This is an appeal from the order of the public utilities commission issued in its docket DF 83-360, authorizing the New Hampshire Electric Cooperative, Inc. (the Coop) to borrow $46,898,000 from the United States Government acting through the Rural Electrification Administration in conjunction with the Federal Financing Bank. The object of this proposed borrowing is the provision of funds to allow the Coop to participate in completing construction of Unit I and “common facilities” at the Seabrook nuclear power plant, of which the Coop owns a 2.17391% share. We recently upheld the commission’s approval of financing in amounts *128up to a total of $525 million to allow Public Service Company of New Hampshire (PSNH) to accomplish the same purpose with respect to its 35.56942% share of the project. Appeal of Conservation Law Foundation of New England, Inc., 127 N.H. 606, 507 A.2d 652 (1986) (Appeal of CLF).
We affirm.
I. Facts and Procedural History
The Coop is an association incorporated for the purpose of engaging “in rural electrification as limited and defined by RSA 301:53.” RSA 301:3, X. See Whitman v. N.H. Electric Cooperative, Inc., 123 N.H. 111, 459 A.2d 224 (1983). (There is no indication in the record that the Coop members have elected to be governed by the provisions of RSA chapter 301-A. See RSA 301-A:36 (Supp. 1985).) The Coop’s members are ratepayers who purchase electricity furnished by the Coop. RSA 301:53, I. Its source of capital, however, is the Rural Electrification Administration (REA) and its lending affiliate, the Federal Financing Bank (FFB), which lends money to this and other cooperatives at attractive interest rates. Prior to its acquisition of an ownership share of Unit I, the Coop functioned as a non-generating utility, purchasing 90% of its members’ electric power needs at wholesale rates from PSNH and the remainder from four smaller generating utilities. Although the Coop maintains its own distribution system, its service territory is non-contiguous, in the sense that the power lines extending into the rural and mountainous service areas are not interconnected. The lack of an internally integrated system limits the ability of the Coop to distribute power received at any of the twenty-seven distinct delivery points, where purchased power enters the Coop’s lines from outside sources, to other points within the Coop’s total service area.
In 1981 the commission authorized the financing necessary for the Coop to acquire the capacity to generate a portion of its power needs by purchasing a 2.17% ownership interest in the Seabrook nuclear power plant. See Re New Hampshire Electric Cooperative Inc., 66 N.H.P.U.C. 139 (1981). This interest will entitle the Coop to receive 25 megawatts (mw) of power from Unit I, or about 20% of its estimated power needs at the time the plant is expected to begin generation. The Coop bought this ownership interest from PSNH, after the latter had been forced by escalating costs to find buyers for a portion of its original share in the plant. See Re Public Service Company of New Hampshire, 64 N.H.P.U.C. 485 (1979); Re Public Service Company of New Hampshire, 64 N.H.P.U.C. 262 (1979).
In acquiring its ownership interest, the Coop entered into several *129contracts of significance for purposes of this appeal, two of which are between the Coop and PSNH. The first is the contract for the sale of the 2.17% interest, which the Coop obtained without assuming or otherwise paying PSNH for any portion of the AFUDC (allowance for funds used during construction) attributable to that share as of the purchase date. (For a discussion of AFUDC, see Appeal of CLF, 127 N.H. at 620, 507 A.2d at 662; Appeal of Public Serv. Co. of N.H., 125 N.H. 46, 50, 480 A.2d 20, 22 (1984).)
The second significant contract is a so-called sell-back agreement. In 1981, PSNH agreed that during the first ten years of commercial operation of Unit I it would purchase the Coop’s “capacity and related energy which is temporarily in excess of the Coop’s needs from Seabrook Unit No. 1 and No. 2 at the [Coop’s] full cost. .. .” On March 8, 1985, the president of PSNH wrote to the manager of the Coop, who countersigned the letter to indicate his agreement, “to clarify the intent of the Parties expressed in the ‘sell back’ . . . agreement.” Under the 1985 clarification, the Coop is entitled “to sell to [PSNH] up to the [Coop’s] full 2.17391% ownership share of Unit No. 1 capacity during the first 10 years of Unit No. l’s commercial operation and [PSNH is obligated] to accept whatever portion the [Coop] has determined to sell to [PSNH].” The commission has treated the sell-back agreement as effectively allowing the Coop to choose a least-cost alternative whenever the cost to the Coop of its share of Seabrook power is greater than the cost of purchasing the same amount of power from PSNH at the wholesale rate set by the Federal Energy Regulatory Commission (FERC).
The Coop’s third significant agreement is with the joint owners of the Seabrook project, including PSNH, under the terms of which the Coop is obligated to contribute its share of construction costs in return for its entitlement to 2.17% of generated power upon completion. In 1981, the Coop and the commission expected that a loan of $75,750,000 would fund both the original purchase from PSNH and the Coop’s share of future expenditures to complete both units. By 1983, however, rising costs forced the Coop to request authority to borrow a further $111 million to fund its share of unfinished construction and to meet its contractual obligation.
Accordingly, on November 18, 1983, the Coop began the proceeding now on appeal when it filed a petition with the commission seeking borrowing authority in the amount of $111 million to meet its interim estimated additional needs for its share of the construction of Seabrook Units I and II. The commission subsequently opened docket DF 83-360 and on February 24, 1984, it granted the full amount of the borrowing authority requested. Roger Easton, Gary McCool, and the Consumer Advocate, Michael Holmes, appealed *130this order, and we remanded the case to the commission. We held that the subjects of a financing proceeding under RSA chapter 369 are not limited to the terms of the proposed financing but, rather, should include a determination of “whether, under all the circumstances, the financing is in the public good — a determination which includes considerations beyond the terms of the proposed borrowing.” Appeal of Easton, 125 N.H. 205, 213, 480 A.2d 88, 91 (1984); see also Appeal of Seacoast Anti-Pollution League, 125 N.H. 465, 472-75, 482 A.2d 509, 515-17 (1984).
After this remand and before the commission began further hearings in this docket, the Coop filed a petition on April 3, 1985, reducing its requested borrowing to $49,580,000, an amount that reflected both substantially changed circumstances from those that existed in 1983 and interim emergency borrowing previously approved. On April 5, 1985, the appellants filed a further appeal raising two issues on which we heard oral argument by the parties and the office of the attorney general. On April 12, 1985, we decided the appeal by an order declaring, first, that in accordance with our ruling in Appeal of SAPL, supra at 468-72, 482 A.2d at 512-15, the then chairman of the commission, Paul McQuade, was disqualified from sitting in the proceedings in this docket. Second, we denied the appellants’ request to suspend a commission order granting the Coop emergency authority to borrow an additional $5,290,484, nearly half of which was for Unit I costs.
After beginning the hearings on remand on April 23, 1985, the commission issued an order delineating the scope of what has come to be known as this Easton inquiry. Commission Report and Fourteenth Supplemental Order No. 17,568 (April 30, 1985); see Appeal of CLF, 127 N.H. at 612-13, 507 A.2d at 657-58. Since many of the questions in this docket were related to questions that the commission previously considered in docket DF 84-200, which dealt with PSNH’s proposed financing of its share of Unit I costs, the commission took administrative notice of the entire record in DF 84-200. See Commission Report No. 17,568, at 3. The commission accepted those findings from docket DF 84-200 that were generic, in that they applied equally to both proceedings, and excluded those findings relevant only to the PSNH request. Id. at 5-6.
The commission held that in the present docket it would examine, inter alia, the following issues: whether the Coop’s share of Unit I would be needed to serve its customers; whether the Coop’s continued participation in Unit I would be in the public interest; whether obtaining alternative energy sources would be preferable to the Coop’s continued investment in Unit I; the reasonableness of the *131Coop’s rates that would result from its continued investment in Unit I; the effect of those rates on demand for electricity by Coop customers; and the effects on the Coop that would follow from a denial of the requested borrowing authority. The commission heard nine days of testimony, which concluded on May 3, 1985. The transcript of that testimony covers 1827 pages, and the record includes 89 exhibits.
On May 31, 1985, a majority of the commission issued a 136-page report and supplemental order, authorizing the Coop to borrow $46,898,000. Commission Report and Seventeenth Supplemental Order No. 17,638, at 136 (May 31, 1985). The difference between this and the amount originally requested is attributable to the borrowing authorized by the third emergency financing order. Commission Report and Sixteenth Supplemental Order No. 17,599 (May 10, 1985). The order attached no conditions to the borrowing. Commissioner Aeschliman filed a 16-page separate report, dissenting in part.
After the commission denied the appellants’ timely motion for rehearing, they brought this appeal. As was true in Appeal of CLF, the present appellants did not question the terms and conditions of the proposed financing, but rather challenged the commission’s methodology and findings in dealing with the broader issues implicated by the public good.
We heard oral argument on the merits on November 12, 1985. On November 15, 1985, 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 [would] be set if the borrowing [occurred] as authorized by the commission . . . .” See Appeal of CLF, 127 N.H. at 613, 507 A.2d at 657. In addition, we asked the commission whether such findings would have any effect on the validity of the conclusions stated in its report and order of May 31, 1985. We based this remand on our view that the commission’s report and order of May 31, 1985, failed to present this court with findings of fact on the issue of future rates sufficient for genuine appellate review, as required by Appeal of Seacoast Anti-Pollution League, 125 N.H. 708, 718, 490 A.2d 1329, 1337 (1984). The commission majority responded on November 22, 1985, with a 45-page report and supplemental order containing rate findings and indicating that those findings had no effect on the validity of the conclusions stated in its original order. Commission Report and Twenty-Second Supplemental Order No. 17,960, at 43-44 (November *13222, 1985). Commissioner Aeschliman again filed a separate supplemental opinion. Thereafter, we allowed the parties to file supplemental briefs.
On March 3, 1986, the appellants filed a letter with this court, calling our attention to a Petition for a Comprehensive Avoided Cost Rate Proceeding filed with the commission by PSNH on February 7, 1986. That petition requested the commission to set rates for the statutorily mandated purchase by all New Hampshire utilities of energy produced by small power producers and cogenerators. See RSA 362-A:4. The appellants expressed concern that the PSNH petition estimated small power production and cogeneration substantially in excess of the amount that the commission had estimated in its consideration of alternative energy sources in this case. We treated the letter as a motion to remand, on which we heard oral argument from the parties, the attorney general’s office, and PSNH on March 7, 1986. By order of even date herewith we deny the motion.
II. Significance of the Appeal of Conservation Law Foundation of New England, Inc.
Since this appeal is similar to Appeal of CLF in dealing with borrowing authority to complete Unit I, a number of the legal issues raised here quite naturally duplicate issues in the preceding case. To avoid repetition we will rely on the opinion in Appeal of CLF where it is possible and will omit discussion, for example, of the appellants’ burden of persuasion, the scope of our review, and the propriety of using an incremental cost analysis to assess the relative economic desirability of alternatives to completing Unit I.
Where the issues presented in this appeal are related but not identical to those raised in Appeal of CLF, we will assume familiarity with that opinion as a basis for our discussion, as in the treatment of need for power. We will trust less to such an assumption, however, in dealing with the commission’s rate projections and the consistency of rates as projected with the statutory standard of the reasonable rate. Without a more extended summary of our prior opinion on this matter, we believe that the distinctions between Appeal of CLF and the present case would not be readily apparent.
III. Issues Requiring No Extended Discussion
Before proceeding, we take note of four issues that merit no extended discussion. First, the record does not support the appellants’ claim that in criticizing the testimony of the witness Christopher Flavin the commission indicated that it had relieved the *133Coop of the burden of proof. The commission’s statement that his testimony “offers no specific program and deals only in generalities” simply explained its limited benefit in the proceeding. Commission Report No. 17,638, at 95-96.
Second, the record does not indicate that in this proceeding the commission approved further financing for construction of Unit II. Commission Report No. 17,638, at 53, 55. The Coop, in its petition seeking authority to borrow $49,580,000,-listed the prior Unit II investment simply for the purpose of presenting a comprehensive statement of existing, as well as anticipated, Seabrook investment.
Third, there is no indication that the commission erred in failing to hold the Coop’s entire Unit I investment ultra vires for failure of the Coop’s management to obtain direct ratepayer approval by means of a referendum. Even assuming that it would be timely to raise such a claim in this appeal, the burden to establish error is on the appellants, who have presented no evidence or authority indicating that any referendum was required. Commission Report and Twentieth Supplemental Order No. 17,699, at 9 (June 28, 1985).
Last, the commission did not err in its evaluation of the significance of a portion of the March 8, 1985 letter from the president of PSNH to the Coop manager, countersigned by the manager to indicate his agreement with its substance. The manager purported to agree on behalf of the Coop to refrain from “actively pursuing] such cogeneration or power from small power producers to replace its Seabrook entitlement or partial requirements service.” Whatever may have been the wisdom of indicating such agreement, the commission found that it is inconsistent with public policy and therefore has no binding legal effect on the Coop. The Coop is in no way restricted in its purchases of power from such electricity producers by the terms of this letter. Commission Report No. 17,638, at 90, 91.
IV. Need for Power
The appellants’ first challenge to the commission’s analysis of the need for power rests on a failure to understand the significance of the Coop’s status as a nongenerating utility. When the appellants assert that “no evidence in the instant record [exists] which would reasonably support a finding that Seabrook or, specifically, the [Coop’s] Seabrook share is ‘reasonably requisite’ for the [Coop’s] ‘present and future use’” (emphasis in original), they fail to recognize that the proper analysis of a nongenerating utility’s power need is significantly different from the analysis that is appropriate in assessing a generating utility’s capacity needs.
*134The proper approach to the analysis in this case begins with the undisputed fact that the Coop needs at least 25 mw of power. Likewise it is undisputed that, as a nongenerating utility, the Coop historically has not been obligated to provide a particular level of generating capacity. Furthermore, as a net purchaser, the Coop “has a secure source of capacity for the full length of the study period [, which] is PSNH’s commitment to meet the [Coop’s] capacity needs through wholesale power sales.” Commission Report No. 17.638, at 112. The commission, therefore, correctly declined to pose the issue of need for Unit I power in absolute terms, or in terms of reliability of source of supply. Its use of a net benefit analysis to evaluate the relative desirability of owning rather than buying power was a reasonable approach for analyzing the need issue. See id. at 108-13.
Although the need analysis was correctly subsumed under the evaluation of ownership and purchase alternatives, it was nonetheless necessary to determine, as in a traditional need analysis, the effect of predicted rate changes on the level of demand. This price-demand or elasticity effect must be considered in calculating the consequences of the requested borrowing on customer rates. (For a more detailed discussion of elasticity of demand, see Appeal of CLF, 127 N.H. at 628-29, 507 A.2d at 668.)
As a second challenge to the need analysis, the appellants contend that the commission failed to evaluate elasticity adequately. We must, therefore, consider the commission’s choice of methodology for calculating the effects of different rate changes on demand for electricity.
Coop witness Lee Smith, chief economist of the economic consulting firm, LaCapra Associates, and Mr. Easton each presented an elasticity analysis. The commission chose to accept the analysis proffered by Ms. Smith, which relied on the load forecast prepared for the Coop by Dalton Associates.
The commission rejected Mr. Easton’s use of price elasticity measures derived from a NEPOOL study for two reasons. The commission found that Mr. Easton’s calculations utilized nominal prices while the NEPOOL price elasticities used were based on real prices. Commission Report No. 17,638, at 74. More significantly, although the NEPOOL standard was successfully employed by Public Service Company of New Hampshire, see Re Public Service Co. of New Hampshire, 66 PUR4th 349, 384 (N.H.P.U.C. 1985), the commission found that this NEPOOL formulation did not reflect the specific characteristics of the Coop’s service area, which is significantly different from the NEPOOL service territory. Commission Report No. 17.638, at 76-77. For example, contrary to the NEPOOL study, the *135Dalton Report assumes an insignificant response by the commercial sector to price changes. Although the Dalton Report presents no historical data on the relationship between price and demand for the commercial sector, id. at 75-76, Ms. Smith explained that the Coop’s commercial class is largely composed of recreational industries. The electricity costs for these industries constitute such a minor part of their total costs that rate increases would have only a minor impact. Id. at 76.
For these reasons we find no error in the commission’s decision to reject Mr. Easton’s calculations, and we likewise find none in its decision to accept Ms. Smith’s position. The commission had both the obligation and the authority to decide upon the proper methodology for calculating price-demand effects, and we find that an evidentiary basis exists for its exercise of judgment in choosing the Dalton methodology espoused by Ms. Smith.
The appellants raise a further challenge to the methodology adopted by the commission, which is implicit in their motion to remand. As we noted in the procedural history, the appellants sought a remand in response to the Petition for a Comprehensive Avoided Cost Rate Proceeding, which PSNH filed with the commission. PSNH asked the commission to establish avoided cost rates applicable to all public utilities when buying power pursuant to the statutory mandate to purchase power from small power producers and cogenerators. See RSA chapter 362-A. PSNH also sought a revision of the current long-term avoided cost rates, because it claimed that the present rates had been set erroneously and were encouraging production far exceeding anything assumed either in docket DF 84-200 and Appeal of CLF, or in the record of the present docket.
The appellants assert that if the PSNH projections are sound, they undermine the commission’s assessment of the significance of cogeneration as an alternative to Unit I power. We think this is wrong. The commission found that small power producers and cogenerators are undependable because they are not subject to any regulatory obligation to produce power and therefore will cease generating electricity whenever it is no longer economically advantageous. The commission concluded that such power sources can only serve as complements to, rather than substitutes for, the Coop’s 25 mw of Unit I capacity. Commission Report No. 17,638, at 88-89. See infra. The most recent PSNH cogeneration projections do nothing to disturb these findings.
If, however, these projections were accepted as sound predictions of the cogeneration and small power production that will occur in *136the future, they would place the validity of the commission’s price-demand calculations in question. The implication would be that in failing to estimate the degree to which projected rates would stimulate cogeneration, the commission must have miscalculated the price-demand effect.
After hearing oral argument on the remand motion, however, we are satisfied that the PSNH claims should not be so read. Instead, we believe that PSNH is arguing that the rates currently set by the commission under RSA 362-A:4 for PSNH’s purchase of cogenerated power reflect an erroneous and inflated application of the “avoided cost” standard of valuation mandated by that section. Failure to correct this misapplication will result in subsidizing, and thus stimulating, cogeneration beyond the point of consumer benefit as intended by the statute. The PSNH figures thus project the effects of what PSNH claims to be a regulatory error in applying RSA 362-A:4. They do not imply an error in measuring price-demand effect given the correct application of RSA 362-A:4.
Furthermore, in both docket DF 84-200 and the present case, the commission relied on staff estimates of cogeneration, which were independent of PSNH’s projections. See DF 84-200 Exhibit 67 attachment 4. Given this independent derivation, the present PSNH claims do not raise a probability of error either in projecting cogenerated power production or in measuring price-demand effects upon which such projections must in part rest. Hence, we will deny the motion to remand.
V. Evaluation of Proposed Alternatives
As the first step in its ultimate need analysis, the commission sought to ascertain whether ownership of any generating alternatives would be more desirable than continued ownership of the Coop’s 2.17% share of Unit I. In making this determination, it employed an incremental cost analysis to evaluate the alternatives proposed. Commission Report No. 17,638, at 102. The appellants claim that it was error to employ this methodology.
“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. We,recently held that the use of an incremental cost analysis is reasonable for the purpose of assessing alternatives to Unit I, Appeal of CLF, 127 N.H. at 620, 507 A.2d at 661-62, and therefore reject the appellants’ challenge without further discussion.
*137Appellants next challenge the commission’s reliance on certain expert judgments, rather than full quantitative cost studies, to determine whether the Coop’s continued participation in Unit I is preferable to various alternatives. Appellants emphasize our language in Appeal of Easton, 125 N.H. at 212, 480 A.2d at 91, that the commission should consider “whether the uses to which the loan will be put can be economically justified compared to other options available to the [Coop] . . . .” Appellants contend that our holding requires formal quantitative studies of all proposed alternatives to the Coop’s participation in Unit I.
Although our language in Easton requires a thorough analysis of alternatives to the Coop’s ownership of 25 mw of Unit I capacity, it does not support the appellants’ present claims. The Coop argued that certain alternatives were impractical on the basis of what it called logical analysis, and the commission accepted this claim. It specifically found that independent generating facilities would be impractical to develop as alternatives to Unit I ownership, because the Coop lacks an internally integrated transmission system, which the development of an independent generating facility would require. The commission agreed with Ms. Smith’s conclusion that construction of an internal generating system would be “financially and logistically unrealistic” for the Coop. Exhibit R-24, at 3. It found that “the geographical terrain throughout... the Cooperative service area does not lend itself easily to an integrated network. The mountainous areas dividing the various small pockets of customers prevent the installation of contiguous distribution facilities .... [T]he . . . distance^] between various small customer loads make it economically impossible to develop an integrated system.” Commission Report No. 17,638, at 83-84. The appellants have failed to demonstrate that this finding is unreasonable.
Once the commission concluded that all independent power generating alternatives required construction of an economically impractical integrated transmission system, incremental cost analyses of these alternatives exclusive of transmission system costs were, in a strict sense, superfluous. The commission nevertheless determined that coal-burning plants, windpower, and wood-fired plants are all more expensive on an incremental cost basis than the cost of continued participation in Unit I. Commission Report No. 17,638, at 114.
The commission then considered non-owned generating alternatives. It rejected both Canadian energy from Hydro-Quebec and cogeneration and small power production, applying the same *138standards described in Re PSNH, 66 PUR4th at 389-90. It concluded that both alternatives would be undependable substitutes for Unit I power, and for the same reasons we gave in Appeal of CLF, 127 N.H. at 622-24, 507 A.2d at 663-64, we do not find the commission’s conclusions unreasonable.
When the commission turned to nongenerating alternatives, it concluded that there was insufficient evidence that conservation resources could be developed at a cost below the incremental cost for Unit I. Commission Report No. 17,638, at 115. Moreover, Ms. Smith testified that customers of distributing, as opposed to generating, utilities benefit from conservation only if they participate in its programs. The cost of conservation programs results in increased rates, and customers benefit from conservation only to the extent that their participation reduces their energy consumption, thereby offsetting the rate increase. Nonparticipating customers, on the other hand, receive no benefit in the form of reduced usage, but face the increased rates. Id. at 94.
The commission concluded generally that all power generating alternatives should be regarded as complements to, rather than replacements for, Unit I power. Commission Report No. 17,638, at 112. Likewise it concluded that nongenerating alternatives could only serve as complements to Unit I power. Although nongenerating conservation alternatives cannot substitute for the 25 mw of base-load capacity offered by Unit I, they could nonetheless reduce the need to purchase power and may therefore be treated as potentially affecting marginal load requirements. Id. at 112-13.
Finally, before ending the discussion of Unit I alternatives, we must address the appellants’ contention that the commission erred in evaluating those alternatives only as a series of sole substitutes for the Coop’s Unit I ownership share. On the contrary, the commission concluded that “there is no combination of alternatives that cost less than Seabrook under an incremental cost standard.” Commission Report No. 17,699, at 21. The commission stated that “[i]n the absence of substantial evidence that the synergism of discrete alternatives and other conservation measures will substitute for Seabrook capacity and energy, we cannot responsibly abandon Seabrook for conjectural and inadequate sources of power to meet demand.” Id. at 21-22. It is thus clear that the commission did not fail to evaluate the combined effect of proposed alternatives.
Once the commission concluded that Unit I alternatives to replace or decrease the demand for owned Unit I power were either impractical or more costly than continued participation in Unit I, it reasoned that only two alternatives remained: continued ownership *139participation in Unit I, or termination of participation and purchase of equivalent power from PSNH. Commission Report No. 17,638, at 84. The commission utilized a quantitative net benefit analysis to compare the cost of continued participation in Unit I to termination of participation. Coop witness Ms. Smith provided 15 scenarios, in each of which the net present value analysis favors continued participation as less expensive than termination. Id. at 111; Exhibits R-21 A, R-21B, R-21C.
The net present value analysis favoring completion reflects two related factors. The first is the likelihood that, upon termination of participation, the fixed market value of the Coop’s share of Unit I would be less than its actual investment in the plant. An appropriate range for that market value was estimated at $12.5 million to $50 million. Commission Report No. 17,638, at 125.
The second related factor affecting the analysis is the Coop’s liability under the Joint Ownership Agreement. That agreement provides that PSNH may make a Unit I installment payment which the Coop fails to pay when due, and is entitled to recover that amount from the Coop with interest at 2% above the prime rate. After five months of continued default, the Coop is responsible not only for the full amount of the defaulted payments with interest at 2% above prime, but also for a penalty equal to 25% of the lesser of the Coop’s net investment or the value of its share in Unit I. The Coop then would be entitled to reimbursement for the lesser of its net investment or the value of its share reduced by the liability. Commission Report No. 17,638, at 124.
Considering the most favorable prediction of fair market value, the $50 million value of the Coop’s share would be reduced by a 25% penalty of $12.5 million. Its return would be further reduced by the accrued Unit I payments and interest for the five month grace period, which would total $6 million. After subtracting these expenses, the Coop would be entitled to a net return of $31.5 million. The Coop would still owe $78 million to the FFB, and the $31.5 million return would leave the Coop with a net liability of $46.5 million even under the most optimistic predictions. Commission Report No. 17,638, at 126. The cost of discharging this debt, when combined with the cost of purchasing power to satisfy all of the Coop’s demand, would produce higher retail rates to Coop members than any rates projected on the basis of continued participation. Id. at 127. The commission therefore concluded that “continued participation in Seabrook Unit I provides lowest cost power to [Coop] ratepayers.” Id. at 112. Those calculations support the commission’s con*140elusion that the Coop’s continued ownership of an interest in Unit I is the economically preferable alternative at this point.
VI. Rate Effects and Reasonableness of Resulting Rates
A. Implications of Financing for Later Ratemaking
As we explained in some detail in Appeal of CLF, 127 N.H. at 640, 507 A.2d at 675-76, a proceeding to evaluate a proposed financing plan under RSA chapter 369 is distinctly different from a proceeding to set rates under RSA 378:27 and :28. If, however, the commission acts favorably on a financing proposal, the utility may be expected to request that the resulting capitalization be supported entirely by customer rates, a request that customers may oppose on the ground that the expenditures for the objects of the financing should not be charged to them. Although these issues of rate support may be resolved only at the later proceeding to set rates, at a financing proceeding the commission cannot ignore the potential effect of its decision on the interests that must be considered when rates are finally set.
Accordingly, the commission’s obligation to consider the public good when evaluating a financing request requires the commission to determine whether the financing, if approved, will result in a capital structure that can be supported by lawful rates. A lawful rate is what the statutes describe as a “reasonable,” RSA 378:27 and :28, or a “just and reasonable,” RSA 378:7, rate. At the financing stage, therefore, the commission’s obligation is to determine whether rates necessary to support the anticipated capitalization will be consistent with the reasonable rate standard.
The concept of the reasonable rate must be understood by reference to the legitimate components of a utility’s revenue requirement and to the process by which those components are specifically approved by the commission. Those components are summarized in the 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 depreciation 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 that rate base.” Appeal of CLF, 127 N.H. at 633-34, 507 A.2d at 671. The resulting return provides “ ‘interest on long-term debt, dividends on preferred stock, and earnings on common stock (including surplus and retained earnings).”’ Id. at 635, 507 A.2d at 672 (quoting C. Phillips, Jr., The Regulation of Public Utilities 332 (1985)). To the extent that the utility’s revenue requirement must be raised through customer rates, a utility’s *141average rate is the amount per unit of energy consumed that will yield the revenue.
In speaking of an average rate, we simply recognize that a utility does not necessarily have a single retail rate that it charges to all of its consumers. Rather the utility may establish different rates for various classes of consumers when circumstances render any lack of uniformity reasonable. RSA 378:11; see Granite State Alarm, Inc. & a. v. New England Tel. & Tel., 111 N.H. 235, 237-38, 279 A.2d 595, 597 (1971). The Coop’s rate projections, therefore, do not purport to forecast one rate that will be charged uniformly to all of the Coop’s members but instead to forecast the average rate the Coop must charge to satisfy its revenue requirement, given the projected level of demand. For the sake of simplicity, in the discussion that follows when we speak of a rate forecast it should be understood that we are referring to an average rate.
That average customer rate is reasonable only if the commission limits the components of the revenue requirement in accordance with certain basic ratemaking standards. The commission must review operating expenses, of course, and it must limit the percentage rate of return on the rate base to a level that is comparable to the return realized on other investments of comparable risk. For present purposes, however, the most significant variable in the revenue-setting and ratemaking process is the rate base.
As in Appeal of CLF, our concern over the rate effects of this financing will concentrate on two related principles that limit both the inclusion of items in the rate base and the value of an item that is included. First, the principle of prudence requires that an investment or a constituent element of an investment that was foreseeably wasteful when made be excluded from the rate base. The corresponding inquiry in a proceeding to set rates is thus commonly spoken of as a prudency hearing. Such a hearing, however, is not concerned solely with prudence in this strict sense; it is also the occasion for applying a second limitation on the rate base, to costs attributable to property that is used and useful in the public service. See RSA 378:27.
Neither of these principles of rate base limitation can be reduced to a precise formula. The principle of prudence entails the uncertainty that is inherent in any backward-looking judgment, and the principle of usefulness is commonly described as allowing a rate-setting commission substantial flexibility for pragmatic judgments about what should or should not be regarded as useful. See Appeal of CLF, 127 N.H. at 637, 507 A.2d at 674. This flexibility mirrors the need to provide an opportunity for the exercise of expert judgment *142in giving due recognition to the two competing interests that come to the fore in any contested rate proceeding, the interests of investors who would like a guaranteed return on any investment and the interests of customers who would like low rates.
In advance of a rate proceeding, neither the investors nor the ratepayers can be entirely certain of the extent to which their interests ultimately will be recognized. They know only that their interests are subject to the commission’s obligation to give appropriate recognition to the interests of the others. Thus, the obligation to limit the rate base and other revenue producing factors precludes a guarantee of survival to the utility, and it is no guarantee of freedom from hardship to the customers. Although a proper rate may be too low to provide a return on actual equity investment or even to meet a utility’s actual debt obligations, that rate may be higher than comparable utility rates elsewhere. In any event, a rate is judged reasonable only insofar as it results from a ratemaking process in which the prudence and usefulness principles of rate base valuation, among others, are applied. See Appeal of CLF, 127 N.H. at 638-39, 507 A.2d at 674-75.
We have accordingly held that although the commission is not obliged to set rates in a financing proceeding, it nonetheless must determine whether at a later rate hearing there will be a genuine opportunity to set rates consistent both with the principles we have discussed and with the solvency of the utility. Thus in this case, as in Appeal of CLF, the history of cost escalation in the Sea-brook project obligated the commission to anticipate the probable range of rates that could result from applying the prudence and usefulness principles after making the investment that is the object of the financing in question. The breadth of the range must reflect the degree of risk that the prudence and usefulness principles will later support rate base exclusions, with the high side representing rates comparatively favorable to investors and the low side comparatively favoring ratepayers. See Appeal of CLF, 127 N.H. at 641, 507 A.2d at 676.
In Appeal of CLF we held that if rates could be set within that range consistently with the utility’s solvency and reasonable customer demand, and if the object of the financing was otherwise desirable, the commission could find the financing to be consistent with the public good. Appeal of CLF, 127 N.H. at 641, 507 A.2d at 676. Conversely, we held that if it appeared that corporate solvency would not be consistent with traditional principles of revenue limitation, the public good standard would require denial of the financ*143ing request, subject to two qualifications. See Appeal of CLF, 127 N.H. at 624-25, 645, 507 A.2d at 664-65, 678-79. If traditional ratemaking principles would call for a rate base too low to support the capitalization that the financing would produce, but corporate insolvency was the only alternative to the otherwise disfavored financing and later rate base inclusions, the commission could consider the policy question whether the financing, and the rate structure necessary to support it, would be preferable to insolvency and bankruptcy. On similar reasoning, we left open the possibility that a partially constructed plant could be completed even when an adequate alternative power source had become cheaper on an incremental cost analysis, if failure to complete the plant already begun would result in bankruptcy. Id. at 624-25, 507 A.2d 664-65.
The facts in Appeal of CLF raised neither of these exceptional situations. The commission had found a range of rate effects of PSNH’s participation in the completion of Unit I that would allow for corporate solvency, while providing genuine scope for the application of traditional principles of revenue limitation. We concluded that the commission’s decision amounted to an adequate compliance with the standards we have summarized, and we upheld the commission’s order approving the financing to complete PSNH’s share of Unit I. See Appeal of CLF, 127 N.H. at 644-45, 507 A.2d at 678.
B. Need for Independent Analysis of Rate Effects in This Case
The commission’s decision and our affirmance of it in Appeal of CLF, however, do not control the result in the present appeal, and it is important to understand why this case must be considered independently. First, the effect of Unit I investment on the average rate is not the same for PSNH and the Coop. The combination of Unit I power attributable to the Coop’s 2.17% share and Unit I power in the Coop’s wholesale power purchases will satisfy a greater proportion of the Coop’s total power needs than PSNH’s share of Unit I power in relation to its total need. Since the unit cost of Unit I power will be greater than the cost of power from existing sources, the Unit I costs will tend to drive the Coop’s rates higher than PSNH’s rates, other things remaining equal.
Second, although the construction cost per kilowatt of Unit I capacity is the same for the Coop and PSNH, the Coop’s cost of debt capital is less than PSNH’s average debt cost. The Coop’s principal source of investment is the FFB, which will lend money at 1/8 of 1% above “the U.S. Treasury cost of borrowing.” Commission Report No. 17,638, at 117. This rate is markedly lower than PSNH’s stated borrowing cost, although it is of course the after-tax cost of borrowing that is significant for comparison purposes. While the Coop’s *144lower cost of capital tends to offset the relatively higher cost of its proportion of Unit I power, it apparently will not be sufficient to do so entirely. The higher cost of Unit I power will still probably raise the average cost of the Coop’s power above that of PSNH for the period covered by the rate projections submitted in this case.
Third, although PSNH has no way to avoid the effect of Unit I ownership in tending to drive rates upward, during the first ten years of Unit I operation the commission has found that the Coop does have a mechanism to avoid such effects to the extent that they result from its ownership share. The commission has interpreted the sell-back agreement with PSNH to provide that the Coop can sell its share of Unit I power back to PSNH at the Coop’s cost if that is higher than the cost at which the Coop can purchase wholesale power from PSNH. See Commission Report No. 17,638, at 123. The commission has found that the contract’s practical effect is that the Coop can transfer fixed Unit I costs to PSNH if that would be advantageous during the first decade of operation. Id. at 130.
Fourth, PSNH and the Coop have different capital structures and correspondingly different sources of revenue need. Even if a given component of revenue need is treated similarly for both PSNH and the Coop, the implications for corporate solvency may be quite different. PSNH’s capital structure comprises both equity and debt capital, whereas the. Coop has no equity investors and its capital consists entirely of debt-.with a surplus (referred to as Patronage Capital) of some six million dollars. Commission Report No. 17,638, at 116; Exhibit R-9. As a consequence of this difference, a rate base exclusion of some Unit I investment could fall on PSNH’s equity investors without causing insolvency, whereas any exclusion of the Coop’s Unit I investment would fall on the Coop’s lenders and threaten default, subject only to the small cushion of surplus.
For these reasons, the commission was obligated in this case to make a separate examination of the effects on the Coop’s rates of the financing to complete Unit I.
C. The Commission’s Findings on Rate Effects
In Commission Report No. 17,960 the commission established an upper limit to the range of probable rates by combining the power cost projections of Exhibit R-21A with the distribution and administration costs implicit in Exhibit R-47. Exhibit R-21A relies on the assumptions utilized in the PSNH financial scenario referred to as KD-NU-R in docket DF 84-200. The major assumptions underlying this scenario are: (1) Unit I construction costs of $882 million to go as of August 1, 1984; (2) full dollar inclusion in rate base of the *145Coop’s share of total Unit I project cost of $4.6 billion; (3) commercial operation date (COD) of October 31, 1986; (4) full recovery of Unit II costs from ratepayers; (5) full loss of wholesale purchases by UNITIL companies (see Appeal of CLF, 127 N.H. at 630-31, 507 A.2d at 669 for a discussion of UNITIL); (6) availability factor of 60% (“availability factor” is “used synonymously” with “capacity factor,” Re PSNH, 66 PUR4th at 403 n.44); and (7) no phase-in of rate increases. Exhibit R-21D. Retail rates in cents per kilowatt hour were established by combining the power supply costs of Exhibit R-21A with the distribution, administrative, and general costs, which were derived by subtracting the power supply costs forecast in Exhibit R-21B from the retail rates forecast in Exhibit 47. See Commission Report No. 17,960, at 12. The resulting non-power supply costs were escalated at 5% per year to compensate for inflationary effects.
Immediately, however, we encounter error. The commission chose to use the power supply costs of Exhibit R-21A, even though the basis of analysis in the earlier Commission Report No. 17,638 was Exhibit R-21B, which projected higher costs. In the later report, the commission reasoned that the purpose of the analysis underlying the previous order was to “assess the maximum exposure of the [Coop] and its ratepayers, rather than [to project] the range of probable rates[,]” Commission Report No. 17,960, at 10, making it reasonable to choose the higher cost construction estimate. The commission concluded that for the purpose of investigating the reasonably probable range of rates, use of Exhibit R-21A was appropriate because of its more likely completion cost of $882 million to go as of January 1, 1985. Id. at 10-11.
This reasoning is open to serious objection, however. First it rests upon an erroneous reading of the exhibits. We have not been cited to any exhibits or testimony in this case indicating that the cost to go assumed in any '•ate forecast was $882 million as of January 1, 1985. To the contrary, the testimony and exhibits explicitly indicate that when they used the figure of $882 million, that figure was an estimate of the cost to go as of August 1984. Witness Frederick Anderson, the Coop’s Assistant Director of Budgets and Finance, presented two scenarios forecasting the cash cost to go for the Coop. Scenario 2, which forecast a need for additional borrowing authority of $32,235,645, relied on a direct completion cost of $882 million as of August 1984. Exhibit R-l. Furthermore, the KD-NU-R scenario upon which the Commission relied for the power supply costs component of its upper level rate forecast specifically states that it reflects construction costs of $882 million to go as of August 1, 1984. This scenario was the basis for Exhibit R-21A.
*146Moreover, the commission itself found that the cost to go was $882 million as of January 1, 1985, rather than August 1, 1984, as assumed in Exhibit R-21A. Because the commission thus concluded that the construction cost to go was greater than the cost assumed in Exhibit R-21A, it could not reasonably rely on the power costs forecast in that exhibit for its projection of the probable high end of the range of retail rates, and it was error to do so.
In fact, the construction cost estimate reflected in Exhibit R-21B appears to be more nearly consistent with the commission’s actual projection of cost of completion, and is also consistent with the amount of the commission’s final financing authorization. Mr. Anderson testified in accordance with this exhibit that based on an assumption of a direct cost of completing Unit I of $1 billion as of January 1, 1985, the Coop needed additional borrowing authority of $49,580,000. See Commission Report No. 17,638, at 121. The commission authorized additional borrowing in the amount of $46,898,000 (the initial request of $49,580,000 less emergency financing of $2,682,000 approved on May 10, 1985), which it found “reasonably requisite for present and future use to supply reliable electric service .. . .” Id. at 122. Exhibit R-21B thus incorporates into its power supply cost the total construction cost of Unit I which provided the basis for the calculation of the additional borrowing authority requested by the Coop and subsequently approved by the commission.
We find, therefore, that the commission miscalculated the probable upper range of rates by utilizing a completion cost that failed to correspond either to the assumptions present in the scenarios forecasting power supply costs or to the amount utilized in calculating the additional borrowing authority the commission granted. Whether this error demands that we remand the case to the commission for a third time depends on how significantly that error affects the projection of the high side of potential rate effect. After a careful consideration of the record, we have concluded that the upper range of rates can be determined consistently with the projection of cost to completion of $1 billion as of January 1, 1985, and with sufficient reliability to ascertain whether such rates are reasonable within the meaning of the statute. Therefore a further remand, though justifiably within our discretion, is not necessary, and we will proceed to correct the commission’s figures.
We rely upon Exhibit R-47. This exhibit incorporates the power cost projections of Exhibit R-21B, which in its KD-NU-R scenario relies on the following assumptions: (1) construction costs of $1.3 billion to go as of August 1984 (the commission reasonably found that a cost to go of $1.3 billion as of August 1984 is equivalent to a $1 *147billion cost to go as of January 1, 1985, Commission Report No. 17,960, at 10); (2) full dollar inclusion in the rate base of both the Coop’s and PSNH’s share of total project cost; (3) COD of October 1987; (4) full recovery of the cost of Unit II from ratepayers; (5) full loss of UNITIL load; (6) availability factor of 60%; (7) no phase-in of rate increases; and (8) the Coop’s exercise of its option to sell its share of Unit I power back to PSNH during the ten years of the sell-back agreement when its Unit I KWH costs are higher than purchased power costs. Exhibit R-47 then estimates retail rates by adding the Coop’s distribution, administrative, and general costs (escalated at 5% per year) to the power supply costs of Exhibit R-21B.
On these assumptions, taking 1985 as the base year, the rates forecast in nominal terms, i.e., in terms adjusted to include anticipated inflation, more than double by the end of 1988 (the first full year of Unit I operation). This reflects the assumption that there will not be a phase-in of rate increases. The largest one year increase in rates occurs in 1988, at which time the rates increase 6.92<P per kilowatt hour to 19.02<P per kilowatt hour. This is a 57.18% increase over the preceding year. The rates reach an initial peak in 1990 and then decline slightly during the remainder of the ten years of the sell-back agreement. When the Coop is required to use its share of owned Unit I power in 1997, the rates again begin to rise and reach their highest level in 2004, the last year of the forecast. At that time the rates will reach 30.25<F per kilowatt hour, which is 234.25% above the 1985 rate.
It is instructive to note the significant differences that result from projecting the rates in so-called real terms; i.e., accounting for the effects of the additions to the PSNH and Coop rate bases without any adjustment for inflation. In real terms, the largest one year increase also occurs in 1988, when the rates increase 5.46<P per kilowatt hour to 16.43<? per kilowatt hour, an increase of 49.77% over the preceding year. The highest rate forecast is 17.86<P per kilowatt hour in 1989 (as compared to a nominal high rate of 30.25<P per kilowatt hour in 2004), which is 97.35% above the 1985 rate. From 1990 through 2001, the retail price steadily declines. The end of the sell-back agreement is not shown to have any immediate impact, as the retail rates do not begin to rise until 2002, well after the termination of the sell-back agreement in October 1997.
A comparison of the rates forecast by the commission in Table 1 of Commission Report No. 17,960, at 15, and the rates forecast in Exhibit R-47, which rest upon a cost to go assumption closer to the commission’s actual assumption, shows that the greatest rate differ*148ence between the two forecasts (ignoring 1987) occurs in 1990 when the Exhibit R-47 forecast is 2.87<P per kilowatt hour higher. (In 1987 the rates in the commission’s order exceed the rates in Exhibit R-47 as a result of the differing assumptions of COD.) By 1992, the Exhibit R-47 rate forecast is less than 2<P per kilowatt hour higher than the commission’s forecast. This differential decreases to less than 1.5<P per kilowatt hour in 1994, and in 2003 it drops to less than 1<P per kilowatt hour.
In deciding whether these differences are so great as to require a remand, we have been mindful that some of the Exhibit R-47 assumptions may be unduly high. For example, the commission has never accepted a COD of October 1987, having found that a COD of December 1986 was reasonable. To the extent the commission’s finding proves to be a more accurate estimate, the effect will be to decrease the power supply costs, which will have the consequent effect of reducing rates. Similarly, RSA 378:30-a would preclude any recovery for Unit II investment. See Appeal of Public Service Company, 125 N.H. 46, 53-54, 480 A.2d 20, 25 (1984). Corresponding modifications in the Exhibit R-47 rate projections would further reduce the differences between the Exhibit R-47 projections and those of the commission. As our later discussion indicates, we believe that these differences in projections are not significant for the purpose of determining whether the actual future rate will fall within the range of what is reasonable. See RSA 378:27, :28. Accordingly, we choose not to remand the case to the commission.
Turning now to the lower end of the probable rate range, the commission claimed to be unable to make such a determination, arguing that any projection of rate base exclusions for the Unit I investment prior to prudency hearings before itself and FERC would be entirely speculative. Commission Report No. 17,960, at 20. To the extent, however, that future rates will reflect the cost of wholesale power purchased from PSNH, the commission found that it could determine a lower limit of rates reflecting the minimum purchased power costs consistent with the survival of PSNH, the Coop’s principal supplier. To do this, the commission relied on the low-end, no phase-in rate projections that it made in docket DF 84-200. See Commission Report No. 17, 960, at 20-25; see also Appeal of CLF, 127 N.H. at 643-44, 507 A.2d at 678.
On the assumption that the rates set by FERC and the commission will be sufficient to raise enough revenue to avoid forcing either PSNH or the Coop into default, the commission developed a ratio which identified the effect on PSNH rates caused by the rate base exclusion considered in docket DF 84-200. This ratio was then applied to the Coop’s cost of purchased power from PSNH, taken from *149Exhibit R-21A, producing a purchased power cost discounted by the PSNH rate base exclusion. After adding the cost of power supplied by other producers and its Unit I ownership interest, as well as the administrative and distribution costs, the commission arrived at an estimate of the lower end of the Coop’s retail rates. Commission Report No. 17,960, at 20-23.
Before summarizing the low end rate calculations, we must again deal with an error which in this instance occurred first in Table 6, listing the low end of the range of retail rates. Commission Report No. 17,960, at 28. One of the assumptions underlying the Table 6 calculations was that the owners would recover Unit II costs. But see RSA 378:30-a. A cost figure for Unit II recovery was mistakenly included in the “Other Power Costs” column for 1985, even though repayment for Unit II was actually assumed to begin in 1986. The “Total Power Costs,” “Power Costs Cents/KWH,” and the “Retail Rate Cents/KWH” columns consequently were miscalculated as a result of the initial error, and this inaccuracy was reflected in Tables 4 and 5 where the 1985 retail rate was used as the base year for calculating percentage increase and real dollar rates. The figures in this opinion have been recalculated to correct this error.
In nominal terms, the greatest one year increase occurs in 1987, when the rates increase 5.20<P per kilowatt hour to 16.03<P per kilowatt hour, a percentage increase of 48.01% over the preceding year. The rates remain constant, with minor variations, through the end of the sell-back agreement in 1997, at which time they begin to rise steadily, reaching a peak in 2004. In this final year of the forecast, the rate forecast is 27.57$ per kilowatt hour, a percentage increase of 203.30% over the 1985 rate.
Again we note that in real, non-inflated terms, the largest one year increase in rates occurs in 1987 when the rates increase 4.23<P per kilowatt hour to 14.54<P per kilowatt hour, a 41.02% increase over the preceding year. The highest rate forecast is 14.88? per kilowatt hour in 1988 (as against a nominal high rate of 27.57<P per kilowatt hour in 2004), which is 63.69% above the 1985 rates. The rates then steadily decline during the course of the sell-back agreement through 1997. In 1998, the rates increase approximately 2<P per kilowatt hour and remain constant for the remainder of the forecast.
D. Reasonableness of Rates Within the Projected Range-
We now consider the challenge to the commission’s further conclusion that a rate set within the projected range will be consistent with the reasonable rate standard. In evaluating the sufficiency of the commission’s treatment of this issue, two points are significant.
*150First, a comparison of rate projections in this case with the projections considered in Appeal of CLF reveals that as a general, if not invariable, rule, the high and low rates projected for the Coop are higher than the corresponding rates projected for PSNH. Standing alone, this is not remarkable. The Coop’s rates are entirely retail, whereas PSNH’s rates include a wholesale component. In 1985, without any reflection of Unit I ownership, the Coop’s rates are at least .55<P per kilowatt hour higher than PSNH’s.
After Unit I operation commences, however, and Unit I ownership interests affect rates, the differential probably will be greater. Commissioner Aeschliman has made the comparison and provided a useful explanation: “the [Coop’s] rates closely track PSNH’s during the period of the [sell-back], as would be expected, but are relatively higher once the [Coop] takes its Seabrook share directly and purchases the balance of its power from PSNH. The apparent reason for this is that when the [Coop] takes its Seabrook share directly, Seabrook power is more heavily weighted in the [Coop’s] power mix than it is in PSNH’s mix of power.” Commission Report No. 17,638, at 5 (Commissioner Aeschliman, dissenting in part). Thus, an increased spread between PSNH and Coop rates may be anticipated as a result of the latter’s 2.17% ownership share in Unit I.
The second significant point in considering the sufficiency of the commission’s consideration of rate effect is that the projection of a range within which the Coop’s rates will be set rests entirely on the possible range of exclusions from PSNH’s rate base considered in docket DF 84-200. In dealing with the Coop’s own rate base as reflecting its investment in the 2.17% share of Unit I, the commission has assumed, without discussion, that there will be full dollar inclusion of the Coop’s entire Unit I investment. Despite the commission’s silence, the reason for this assumption is readily apparent from the Coop’s capital structure. As we have noted, all outside investment in the Coop is debt capital, and virtually any loss of revenue resulting from rate base exclusion must fall upon the lenders. While Exhibit R-9 indicates that as of December 31, 1984, the Coop has a $6.7 million surplus, this would be insufficient to provide a cushion for any significant rate base exclusion. In fact, it is questionable whether the surplus could provide any mitigation for a rate base exclusion of investment in Unit I, when we consider the $11 million investment attributable to Unit II that will be irrecoverable through rates under the terms of RSA 378:30-a if Unit II is not completed.
Accordingly, we read the commission’s decision to reflect an understanding that the Coop must have full dollar rate base inclu*151sion of its Unit I investment, if it is to avoid default to its lenders and the possibility of bankruptcy. The issue before us is, therefore, whether the present record sustains a finding that rates resulting in part from full dollar inclusion of the Coop’s investment will be reasonable within the meaning of the statute. We conclude that the record is sufficient.
The appellants’ argument to the contrary rests on their claim that the commission gave insufficient attention to the effects of default and bankruptcy as alternatives to full rate base inclusion of the Coop’s Unit I investment. At first sight this claim may appear to have merit in light of our considered dictum in Appeal of CLF that the commission must consider the desirability of bankruptcy when increased financial demands leave a utility with bankruptcy as the sole alternative to full dollar recognition of arguably excludable elements of its revenue requirement. See Appeal of CLF, 127 N.H. at 625, 507 A.2d at 664-65.
In Appeal of CLF, however, we held that the commission must consider bankruptcy as an alternative only if corporate survival required that the full investment be included in the rate base, even though some of that investment was potentially excludable under the principles of prudence and usefulness. If, on the contrary, it could reasonably be assumed in advance that the investment to be financed would receive full rate base inclusion, and would be commercially feasible, our statements in Appeal of CLF would not require consideration of a bankruptcy alternative. In fact, the commission has made just such an assumption here. Hence, unless the appellants have demonstrated that it was unreasonable to assume that no part of Unit I investment would be excluded from rate base, the commission was not required to consider bankruptcy as an alternative to the financing, even if corporate survival requires that the investment receive full rate base inclusion.
The appellants have made no such demonstration, and we accordingly find no reversible error in the commission’s assumption that the Coop will be entitled to full rate base inclusion of the investment resulting from the present financing. The appellants’ failure to demonstrate error is apparent from a consideration of the way in which customary principles of rate base limitation might be applied to the anticipated investment in question.
Looking first at the possible applications of the used and useful standard, traditional examples provide no precedent for possible rate base exclusion of any of the Coop’s Unit I investment. It is undisputed that the Coop will have a demand at least as great as its 25 mw entitlement to Unit I power, and thus none of the Coop’s *152power-generating capacity could be excluded as excess capacity in the traditional sense. Whether or not the availability of adequate power from wholesale purchases could provide any opportunity for a “pragmatic” application of the used and useful standard to exclude generating capacity is an issue that has not been raised, and there is no claim or indication that the commission erred in assuming that there would be no exclusion on the usefulness principle.
The possible application of the prudence standard presents issues of greater complexity, however. Initially, there does not appear to be any basis to attribute any imprudence to the Coop for expenditures by PSNH prior to the Coop’s purchase of the 2.17% share in 1981. While the amount paid for that share could be the subject of an imprudence claim, that is not in fact a serious possibility. The Coop acquired its share free of any AFUDC then attributable to it, and the commission approved both the purchase price and the financing originally peeded. Indeed, there was evidence that the then governor publicly endorsed the Coop’s purchase, and it appears that the commission encouraged the Coop to acquire its ownership interest once it became evident that PSNH could not support its original ownership share in the face of rising costs. See Re New Hampshire Electric Cooperative, Inc., 66 N.H.P.U.C. 139 (1981); Re Public Service Company of New Hampshire, 64 N.H.P.U.C. 262, 265 (1979). Although we do not hold that the commission’s action has a legally preclusive effect, we do face reality when we recognize that with this history a serious question about the prudence of the Coop’s purchase is unlikely.
Although it does not necessarily follow from this that the Coop’s responsibility for the prudence of expenditures made after the 1981 purchase may not be questioned, we believe that Commissioner Aeschliman is correct in observing as a practical matter that the Coop is in a significantly different position from PSNH even as to this period. See Commission Report No. 17,638, at 1-3 (Commissioner Aeschliman, dissenting in part). We must face the fact again that the commission emphatically encouraged the Coop’s acquisition of a fractional interest in an increasingly expensive project, without any substantial ability to manage the project or otherwise to control expenses. It would simply be unrealistic to expect the same commission to place blame on the Coop for failing to control construction expenditures during this period. For practical purposes, the Coop stands in a different position from that of PSNH, both in its relationship to the project and in its relationship to the commission.
Last, with respect to the period from the commission’s present review to the expected completion of the project, the record *153contains substantial evidence that the level of construction expenditure has been brought under control, and the appellants have not carried their burden to demonstrate that the commission erred in apparently presupposing that prudence would not be a serious issue. We therefore believe that Commissioner Aeschliman was correct to conclude that rate base exclusions on principles of prudence and usefulness are practically out of the question. We hold that the appellants have not demonstrated error in the commission’s anticipation that none of the Coop’s Unit I investment will be excludable from the rate base.
Thus, it is apparent that the commission did not assume that rate base inclusion of otherwise excludable elements of Unit I investment would be required in order to avoid the Coop’s insolvency and possible bankruptcy. It appears, rather, that the commission correctly declined to consider bankruptcy because it could foresee no potential need to choose between bankruptcy and rate base inclusions that would be potentially improper on traditional rate-making principles. This was not a violation of the standards that we discussed in Appeal of CLF, 127 N.H. at 633-40, 507 A.2d 671-75.
For clarity’s sake, however, it is worth noting the implication of any claim that in these circumstances the commission should nonetheless have considered the Coop’s insolvency and possible bankruptcy as alternatives to its participation in the remaining Unit I construction, or to full rate base inclusion of investment from this financing if Unit I is completed. Such a claim implies that the commission should consider forcing a utility into bankruptcy for no other reason than to extricate the utility and its ratepayers from the consequences of a costly agreement, even though the commission finds that the consequences are commercially feasible. A utility’s contracts thus would be subject to nullification by commission action without reference either to market feasibility or to the traditional prudence or usefulness principles of ratemaking equity, and the commission could place the risk of any subsequent dissatisfaction with a corporate agreement entirely upon the Coop’s lenders. It is sufficient to note in this case that the appellants have made no such explicit argument, and no issue has been raised about the merits of such a position.
In concluding, we recognize that the Coop’s ratepayers may be required to bear the burden of higher rates than they would pay without the Coop’s ownership of a share of Unit I. As Commissioner Aeschliman observed, “[c]ertainly it was not contemplated that the [Coop] would be disadvantaged as a result of its Seabrook purchase in comparison to continuing to purchase [90% of] its power wholesale *154from PSNH .... And yet that is precisely what is in danger of happening.” Commission Report No. 17,638, at 4-5 (Commissioner Aeschliman, dissenting in part). In view of this possibility, it is perhaps worth noting that Coop ratepayers have one mechanism to influence management decisions affecting rates, which ratepayers of investor-owned utilities do not have. Unlike a PSNH customer, a Coop ratepayer has a vote in electing directors, who in turn select officers. See RSA 301:16. The ratepayers of the Coop elected the management that chose to purchase the ownership share in Unit I. Although that choice may prove to be unnecessarily costly, the Coop ratepayers are not without some responsibility for it.
Affirmed.
Brock, and Batchelder, JJ., did not sit; Smith and Dickson, JJ., superior court justices, sat by special assignment under RSA 490:3; King, C.J., dissented.King, C.J., dissenting: This case, like much of the litigation engendered by the Seabrook project, attests to the importance of commission review of management decisions relating to capitalization and indebtedness of a utility. The broad powers of oversight, see RSA chs. 369 and 378, and investigation, see RSA 365:5, :19, conferred by the legislature upon the PUC enable and require the commission to protect the utility ratepayers of this State, regardless of their ability or willingness to act on their own behalf.
The Coop’s financing proposal presented the PUC with two alternatives. It could approve the financing to permit continued participation in Seabrook, or it could reject the proposal, thereby precluding further Coop involvement. I believe the commission has failed to fully assess the latter option and would remand for an investigation of the consequences of bankruptcy, as required by Appeal of Seacoast Anti-Pollution League, 125 N.H. 708, 718, 490 A.2d 1329, 1336 (1984) and Appeal of Conservation Law Foundation of New England, Inc., 127 N.H. 606, 624-25, 507 A.2d 652, 664-65 (1986).
I. History of Coop Involvement in Seabrook
At the time the Coop purchased its interest in Seabrook, the PUC did not make a detailed inquiry into the proper level of investment or ownership for the Coop. PSNH had applied to divest itself of 22 per cent ownership in Seabrook. The PUC examined PSNH’s revenue sources and found that due to a disparity in rates, PSNH’s retail customers were subsidizing its wholesale customers, including the Coop. The commission indicated that PSNH could ameliorate this inequitable situation by allowing wholesale customers to own a *155portion of Seabrook and satisfy their own generation needs. This approach, according to the PUC, was “extremely important given that the [Coop] could avail itself of less expensive financing through [Rural Electrification Administration (REA)] financing than can PSNH. . . .” Re Public Service Company of New Hampshire, 64 N.H.P.U.C. 262, 265 (1979).
In 1981, the Coop petitioned for authority to borrow $75,750,000 through the REA to purchase its Seabrook interest and fund its share of future construction. Intervenors requested that the commission complete “an analysis of demand and how best to meet that demand,” seek a vote of the Coop’s membership, and make an “examination of other energy options.” In response, the PUC stated that it had already
“found that ownership by the cooperative was superior to ownership by PSNH due to the cooperative’s ability to avail itself of lower cost REA financing. (64 NHPUC 262, 265.) [This ‘finding’ is quoted in full supra.] In fact the record in this proceeding reveals that the rate the cooperative can receive from the REA is almost half the rate paid by PSNH to do comparable financing. Clearly, the public good as well as the purpose of this financing has been found by this commission to lie in the cooperative having an ownership interest in Seabrook.
Recently, the Maine and Massachusetts commissions have conducted studies which reveal that investment in Seabrook is superior to initiating construction of other base load plants and better than continued reliance upon oil. This action outside our borders lends further support to our findings.”
Re New Hampshire Electric Cooperative, Inc., 66 N.H.P.U.C. 139, 140 (1981).
The intervenors also challenged the cost and amount of the financing. The PUC responded:
“[n]o party disputes the testimony submitted by the cooperative which states that the cooperative can finance a given level of Seabrook at a much lower rate than it can buy it from PSNH. The cooperative does not have to pay a return to stockholders, income taxes, or debt service in excess of the prime. Therefore, the commission finds the cost rate to be in the public good.”
Re NHEC, 66 N.H.P.U.C. at 140.
Finally, the intervenors posited that the amount of financing *156requested might not be sufficient to cover the Coop’s responsibility for a 2.17% ownership interest, suggesting that the escalating cost of Seabrook might necessitate further financing. According to the PUC, the finding of public good was not negated by this possibility:
“[t]he standard for financings can never guarantee protection from all factors. No one has a crystal ball to determine the actual final cost of any plant. However, this $75,750,000 is a reasonable level at this time and at a cost rate significantly below that of any other New England utility.”
Re NHEC, 66 N.H.P.U.C. at 140.
The above-quoted passages constitute the entirety of the commission’s findings on the advisability of the Coop’s involvement in Sea-brook. Seizing upon the attractive circumstance of the Coop’s access to low-cost financing, the PUC seemingly ignored another factor that counseled greater caution. Direct ownership of Seabrook power was expected to meet the Coop’s power requirements at lower cost than if the Coop purchased an equivalent amount of wholesale power from PSNH. Yet Seabrook also was intended to satisfy a greater proportion of the Coop’s total power needs than it would for PSNH. The significance of the difference in power mix is that any escalation in Seabrook costs was bound to have a greater effect on the Coop’s rates than on PSNH’s. By taking a direct ownership interest, the Coop lost the advantage of PSNH’s power mix, which would have mollified the impact of the high cost of Seabrook on Coop rates.
The Coop originally might have arranged for a sell-back agreement covering the full term of Seabrook power production. Furthermore, it might have ensured that its 2.17319% interest would be obtained at a specified maximum cost, so that any escalation in cost would not require further contribution by the Coop to maintain its ownership interest. The PUC instead contemplated that the Coop would provide low-cost financing for Seabrook, alleviate PSNH’s financial difficulties, and at the same time obtain a portion of its power supply directly, at lower cost. This “no losers” plan has miscarried, apparently because of the project cost escalations envisioned by the intervenors in 1981. The PUC now estimates that when Seabrook comes on line, the Coop’s retail rates will be even higher than PSNH’s.
II. Bankruptcy
Although the Coop’s options are limited by its contracts with the Seabrook joint owners and the REA, the PUC’s responsibility to *157assess the propriety of continued participation was in no way diminished. Consideration of bankruptcy was imperative, not because the Coop might be extricated from its “bad deal,” but to determine “whether, under all the circumstances, the financing is in the public good.” Appeal of Easton, 125 N.H. 205, 213, 480 A.2d 88, 91 (1984).
The commission should be required to evaluate the effects of bankruptcy where, as here, bankruptcy may result from denial of the requested financing. See Appeal of CLF, 127 N.H. at 670, 507 A.2d at 695-96 (King, C.J., and Batchelder, J., dissenting). The commission undertook this task by performing a “net present value” analysis of the “aggregate benefits of Seabrook participation versus cancellation.” Commission Report and Seventeenth Supplemental Order No. 17,638, at 123-29 (May 31,1985). It was understood by all parties that failure of the Coop to continue funding its share of Sea-brook construction would constitute a default on its obligations to the joint owners. Accordingly, the PUC endeavored to calculate the potential financial consequences of default under the joint ownership agreement, which contemplates assessment of a penalty and purchase of the Coop’s Seabrook interest by the other joint owners. The commission then determined the range of the Coop’s potential liability to the Federal Financing Bank (FFB), assuming application of the proceeds of the sale to the FFB debt.
Nonetheless, this analysis was incomplete, because it failed to consider that bankruptcy would most likely follow a default to the joint owners. Under all cancellation scenarios, the Coop would owe a significant sum to the FFB. Although the PUC apparently assumes that the debt would find inclusion in the Coop’s rate base, the record contains no findings or discussion that reconcile this assumption and the language of the utilities statutes. Under RSA 378:27, “rates shall be sufficient to yield not less than a reasonable return on the cost of the property of the utility used and useful in the public service” (emphasis added). The PUC assumes that upon default, the remaining owners will purchase the Coop’s interest. Hence, the Coop would have a debt arising out of a former, rather than a present, property interest, which could not properly be included in rate base.
The commission should have discussed the basis for its assumption. In the absence of a new interpretation, it appears that the rate-setting statutes have no provision for recovery of a debt of this type from ratepayers. Accordingly, the commission’s net present value analysis should have included an evaluation of the effects of bankruptcy on the Coop’s obligations. Denial of the proposed financing would precipitate a default on the Coop’s obligations to the FFB because there would be no property to include in rate base to raise the revenue necessary to discharge the FFB debt. It is reasonable to *158conclude that bankruptcy would follow. Thus, in order to evaluate the likely consequences of denying the financing, the PUC should have considered both default and bankruptcy.
Even under the standard adopted by the majority in Appeal of CLF, this case required some investigation of the effects of reorganization. In Appeal of CLF, the court stated:
“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.”
Appeal of CLF, 127 N.H. at 625, 507 A.2d at 665.
The rationale for this rule is that where full investment must be included in rate base, there can be no “genuine opportunity to set rates consistent both with the principles [of prudence and usefulness] and with the solvency of the utility.” Ordinarily, a utility may include in the rate base only the prudently incurred cost of property used and useful in the public service. See RSA 378:27, :28. Full cost recovery implies rates set in derogation of this statutory mandate if any rate base exclusion is appropriate. When the capitalization of the utility can be supported only by full dollar recovery, the commission is required to evaluate at the financing stage whether charging ratepayers for imprudently incurred costs or the cost of property not used and useful would be preferable to the effects of bankruptcy. If bankruptcy were found preferable, the commission could deny the financing. This approach prevents the economic waste that would occur if the financing were approved but an evaluation of interests at the rate-setting stage mandated utility bankruptcy.
The commission assumed that the Coop would require full dollar inclusion of investment in its rate base to avoid default; thus, it should have estimated the probable amount of allowable exclusions under principles of prudence and usefulness. See Appeal of CLF, 127 N.H. at 642, 507 A.2d at 676. Using this figure for comparison, the PUC ought to have determined whether the rate resulting from full dollar inclusion would be reasonable as the only alternative to default and bankruptcy. The commission did not estimate the amount of possible exclusions, nor did it investigate the effects of bankruptcy. I would remand for examination of these matters.
*159According to the majority, the PUC was not required to perform this analysis because no basis exists for later exclusions: “it could be assumed in advance that the investment to be financed would receive full rate base inclusion.” The court’s analysis of this issue is procedurally improper. Whether property is “used and useful” and whether costs were prudently incurred are decisions to be made in the first instance by the PUC. Contrary to the majority’s assertion that the commission assumed full rate base inclusion, the commission explicitly stated that it “did not in this proceeding engage in an assessment of how much, if any, of the [Coop’s] investment in Sea-brook [Unit] I was or will be prudently incurred.” Commission Report and Twenty-Second Supplemental Order No. 17,960, at 41 (November 22,1985). The commission apparently heard no evidence on the issue. No findings or conclusions as to the probable amount of exclusions appear in the record.
The court’s analysis is inappropriate for another reason. Traditional application of the “usefulness” concept may not point to a likely exclusion. But, as this court has observed, “[i]n 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.” Appeal of CLF, 127 N.H. at 647, 507 A.2d at 680. The court ought to have given the PUC an opportunity to apply its expertise in this area.
With respect to prudence, it may be true that few areas of the Coop’s involvement in Seabrook lend themselves to review. But a prediction of the likely result of a prudence investigation cannot rest on the fact that the PUC previously approved the Coop’s actions. In a prudence review, the PUC takes a fresh look at what was done in the past. New Hampshire’s statutory scheme obligates a utility to get commission approval to finance plant construction; this approval does not preclude a later finding of imprudence with respect to management’s application of the funds, nor does it imply that no imprudence will be found. The PUC must investigate prudence issues even though it evaluated the project for the purposes of financing. Moreover, a prediction of the likely amount of disallowance cannot be based on “political realities.” If the PUC were to make its decision on an improper basis, it would be this court’s duty to ensure that political realities yield to statutory compliance.
The fact that Coop management, by virtue of the Coop’s small percentage interest in Seabrook, could not control project management decisions does not preclude a prudence disallowance for Coop ratepayers. The imprudence of PSNH management, if any, will not affect PSNH customers alone. Hence, if the PUC makes a prudence disallowance against PSNH, it could find that Coop customers are *160entitled to a disallowance for the same cause in an amount proportional to the Coop’s direct ownership share. It is irrelevant that the Coop’s managers were not the controlling decision-makers, because prudence disallowances are not intended to punish any particular management team; the disallowance simply reflects the consumer’s statutory right not to pay for imprudent expenditures connected with a project. See RSA 378:27, :28. In addition, there is no indication that the joint ownership agreement permits the Coop to disavow decisions of the joint owners simply because it was not in the majority who voted for the action.
Finally, I take issue with the court’s statement that, “with respect to the period from the commission’s present review to the expected completion of the project, the record contains substantial evidence that the level of construction expenditure has been brought under control, and the appellants have not carried their burden to demonstrate that the commission erred in apparently presupposing that prudence would not be a serious issue.” The PUC did not consider the issue, but rather assumed that the determination would be made at a later time. On this record, the court cannot conclude as a matter of law that there will be no disallowance for imprudence, nor can it infer to a near factual certainty that this result will obtain. Moreover, the court incorrectly places the burden on the intervenors to prove that there will be some imprudent expenditures. “[R]ate-making or prudency proceedings are distinct both in time and in objectives from Easton [financing] proceedings.” Appeal of CLF, 127 N.H. at 640, 507 A.2d at 675. Only in the context of a properly-noticed hearing, when the expenditures under review have actually been made, do intervenors have to prove that an investment was imprudent. No statute or case imposes a burden of proof on intervenors in a financing proceeding to establish future imprudence. At this stage, the PUC simply should have estimated the probable amount of exclusions.
Because some of the investment is arguably excludable, the commission was required to consider the relative desirability of rates based on full inclusion of investment, on the one hand, and bankruptcy, on the other. The majority has indicated that this determination involves a “balancing” of the “competing interests” of investors, who “would like a guaranteed return on any investment” and customers, who “would like low rates” (emphasis added). Our statutes do not support this characterization. Rather, the consumers’ interest is the consumers’ statutory right not to pay for imprudent expenditures or utility property not used and useful. See RSA 378:27, :28. This interest is protected by including in rate base only *161that amount of investment prudently expended for used and useful property.
Likewise, investors have no legitimate expectation in a return on imprudently expended investment or property not used and useful. (This proposition holds true whether the investors are equity shareholders or debt investors: the difference between them lies not in their expectation of return on investment but in their right to a remedy upon the utility’s failure to give them that return.) The PUC gives proper recognition to the investors’ interest by allowing full inclusion in rate base of prudent investment in used and useful property.
In the ordinary situation, then, the consumers’ and investors’ interests neither compete nor need to be balanced because our rate-setting statutes specify what each party is due. Nonetheless, when some investment is excludable but full cost inclusion in rate base is necessary for utility solvency, the commission may be justified in setting rates based on full cost. In that situation, unless the consumers’ legitimate expectation in not paying for excludable costs is compromised, the utility will go bankrupt. Still, the interests cannot be said to “compete” because investors have no legitimate expectation of a return on excludable investment.
To decide whether the rates resulting from full inclusion are “reasonable” as the alternative to bankruptcy, the PUC must consider what benefits ratepayers would receive for compromising their legitimate expectations. If bankruptcy would result in higher rates or inadequate service, consumers would benefit by paying for excludable costs. In this situation, the PUC would be justified in compromising the consumers’ interest. If rates and service would be comparable under either course of action, the PUC might properly choose to avoid bankruptcy, on the rationale that if the consumer is not affected, it is reasonable to consider what is best for the investor. But if consumers would be benefited by bankruptcy, no reason would exist for requiring them to compromise their legitimate expectation in not paying for excludable costs.
The rate-setting statute, RSA 378:27, states that “rates shall be sufficient to yield not less than a reasonable return on the cost of the property of the utility used and useful in the public service . . . .” There is no statutory limitation prohibiting the PUC from making an otherwise permissible disallowance simply because recognition of the ratepayers’ interest may result in default. Nor must rates be set to enable a utility to stay solvent when the interests of utility customers are best served by reorganization, and the investors’ interest is not abridged. Cf. Appeal of CLF, 127 N.H. at 635-36, 507 A.2d at 672.
*162Although the high cost of Seabrook, the penalties of default and the infeasibility of power supply alternatives limit the Coop’s choices at this juncture, the commission was required to evaluate all possibilities presented by this financing, including bankruptcy. The majority opinion concludes that “[t]he ratepayers of the Coop elected the management that chose to purchase the ownership share of Unit I. Although that choice may prove to be unnecessarily costly, the Coop ratepayers are not without some responsibility for it.” I would place the greatest responsibility on the commission.