Souter, J.
In this appeal from a rate order of the public utilities commission, the utility assigns error in setting the rate of return on common equity capital, and in recognizing an addition to the accumulated deferred tax account by a greater amount with respect to the test year than the total deferred taxes projected annually for collection from ratepayers. We affirm.
On May 29, 1986, Public Service Company of New Hampshire began proceedings before the public utilities commission under RSA chapter 378, to obtain authorization for increasing rates charged to customers for electricity. Although the rates originally proposed would have produced an annual revenue increase expected to be nearly $59 million, the enactment of the Tax Reform [750]*750Act of 1986, Pub. L. No. 99-514, 100 Stat. 2085, prompted the company to reduce its request to provide for an increase of something under $39 million.
After hearings were completed on June 29, 1987, the commission issued order No. 18,726 in its docket DR 86-122, authorizing rates sufficient to generate an additional $20.5 million to be collected by billing customers during a period beginning July 1, 1987. The commission issued a further report and order No. 18,775 on the company’s motion for rehearing under RSA 541:3, whereupon this appeal followed under RSA 541:6. Between the filing of the appeal and argument, the commission transferred questions of law that had arisen on a subsequent and independent request for rate increases, to which we responded by sustaining the constitutionality of the anti-CWIP statute, RSA 378:30-a, as reported in Petition of Public Service Co. of New Hampshire, 130 N.H. 265, 539 A.2d 263 (1988). Following this decision, the company filed a petition in the United States District Court for the District of New Hampshire seeking protection under the bankruptcy act, 11 U.S.C. § 1101 et seq.
The company’s burden in this appeal is imposed by RSA. 541:13, under which we “will not sustain [the] appeal, except for errors of law, unless [the utility] demonstrates by a clear preponderance of the evidence that the [commission’s] decision was unjust or unreasonable or reflects an abuse of [the commission’s] discretion.” Appeal of Granite State Electric Co., 124 N.H. 144, 146, 467 A.2d 252, 253 (1983). The company’s efforts to carry this familiar burden can be evaluated best in light of some equally familiar standards of public utility ratemaking.
In setting rates, a regulatory commission follows a process of identifying consumer and producer interests competing for recognition, with an ultimate goal of striking a fair balance or accommodation between them, to be reflected in charges to customers that may be described as just and reasonable both to the customer and to the utility. See RSA 378:27, :28; Appeal of Conservation Law Foundation, 127 N.H. 606, 633, 638, 507 A.2d 652, 671, 674 (1986). Because there is no one single point at which the balance must necessarily be struck in order to achieve a just accommodation, it is said, in conclusory terms, that a utility’s charges to customers are appropriate if they fall within a zone of reasonableness between the extremes of confiscating a utility’s property, at one end, and exploiting customers for the utility’s benefit, at the other. See Permian Basin Area Rate Cases, 390 U.S. [751]*751747, 770 (1968); Petition of Public Serv. Co. of N.H., supra at 274, 539 A.2d at 268.
The conceptual structure in which a regulatory commission attempts to achieve this result equates the total revenue to be raised by rates charged to customers with the sum of approved operating expenses plus an amount computed by applying a percentage rate of return to the depreciated value of the company’s rate base, comprising its investment in property used and useful in the production of energy for its customers. See RSA 378:27; Appeal of Conservation Law Foundation, supra at 633-34, 507 A.2d at 671.
Of the formula’s three variables, operating expense, value of rate base and rate of return, only the last is a matter of concern in dealing with the first issue in this appeal. The objectives of setting a reasonable rate of return on a utility’s rate base, see RSA 378:27, :28, include compensating the company’s investors for the risks they assume when they lend to the company and buy its stock. See, e.g., Power Comm’n v. Hope Gas Co., 320 U.S. 591, 603 (1944); Petition of Public Serv. Co. of N.H., 130 N.H. at 275, 539 A.2d at 269. The anti-CWIP statute, for example, places the entire risk of loss from an uncompleted plant on the company’s investors, Petition of Public Serv. Co. of N.H., 125 N.H. 595, 484 A.2d 1139 (1984), and the same is true when a plant has been completed but never placed in operation, Petition of Public Serv. Co. of N.H., 130 N.H. at 276, 539 A.2d at 267-68. The “constitutional consequence of this type of risk allocation is that those who bear the risk must be compensated by a return on their investment that reflects the risk that the statute places upon them.” Id. at 275, 539 A.2d at 269.
This general standard has been reflected over the years in the rule that a utility’s threshold entitlement is to a rate of return equal to the cost of capital. See, e.g., Power Comm’n v. Hope Gas Co., supra at 603; Bluefield Co. v. Pub. Serv. Comm., 262 U.S. 679, 692 (1923); Appeal of Conservation Law Foundation, 127 N.H. at 635, 507 A.2d at 672; Company v. State, 95 N.H. 353, 361, 64 A.2d 9, 16 (1949). The cost of capital is understood to be what “a utility must receive to maintain its credit, to pay a return to the owners of the enterprise, and to insure the attraction of capital in amounts adequate to meet future needs.” C. Phillips, Jr., The Regulation of Public Utilities 346 (1985); see also S. W. Tel. Co. v. Pub. Serv. Comm., 262 U.S. 276, 306 (1923) (Brandeis, J., dissenting); New Eng. Tel. & Tel. Co. v. State, 104 N.H. 229, 234, 183 A.2d 237, 241 (1962). Any overall capital cost must necessarily be a weighted average of the costs of the various debt and equity components of [752]*752the utility’s capital structure, see Appeal of Conservation Law Foundation, supra at 635, 507 A.2d at 672, with the consequence that in a rate proceeding the cost of each such component must be separately identified. It is in focusing on the cost of the company’s common equity capital that the first issue arises in this appeal.
Three witnesses addressed the cost of common equity as of the time of the rate hearing. The commission staff presented Dr. Sarah Voll, who employed a discounted cash flow (DCF) methodology to compute the cost of common equity for thirty-five utilities, about half of which were constructing or had completed nuclear plants. Based on these computations, Voll’s opinion was that the appropriate cost of common equity might range from 11.63% to 13%, with the higher rate applicable to utilities building or employing nuclear powered generators. She also testified about the character of the company’s stock as an investment, and her judgment was that the company “can be characterized as a speculative venture as long as reputable investment services such as Value Line contain statements [as of September 26, 1986] like ‘This stock is best avoided by all but the most speculative of accounts.’” She stated flatly that “[investors who purchase Public Service’s stock in hopes of realizing a substantial capital gain are in fact speculating.”
The company presented two witnesses on the issue, Robert Rosenberg and J. Peter Williamson, who agreed that the DCF methodology could not be applied directly to value the company’s stock, since the company was paying no dividend. Rosenberg applied DCF to thirty utilities with nuclear plants under construction, from which he derived an average common equity cost of 13.3%. He then compared the risks represented by the sample stocks with the company’s risks, which he found higher for reasons that included uncertainty over the outcome of Seabrook licensing proceedings. He concluded that the higher risk indicated a real equity cost of 19% for the company. Williamson reached the same result based on inferences drawn from comparing yields on the company’s capital instruments with yields on a broad market sample.
The commission accepted neither Voll’s 13% benchmark nor Rosenberg’s and Williamson’s 19%. It apparently viewed Voll’s upper limit as inadequate to reflect the company’s risks, and accordingly added a premium equal to the difference between the upper limit of common equity cost revealed by Rosenberg’s market sample, i.e., 13.3%, and the figure of 11.43%, identified as the average common equity cost computed by the Federal Energy Regulatory Commission (FERC) on a sample of diverse companies, [753]*753including those with nuclear plants under construction. The commission rounded this difference of 1.87% up to an even 2%, which it added to Voll’s upper limit of 13%, and thus set a rate of return on common equity of 15%.
The company contested this figure without success in its motion for rehearing. While much of the commission’s response to the company’s challenge had to do with articulating constitutional ratemaking requirements, which we will address below, two of the commission’s factual conclusions deserve attention.
“The Commission finds that the 19 percent return on equity [proposed by Rosenberg and Williamson] is one reasonable measure compensating the investor for all risks inherent in [the company]. The Commission further finds that the 19 percent return on equity is a level of return that is generally out of line with a utility type of business and is the type of return of a highly profitable enterprise or speculative venture.”
The commission proceeded to explain the difference between the normal utility’s return and the 19% level by reference to the uncertainty that the Seabrook plant would ever operate and the corresponding doubt that the company would ever recover any of its Seabrook investment, equal to four times the value of its existing rate base.
Although the company argues here as well that the order setting the common equity return at 15% is reversible for an error of law, identifying the error as the company conceives it requires some care, for there are three separable strands in the line of the company’s thinking. First, the company argues that the commission failed to consider those components of risk that were peculiar to the company, that is, those resulting from the Seabrook investment, which have impelled the company into bankruptcy since the date of the commission’s order. Second, the company argues that the commission has patently failed in its duty to articulate a methodology and findings to support the specific result it reached. The company notes that fulfillment of these obligations of administrative law are essential for any significant judicial review of rate orders, which depend for their validity upon the integrity of the reasoning process by which a series of relevant competing factors are evaluated in relation to each other. See, e.g., Permian Basin Area, Rate Cases, 390 U.S. at 792; Company v. State, 95 N.H. at 357, 359, 64 A.2d at 14, 15. Finally, the company argues that on the findings and record before us, as a matter of law the only supportable result is a return on common equity of 19%. The [754]*754company has expressed this last view through its prayer that the case be remanded “with instructions that the rates allowed to [the company] in this case be adjusted to reflect a return on common equity capital of 19%”
The first and second of these supposed legal errors can be eliminated, however. It seems clear to us that the company does not rest its request for relief on the commission’s claimed failure merely to consider risks peculiar to the company. Far from ignoring the risk factors flowing from the Seabrook investment, the commission obviously had them very much in mind; it simply refused to conclude that they entitled the company to a 19% rate of return.
It is equally clear that the company would not have us predicate any relief on the commission’s claimed failure to articulate its methodology and findings better than it did. At oral argument, the court showed some inclination to agree that the commission’s reasoning might be insufficiently clear and suggested through questions that a remand might be in order. At the close of the argument, however, the company’s counsel was candid in saying that the commission’s failure to recognize compensable risk was clear from the record, and he urged the court not to remand for a more articulate elucidation of what the commission meant to be doing.
That leaves the third view of legal error, that the company is entitled as a matter of law to 19%, which this court is asked to instruct the commission to allow. The company rests this position on two premises: (1) Hope and our recent opinion in Petition of Public Service Co. of New Hampshire recognize a constitutional mandate to compensate the company’s investors with a return “that reflects the risk that the [anti-CWIP] statute places upon them,” Petition of Public Serv. Co. of N.H., 130 N.H. at 275, 539 A.2d at 269; and (2) the commission found, as stated in the rehearing order, that “the 19 percent return on equity is one reasonable measure compensating the investor for all risks inherent in” the company. The company views these as the major and minor premises of a syllogism demonstrating a constitutional entitlement to 19% on common equity.
The syllogism is only as sound as its premises, however, and the language quoted from Petition of Public Service Co. of New Hampshire supra does not purport to be a complete statement of the relevant law, any more than the quoted finding on rehearing states all of the relevant facts. Once the law and the facts are [755]*755adequately accounted for, the company is left in no position to demand 19% as a matter of law.
There are two flaws in the legal component of the company’s argument. The first is the assumption that 19% as “one reasonable measure” of risk must as a matter of law be the only reasonable measure. But as we have already seen, what is reasonable in ratemaking can not be pinpointed. There is a range of what is reasonable, and from the commission’s finding it would not follow that 19% was anything more than one point in the range.
The company’s second erroneous assumption is that Hope and Petition of Public Service Co. of New Hampshire supra hold that a rising probability that risk will be realized must be followed indefinitely by a rising rate of return to the investor upon whom the loss may fall. On the company’s reasoning, an equity investor’s entitlement to 19% at the time this petition was heard would have grown, for example, to an entitlement to 34% by the time the commission transferred the questions that led to our opinion in Petition of Public Service Co. of New Hampshire, for such was computed to be the consequence of the company’s need for revenue on its existing rate base by the summer of 1987. Petition of Public Serv. Co. of N.H., 130 N.H. at 271, 539 A.2d at 266. Such logic, indeed, would provide the investors not with a reasonable rate of return but with plenary indemnification, for the ultimate consequence of automatically raising the return as the prospects for recovering the investment fell would be nothing less than a shifting of the entire risk from the investors to the ratepayers.
It is needless to say, as we held in Petition of Public Service Co. of New Hampshire, that Hope is no authority to require such shifting of risk after risk has materialized, Petition of Public Serv. Co. of N.H., supra at 277, 539 A.2d at 269, nor is it authority for indefinitely augmenting return up to the moment of an increasingly probable disaster. Providing such an upward spiral of return would, indeed, represent a patent failure to weigh the interests of the customers in the balancing process. See Appeal of Conservation Law Foundation, 127 N.H. at 636, 507 A.2d at 672. The company’s view thus runs counter to the settled law that a regulated utility has no abstract constitutional right to make a profit, Power Comm’n v. Hope Gas Co., 320 U.S. at 603; Power Comm’n v. Pipeline Co., 315 U.S. 575, 590 (1947), and no right to the rehabilitation of financial integrity that market forces have hopelessly undermined, Market Street R. Co. v. Comm’n, 324 U.S. 548, 566 (1945).
[756]*756Having said that, we also have to recognize that there is no constitutional or other basis for holding the commission to be without authority to increase a rate of return, once a utility’s plans are understood to carry a more onerous risk than originally perceived. The relationship between investors and ratepayers is not static, and regulators must at least consider restriking old balances between the competing interests as conditions change. And although the difficulty lies in recognizing where legitimate adjustment ends and impermissible risk-shifting begins, there is at least one general rule that effectively answers the company’s position in this case. The source of that rule is an opinion antedating Hope by twenty years.
Hope was not the first case to hold that risks undertaken by a regulated utility should be compensated by returns commensurate with similar risks in other businesses, for that standard had found effective articulation in Bluefield Co. v. Pub. Serv. Comm., 262 U.S. 679, 692 (1923). In Bluefield, however, the Court stated not only this measure of a utility’s entitlement, but also the limit of a utility’s legitimate claim, in holding that a utility has no constitutional right to returns commensurate with “highly profitable enterprises or speculative ventures.” Id. at 693. The implication, of course, is that utilities and their investors are responsible for controlling risks, and while the probability of suffering the consequences of a risk may grow to the point where only a speculator would accept it, ratepayers need not support a utility’s capital with premiums that a speculator would demand.
The dispositive significance of this limitation on the company’s constitutional entitlement is apparent once we recall that the commission found not only that 19% was what the equity investor would require, but that “the 19% return . . . [was] the type of return of a highly profitable enterprise or speculative venture.” To support this conclusion it had, of course, not only the surveys indicating generally lower returns in the nuclear power industry, but also Voll’s explicit expert testimony reflecting widely disseminated investment opinion that the company’s common stock was a speculative vehicle. Because 19% was an appropriate return on stock that was only a speculative investment, it does not follow that the company was entitled to 19%.
While this disposes of the company’s call for remand with instructions to order a 19% return on equity, some further observations on the commission’s treatment of the issue are nonetheless in order. It will be recalled that the commission set the [757]*75715% rate of return by adding 2% to the upper end of the range disclosed in Voll’s sample. The 2% represented the difference, rounded off, between the upper limit of the returns listed in Rosenberg’s sample of companies with nuclear generating plant construction, and the average return shown in the broad FERC sample.
The commission explained that it added the 2% to Volks high number in order to compensate the company for the risk of its “nuclear construction program[ ].” This stated reason is rationally inadequate. As Commissioner Aeschliman pointed out in her dissent, the upper figure in the Voll sample represented a return to companies with nuclear plants under construction; there was no ground to add anything more simply for the risk of nuclear construction already reflected at Volks 13%. Nor, for that matter, did the difference between the FERC average and Rosenberg’s 13.3% even represent the difference between returns for companies without nuclear plants and those with them, since the FERC sample itself included companies with nuclear generators. Whatever the 2% represented, it could not have been found on this record to measure the incremental risk inherent in nuclear construction. The company, naturally, seeks no relief on the basis of these methodological deficiencies, and the company is the only party seeking relief. But the deficiencies are there, and they indicate the need for more critical scrutiny by the commission in the future.
The company’s second assignment of error goes to the commission’s treatment of the effect on the value of the company’s rate base of deferred federal income taxes accumulated during the test year. Because the company’s financial accounting for ratemaking purposes may be different from its tax accounting, the company is customarily allowed to charge customers for tax expenses prior to the actual accrual of the company’s tax liability. The classic example of this practice cited by the parties in the instant case occurs when allowable tax expense is calculated on the assumption of straight-line depreciation, but taxes are accrued and paid on the basis of accelerated depreciation. The resulting funds collected from customers and held for later payment of taxes as they accrue are known as deferred taxes.
To compensate the customers for a utility’s advantage in holding such free funds, the commission requires the utility to deduct the total of accumulated deferred taxes from the value of the rate base. The result is a decrease in rate base as deferred taxes are collected [758]*758and an increase as the tax liability accrues and the utility pays the tax.
Changes in a utility’s tax rate, other things remaining equal, will in obvious ways modify the results of charging customers for deferred taxes. If, for example, the tax rate falls, the amounts of taxes deferred will fall along with the total tax burden, and accumulated deferred taxes will thereafter diminish the rate base by a lesser amount than would have been the case under the old, higher tax rates. When, however, such a change in tax rate occurs shortly before a ratemaking proceeding, unusual problems can arise in reflecting the effects of deferred taxes and accumulated deferred taxes among the items included in so-called test year data.
The concept of the test ye^r provides a methodology for predicting a utility’s future revenue requirement based on past experience. A regulatory commission selects a representative year as a model of the company’s operation, on the basis of which it hopes to predict future performance. See New England Tel. & Tel. Co. v. State, 113 N.H. 92, 95-96, 302 A.2d 814, 817 (1973).
The model, however, is destined to be imperfect. Often enough, even before the ratemaking proceeding has begun, elements of income and expense are known to have changed from their levels in the test year, and the commission must then decide whether to modify the test year data in order to obtain a more accurate prediction. One commentator has observed that “‘[philosophically, the strict test year assumes the past relationship among revenues, costs, and net investment during the test year will continue into the future.’ To the extent that these relationships are not constant, the actual rate of return earned by a utility may be quite different from the rate allowed by the commission. For many years, commissions have adjusted test-year data for ‘known changes,’ i.e., a change that actually took place during or after the test period ....” C. Phillips, Jr., The Regulation op Public Utilities 182 (1985) (footnotes omitted) (emphasis in original); see Appeal of Manchester Gas Co., 129 N.H. 800, 806, 533 A.2d 366, 370 (1987) (commission may limit recognition of changes to those occurring within twelve months of test year’s end, in order to conform to matching principle that adjustments should include both revenue and expense changes). Thus, the test year data are modified to conform to actual experience, for the sake of furnishing the most accurate possible prediction of the utility’s fortunes in the period following the rate order.
[759]*759In the proceeding now before us, the commission selected a test year ending September 30,1985, one consequence of which was that the historical data failed to reflect the decrease in the rate of the company’s federal income taxation provided by the Tax Reform Act of 1986. The company therefore proffered two adjustments to the test year data (and to its own submission to the commission, which had been prepared using the old tax rate). It proposed to decrease its claimed operating expense for deferred taxes. And it proposed to decrease by a like amount the test year addition to accumulated deferred taxes, with the result that the average deduction from the test year rate base would likewise be decreased. The object of the first proposal was to reduce an operating expense to be charged to customers; the object of the second was to increase the average test year rate base, and to increase revenues accordingly. The commission accepted the first proposal but declined the second.
The company claims the latter decision was erroneous and describes it as violating double entry bookkeeping concepts, as well as the matching principle of recognizing both revenue and expense changes to test year data. What the company overlooks, however, is that the commission was not prescribing bookkeeping methods; it was adjusting test year data to conform to known fact, and in so doing it accurately reflected known fact in each instance.
The tax rate fell after the close of the test year, resulting in lower projections of deferred tax expense under financial accounting principles for the period to be covered by the rate order. The commission therefore properly adjusted the test year deferred tax expense item to reflect the effect of the tax law change on the company’s future taxation. The commission thus recognized that a future event would be different from corresponding events in the past and modified the data to reflect this known and predictable difference.
Because, however, the lower tax rate had not in fact applied to the period selected as the test year, the company had actually received revenue representing deferred taxes computed at the higher rate then in effect, and the company had increased its accumulated deferred taxes by that amount. Having thus collected enough money to satisfy its anticipated tax burden at the old, higher rate, the company ended up with more money than it would need to pay the taxes that would accrue at the new, lower rate. Consequently, the company had the use of the excess along with all of the other funds represented by the accumulated deferred tax account, and it would continue to benefit from the favorable consequences of holding that excess until it should be drawn down [760]*760in a manner prescribed by the commission (a matter that is not before us). If, therefore, the commission had decreased the figure indicating the test year addition to accumulated deferred taxes and had set rates accordingly, the commission would not have given customers credit for a portion of deferred tax expense they had actually paid during the historical test year.
To have done as the company requested, the commission would have lost sight entirely of the reason for permitting changes to test year data, which is not to falsify the past, but to increase the accuracy of predicting the future. Thus, test year data is properly modified to reflect changes that will predictably cause future experience to differ from past experience, e.g., in the rates at which employees are paid or taxes collected. But test year data may not be modified to pretend that past experience with predictable effects on the future did not occur, e.g., that deferred taxes were accumulated in different amounts from those actually shown on the books. The company’s second proposal would have falsified the past and would have produced a false prediction of the company’s future revenue need, by pretending the company would not enjoy the benefit of some revenue it had in fact collected.
Nothing, of course, in the commission’s order assumes that the company will continue to add to accumulated deferred taxes at the old rates. Its deferred tax expense will be computed at the new rates, as will its annual additions to the account for accumulations. The order merely assumes, correctly, that the company will benefit from holding what we have called the excess, which it actually received. In setting the company’s rate of return on that assumption, the commission acted true to the facts.
We are aware of at least three other regulatory agencies that have taken the same position. See Re Carolina Power & Light Co., 87 P.U.R. 64, 104-05 (1987); Pennsylvania Public Utility Comm. v. Pennsylvania Power Co., 85 P.U.R. 4th 323, 380-81 (1987); Re Potomac Electric Power Co., 83 P.U.R. 4th 219, 229-30 (1987). It would be surprising to find contrary authority, of which the company has cited none.
Affirmed.
All concurred.