On September 23, 1981, the board of directors of the Boston Edison Company (Edison) voted to cancel the construction in Plymouth of a second nuclear power plant, called Pilgrim II.2 On October 16, 1981, Edison filed with the Department of Public Utilities (department) revised electric rate schedules providing for a general rate increase, effective November 1, 1981. Edison proposed to recover, as a cost of service reflected in its rates, its investment in Pilgrim II, that is, its share of the net costs of the project, estimated (after salvage) to exceed $278,000,000.3 Pursuant to its authority under G. L. c. 25, § 18, the department suspended the proposed rates until the earlier of May 1, 1982, or its further order. The department then undertook what it characterized in its decision of April 30, 1982, as “probably the lengthiest and most complex rate case to have come before the Department in recent years.”
All the appeals are from the department’s April 30, 1982, decision (G. L. c. 25, § 5) and are before us on a reservation and report by a single justice. In large measure, these various appeals are challenges to the department’s determination to permit Edison through its rates to amortize a portion of its investment in Pilgrim II. The department said in its decision that “the question presented by the demise of the Pilgrim II project is one of the most serious and the most controversial regulatory issues that the Department has ever confronted.” There are collateral issues concerning (a) the department’s treatment of tax benefits resulting from the abandonment of the project and (b) the department’s approval of carrying charges on a por*211tian of the amortized recovery. The city of Boston presents an entirely separate challenge to its street lighting rates. Before we discuss these substantive issues, we shall deal with procedural questions related to certain of the appeals. We shall then consider the principal issue presented in these appeals. We conclude that the department committed no error of law in its decision to permit Edison to recover as a cost of service a portion of its investment in Pilgrim II and to collect certain carrying charges. Finally, we shall discuss and reject the city of Boston’s challenge to the street lighting rates.
Procedural Issues
The department and Edison argue that the appeal of Massachusetts Public Interest Research Group, Inc. (MASSPIRG), was not seasonably filed and that it must be dismissed. We need not pass on this question to the extent the issues MASSPIRG raises are also raised by appellants whose appeals were timely. We could, in such a situation, simply treat MASSPIRG’s brief as one filed by an amicus curiae.4 There is, however, one issue raised by MASSPIRG not advanced by any other party in these appeals and, because of that issue, we must consider the timeliness of MASSPIRG’s appeal.
Following the department’s April 30, 1982, decision, various motions for reconsideration were filed. On May 20, pursuant to its authority under G. L. c. 25, § 5, the department granted an extension of time for the filing of any appeal from its April 30 decision until ten days after its final *212order on the motions for reconsideration. Among the motions filed was a motion of MASSPIRG that Commissioner Edward L. Belgrade (Commissioner Belgrade) recuse himself from further participation on the question of Edison’s recovery of the cost of Pilgrim II. The department issued an order on July 20, 1982, concerning the various pending motions, including MASSPIRG’s motion that Commissioner Belgrade recuse himself. Commissioner Belgrade declined to recuse himself from any portion of the proceeding, and the other two commissioners concurred in his decision. MASSPIRG states that the order was mailed on July 21, 1982.
On August 13, 1982, MASSPIRG filed its petition for appeal with the secretary of the department. That petition was not filed within ten days of the department’s final order concerning the motions filed following the April 30 decision. In the circumstances, the April 30 decision was probably the final order from which, under G. L. c. 25, an appeal had to be taken within twenty days but for the ten-day extension granted by the department. See Plymouth County Nuclear Information Comm., Inc. v. Energy Facilities Siting Council, 374 Mass. 236, 244-245 (1978). But we need not decide that point because the motion seeking Commissioner Belgrade's recusal was a new matter decided on July 20, 1982, and as to that matter MASSPIRG had twenty days from the date of service in which to appeal. G. L. c. 25, § 5.
MASSPIRG’s petition for appeal, however, was not filed with the department’s secretary within twenty days of service (on July 21) of notice of the July 20 denial of its motion to recuse, but rather it was filed on August 13. MASSPIRG seeks to bridge the three day gap (from August 10 to August 13) by reliance on Mass. R. Civ. P. 6 (d), 365 Mass. 747 (1974), which adds three days to any prescribed period within which action must be taken when notice is served by mail. To apply the rule to this situation would contradict the provisions of G. L. c. 25, § 5, specifically governing the timing of appeals. Moreover, rules of court do not govern procedures in the Executive Department. They apply to *213proceedings in courts. See Mass. R. Civ. P. 1, as appearing in 385 Mass. 1213 (1982); Mass. R. A. P. 1, as amended, 378 Mass. 925 (1979). The late filing of MASSPIRG’s notice of appeal with the department’s secretary is an error of a type that calls for dismissal of MASSPIRG’s appeal. Schulte v. Director of the Div. of Employment Sec., 369 Mass. 74, 79 (1975).
Although we conclude that MASSPIRG’s appeal must be dismissed, we think it appropriate to comment on the underlying claim that Commissioner Belgrade should have disqualified himself in this proceeding because he appeared in the prior, related proceeding (D.P.U. 19494) on behalf of the Massachusetts Executive Office of Energy Resources (MEOER) of which he was the Director of Resource Development. Although his involvement was not extensive, he did participate in writing and in filing a brief in D.P.U. 19494 on behalf of MEOER in support of a determination that the Pilgrim II project was reasonable. The brief expressed unenthusiastic approval of the project.5 Commissioner Belgrade was appointed a commissioner of the department on September 15, 1981, while D.P.U. 19494 was pending. He recused himself from that matter, which was decided by the other two commissioners one week later.
The earlier proceeding (D.P.U. 19494) was an investigation by the department of future demand for electricity in Edison’s service territory, the appropriate level of Edison’s reserve capacity, and alternatives to Edison’s program for the construction of Pilgrim II. The hearings in D.P.U. 19494 concluded on September 28, 1979, after 121 days of *214hearings. Almost two years later the department concluded, on September 22, 1981, that oil as a source of electric generation was not an acceptable option and “that nuclear power, despite its recent problems [was] most likely to be the least expensive generation alternative available to the Company.” The department concluded that Edison’s generation construction program was reasonable. The decision in D.P.U. 19494 did not determine whether Edison could recover costs associated with Pilgrim II, whether Edison reasonably could finance the project, and, of course, whether Edison acted prudently in continuing with the project based on events occurring after the record closed in D.P.U. 19494.
On March 16, 1982, responding to Commissioner Selgrade’s request, the State Ethics Commission advised Commissioner Selgrade that, on the facts presented by him, he would not violate the conflict of interest law (G. L. c. 268A) by participating in deciding Edison’s pending rate proceeding. The Ethics Commission declined to pass on whether Commissioner Belgrade’s participation would warrant the conclusion that, on common law principles, he should disqualify himself for bias. The Ethics Commission noted that Commissioner Belgrade’s “previous participation, although advocating a position favorable to the Company was, in fact, a part of [his] role as a state employee.”
MASSPIRG was a party to the earlier proceeding (D.P.U. 19494) and had notice, therefore, of Commissioner Belgrade’s involvement in the earlier proceeding. One of MASSPIRG’s current counsel filed an affidavit with the motion to recuse stating that he did not know of the commissioner’s involvement in the earlier proceeding until May 28, 1982. We do not know when other representatives of MASSPIRG learned of the commissioner’s participation in the earlier proceeding, but the inference is unavoidable that some representatives of MASSPIRG knew or should have known of that involvement at the time of Commissioner Belgrade’s participation.
*215We comment on this issue because it presents the question whether the proceeding was essentially unfair because of Commissioner Belgrade’s involvement in it, a circumstance which, if proved, we might regard as appropriate for our consideration on our own motion.6 We conclude that the proceeding was not tainted with unfairness justifying the reversal of the department’s decision.
There are several reasons for our conclusion. MASSPIRG’s motion that Commissioner Belgrade recuse himself sought recusal only as to his further participation in the proceeding on the question of Edison’s recovery of the cost of Pilgrim II. All action after the recusal motion was filed was taken by all three commissioners, including the commissioner who dissented on the Pilgrim II issue. None of that subsequent action involved the Pilgrim II recovery issue directly. The motion was made after the proceeding had been substantially concluded following many days of hearings and the filing of a decision of over 250 pages. Other parties, including the Attorney General, have not argued this issue to us. Commissioner Belgrade participated in the earlier proceeding as a State employee advocating policies of the Executive Department which may or may not have been consistent with his own views. MASSPIRG was served a copy of the MEOER brief signed by Commissioner Belgrade and thus had notice of his involvement in that matter well before the present proceeding began. The State Ethics Commission considered the matter and ruled that Commis*216sio ner Belgrade’s participation would not violate the conflict of interest law, including G. L. c. 268A, § 23 (e), as then amended, concerning the appearance of impropriety. The earlier proceeding, although related to this subsequent rate proceeding, focused on different issues, and the determination of the first matter in Edison’s favor did not preordain Edison’s recovery of all or part of the costs of Pilgrim II. That conclusion is shown dramatically by the decision of one of the commissioners who in the first proceeding decided in Edison’s favor but in the second proceeding voted against Edison’s recovery of any portion of the cost of Pilgrim II. Finally, if Commissioner Belgrade had not participated in the Pilgrim II decision, the department would have been deadlocked on that issue and no decision at all might have been possible, with the result that no final action would have been taken on Edison’s rate filing (at least on this issue) within six months and the rates as filed would have become effective in full on May 1, 1982, when the statutory suspension of the rate filing expired. See G. L. c. 25, § 18. If such a deadlock had occurred, Commissioner Belgrade’s involvement might have been justified under the rule of necessity. See Moran v. School Comm. of Littleton, 317 Mass. 591, 593-594 (1945); United States v. Will, 449 U.S. 200, 213-216 (1980).
Stanley U. Robinson, Third, a residential ratepayer, filed two appeals from determinations of the department in the rate proceeding. The first appeal challenges the department’s denial of his petition to intervene in the proceeding as a party. The other is an appeal from the department’s final order.
The department denied Robinson’s motion to intervene as a party but granted him “limited participant status,” stating that he did not have status as a party but that he could conduct discovery and file memoranda or briefs. The Attorney General was designated as Robinson’s “lead counsel.” The issue of Robinson’s participation as a party was within the broad discretion of the department under G. L. c. 30A, § 10, cl. (4), as appearing in St. 1978, c. 60, § 1. *217See Boston Edison Co. v. Department of Pub. Utils., 375 Mass. 1, 45-46, cert. denied, 439 U.S. 921 (1978). The department did not abuse its discretion. It did not have to hold a hearing on the motion to intervene, nor did it, in the circumstances, have to state its reasons. A judgment shall be entered affirming the department’s order denying Robinson’s motion to intervene.7
Edison moved to dismiss Robinson’s appeal from the department’s final decision on the ground that under G. L. c. 25, § 5, only an “aggrieved party in interest” may appeal from a decision of the department (emphasis supplied). In Save the Bay, Inc. v. Department of Pub. Utils., 366 Mass. 667, 673 (1975), we defined those persons who could qualify as an “aggrieved party in interest.” Robinson was not admitted as an intervener or party; he had no right to intervene as a matter of law; and, as a residential customer alleging no peculiar damage to himself, he had no constitutional or statutory right to participate fully in the proceedings. Id. Consequently, Edison’s motion to dismiss Robinson’s appeal should be allowed. Judgment shall be entered dismissing Robinson’s appeal from the final decision of the department.
Abandonment of Pilgrim II
The principal dispute in these appeals involves the propriety of the department’s determination to permit Edison to recover, by amortization over thirteen years, a portion of its investment in Pilgrim II and to permit Edison to collect a carrying charge on a portion of the unamortized balance. We conclude that there is no legal impediment to the department’s permitting Edison, or any other utility, to recover its prudent investment in a nuclear power plant reasonably abandoned before completion and to recover carrying charges on amounts amortized. We further con-*218elude that the department’s decision conforms to the requirements of the State Administrative Procedure Act (G. L. c. 30A), in terms both of the adequacy of the statement of reasons for the decision and of the sufficiency of substantial evidence to support the result.
The department’s April 30, 1982, decision in this case devoted almost 100 pages to a discussion of “Pilgrim II Cost Recovery.” The department began by considering the background of Pilgrim II, including its decision in D.P.U. 19494. The department concluded that it was not barred by any rule of law from allowing recovery of costs of Pilgrim II even though Pilgrim II never generated any electricity. It characterized the abandonment of Pilgrim II as an extraordinary loss, a loss of a magnitude never before considered by the department. This loss required “separate, independent and unique rules that properly reflect the just balance of the affected interests” (ratepayers and investors). The department looked for guidance to decisions in other jurisdictions, including a decision of the Federal Energy Regulatory Commission, and saw the results as overwhelmingly favoring recovery of some part of the loss.
Edison, the department said, had the burden of proving that Pilgrim II was a prudent investment from its inception to cancellation. D.P.U. 19494 was dispositive of the issue of the reasonableness of Edison’s construction program, based on the record in that proceeding. However, D.P.U. 19494 did not consider events occurring after the record closed in that proceeding, and it did not determine Edison’s ability to finance Pilgrim II. The department’s decision then discussed the evidence on which the Attorney General and other parties relied in arguing that Edison’s investment in Pilgrim II was imprudent, or at least became so, long before Edison determined to cancel the project on September 23, 1981. Next, the department considered Edison’s contentions in support of its claim that Pilgrim II was a reasonable, prudent investment up to the date Edison cancelled the project.
The department stated certain general conclusions. The Pilgrim II loss was large compared to the size of Edison and, *219in this respect, unprecedented elsewhere in the country. The loss approximated two-thirds of Edison’s entire net worth, more than 25% of its permanently invested capital, and more than earnings retained and reinvested over its eighty-six year history. Pilgrim II was originally intended to meet a growth in customer demand. Although that demand never materialized, Edison needed to reduce its dependence on increasingly expensive and unreliable supplies of oil. Rate regulatory practices in the Commonwealth required Edison to exclude the costs of Pilgrim II from its rate structure during the construction process. By these practices, capitalized costs — costs of construction work in progress (CWIP) and the cost of financing the construction, allowance for funds used during construction (AFUDC) — are not recoverable from ratepayers until the project is completed. When it is completed, capitalized costs may be recovered through depreciation allowances reflected in the rates. As a final general principle, the department noted that the risk that Pilgrim II would not be completed was not reflected in the return allowed investors in Edison’s previous rates.
The department then stated that, if Edison were to absorb the Pilgrim II loss, “regulatory policies, and the returns they dictate, appear to us to be inadequate to compensate investors for the new level of risk. Investors who are inadequately compensated do not remain investors for long. . . . The disdain of the financial markets for this Company will be formidable . . The cost of new capital would increase and service would deteriorate unavoidably because of the scarcity of reasonably priced capital. The question “[i]n a very real sense,” as the department saw it, was not who should bear the costs of the abandonment of Pilgrim II but “rather, when should those costs be faced.” Edison, it said, must have the ability to raise capital to invest in facilities to reduce its 65 % dependency on oil for generating electricity. The controlling consideration is that the company must be able to meet its service obligation. The department saw “no realistic way” to make capital available *220for oil displacement if the Pilgrim II loss were to be fully absorbed by Edison’s shareholders.
The department then considered the question of the prudence of Edison’s actions during the history of the project. It noted particular factors that affected the reasonableness of the project from time to time. The cost of oil was increasing, and Pilgrim I had saved Edison’s customers millions of dollars. However, the cost of construction, regulatory delays in obtaining approval of the construction of Pilgrim II, and a decline in projected demand for electricity presented problems. The accident at Three Mile Island occurred in the spring of 1979. Through 1978 and 1979 the propriety of continuing with the project became less clear, but, in the department’s view, not to the point where prudence required abandonment. In the first half of 1980, problems arose from the delay in obtaining a construction permit from the Nuclear Regulatory Commission and from projections of increased costs of construction over an extended and uncertain construction schedule. At a June, 1980, meeting, Edison’s directors voted to proceed with licensing efforts but to limit project expenditures. The department concluded that Edison should have cancelled the project in June, 1980, because “project uncertainty had become intolerably high by June, 1980.” Cancellation was then, in the department’s view, “the only prudent course of action.”8
*221The department then considered what amount of Edison’s Pilgrim II loss should be reflected in the rates. Edison’s share of the estimated costs of Pilgrim II was approximately $278,000,000. All expenditures after July 1, 1980, were to be eliminated from the amount to be recovered. The department further disallowed recovery of the equity rate of return portion of AFUDC, as it had in Western Mass. Elec. Co., D.P.U. 558, at 40 (July 31, 1981). The department then determined that the amount to be recovered, after deduction of tax benefits arising from abandonment of Pilgrim II, should be amortized over thirteen years, with an annual carrying cost rate of 14 % applied only to the non-AFUDC component of the amount to be recovered. The amount to be recovered was made uniform for each year by a levelizing factor formula. The annual recovery allowed to be reflected in Edison’s rates was approximately $12,500,000, as shown (in approximations) by the following calculations:
Edison request $278,300,000
Less disallowed expenses (Post-June,
1980, expenses of Pilgrim II [26.9% of all Pilgrim II expenses], and AFUDC attributable to common equity) 76,700,000
$201,600,000
Less tax benefit attributable
to amount for which recovery is allowed
(73.1% of the total tax benefit) 84,800,000
Amount allowed to be recovered
amortized over thirteen years $116,800,000
Amount subject to carrying charges, determined by deducting pre-July, 1980, debt and preferred AFUDC from $116,800,000 $ 34,200,000
Amount subject to carrying charges of 14% per annum $ 82,600,000
*222Annual amortization amount (excluding income taxes)
Annual amount with carrying charges $ 9,900,000
Annual amount of AFUDC 2,600,000
$ 12,500,000
This court has not previously dealt with the question whether the department may lawfully permit a utility to recover prudently incurred costs of a construction project reasonably abandoned prior to its completion. The appellants point out that Pilgrim II never generated any electricity and argue that, until such an asset has “come on line,” its costs may not be reflected in rates. They point to no statutory basis for this argument. We have considered the right of a utility to reflect in its rates the prudently incurred cost of plant reasonably abandoned before it was fully depreciated. We have deferred to the policy of the department concerning recovery of the cost of such abandoned facilities. See Boston Edison Co. v. Department of Pub. Utils., 375 Mass. 1, 20-21 (1978) (experimental, abandoned pollution control device amortized over ten years; no error in excluding the unamortized retired plant from the rate base); Fitchburg Gas & Elec. Light Co. v. Department of Pub. Utils., 371 Mass. 881, 886-887 (1977) (same result as to reasonably abandoned production facilities). We have required that the department be consistent in its treatment of particular items of unamortized retired plant of a company, in the absence of an appropriate statement of reasons for any change in treatment. See Boston Gas Co. v. Department of Pub. Utils., 367 Mass. 92, 104 (1975). Thus, a department determination to permit unamortized retired plant in the rate base could not be changed in a subsequent proceeding without some justification. Id. See Boston Gas Co. v. Department of Pub. Utils., 368 Mass. 780, 802 (1975). In none of these opinions is there even a suggestion that the department might lack statutory authority to permit recovery of the undepreciated cost of prudently aban*223doned plant and a carrying charge during any period of amortization.
Judicial decisions elsewhere concerning the recovery of the cost of nuclear plants abandoned before completion are rare and not particularly instructive for the purposes of this case. In Central Me. Power Co. v. Public Utils. Comm'n, 433 A.2d 331, 344-345 (Me. 1981), the appeal was presented on the assumption that the cost of abandoned plant could be recovered. The utility argued unsuccessfully that the commission, in allowing an amortization of expenses (exclusive of associated AFUDC) over five years, had not engaged in a reasonable balancing of the interests of ratepayers and shareholders. The assumption that abandoned plant could be recovered also was made in Wisconsin Pub. Serv. Corp. v. Madison Gas & Elec. Co., 109 Wis. 2d 256, 265 (1982), where the utility was able to persuade the court that the procedure the commission designated for the recovery of costs in fact prevented recovery of the allowed costs and was arbitrary. In Office of Consumers’ Counsel v. Public Utils. Comm’n, 67 Ohio St. 2d 153, 166 (1981), appeal dismissed sub nom. Cleveland Elec. Illuminating Co. v. Office of Consumers’ Counsel, 455 U.S. 914 (1982), a divided court, relying strictly on special provisions of an Ohio statute, reversed an agency decision that allowed amortization of the costs of abandoned nuclear facilities as beyond the agency’s statutory authority.9 Except for this Ohio case, which is of no guidance to us, we are aware of no appellate judicial opinion in which the authority of a State regulatory agency to allow recovery of the prudent costs of an abandoned nuclear plant was even raised, much less successfully challenged.
*224A substantial majority of the public utility regulatory agencies that have considered the question have permitted a utility to recover all or some portion of the prudently incurred costs of a nuclear power plant reasonably abandoned before completion.10 See Office of Consumers’ Counsel v. Public Utils. Comm’n, supra at 162 n. 6. Agency decisions that have denied recovery have relied on circumstances not present in the case before us.11 Among the agency decisions *225allowing recovery of costs of abandoned plant, a variety of conclusions have been reached concerning the allowance *226of a carrying charge or a return on those costs when they have been amortized.12 Those agency decisions denying the recovery of costs or a return on amortized costs generally *227have not rested on the ground that the agency lacked authority to allow recovery of these items. The question decided in almost all cases, allowing or disallowing costs and carrying charges, was treated, expressly or impliedly, as falling within the discretion and judgment of the agency.13 The argument that in Massachusetts recovery of the prudently incurred costs of an abandoned nuclear plant is unlawful is apparently a recent one.14
The department had the authority to permit the recovery of a company’s prudent investment in plant reasonably abandoned before completion. The general statutory authority of the department permits the department to approve rates reflecting such an investment (G. L. c. 164, § 94).15 The fact that Pilgrim II never produced any electricity, never became “used or useful,” does not prohibit *228recognition of its prudent costs in ratemaking. The same conclusion is appropriate concerning the department’s allowance of carrying charges on amounts amortized and charged to the cost of operations.
We do not substitute our judgment for that of an administrative agency where no constitutional question is presented. We decline to prescribe a common law of utility rate-making. Ratemaking is a legislative, not a judicial, function. Our involvement, beyond constitutional questions, has been to assure that the agency has adhered to statutory requirements. Where agency action is irrational, we will find an error of law. See Southbridge Water Supply Co. v. Department of Pub. Utils., 368 Mass. 300, 308-309 (1975). But, as we have said and ruled in a variety of contexts, questions of policy are for an administrative agency. See, e.g., School Comm. of Newton v. Labor Relations Comm’n, 388 Mass. 557, 573 (1983); Boston Edison Co. v. Assessors of Watertown, 387 Mass. 298, 302 (1982); Massachusetts Elec. Co. v. Department of Pub. Utils., 376 Mass. 294, 302 (1978); New England Tel. & Tel. Co. v. Department of Pub. Utils., 372 Mass. 678, 684 (1977); Boston Gas Co. v. Department of Pub. Utils., 367 Mass. 92, 98 (1975). We come then to the appellants’ challenges to the reasoning of the department in its decision and to their claim that the department’s decision to allow recovery of a portion of Edison’s investment in Pilgrim II and a carrying charge on a portion of the amortized balance was not supported by substantial evidence.
We conclude that the department’s decision was reasonable in its analysis and thus not arbitrary or capricious and that there was substantial evidence to support its findings. Many of the appellants’ arguments present logical reasons, with supporting evidence, for arriving at conclusions contrary to the department’s conclusions on particular issues, such as the prudence of Edison’s investment decisions and the financeability of the project at various times. Qur task under the State Administrative Procedure Act (G. L. c. 30A, § 14 [7]) is not to decide the issues anew but to *229determine whether there was any error of law. We forbear from reciting the substantial evidence in support of the department’s conclusions, and certainly we will not burden this opinion with the contrary evidence on which the appellants rely. There were some moments when the feasibility and prudence of Edison’s investment in Pilgrim II looked better than they did at other moments. However, the department was warranted in concluding that the investment was a prudent one up to June 30, 1980. The date that represented the dividing line between prudence and imprudence required a judgment call which, on all the evidence, the department was not required to place any earlier than June 30, 1980. See Boston Edison Co. v. Department of Pub. Utils., 375 Mass. 1, 12 (1978). On the issue of prudence, the department had no authority to substitute its judgment for the reasonably exercised prerogatives of Edison’s business managers. New England Tel. & Tel. Co. v. Department of Pub. Utils., 360 Mass. 443, 484 (1971). See Mystic Valley Gas. Co. v. Department of Pub. Utils., 359 Mass. 420, 431-432 (1971). The fact that neither the department nor the ratepayers had a direct voice in Edison’s business decisions is not relevant to the issues in this case.16
We accept as within the department’s discretion the conclusion to allow some direct recovery of the costs of Pilgrim II, and carrying charges on a portion of that recovery. The department’s judgment as to the consequences of its allowance or disallowance of such recoveries was logically expressed. Investor confidence is not an insignificant element in utility regulation. A judgment that the adverse consequences of disallowing any recovery would be a serious threat to the company’s financial integrity, and indirectly to its customers, was warranted. Obviously, many future *230events cannot be determined with certainty and must be forecast based on reasonable judgment. There is substantial evidence in the record to support the department’s conclusions concerning the consequences of the disallowance of any recovery.
We comment briefly on certain challenges made to the department’s decision. The department correctly placed the burden on Edison to demonstrate the prudence of its decision to commence and to continue with the Pilgrim II project. We do not see a fatal shift in this burden in the department’s comments on the inadequacy of certain arguments made by the interveners or in the department’s statement at one point that on the basis of particular facts it could not find “imprudence.”
Once in its decision the department suggests that the project was financeable in part because of “the state regulatory commitment to maintain [Edison’s] financial integrity in the face of the need for that particular facility.” It is no doubt true that a rate regulatory agency’s determination of the feasibility of a project will tend to establish both investor confidence and the financeability of the project through future rates approved by that same agency. However, the department made no explicit statement on the reasonableness of Edison’s investment in Pilgrim II until September 22, 1981, the day before Edison cancelled the project. On the other hand, the record in D.P.U. 19494 closed in February, 1980, and the expectation of departmental approval was reasonable considering the substantial savings to consumers shown on that record if Pilgrim II were completed. The department’s decision that the project was financeable at various times was not based solely on the conclusion in D.P.U. 19494 that the project was reasonable. The prospects of financing.Pilgrim II at various times were enhanced by the reasonable expectation that the department’s decision in D.P.U. 19494 would be favorable.
The appellants argue that, if Edison could finance billions of dollars for Pilgrim II, it was plainly wrong for the department to conclude that Edison could not absorb the *231relatively small losses from the abandonment of the project. The financing of the project depended on reflecting the investment in Pilgrim II in Edison’s rates. Thus the appellants’ comparison is inappropriate.
The Attorney General has argued separately that the department’s decision was arbitrary and capricious in its allowance of carrying charges and in its treatment of the income tax benefits resulting from the abandonment of the project. The city of Boston also argues against the allowance of carrying charges.
We have discussed carrying charges in our general consideration of the department’s authority and discretion. We see no error in the department’s allowance of carrying charges on the non-AFUDC portion of the amount being amortized. The amount of the carrying charge is subject to adjustment in future rate proceedings.17 The circumstances of this case are sufficiently different from the Western Mass. Elec. Co. case, D.P.U. 558 (July 31, 1981), so that the disallowance of a carrying charge there (where the recovery of an investment of approximately $2,700,000 in two abandoned nuclear plants in Montague was allowed over four years) and the allowance of a carrying charge here do not demonstrate arbitrary or capricious action. The department’s judgment that Edison would need the funds to finance future plans is not unreasonable on this record.
The Attorney General argues that the department acted arbitrarily in its treatment of the Federal tax benefits derived from the abandonment of the project. The department allocated Edison’s loss after taxes between ratepayers and shareholders, with June 30, 1980, marking the dividing line.18 This allocation was within the department’s discretion. In the absence of any contest on this issue raised be*232fore the department, we conclude that the department’s determination, without extended discussion, to permit the tax benefits to be distributed in proportion to the allocation of costs is not arbitrary. In fact, the city of Boston agrees with the department’s method of allocation.19
We, therefore, conclude that the department had the authority to allow Edison’s rates to reflect the amortization of a portion of the after-tax cost of Pilgrim II and carrying charges on a portion of those amortized costs. We further conclude that the department’s decision complied with the requirements of G. L. c. 30A, § 14 (7).
Street Lighting Rates
The city of Boston challenges the increase in the street lighting rates as “unduly or irrationally discriminatory, ” citing American Hoechest Corp. v. Department of Pub. Utils., 379 Mass. 408, 411 (1980). Edison proposed, and the department agreed, that the revenue to be generated by the rate increase would be allocated among the various classes of customers so that those classes currently providing higher than average rates of return would receive lower than average percentage increases.20 In the test year used in this case, the street lighting class generated a rate of return to Edison above the average. The average rate of return for *233all classes was 8.27%. Edison’s study of the cost of service by class showed the street lighting rate of return to be 13.4% , or more than 60% above the average. The company proposed an increase for the street lighting class of 18.9% and an increase for all classes averaging 23.88%. The department noted that the allocations proposed by Edison would reduce the differentials but would not eliminate them.
Because the department has now acted on an Edison rate filing (Boston Edison Co., D.P.U. 1350 [May 31, 1983]) subsequent to the one under consideration here, the issue of-the street lighting rates contained in this proceeding may well be moot because retroactive adjustments are not permissible. See Boston Edison Co. v. Department of Pub. Utils., 375 Mass. 1, 38 (1978). We nevertheless comment on the propriety of the department’s action because the issue of allocation of revenue needs among classes of customers will be a recurring one.
A choice between alternative methods of allocation of revenue needs made by Edison and approved by the department is appropriate as long as it does not have a confiscatory effect and is not otherwise illegal. See Massachusetts Elec. Co. v. Department of Pub. Utils., 376 Mass. 294, 302 (1978). The issue is the propriety of the department’s decision on Edison’s filing and not the merits of any alternative proposal advanced by an intervener in the proceeding. See Trustees of Clark Univ. v. Department of Pub. Utils., 372 Mass. 331, 335 (1977).
We conclude that the department acted within its discretion in approving Edison’s proposal even though the proposal only narrowed the differences in class rates of return. Cost of service need not be the sole criterion used in establishing rate classifications. American Hoechest Corp. v. Department of Pub. Utils., 379 Mass. 408, 411-412 (1980). The department was warranted in its discretion in concluding that a “correction” of differentials in one step would be more disruptive than a series of a gradual adjustments over time. A class generating a disproportionately large *234rate of return, greater than the street lighting class, was exempted from any increase, but it was not unlawful to make that distinction. Edison, with the department’s approval, could reasonably draw the line where it did. We add, however, that unless there is a reasonable justification for significant differentials in the rates of return of classes, perhaps based on differences in usage or on public policy considerations (see American Hoechest Corp. v. Department of Pub. Utils., supra at 412-413), continuing substantial differences in the rates of return of classes may result in “unduly or irrationally discriminatory” rate classifications. Id. at 411.
Conclusion
In the appeals of .the Attorney General and of the city of Boston, judgments shall be entered affirming the decision of the department. The appeals of MASSPIRG and SMOG shall be dismissed for the separate reasons stated in this opinion. Judgments shall be entered, respectively, affirming the determination of the department not to permit Stanley U. Robinson, Third, to intervene as a party and dismissing his appeal from the department’s decision.
So ordered.
Planning for Pilgrim II, a 1,150 megawatt plant, began in 1971. Edison, the lead participant, owned approximately 59% of the project and was joined in intended ownership by other Massachusetts utility companies and utility companies in Maine, New Hampshire, and Vermont.
Edison specifically proposed that its investment be amortized in depreciation and amortization expense over ten years, with a carrying charge to compensate it for the delay in recovery of its investment.
This is the position we take with respect to Edison’s motion that the appeal of South Middlesex Opportunity Council (SMOG) be dismissed because of an alleged significant discrepancy between SMOC’s petition for appeal filed with the department and SMOC’s petition filed in the Supreme Judicial Court for the county of Suffolk. SMOG has only argued issues advanced by at least one party properly before us, and, in deciding those issues, we have considered the arguments advanced by SMOG. We simply conclude that, in the circumstances, there is no need to pass on Edison’s motion to dismiss SMOC’s appeal. We dismiss SMOC’s appeal because the issues raised by it are resolved in the appeals that are before us.
That brief stated in conclusion:
“The Office cannot be enthusiastic about a decision supporting the expenditure of $2.0 billion. We reach our judgment against the sober backdrop of today’s price of oil. That back-drop adds a good deal of clarity to our judgment but provides no joy. In the final analysis, we fear that the decision to forego [sic] the Company’s construction program will mean a far greater direct and indirect economic cost.
“The Massachusetts Office of Energy Resources urges the Commission to approve the construction of Pilgrim II on the earliest date the Nuclear Regulatory Commission resumes licensing.”
There is no doubt that, in particular circumstances, an agency decision maker’s prior participation in a matter as counsel would involve a denial of due process if he were to participate in deciding the matter. See American Cyanamid Co. v. FTC, 363 F.2d 757, 767 (6th Cir. 1966) (on seasonable objection, Federal Trade Commission chairman should not have participated in a decision on a subject as to which he had participated extensively as chief counsel and staff director of a Senate subcommittee); American Gen. Ins. Co. v. FTC, 589 F.2d 462, 463 (9th Cir. 1979) (member of Federal Trade Commission should have disqualified himself from participation in deciding an issue in a proceeding in which he had earlier appeared as general counsel for the Federal Trade Commission and filed a brief arguing the same issue).
If Robinson had been entitled to intervene, he says he would have dealt with certain issues not pressed, at least before us, by any other party.
The department’s decision was not unanimous on the issue of Edison’s recovery of the costs of Pilgrim II. Commissioner George R. Sprague disagreed with his two colleagues. He believed that Edison’s shareholders should bear all the expenses of Pilgrim II. He said in his dissent “that it is not proper for a utility to charge its customers, by whatever means, for investments which have not become used and useful.” He further concluded that Edison should not be entitled to any recovery of Pilgrim II costs because it made the decision to proceed with the project on its own and resisted attempts at public, regulatory involvement in the question of the prudence of Pilgrim II. We see no basis for disallowing prudently incurred costs on the ground that the utility alone made the investment decision. He also believed that Edison had failed to meet its burden to show its ability to finance Pilgrim II and that Edison could absorb the loss without significant adverse consequences. On these points, Commissioner Sprague’s conclusions may or may not be supported on the record. Our concern is the lawfulness of the decision of the department, in which a majority of the commissioners joined.
The Ohio statute (Ohio Rev. Code Ann. § 4909.15 [A] [4] [Baldwin 1978 & Supp. 1981]), provided that when fixing and determining just and reasonable rates the public utilities commission “shall determine ... (4) [t]he cost to the utility of rendering the public utility service for the test period.” A majority of the Ohio court concluded that the construction costs of the abandoned nuclear plant were not a cost of rendering service for the test period. Office of Consumers’ Counsel v. Public Utils. Comm’n, supra at 163-164. Massachusetts has no similar statutory provision.
See, e.g., Gulf States Utils Co., 40 P.U.R.4th 593, 595 (La. Pub. Serv. Comm’n 1980) (abandoned nuclear plant amortized over ten years); Bangor Hydro-Elec. Co., 46 P.U.R.4th 503, 556-557 (Me. Pub. Utils. Comm’n 1982) (cost of cancelled nuclear plant, exclusive of AFUDC, amortized over five years); Detroit Edison, Nuclear Reg. Rep. (CCH) ¶ 20,035, at 16,223-16,224 (Mich. Pub. Serv. Comm’n 1976) (cost of abandoned plant amortized over ten years); Atlantic City Elec. Co., 51 P.U.R.4th 109, 115 (N.J. Bd. of Pub. Utils. 1983) (utility’s 5% ownership in abandoned nuclear plant amortized over fifteen years on a straight-line basis); Caroline Power & Light Co., 49 P.U.R.4th 188, 217-219 (N.C. Utils. Comm’n 1982) (investment in cancelled plant recoverable over a ten-year period); Pub. Serv. Co., Cause No. 27068, Order No. 206560, at 59 (Okla. Corp. Comm’n, Jan. 15, 1982) (utility entitled to recovery over ten years “on a straight line basis” of net prudent investment in abandoned nuclear plant); Virginia Elec. & Power Co., 29 P.U.R.4th 65, 76-77, 81 (Va. St. Corp. Comm’n 1979) (prudent costs of cancelled nuclear plant, including capitalized AFUDC, amortized over ten years); Central Vt. Pub. Serv. Corp., 49 P.U.R.4th 372, 391-392, 394 (Vt. Pub. Serv. Bd. 1982) (net loss in Pilgrim II, 1.73 % of which was owned by the utility, amortized as a cost of service over ten-year period). For opinions contra, see n.11 below.
The Federal Energy Regulatory Commission has allowed utilities to amortize over ten years the prudent costs, including the equity portion AFUDC, of an abandoned nuclear plant, without any return on the amortized balance. Northern States Power Co., 17 F.E.R.C. ¶ 61,196, at 61,379 (1981). A utility seeking Federal Energy Regulatory Commission approval of the cost of abandoned nuclear plant must show the expenses were prudent, and, if they were, the Federal Energy Regulatory Commission has allowed recovery. See Anaheim, Riverside, Banning, Colton, & Azusa v. Federal Energy Regulatory Comm’n, 669 F.2d 799, 809 n.17 (D.C. Cir. 1981).
Following the decision of the Supreme Court of Ohio construing the Ohio statutes as forbidding direct recovery of costs of cancelled nuclear plants (Office of Consumers’ Counsel v. Public Utils. Comm’n, supra), the Ohio Public Utilities Commission, of course, had to follow the court’s decision. However, the commission indicated that but for the constraint *225of the court’s interpretation it would approve continued amortization. Cleveland Elec. Illuminating Co., No. 81-146-EL-AIR, No. 81-1565-EL-UNC at 27 (Pub. Utils. Comm’n of Ohio 1982). In deciding this subsequent rate proceeding, the commission further stated that the return it allowed “provides revenues sufficient to provide for amortization of [the unamortized] balance over a reasonable period of time.” Id. at 28.
The Public Service Commission of Wyoming denied recovery of an investment in abandoned nuclear power plants, relying on a statutory provision that it consider “the property and business of any public utility, used and useful for the convenience of the public” (emphasis supplied). Wyo. Stat. § 37-2-119 (1977). Lower Valley Power & Light, Inc., No. 9617-SUB11, at 9 (Pub. Serv. Comm’n of Wyo. Dec. 2, 1982). It concluded, both as a matter of statutory construction and in its discretion, that, if utility plant is not used and useful in rendering service to the public, the investment cannot be included in the rate base or recovered in some other way, even if the investment was prudent. Id. at 11, 16. In Portland Gen. Elec. Co., 49 P.U.R.4th 274 (Or. Pub. Util. Comm’r 1982), the company apparently agreed that no costs attributable to the abandonment of a nuclear power plant would be reflected in the rates. Id. at 277. When the issue came up in the State of Washington whether to allow amortization of the Washington share of the Pacific Power and Light Company’s minority interest in the abandoned plants which had been the subject of the decisions in Oregon and Wyoming, the Washington commission concluded that the investors’ rate of return would have to be increased because of the perceived greater risk to investors and, therefore, to allow amortization of costs of the terminated project would “constitute a double recovery of the Washington ratepayers’ portion of investment.” Washington Utils. & Transp. Comm’n v. Pacific Power & Light Co., 51 P.U.R.4th 158, 168 (Wash. Utils. & Transp. Comm’n 1983). However, the commission allowed an additional 2.5% return on common equity above the 16% otherwise approved because of the “lack of direct recovery for terminated projects, the financial health of the company, and . . . investor assessment of the risk.” Id. at 182.
Other agency decisions denying recovery of the costs of abandoned nuclear facilities have also involved special circumstances. Arizona Pub. Serv. Co., 38 P.U.R.4th 547, 556 (Ariz. Corp. Comm’n 1980) (company did not meet burden of proving prudence of expense; Arizona ratepayers should not bear expense caused by regulatory conditions in California leading to abandonment of the project); Northern States Power Co., 42 P.U.R.4th 339, 360-364 (Minn. Pub. Utils. Comm’n 1981) (recovery of costs of abandoned plant in Minnesota allowed; no recovery allowed for costs of abandoned plant in Wisconsin which the Wisconsin Public Service Commission, acting “in a parochial fashion,” caused to be abandoned); *226Nevada Power Co., 41 P.U.R.4th 367, 386 (Nev. Pub. Serv. Comm’n 1980) (costs of abandoned plant already set off and recovered against excessive earnings should not be amortized); Northern States Power Co., No. 10,097, at 7-8, 10 (Pub. Serv. Comm’n of N.D. Dec. 31,1980) (losses from cancellation of generating plant in Minnesota amortized; cost of abandoned plant in Wisconsin not to be amortized while decision of Federal Energy Regulatory Commission awaited as to allocation of costs among utilities).
Some State regulatory agencies have allowed a carrying charge, or even rate base treatment, on the unamortized balance. See, e.g., Gulf States Utils. Co., 40 P.U.R.4th 593, 595 (La. Pub. Serv. Comm’n 1980) (unamortized cost of abandoned plant included in rate base); Rochester Gas & Elec. Corp., Case No. 27794, Op. No. 82-1, at 32, 71 (N.Y. Pub. Serv. Comm’n Jan. 12, 1982) (utilities entitled to carrying charge on the unamortized prudent investment in abandoned nuclear plant, measured by each utility’s over-all cost of capital); Carolina Power & Light Co., 49 P.U.R.4th 188, 217 (N.C. Utils. Comm’n 1982) (portion of unamortized balance of the cost of cancelled plant supported by long term debt included in rate base); Public Serv. Co., Cause No. 27068, Order No. 206560, at 60 (Okla. Corp. Comm’n Jan. 15, 1982) (return allowed on debt and preferred portion of the amortized loss, but not on the equity portion).
Other State regulatory agencies have denied rate base treatment or a carrying charge on the unamortized portion of the allowed prudent costs. See, e.g., San Diego Gas & Elec. Co., 31 P.U.R.4th 435, 449 (Cal. Pub. Utils. Comm’n 1979) (rate base treatment denied for abandonment costs of nuclear plant); Bangor Hydro-Elec. Co., 46 P.U.R.4th 503, 557 (Me. Pub. Utils. Comm’n 1982) (no rate base treatment of amortized cost of abandoned plant); Detroit Edison Co., Nuclear Reg. Rep. (CCH) 20,035, at 16,224 (Mich. Pub. Serv. Comm’n 1976) (same); Atlantic City Elec. Co., 51 P.U.R.4th 109, 115 (N.J. Bd. of Pub. Utils. 1983) (no carrying costs allowed on net loss amortized over fifteen years); Central Vt. Pub. Serv. Corp., 49 P.U.R.4th 372, 394 (Vt. Pub. Serv. Bd. 1982) (amortized net loss not to be included in rate base); Virginia Elec. & Power Co., 29 P.U.R.4th 65, 81 (Va. St. Corp. Comm’n 1979) (no return allowed on net loss amortized over ten years); NEPCO Mun. Rate Comm’n v. FERC, 668 F.2d 1327, 1332-1335 (D.C. Cir. 1981), cert. denied, 457 U.S. 1117 (1982) (court upheld the Federal Energy Regulatory Commission’s decision not to allow inclusion in the rate base of the costs of an abandoned oil-fired electric generating facility, to be amortized in the rates over five years).
The only exceptions of which we are aware are those decisions in which the agency regarded the result as dictated by statute. See the Wyoming and Ohio decisions cited in n.11 above.
In Western Mass. Elec. Co., D.P.U. 558, at 39-41 (July 31, 1981), the department allowed the utility to recover over four years the prudently incurred costs of an abandoned nuclear plant, excluding the equity rate of return portion of AFUDC. The Attorney General, a party in that proceeding, did not argue that the recovery of such costs was unlawful. At least, the department’s discussion of the problem proceeds on the assumption that all parties agreed that a Massachusetts utility could lawfully recover at least some portion of the prudent costs of an abandoned nuclear plant. In Commonwealth Elec. Co., D.P.U. 956, decided on May 28, 1982, after the department’s decision in this case, the Attorney General apparently did not argue that recovery of some portion of the utility’s prudent costs of abandoned nuclear plants (including Pilgrim II) could not lawfully be allowed. In the Commonwealth Elec. Co. case, the company did not request carrying costs at least on its investment in Pilgrim II, and the department, considering the company’s particular circumstances, allowed the write-off of the costs over two years for Pilgrim II and over three years for nuclear units abandoned in Montague. Id. at 7-9, 15-16.
We find nothing in G. L. c. 164, § 94G½, inserted by St. 1980, c. 464, concerning the allowance of an oil conservation adjustment in the rates of an electric company that fairly can be construed as providing a restriction on the broad power of the department to exercise its discretion to allow recovery of its prudent investment in a prudently abandoned nuclear power plant.
Of course, the department’s decision hardly constitutes a guaranty that a utility will sustain no loss in abandoning such a project. The project must be prudent at its inception and prudently abandoned. At the most, a utility may recover for investments that were not mistaken at the time they were made. Even then, there is a risk that the department will disallow recovery of all or, as here, some part of such an investment.
The 14% return allowed in this proceeding was reduced to 9.3% in the recently concluded rate proceeding. See Boston Edison Co., D.P.U. 1350, at 64 (May 31, 1983).
This is another way of stating that 73.1% of the after tax loss was allocated to ratepayers and 26.9% of the after tax loss was allocated to shareholders.
Boston advances a separate argument. It challenges the deduction from the deferred tax reserve of all the deferred tax benefits resulting from the abandonment of the project, thus increasing the rate base. It contends that the portion of deferred tax benefits attributable to the shareholders (26.9 %) should not have been removed from the rate base calculation. If, however, that portion of deferred tax benefits had been left in the rate base calculation, the after-tax loss to be allocated between ratepayers and shareholders in the calculation of the amount to be amortized would have had to have been correspondingly adjusted. Boston has not demonstrated that the change it proposes would have made a significant difference if all the appropriate adjustments were made. In any event, the department’s treatment of the deferred taxes as part of the total tax benefits was reasonable.
One combined class was assigned no increase because its rate of return was 20% above the average.