study would have been entirely fruitless, it was not required in the light of the commission’s other conclusions.
It is likewise noteworthy that Commissioner Aeschliman, in her separate opinion, agreed that “bankruptcy entails great uncertainty and risk for ratepayers as well as enormous administrative expense .... [A] PSNH bankruptcy would ... be contrary to the public good . . . .” Re PSNH, 66 PUR4th at 470. All commissioners thus agreed in the specific context of this case with the more general dictum of this court, that “[a] bankrupt utility is not in the public interest.” Appeal of Legislative Utility Consumers’ Council, 120 N.H. 173,174, 412 A.2d 738, 739 (1980). We cannot hold the commission’s conclusions on the issue of bankruptcy to be unreasonable.
V. Need for Power
Just as the completion cost figure governs the relative desirability of proposed alternatives, the total projected cost of Unit I governs the disposition of the two remaining issues in this appeal: the calculation of the demand for power at rates necessary to support the company’s capitalization that will result if Unit I is completed, and the compatibility of that capitalization with the obligation to set customers’ rates at a just and reasonable level. This section examines the commission’s determination that there would be a need for Unit *628I as a source of power. The commission affirmed this determination in its Report and Fifteenth Supplemental Order No. 17,939 (November 8, 1985), when it certified that its earlier conclusions remained valid in light of its specific calculations of the probable range of ultimate customer rates. That range was limited at its high end by a rate calculated on the assumption that there would be total dollar recognition of all of the company’s share of the total Unit I investment, projected at a cost to all joint owners of $4.6 billion.
On several grounds, the appellants CLF, CRR, and Seacoast AntiPollution League (SAPL) challenge the commission’s determination of the need for the power to be generated by Unit I. We should note that the company contests the appellants’ right to raise this issue at all on appeal, asserting that it was not properly raised below. We conclude, however, that the appellants’ motions for rehearing implicitly raise the need for power issue, and we will therefore consider it. See Consumer Advocate Motion for Rehearing on Report & Order No. 17,558 at 2, ¶ 4 (joined by SAPL, see id. at 5).
As a threshold matter, we observe that the commission was unanimous in its conclusion that there will be a need for additional electrical power in the foreseeable future. Commissioner Aeschliman, in her separate opinion, states that while there would be no actual physical shortages of power through the end of the century should Unit I not be completed, Re PSNH, 66 PUR4th at 452, “the company would have a need for power in addition to present generating sources without Seabrook[.]” Id. at 446. This finding of a need for power, of course, reflects the company’s responsibility to provide generating capacity before the moment of physical shortage.
On the record before us, we believe that the commission could reasonably rely on the company’s 1984 load forecast as a starting point for its analysis of and findings on the future demand and need for power. The commission emphasized the reasonableness of the company’s methodology, Re PSNH, 66 PUR4th at 379, while aptly stating that such a forecast, by its very nature, can only be a guide to, and not a perfectly accurate predictor of, future demand. Id. In other words, fallibility is inherent in the forecasting process because of the huge number of variables involved and the complexity of interactions among them. Having said that, however, we must add that the commission also found the 1984 forecast to be consistent with, if not exactly identical to, the commission’s own findings in a prior docket, as well as a conservative prediction of what had actually transpired in 1984. Id. at 380.
Looking to the future, in which price or rate changes will affect demand, the commission observed that the company took price elasticities properly into account in its load forecasts. See Re PSNH, 66 *629PUR4th at 381-87; Exhibits 31 and 42, and see infra. According to the company’s Exhibit 42, the 1984 load forecast takes price-demand effect into account in two ways: “[fjirst, the end use portion [of the 1984 forecast, which predicts future demand by projecting actual past electricity use into the future, see Re PSNH, 66 PUR4th at 381] captures customer response by utilizing time trends of data, explicit recognition of appliance efficiency improvements and explicit recognition of certain energy management programs, such as the ‘One Stop’ service program. Second, additional customer response is captured through econometric modeling [a methodology which predicts demand ‘by examining how past demand was influenced by historic economic and demographic conditions^]’ id.] using price elasticity procedures. . . . Short and long run elasticities and the electricity prices are specified for each end use of electricity to reflect the time lag response by customers to price changes through the use of an elasticity aging function.” Exhibit 42.
Price elasticity of demand may be defined as “‘the per cent change in the quantity [of electricity] consumed divided by the per cent change in the real price of the electricity.’ ” Re PSNH, 66 PUR4th at 382 (quoting Exhibit 42). The company’s load forecasts assumed short- and long-run price elasticities of approximately -.2 and -.5, respectively. Thus, according to the company, “a 10% real price increase will reduce loads [in the long run] by 5% from the levels that would otherwise occur.” Exhibit 42. The commission exercised expert judgment in accepting these numbers after analysis, and we see no opportunity to claim that they lack a supporting evidentiary basis.
We may consider further details of the need calculation in light of the appellants’ objections, beginning with the assertion of CLF that the commission failed to make clear whether Unit I is needed to meet future capacity shortages or merely to avoid economic penalties. We find no merit in this argument. The commission states that “to meet its capacity needs for present and future customers [the company] must have additional generation capacity and Seabrook is the only reliable project.” Re PSNH, 66 PUR4th at 393-94. Without Unit I, the company’s load growth, assuming a 1.5% annual growth rate, shows that prospective customer demand will exceed the company’s capacity to serve that demand in the power year 1988-89. Id. at 388. Thus, the commission clearly found that Unit I is needed to provide the capacity to meet demand.
CLF further claims that the finding of need is specious, because the evidence indicates that even without Unit I, there will be no physical shortages of electricity. Depending on which disputed excess capacity or capacity deficiency calculation one accepts, this *630argument is sound to some degree for some period of time because sufficient alternative sources of power may be available so that blackouts will not occur. It is, however, not to the point. The commission reasonably found that there will be a need for power over time and that Unit I is the most reliable means to meet that projected need.
We should note here that one contested element of the commission’s methodology for calculating need was the reserve factor. Need is defined as “capability responsibility,” which is equal to load or usage plus a “reserve” factor. The appellants have challenged the commission’s finding that a 25% reserve factor was not unreasonable. Re PSNH, 66 PUR4th at 389. We find no error in the commission’s conclusion, not only because a factor of that magnitude is equal to the company’s largest generating unit, see Re PSNH, 41 N.H.P.S.C. 16, 29, affd on other grounds, Public Service Co. v. State, 102 N.H. 150, 153 A.2d 801 (1959), and thus would offset an outage of that unit, but also because the witness John Eichorn testified in post-April hearings that“[i]n [NEPOOL] a participant must maintain in accordance with the agreement approximately 23% reserve ... [.]” (For a description of NEPOOL, see infra.) Mr. Eichorn’s statement bolsters the commission’s conclusion that a 25% reserve was reasonable. We conclude that, even if the commission’s original justification for a 25% figure was insufficient, a two percent difference between what, according to Mr. Eichorn, is required of NEPOOL members, and what the commission found to be not unreasonable, will not cause us to reject the commission’s finding.
Next, we consider the argument of CLF that the commission applied a “build to demand” standard in an alleged misinterpretation of RSA chapter 162-F and RSA 374:1. In its reply brief, CLF modified that position by claiming that, while the commission did not explicitly adopt such a standard, it implicitly did so by considering only new supply sources. We believe that the CLF assertion is simply inaccurate. The commission expressly utilized a “plan for ... demand” standard, Re PSNH, 66 PUR4th at 377, and sufficiently considered options such as conservation and other new supply sources, that arguably could meet the company’s obligation to plan for demand without building Unit I or any other generating plant.
One of the more arcane subjects of contention in this case is the issue of the effect of the loss of the UNITIL load on the demand and calculation of need for power. UNITIL is a holding company parent of Concord Electric Company and Exeter and Hampton Electric Company, each of which has a contract to purchase wholesale power from the company. UNITIL has chosen to terminate its contracts *631with the eompany, effective September 30, 1986. CLF claims that the commission failed to consider the loss of the UNITIL load, arguing: “[t]he majority does not deal with the loss of UNITIL load on the claimed need for power. Against all evidence, and logic, it continues to find a need for Seabrook to meet projected demand— including this load,” (Emphasis in original.)
We observe, however, that the commission did consider the potential loss of the UNITIL load in several ways. First, it noted that the entire load may not ultimately be lost. Even though the contracts have been cancelled prospectively, the commission found that UNI-TIL still could elect to purchase some or all of its power from the company. Nevertheless, even if UNITIL ultimately purchased all of its power elsewhere, the commission found that a demand for the power would still exist, although the power itself might, as a result, be sold at a lower price. Re PSNH, 66 PUR4th at 412.
Moreover, the commission accepted several of the company’s computer scenarios that took the loss of the UNITIL load into account. See Exhibits 124D and 124F. It stated that “[t]he PSNH load model substantiates that loads in excess of those based on the 1984 load forecast are justified using prices of the ‘rate shock’ scenarios without the UNITIL load.” Re PSNH, 66 PUR4th at 387. We cannot say that the commission’s exercise of its discretion and expertise here was error or otherwise inappropriate. Its analysis, while less clear than it could have been, was thorough and we will not upset its conclusions.
As to the related issue of whether the load forecast assumes unrealistically low projected rates for electricity, we conclude that the commission could properly find as it did on this issue. The commission did in fact recognize the price-demand interaction, which appears to be the basic issue CLF raises when it claims that the forecast assumes prices that are too low. The commission stated, “[n]either the 1984 nor the 1985 load forecast assumed sales explicitly based on price behavior under a so-called rate shock scenario. Rather, both load forecasts explicitly modeled sales based on price assumptions in relation to a so-called phase-in scenario. However, comparative analysis reveals that use of the 1984 forecast does not result in overstating sales in Mrs. Hadley’s rate shock scenarios .... Instead, the analysis indicates that the 1984 load forecast consistently provides sales lower than sales which are produced when the prices of the Hadley rate shock scenarios are assumed in the load forecast model.” Re PSNH, 66 PUR4th at 384; see Exhibit 143.
Further, Mr. Staszowski’s scenarios 4, 7, and 8 assume low peak demand growth, energy use, arid sales growth. See Exhibit 136, Attachment A. As we have noted before, the commission observed that *632“[i]n virtually all [of Mr. Staszowski’s] alternative cases, there was a benefit to completing Seabrook Unit No. I.” Re PSNH, 66 PUR4th at 410-11 (footnote omitted). Accordingly, we cannot say that the commission’s conclusions as to the relationship between price levels and demand were unreasonable.
CLF raises a different objection when it asserts that the commission erred in considering NEPOOL’s need for Unit I, because the commission “is under a duty to determine the needs of New Hampshire, and not of NEPOOL.” CLF claims in the alternative that “if the [commission was going to look at New England needs, it should have considered not only the opinions of NEPOOL, .. . but [those of] other New England regulatory bodies, who are charged in the various states served by NEPOOL utilities with protecting the public interests.”
NEPOOL has been described as “a regional power-pooling system” with a membership of approximately sixty New England utilities which collectively contain roughly ninety-eight percent of New England’s generation capacity. See Appeal of New England Power Co., 120 N.H. 866, 870, 424 A.2d 807, 810 (1980), rev’d on other grounds sub nom., New England Power Co. v. New Hampshire, 455 U.S. 331 (1982). NEPOOL’s objectives “are to assure the reliability of the region’s bulk power supply and to attain ‘maximum prácticable economy’ through, inter alia, ‘joint planning, central dispatching . . . and coordinated construction, operation and maintenance of electric generation and transmission facilities owned or controlled by the Participants ....’” New England Power Co. v. New Hampshire, 455 U.S. at 334 (quoting NEPOOL Agreement § 4.1, App. 31 A).
We observe that this issue was not properly raised below in any motion for rehearing. RSA 541:4 provides that “[n]o appeal from any order or decision of the commission shall be taken unless the appellant shall have made application for rehearing as herein provided, and when such application shall have been made, no ground not set forth therein shall be urged, relied on, or given any consideration by the court, unless the court for good cause shown shall allow the appellant to specify additional grounds.” As we stated in Appeal of White Mts. Educ. Ass’n, 125 N.H. 771, 774, 486 A.2d 283, 286 (1984), RSA 541:4 “restricts the scope of the appeal [from an administrative agency] to the grounds, and hence the issues, raised in the motion for rehearing, unless this court expands the scope for good cause shown. The reason for these requirements is obvious: administrative agencies should have a chance to correct their own alleged mistakes before time is spent appealing from them.” (Citation omitted.)
*633In any event, we note that the commission’s discussion of NEPOOL’s need for Unit I was not a basis for its decision, but rather a minor point raised in its discussion of New Hampshire’s need for Unit I power. Moreover, looking to the substantive merits of the argument, it is doubtful that one can calculate New Hampshire’s need for Unit I without reference to NEPOOL; in other words, no bright line may be easily drawn between State and regional needs. NEPOOL’s projected capacity shortages have an important narrowing effect on the company’s choice of alternatives to Unit I power. Thus, it was proper for the commission to look at NEPOOL’s need for Unit I in the context of analyzing New Hampshire’s future power needs.
Since we conclude that the commission’s findings of a need for power are sufficient to withstand these challenges, we now take up the final issue, whether the capitalization that would result from this financing would be consistent with the statutory requirement to set customer rates at a reasonable level.
VI. Reasonableness of Resulting Rates
Statutory law limits customer rates to a level that is “reasonable,” RSA 378:27 and :28, or “just and reasonable,” RSA 378:7. As previously explained, Easton and its progeny mandate that the commission’s consideration of the company’s financing plans must cover the possible results of the ratemaking process that will follow approval of the financing request. This anticipation of rate effect occasioned disagreement between the majority of the commission and Commissioner Aeschliman, and it is the focus of differences between the majority and dissenting members of this court. Our inquiry into the sufficiency of the commission’s treatment of the rate issue requires articulation of both the standard of reasonable rates and the commission’s responsibility in the light of that standard, followed by an examination of the scope of, and justification for, the commission’s findings.
A. The Meaning of “Reasonable Rate”
The term “reasonable rate” must be understood as referring to the result of the ratemaking process. That process appropriately balances the competing interests of ratepayers who desire the lowest possible rates and investors who desire rates that are higher. The rate-making process fixes rates that when charged to customers will satisfy a utility’s revenue requirement. Reduced to its essentials, this revenue requirement may be expressed as a formula: R = 0 + (B x r), where R is the utility’s allowed revenue requirement; 0 is its allowed operating expense; B is its rate base, defined as cost less *634depreciation of the utility’s property that is used and useful in the public service, see RSA 378:27; and r is the rate of return allowed on the rate base. See, e.g., Appeal of Public Serv. Co. of N.H., 125 N.H. at 49, 480 A.2d at 22; Pierce, supra at 511; 1 A. Priest, Principles of Public Utility Regulation 44-47 (1969). One recent commentator has put flesh on the formula in these words:
“This revenue requirement permits the utility to recover from its customers operating expenses (like labor, fuel, and maintenance cost) that it has prudently incurred in providing service that directly benefits the utility’s customers. In addition, the revenue requirement affords the utility the opportunity to make a profit on its investment, in an amount equal to its rate base multiplied by a specified rate of return. The rate base is the amount of money that the utility has invested in capital assets (like generating plant and transmission lines) that it uses to provide service to its customers. Each year the utility depreciates its capital assets to reflect their deterioration in value over time. The amount depreciated is passed through to the utility’s customers as a recoverable operating expense, which permits the utility to recover the amounts that it has actually invested in service-producing assets. The utility is also given the right to make a profit on its investments by including in the revenue requirement the product of the value of the utility’s capital assets remaining after each year’s depreciation and the applicable rate of return. A utility’s rate of return is a composite figure based on the value of outstanding bonds, the rate of return due to preferred stockholders, and the rate of return presently being paid on common equity.”
Glicksman, Allocating the Cost of Construction Excess Capacity: “Who Will Have To Pay For It All?’, 33 Kan. L. Rev. 429, 432 (1985) (footnote omitted). Leaving aside the rules governing permissible variations among rates charged, see RSA 378:10 and :11, “the commission [thus] controls three variables in regulating rates to provide revenue to [a utility]: operating expenses, rate base [cost of used and useful property] and rate or percentage of return allowed on the rate base.” Appeal of Public Serv. Co. of N.H., 125 N.H. at 49, 480 A.2d at 22.
A full understanding of the reasonable rate concept necessitates consideration of the commission’s discretion in setting each of these variables. This discussion, however, will not dwell on the process of determining allowable operating expenses, since the parties’ dis*635agreements regarding the meaning of “reasonable rates” do not turn on anticipated difficulties in deciding the legitimacy of the operating expenses incurred. We will concentrate instead on the process by which the commission determines the rate of return and the rate base to which the rate of return must be applied.
As we have seen, the rate of return is a percentage applied to the rate base expressed as a dollar amount in order to produce “interest on long-term debt, dividends on preferred stock, and earnings on common stock (including surplus or retained earnings).” C. Phillips, Jr., The Regulation of Public Utilities 332 (1985). The commission is bound to set a rate of return that falls within a zone of reasonableness, neither so low as to result in a confiscation of company property, nor so high as to result in extortionate charges to customers. LUCC, 119 N.H. at 341-42, 402 A.2d at 632. A rate falling within that zone should, at a minimum, be sufficient to yield the cost of the debt and equity capital necessary to provide the assets required for the discharge of the company’s responsibility. See New Eng. Tel. & Tel. Co. v. State, 104 N.H. at 232, 234, 183 A.2d at 240, 241; Company v. State, 95 N.H. 353, 361, 64 A.2d 9, 16 (1949).
Subject to the qualifications that follow, the commission should set a rate sufficient to yield a return comparable “to that generally being made at the same time and in the same general part of the country on investments in other business undertakings which are attended by corresponding risks and uncertainties[.]” Bluefield Co. v. Pub. Serv. Comm., 262 U.S. 679, 692 (1923); New England Tel. & Tel. Co. v. State, 113 N.H. 92, 95, 302 A.2d 814, 817 (1973); New Eng. Tel. & Tel. Co. v. State, 104 N.H. at 234, 183 A.2d at 241; New Eng. Tel. & Tel. Co. v. State, 98 N.H. at 221, 97 A.2d at 221; Chicopee Mfg. Co. v. Company, 98 N.H. 5, 13, 93 A.2d 820, 826 (1953), overruled on other grounds, 119 N.H. 359, 366, 402 A.2d 644, 649 (1979); C. Phillips, Jr., supra at 339.
Although these standards look outward to capital costs and comparable risks, it is important to recognize their relationships to the actual circumstances of a utility whose rates are under consideration. On the one hand, the commission has authority to recognize the efficiency or inefficiency of a management when it sets the rate of return within the zone of reasonableness. New England Tel. & Tel. Co. v. State, 104 N.H. 229, 239, 183 A.2d 237, 244 (1962) (citing S.W. Tel. Co. v. Pub. Serv. Comm., 262 U.S. 276, 291 (1923) (Brandéis, J., concurring)). On the other hand, the actual needs of the company do not control what the commission may do when it sets the rate of return and the other variables that deter*636mine allowable revenue. The commission may set the “sufficient” rate of return by reference to a capital structure that it finds appropriate, rather than the actual capital structure of the company, as well as by reference to a rate base that does not necessarily reflect all the cost of the company’s actual assets. See New Eng. Tel. & Tel. Co. v. State, supra at 236, 183 A.2d at 243; New Eng. Tel. & Tel. Co. v. State, 98 N.H. 211, 220, 97 A.2d 213, 220 (1953); C. Phillips, Jr., supra at 347-52; Moline, Wolf Creek and the Rate-Making Process, 33 Kan. L. Rev. 509, 512 (1985); see also Appeal of SAPL, 125 N.H. at 713, 490 A.2d at 1333-34.
The commission has this authority to set the rate of return by reference to appropriate, as distinguished from actual, capital structure because the object of the process is to strike a fair balance between recognizing the interests of the customer and those of the investor, Power Comm’n v. Hope Gas Co., 320 U.S. 591, 603 (1944); Chicopee Mfg. Co. v. Company, supra at 11, 93 A.2d at 825, rather than necessarily to guarantee bondholders their interest or stockholders their dividends. See 1 A. Priest, supra at 202. Thus it is realistic to stress the role that judgment must play in setting a rate of return. See Bluefield Co. v. Pub. Serv. Comm., 262 U.S. at 692; New Eng. Tel. & Tel. Co. v. State, 113 N.H. at 95, 302 A.2d at 817; New Eng. Tel. & Tel. Co. v. State, 104 N.H. at 234, 236-37, 183 A.2d at 241, 243.
The same point is true about the process of setting the remaining variable, the rate base, which can be a task of the greatest complexity. For present purposes we may say that two issues are central in a rate base determination: how to identify property which should be included in the rate base, and how to place a value on that property once it is identified.
Much case law in this jurisdiction and elsewhere has addressed one basic aspect of the latter valuation problem, the relative significance to be accorded to original cost as distinguished from replacement cost reflecting inflation or deflation since date of acquisition. See, e.g., Company v. State, 95 N.H. 353, 64 A.2d 9; 2 A. PRIEST, supra at 503-04. Compare Smyth v. Ames, 169 U.S. 466 (1898), with Power Comm’n v. Hope Gas Co., 320 U.S. at 591, and S.W. Tel. Co. v. Pub. Serv. Comm., 262 U.S. 276, 290 (1923) (Brandéis, J., concurring). We may say at the outset that for our present purposes we need not discuss this basic valuation problem. Instead, we may concentrate on the process governing rate base inclusion and will consider the rate base valuation only to the limited extent that the value of investment in rate base property may be reduced to reflect any lack of corporate foresight.
*637It is a constant in the law of ratemaking that there is no single formulation sufficient to express constitutional, statutory, or judicially derived standards for determining rate base inclusion. See Power Comm’n v. Hope Gas Co., supra at 602; Bluefield Co. v. Pub. Serv. Comm., 262 U.S. at 690-91; LUCC, 119 N.H. at 343, 402 A.2d at 633; New Eng. Tel. & Tel. Co. v. State, 98 N.H. at 218-19, 97 A.2d at 219. Such standards are said to be flexible, LUCC, supra at 343-44, 402 A.2d at 634, and their application subject to “pragmatic” adjustment, New Eng. Tel. & Tel. Co. v. State, supra at 219, 97 A.2d at 219.
Attempts to place appropriate limits on the exercise of such pragmatic flexibility have led to the development of two broad principles governing inclusion or exclusion. The first is that of prudence, which essentially applies an analogue of the common law negligence standard for determining whether to exclude value from rate base. Pierce, supra at 511 (citing, inter alia, S.W. Tel. Co. v. Pub. Serv. Comm., 262 U.S. at 289 n.1 (Brandéis, J., concurring)). While the scope of the prudence principle is by no means clear, see Pierce, supra at 513, it at least requires the exclusion from rate base of costs that should have been foreseen as wasteful. See, e.g., LUCC, 119 N.H. at 343, 402 A.2d at 633-34; Company v. State, 95 N.H. at 360, 64 A.2d at 15; see also S.W. Tel. Co. v. Pub. Serv. Comm., 262 U.S. at 289 (Brandéis, J., concurring). If the entire investment in a given asset was foreseeably wasteful, the entire investment must be excluded; if only some of the constituent costs attributable to a given asset were foreseeably wasteful, the value for rate base purposes of the investment in this asset must be reduced accordingly. See Glicksman, supra at 445. Thus, prudence is a principle not only governing inclusion and exclusion, but governing value as well.
The second principle of rate base inclusion or exclusion derives directly from the statutory description of allowable rate base property as “used and useful.” RSA 378:27, :28. Here again, there is no simple formulation that describes the standard of usefulness, Bluefield Co. v. Pub. Serv. Comm., 262 U.S. at 690-91; New Eng. Tel. & Tel. Co. v. State, 98 N.H. at 218-19, 97 A.2d at 219. Prior case law has invested the commission with flexibility in determining what may qualify as used and useful, LUCC, 119 N.H. at 343-44, 402 A.2d at 633-34, thus necessarily providing scope for policy judgments.
Although the relationship between the principles of prudence and usefulness is certainly obscure in existing case law, see Pierce, supra at 513, and although in this jurisdiction the principles *638seem to have been treated as forming a continuum rather than as wholly distinct criteria, see, e.g., LUCC, 119 N.H. at 341, 344, 402 A.2d at 632, 634, the principles are significantly different in at least one respect that is of great potential significance for the treatment of Unit I costs. While prudence judges an investment or expenditure in the light of what due care required at the time an investment or expenditure was planned and made, usefulness judges its value at the time its reflection in the rate base is under consideration. Under the “used and useful” principle, the commission is not asked to second-guess what was reasonable at some time in the past, but rather to determine what can reasonably be done now with the fruits of investment. It is therefore not surprising that the commission’s flexibility in applying the usefulness principle extends to judgments about the inclusion or not of investment in property held for future use. See LUCC, 119 N.H. at 343-44, 402 A.2d at 633-34; N.H. Gas & Elec., 88 N.H. at 55, 184 A. at 605; C. Phillips, Jr., supra at 316. We will return to this point below, when we consider the possible treatment of excess generating capacity.
To summarize the results of the ratemaking process that we have considered significant for this case, we may say that in a proceeding to set rates the commission must set a reasonable rate of return to be allowed on cost-less-depreciation of used and useful property, provided that cost may not include anything imprudently wasteful. The determinations of reasonable rate of return, prudence, and usefulness alike require the exercise of judgment and discretion in determining the recognition that is appropriately due to the competing interests of the company and its investors and of the customers who must pay the rates to provide the revenue permitted.
It should now be apparent that a rate or structure of rates charged to customers is reasonable within the meaning of the statute when it will produce an amount of revenue that has been determined, and limited, by balancing or relatively weighing investor and customer interests. The commission must exercise its judgment in balancing those interests when it determines the allowable extent of operating expenses, when it identifies the property whose prudently incurred cost is included in the rate base, and when it sets a reasonable rate of return on that rate base. Thus a reasonable rate is the rate resulting from a process that must consider the competing interests of investor and customer and must determine the appropriate recognition that each deserves.
Conversely, given the existing statutes that we have cited, the reasonableness of a rate should not be determined either inde*639pendently of the process by which expenses, rate base, and rate of return are set, or after that process has been completed. Although our cases have often referred to the standard of just and reasonable rates as the “ultimate test” of a commission’s rate determination, see, e.g., LUCC, 119 N.H. at 341, 402 A.2d at 632, the statutes provide neither a procedural nor a conceptual basis for judging reasonableness apart from the process that demands recognition of the customers’, as well as the investors’, interests when passing on expenses, rate base, and rate of return.
Indeed, any attempt to judge reasonableness apart from that process would entail redundancy and risk both illegality and unconstitutionality. Redundancy would be entailed because it is difficult to think of any consideration bearing on reasonableness that may not be raised appropriately in the course of the process we have described. We will see an example of what we believe to be such redundancy when we examine below the implications of Commissioner Aeschliman’s separate opinion.
Legality would at least be open to question simply because the relevant statutes mandate the provision of a reasonable return on net cost of used and useful property. The application of any rate-making standard without reference to such a return would be inconsistent with the statutory mandate. Thus, the customer’s interest may not be recognized to the derogation of a reasonable return, any more than the investor’s interest may be recognized to the extent of guaranteeing a return on actual investment and capital structure sufficient for corporate survival. See Market Street R. Co. v. Comm’n, 324 U.S. 548, 566-67 (1945).
A risk of unconstitutionality would likewise arise from any attempt to determine the reasonableness of a rate apart from the process that we have described. This is so, not because the State or Federal Constitution guarantees a particular rate, but because existing concepts of the constitutional limits of ratemaking have been developed in the context of a process that does not determine how far to recognize one competing interest in isolation from the other. That process has been described metaphorically as a “constitutional calculus” in which the interests of investors, like the interests of customers, are variables. Permian Basin Area Rate Cases, 390 U.S. at 769; see also, e.g., Power Comm’n v. Pipeline Co., 315 U.S. 575, 606-07 (1942) (Black, Douglas and Murphy, JJ., concurring). Consequently, any criterion of reasonableness that might be applied independently from the balancing process that does reflect such interests would run the risk of unconstitutionality by inviting the fixing of rates without regard to the balancing of interests. *640Since concerns about these risks of illegality and unconstitutionality have not been briefed or argued in this appeal, we go no further than to note that the risks are there.
We may conclude the discussion by reaffirming an earlier statement quoted above: “the commission controls three variables in regulating rates to provide revenue to [a utility]: operating expenses, rate base and rate ... of return allowed on the rate base.” Appeal of Public Serv. Co. of N.H., 125 N.H. at 49, 480 A.2d at 22. The commission does not control rates by the application of further criteria independently of its consideration of expenses, rate base, and rate of return. A reasonable rate by definition reflects the values placed on those variables and is the result of the process by which those values are derived in balancing customer and investor interests. The obligation to engage in a balancing process guarantees that those values are not captive either to the investor or to the customer alone.
B. The Commission’s Responsibility to Anticipate the Rate Effect of this Financing
An understanding of the commission’s responsibility must begin with RSA 378:27 and :28, which provide authority to hold a rate-making proceeding at which the variables limiting rates to reasonable levels are to be set. This proceeding includes what is colloquially referred to as the prudency hearing, at which disputes about the usefulness of property and claims of imprudence are heard. As a matter of statutory law, the ratemaking proceeding is distinct from the prior proceedings at which the public good must be considered, such as the original certification proceeding under RSA chapter 162-F (1977 and Supp. 1985) in this case, and later financing hearings under RSA chapter 369, at which the public good must be considered in accordance with Easton. It is clear from the statutes cited that ratemaking or prudency proceedings are distinct both in time and in objectives from Easton proceedings, and nothing that we have said or held about Easton hearings implies otherwise. See Appeal of SAPL, 125 N.H. 708, 490 A.2d 1329; Appeal of SAPL, 125 N.H. 465, 482 A.2d 509; Appeal of Easton, 125 N.H. 205, 480 A.2d 88.
It follows that in an Easton hearing the commission’s responsibility to address the rate implications of a decision approving a utility’s financing request is not a responsibility to determine what these rates will actually be if the financing is allowed. (To this conclusion there is one exception, which we will mention below.) Rather, the commission’s responsibility is to determine whether at a *641later ratemaking proceeding a reasonable rate can be set that will allow the company to support the capitalization that will result from use of the proceeds of the proposed financing. Since a reasonable rate is, by definition, a rate derived from a process that balances investor and customer interests, the commission may find that a reasonable rate can be set in the future if it finds that there will be a genuine opportunity to recognize the interests of customers as well as the interests of investors without bankrupting the utility. To support such a finding, the commission should predict a range of rates within which a balance may be struck, with the high side representing the nearest probable reflection of the investors’ interests and the low side representing the closest probable approach to the interests of the customers. Hence, in this proceeding, the commission was obliged to determine whether the probable range of rates would provide genuine scope to resolve the competition between the interests and to determine whether a rate set within the range would allow the company to support the anticipated capitalization. Upon an affirmative determination, the commission could grant the company’s request consistently with its obligation to set reasonable rates at the later ratemaking proceeding.
The one exceptional instance in which the commission would be effectively obligated to determine a rate, rather than a range, in an Easton proceeding like this, would occur if the commission were to find that a requested financing would require virtually full dollar inclusion in rate base as the only alternative to corporate bankruptcy. In that case, the commission would be obligated to determine the particular rate effect that would result. If it found that rate commercially feasible, the commission would then be required to determine whether it would be reasonable as the only alternative to the probable effects of bankruptcy on the customers. This possible exception is not before us in the present case.
C. The Commission’s Findings as to Rate Effect
We conclude that in its supplemental order the commission made sustainable findings that satisfy its obligation to consider rate effects and the ability of the company to support the resulting capital structure with reasonable rates. Commission Report and Fifteenth Supplemental Order No. 17,939 (November 8,1985).
The commission established the upper range of probable rates by making three distinct projections utilizing the so-called base case scenario presented in Exhibit 99A. The base case relies on the undisputed current rate base, to which is added the company’s share of projected Unit I costs. The major assumptions underlying the *642base case are: (1) full dollar inclusion in rate base of the company’s share of total project cost of $4.6 billion; (2) COD of October 31, 1986; (3) no write-off or recovery of Unit II; (4) no loss of UNITIL load; (5) presently requested financing of $525 million secured by third mortgage bonds; (6) availability factor of 59% maturing to 72% (“availability factor” is “used synonymously” with “capacity factor,” Re PSNH, 66 PUR4th at 403 n.44); and (7) phase-in of rate increases over seven years.
In Exhibit 99B, certain sensitive assumptions were then altered to determine their effect on the rates forecast in Exhibit 99A. Exhibit 99B generated a different rate forecast by eliminating the phase-in of the rate increase, resulting in abrupt, or “rate-shock,” increases. Another exhibit, 124D, also eliminated the rate phase-in, but went further to assume to total loss of the UNITIL load, constant availability factor of 60%, and a write-off of Unit II costs against retained earnings.
Rates were then projected using these assumptions. In order to summarize the results here, we have chosen to emphasize the rates for four distinct years or sets of circumstances. The first rate considered is for the year in which the largest percentage increase occurs; the second is for 1991, in which the three upper range rates are in closest proximity; the third is for 1994, in which the three upper range rates diverge to the greatest extent; and the fourth is for the year in which each rate is the highest. Stated in nominal terms, i.e., in terms adjusted to include anticipated inflation as well as the effect of additions to rate base, the largest percentage increase for Exhibit 99A occurs in 1988 and results in a 21.44% increase from the previous year (a rise of 2.05<P/kwh to 11.6<P/kwh); for Exhibit 99B this occurs in 1987 and results in a 50.57% increase (a rise of 4.85<P/kwh to 14.44<P/kwh); and for Exhibit 124D this occurs in 1987 and results in a 67.48% increase (a rise of 6.39<P/kwh to 15.86<P/kwh). The rates forecast for 1991 are 17.65<!:/kwh for Exhibit 99A, 16.62<P/kwh for Exhibit 99B, and 18.62<P/kwh for Exhibit 124D. The rates forecast for 1994 are 24.26<P/kwh for Exhibit 99A, 17.23<P/kwh for Exhibit 99B, and 17.65<P/kwh for Exhibit 124D. The highest rate forecast for Exhibit 99A occurs in 2002 and is 25.08(P/kwh or 195.4% above the rate for 1985; for Exhibit 99B it occurs in 2003 and is 24.89<P/kwh or 193.1% above 1985; and for Exhibit 124D it occurs in 2003 and is 25.18<P/kwh or 204.8% above 1985.
This comparison indicates that when rates are phased in, they continue to rise above the rates in the rate-shock scenarios because the company accumulates deferred revenues during the phase-in period, which must be recovered through substantially higher rates *643in later years. The rates in the two rate-shock scenarios are roughly parallel, but the rates resulting from the assumptions in Exhibit 124D are slightly higher because the scenario assumes a loss of the UNITIL load and a lower capacity factor, resulting in lower revenues and higher unit costs.
It is instructive to note the significant differences when the rates are projected in so-called real terms, that is, without adjustments for inflation, reflecting only the effects of the rate base additions. In real terms, the highest rate on the base ease scenario of Exhibit 99A is 14.89<P/kwh in 1994 (as against a nominal high rate of 25.08<P/kwh in 2002); this is 84.2% above the 1985 rate. On the Exhibit 99B scenario the highest rate is 13.57<P/kwh in 1988 (as against a nominal high rate of 24.89<P/kwh in 2003); this is 67.8% above the 1985 rate. And on the Exhibit 124D scenario the rate is 14.09<P/kwh in 1990 (as against the nominal high rate of 25.18iP/kwh in 2003); this is 79.1% above the 1985 rate.
In each of these instances, there is a clear evidentiary basis for the calculation. It consists of the value of the present rate base, which is not in dispute, plus plant additions and projections of usage that we have reviewed above and found to be based on substantial evidence and within the scope of the commission’s expert judgment.
The commission did not determine a lower end of the probable rate range, for it took the position that it would be entirely speculative to project exclusions from rate base for any portion of the Unit I investment without first holding a prudency hearing. We do not wholly accept the commission’s position on this point. Although there was little evidence in this Easton record of imprudent construction expenditures, the commission did make findings on the extent to which Unit I would provide excess capacity within the first decade of operation. The commission could have considered exclusion from rate base at least on the usefulness principle, and could have derived a probable low range of rates accordingly.
Although the commission did not do this, we do not consider that the failure warranted reversal, because we believe that the commission effectively satisfied the Easton requirement by finding the maximum extent of rate base exclusion that would nonetheless allow the company to avoid bankruptcy. It did this on the basis of an alternative rate scenario prepared by the commission’s consultant, Touche Ross & Co., and introduced by its witness, Mr. Trawicki, as Exhibit 95, schedule 9. This schedule rests, in turn, on the figures introduced by the company in Exhibit 6.
The major assumptions are: (1) no phase-in of Unit I costs; (2) total project cost of $4.5 billion; (3) in-service date of August 1, 1986; (4) *644no write-off or recovery of Unit II costs; (5) no loss of UNITIL load; (6) $480 million Newbrook financing; and (7) availability factor of 72%. Schedule 9 then forecasts the rates resulting from an exclusion from rate base of $1.0 billion of the company’s $1.7 billion investment, following the prudency hearing. The consultant calculated the $1.0 billion figure to be the maximum exclusion which would still allow for rates compatible with the company’s financial survival, in the sense that they would generate cash flow “sufficient to fund operating expenses, debt service and construction requirements when due.” Exhibit 95 at 31. If we then select rates to compare against those highlighted from the earlier scenarios, the largest percentage increase occurs in 1987 in nominal, inflation-adjusted terms of 25.77%, from 9.74<P/kwh to 12.25<P/kwh. The rate forecast for 1991 is 14.15<P/kwh and for 1994 is 15.14<P/kwh. The highest rate forecast occurs in 2000 (the last year of the forecast) and is 17.77<P/kwh or 108.1% above the 1985 rate. In real, non-inflated terms, the highest rate would be lONS^/kwh in 1987, or 30.1% above the 1985 rate. The rates forecast in Exhibit 119 are approximately parallel to the rates forecast in Exhibit 99B but are lower in accordance with the lower rate base resulting from the exclusion of $1.0 billion of the company’s investment.
Although the commission was careful not to suggest that a reasonable rate would probably be found to be as low as the survival level, and while the survival level itself would require adjustment in the light of assumptions about UNITIL load and availability factor, inter alia, the projection of survival level did establish a significant datum for purposes of the commission’s Easton responsibility. By projecting survival despite the possible exclusion of nearly 60% of the company’s Unit I investment, the commission supported its conclusion that there will be substantial opportunity to consider possible rate base exclusions to reflect the customers’ interest if the financing is approved and the company participates in the completion of Unit I. As the commission expressed it, “[t]here is substantial economic leverage to establish a rate level that will not be oppressive to consumers or the New Hampshire economy or which is unfair to stockholders in the event of disallowance of any portion of the capital investment on the basis of imprudence[,]” Re PSNH, 66 PUR4th at 423, or, we should add, on the basis of the used and useful principle.
This adequately addresses the question whether the company’s continued participation will result in a capital structure that probably can be supported by reasonable rates. We therefore hold that the commission’s findings and conclusions on the rate implica*645tions of the proposed financing satisfy the Easton requirements and withstand the appellants’ challenges.
As will be seen, we are at odds here with our brothers who dissent today. We respectfully but emphatically disagree with their position that the commission has failed to deal adequately with the rate implications of the proposed financing. Our prior cases hold that in this Easton proceeding an inquiry into rate effects should focus on the question whether the investment to complete Unit I would result in a capitalization that could be supported by reasonable rates. See Appeal of SAPL, 125 N.H. at 718, 490 A.2d at 1337; Appeal of Easton, 125 N.H. at 212-13, 480 A.2d at 91. Since the commission has addressed this issue and has made findings that support the financing request, it is not obligated to go further in this proceeding. To demand more would assume that in the context of this case the commission could find the nearly completed plant supportable by reasonable rates and nonetheless properly deny the financing with the result of corporate bankruptcy. In our judgment the commission could properly find such an option inconsistent with the public interest.
Although we therefore sustain the commission’s treatment of rate effects, Commissioner Aeschliman’s thoughtful separate opinion on this issue nonetheless deserves comment. It is an occasion not only to underscore some of the preceding analysis of the process by which reasonable rates are derived, but also to confirm the assumptions that the commission has made in referring to its authority to allocate the unprecedented burden of Unit I costs when it deals with the rate base in the approaching rate and prudency hearing.
Commissioner Aeschliman accepts the position taken by the appellants’ witness Gregory Palast that a company rate more than 4-5<P/kwh above the average NEPOOL rate would result in significantly lessened demand and substantial loss of industry to the State, beyond what the commission majority projected in evaluating the price-demand effect. Re PSNH, 66 PUR4th at 460-61. Commissioner Aeschliman therefore believes that it is imperative to set rates within the tolerable 4-5<P/kwh differential. She would reach this result by initially excluding from rate base some $500 million of equity AFUDC attributable to the company’s Unit I investment, on the ground that if represents the accrued capitalized cost of equity capital invested in excess capacity that is not initially used and useful. Id. at 472-73. (Equity AFUDC is the capitalized value of the return on equity investment in Unit I, which must be deferred until the plant is commercially operating, as required by RSA 378:30-a, the “anti-CWIP” law. See Appeal of Public Serv. Co. of N.H., 125 N.H. at 46, 50, 480 A.2d at 20, 22.) Her approach thus has two steps, *646the application of a criterion of reasonableness dependent on the level of regional rates and the application of the usefulness principle to authorize the allocation to the investors of a part of the burden of excess capacity.
Although we believe that the first step in Commissioner Aeschliman’s approach suffers from an analytical flaw, if regarded as a ratemaking device, we nonetheless recognize its value as a tool of criticism. We believe that her second step is one that the full commission may properly consider in the coming ratemaking proceeding.
Dealing first with what we see as the flaw, Commissioner Aeschliman seems to take the position that a given differential between projected company rates and the projected NEPOOL average should itself function as a criterion of what is just and reasonable. Our concern is that this subjection of a proposed rate to a standard of reasonableness independent of the balancing process by which the commission sets allowable expense, rate base, and rate of return is open to the criticisms that we expressed earlier in section VI A supra. The 4-5<P/kwh rate differential is really a standard for predicting tolerable price-demand effect. That is, it would function not merely to predict market behavior, but also to identify economically and socially acceptable limits to the energy prices that cause such market behavior. If it is a sound measure, it should be incorporated into the analysis of need for power. If it were so incorporated, its reappearance as a separate criterion of reasonableness would be redundant. Moreover, as a separate criterion for judging the level of reasonable rates, this standard would evaluate the customers’ interest by reference to the circumstances of other NEPOOL company ratepayers rather than of company investors, and would, therefore, be subject at least to legal and constitutional question. But see Drobak, From Turnpike to Nuclear Power: The Constitutional Limits on Utility Rate Regulation, 65 B.U.L. Rev. 65 (1985).
These considerations support the view expressed by this court over thirty years ago, that “[o]f itself, the evidence relating to rates elsewhere has no conclusive probative force. Its affirmative effect depends upon other evidence to which it may lead.” Company v. State, 95 N.H. at 363, 64 A.2d at 17. Consequently, we seriously question the use of rate differential as an independent criterion of reasonableness, but we also recognize that such a standard, as Commissioner Aeschliman has used it, has value as a critical tool to test the soundness of the process that has produced a proposed revenue level and rate requirement. When such a critical tool suggests the problem that Commissioner Aeschliman infers would follow from fixing rates at the high side of the range considered in this *647case, it is well to consider the commission’s legitimate flexibility in dealing with rate base inclusion and in valuing rate base property on a standard of prudence.
We may start with the commission’s authority to exclude investment or to reduce the value of otherwise recognizable investment on the ground of imprudence. It has to be admitted, on the one hand, that reliance on claims of imprudence can be an expensive way of asserting ratepayer interests. See Pierce, supra at 511-12. Claims of imprudence rest on evidence of what the utility’s responsible officials knew or should have known at some time in the past. In this case, moreover, the appellants are surely correct in pointing out that their assertion of imprudence would be limited in a practical way, at least, by the fact that the commission issued a certificate of site and facility for this construction in 1973 and has approved completion of Unit I now. These facts are no impediment, on the other hand, to the consideration of evidence that the company managed the construction imprudently. We of course express no opinion on the merits that such an assertion might have.
In any event, it is important to bear in mind, as Commissioner Aeschliman’s separate opinion indicates, that the principle of used and useful property will also be applicable in determining rate base. In the face of rate issues that are unparalleled in the State’s history, we should recall that the usefulness principle lends itself to development over time and under new conditions. See Company v. State, 95 N.H. at 358, 64 A.2d at 14. We therefore attend seriously to the suggestions of the separate opinion, that the burden of excess capacity that may be created by such giant projects may appropriately be shared as between investors and customers, see Glicksman, supra at 439-40, and that the usefulness principle may be applied to effect such a shared allocation.
In noting this, we do not imply that we would necessarily agree with Commissioner Aeschliman’s calculation that without UNITIL the entire 409 megawatts, representing the company’s share of the capacity of Unit I, will be excess capacity until 1997-98. Re PSNH, 66 PUR4th at 454. Her conclusion differs from that of the commission majority, which we have already found to rest on a sufficient evidentiary basis. There is no dispute, however, that in its early years of operation the addition of Unit I will result in excess capacity; if the commission’s calculations are adjusted to reflect the loss of capacity responsibility for the UNITIL load, they indicate that the excess capacity will be substantial. Id. at 389.
Nor do we imply that Commissioner Aeschliman’s particular choice of an equity AFUDC exclusion attributable to the new plant is *648necessarily the best method, or even an acceptable mechanism, for recognizing excess capacity under the principle of usefulness. Such a choice is one to be made by expert policymakers, not by this court. Here we confine ourselves to noting that Commissioner Aeschliman’s proposal is one of a variety of regulatory treatments that commissions have devised in order to allocate burdens between investors and customers. See Pierce, supra at 514 et seq.) Colton, Excess Capacity: Who Gets the Charge From the Power Plant?, 34 Hastings L.J. 1133, 1153 et seq. (1983); Glicksman, supra at 453 et seq.
The variety of these treatments reflects not only pragmatic responses to different facts, but different policy choices as well. For example, a rate base exclusion tied to usefulness, without more, is an incentive to stimulate demand, just as the “anti-CWIP” law is an incentive to complete a plant. See Pierce, supra at 516. Thus, in order to limit the force of such an incentive, the Connecticut Division of Public Utilities Control has provided for relief from excess capacity exclusions upon the utility’s demonstration of effective conservation and efficiency measures. See Re Connecticut Light & Power Co., 30 PUR4th 67, 92-93 (Conn. Div. of Publ. Util. Control 1979). Like concerns prompted the Iowa State Commerce Commission to develop a formula for reflecting excess capacity in rate of return, rather than in rate base. See Re Iowa Pub. Service Co., 46 PUR4th 339, 370-71 (Iowa State Commerce Comm’n 1982).
Whether or not Commissioner Aeschliman’s approach is ultimately adopted in the coming rate proceeding, her separate opinion is a reminder both that regulatory concepts are subject to development in the light of new conditions and that the traditional ratemaking process gives the commission flexibility to accommodate the legitimate interests of both customers and investors in responding to the extraordinary issues disclosed by this case. Although in this Easton proceeding Commissioner Aeschliman’s conclusion would be more restrictive than that of the commission majority, her approach is further support for the majority view, which we have quoted before, that “[t]here is substantial economic leverage to establish a rate level that will not be oppressive to consumers or the New Hampshire economy or which is unfair to stockholders. .. .” Re PSNH, 66 PUR4th at 432.
Affirmed.
King, C.J., and Batchelder, J., dissented.