Boston Gas Co. v. Department of Public Utilities

Cuttee, J. Boston Gas Company (Company) filed with the Department of Public Utilities (D.P.U.) on February 12, 1969, new rate schedules to become effective March 1, 1969. The D.P.U. suspended these rates until January 9, 1970, pending investigation. On January 8, 1970, it disallowed the rates, which originally had been designed to produce additional gross annual revenues of $3,493,600. 1 It, however, permitted Company to file hew rates to produce additional annual gross revenues . of $1,974,000. Later (because of D.P.U. mistakes in computations) a new order (March 4, 1970) permitted the additional gross revenues to be increased to $2,021,022 *295 (see fn. 1). Company then filed its appeal in the county court under G. L. c. 25, § 5 (as amended through St. 1956, c. 190). 2 A single justice reported the case for the determination of the full court upon the record before the D.P.U. and certain papers filed in the county court. Upon Company’s petition, another single justice (June 24, 1970) directed that additional evidence 3 be transmitted to this court (but without any “determination by . . . [the county c]curt that . . . [such] additional evidence is properly includible in” the record). General Background. Company, a wholly owned subsidiary of Eastern Gas & Fuel Associates, serves thirty-two cities and towns (including Boston) running in a strip from Groton and Shirley on the northwest to Rockland and Whitman on the southeast. It obtains natural gas from the pipelines of Algonquin Gas Transmission Company. All but about two per cent of the gas sold by Company in 1968 was natural gas. The small amount of manufactured gas was used to meet winter “peak” demand. In the future, for such “peak” needs, Company plans to rely on liquefied natural gas (LNG) stored by it. Company’s last rate increase was in 1956. It avoided increases from 1956 to 1969 (at a time when most other prices were rising rapidly) by greater use of labor saving machinery, by increased employee productivity, by efforts to increase sales, and because of the savings effected by the introduction of natural gas in the 1950’s. In 1969, an application for higher rates was induced by continuing in *296 flationary pressures, increases in wages and operating expenses, the need of additional plant investment, and other factors. The amount of the requested increase was influenced (see fn. 1) by Company’s competition with other utilities and (especially in the home heating market) with the “unregulated fuel oil industry.” Home heating has become an increasingly large part of Company’s business. Before the D.P.U. both Company and counsel for the department dealt with the issues as being (a) Company’s proper rate of return and its rate base; and (b) Company’s cost of rendering the services which produce its revenues. In the late autumn of 1969, at the hearings, 1968 was the most recent year for which complete figures were available and was used as a “test year” (see fn. 23, infra). The D.P.U., however, made certain adjustments of the 1968 figures, the propriety of which Company disputes. Rate of Return and Cost of Capital. 1. On December 31, 1968, Company’s long-term debt consisted of $27,305,000 of first mortgage bonds, due in. 1990, and $16,000,000 of short-term debt. In 1969 the short-term debt was converted into $15,000,000 of three-year notes. This conversion gave Company, at the time of the hearings, a capital structure of 53.7% debt and 46.3% common stock equity. It was then anticipated that soon there would be $14,000,000 of new debt financing to meet Company’s planned construction expenditures in 1969-1970. If this amount of financing had taken place, Company would have had 60% debt and 40% equity. 4 The D.P.U. determined that it was prudent for Company to continue a “policy of financing by debt” and proceeded *297 to fix fair return by adjusting Company’s outstanding debt ratio to an assumed 60% (with a 40% equity ratio) in order to reflect the expected cost of additional debt capital soon to be acquired. This the D.P.U. said was so as not to “predicate a finding of fair rate of return on a capital structure and cost of capital which will not actually exist during the period when the proposed rates are expected to be in effect.” The D.P.U. thereupon computed the composite cost of debt to Company as follows: Issue Net Cost Rate Annual Charges First Mortgage 4.68% (Actual) $1,228,734 Bonds issued 1965, due 1990 ($26,255,000) 3-year Notes 8.00% 1,200,000 ($15,000,000) Anticipated 1969-70 8.00% (Estimated) 1,120,000 ($14,000,000) On this expected debt of $55,255,000, the annual charges anticipated by the D.P.U. were $3,548,734 with an expected composite debt cost rate of 6.42%. On January 8, 1970, when the D.P.U. announced its rate decision, it thus expected that Company would be able to borrow at 8%. Later financing, however, did not work out as well as had been anticipated. Company in fact was authorized by the D.P.U. on March 25, 1970, to redeem its $15,000,000 of notes from the proceeds of $20,000,000 of additional first mortgage bonds, payable in not more than twenty-five years from the date of issue. On April 23, 1970, the D.P.U. approved additional provisions of the new bonds, including an “effective interest cost of the bonds to” Company of 9.73%. 5 *298 Company thereupon (as has been stated) obtained in the county court an order directing the D.P.U. to transmit to this court the evidence concerning the unexpectedly high cost and more limited amount of the new 1970 debt. If Company’s actual outstanding debt, after the new bond issue, had been employed in arriving at the rate decision (in computing Company’s cost of debt and debt ratio) the computation would have been as follows: Mortgage Debt Net Cost Rate 1965 issue $26,255,000 4.68% 1970 issue $20,000,000 9.73% or a composite debt rate of 6.86% on $46,255,000 of mortgage debt (constituting 56% of Company’s capital), instead of the 6.42% composite rate (on $55,255,000) estimated by the D.P.U. in its rate decision. 2. Company contends that, in fixing the cost of debt for rate purpose, the new composite 6.86% cost of debt rate on Company’s total mortgage debt of $46,255,000 should be used, and that Company should be treated as having the 56%-44% debt-equity capital structure then in effect (with D.P.U. acquiescence) after the 1970 bond issue. The D.P.U. now argues that this court should not consider the changed circumstances which arose after the D.P.U.’s rate decision of January 8, 1970, and, indeed, even after the modification of that decision on March 4, 1970. 6 For us to remand the case for new hearings, to adduce the new evidence before the D.P.U., would result in un- *299 desirable additional delay of these proceedings already more than two years old. 7 The limited additional evidence (fn. 3, supra) relates to matters of public record contained in the D.P.U.’s own files. The 1969 Company annual return (see G. L. [Ter. Ed.] c. 164, § 83) merely adds the return of that year to the series of such returns (beginning in 1964) already in evidence before the D.P.U. The two 1970 financing decisions reveal that both parties at the rate hearings, and the D.P.U. itself in its rate decision, in fact made an inaccurate “prophecy” (see Lowell Gas Co. v. Department of Pub. Util. 324 Mass. 80, 89), concerning the amount of the new bond issue and the necessary interest rate. It is now possible and proper for us (since the “rate order is challenged on constitutional grounds in this court”) and for the D.P.U., to examine and evaluate the current facts with the aid of “the improved vision of intervening experience.” See Boston Consol. Gas Co. v. Department of Pub. Util. 327 Mass. 103, 104-107; Opinion of the Justices, 328 Mass. 679, 686-690 (where the Justices said that this court, with respect to constitutional issues, in performing its duty to afford “a utility claiming confiscation ... an independent judicial review as to both law and fact,” must have authority to require the production of “new evidence necessary to bring the proof as nearly as reasonably possible down to the date of final decision” 8 ). See also New England Tel. & Tel. Co. v. *300 Department of Pub. Util. 331 Mass. 604, 606 (additional evidence submitted to the county court by stipulation). Federal cases, considering to some extent evidence becoming available during the pendency of an appeal from an administrative decision, include Pittsburgh v. Federal Power Commn. 237 F. 2d 741, 751-753 (Ct. App. D. C.), and Scenic Hudson Preservation Conference v. Federal Power Commn. 354 F. 2d 608, 617-620 (2d Cir.). See Algonquin Gas Transmission Co. 43 F. P. C. 53, 59-61. Nothing in G. L. c. 25, § 5, 9 compels us to delay proceedings by remanding the case (in advance of our decision) to the D.P.U. for reopened hearings on the additional evidence (fn. 3). Particularly is this so in view of the character of that additional evidence and the ease with which it may be verified. To the extent that our decision is affected by the additional evidence, 10 we remand the case to the D.P.U. to verify its accuracy, and to make appropriate corrections in the rate decision based on that evidence. We assume that, as a practical matter, the existence of the D.P.U. *301 decisions on the 1970 bond issue (and of the facts thereby determined) cannot be questioned. On that assumption, we state the applicable principles of decision. 3. Company’s cos£ of long-term debt is now fixed, with the D.P.U.’s approval, for a considerable period of time. On its incurred debt, it must pay a composite interest rate of 6.86% on $46,255,000 of bonds instead of a composite rate of 6.42% on $55,255,000 of bonds and notes. The changes in Federal income tax and in the Massachusetts franchise tax caused by the new circumstances can easily be computed and are to be reflected in the new rates. Company prudently did not attempt to put out (at the high interest rates and unfavorable market conditions prevailing in April, 1970) new bonds sufficient to increase its debt ratio to 60%. 11 The record shows that Company increased its debt ratio (largely through retirement of 75,000 shares of common stock) from 20% in 1963 to 56% in 1970. It would be unreasonable, and an undue interference with reasonable Company judgment, for the D.P.U. to insist that Company’s rate of return conform with precision to what the D.P.U. regards as an optimum 60% debt ratio. Within a substantial range this is a matter for Company’s determination. Market conditions (recognized by the D.P.U. itself, see fn. 11) have prevented Company from reaching a 60% figure. There is no evidence that (in disregard of fair and prevailing utility practice or of D.P.U. admonitions) Com- *302 pony has adopted an unreasonably low debt ratio which may be regarded as a “company luxury” imposing an undue burden on consumers. Cf. New England Tel. & Tel. Co. v. Department of Pub. Util. 331 Mass. 604, 618-619. Company’s debt cost appears to have been prudently incurred. In the circumstances we conclude that “[Y]s to debt capital . . . [C]ompany is entitled to earnings which will pay its actual interest obligations ... on existing debt capital and upon new debt capital required in the immediate future.” See New England Tel. & Tel. Co. v. Department of Pub. Util. 327 Mass. 81, 88. 12 The D.P.U. is to adjust its calculation of the cost of Company’s debt in a manner consistent with the foregoing discussion. Cost of Equity. 4. An expert (C. J. McCarthy) called by Company estimated the reasonable cost of its equity capital to be a return of 12% to 12.5%. The D.P.U. called a witness (David A. Kosh) whose estimate was a return of 10.5%. The D.P.U. found that cost to be 10.75%. Because Company’s stock is all held by its parent association, no assistance is rendered by any appraisal of the shares in the open market for, of course, no shares are bought and sold. The problem is one of determining, as fairly as can be done, the return on the basis of which Company can sell shares of its common capital stock (currently or within a reasonable time). Such a determination must take into account (a) book value, (b) current and prospective earnings and dividends, and (c) other relevant factors. *303 The difficulties of such a determination are great. See Wannacomet Water Co. v. Department of Pub. Util. 346 Mass. 453, 465-470. It is hard to find other companies which provide a satisfactory basis of comparison with a wholly owned subsidiary of a large company. Moreover, in cases like this one, there cannot be undue reliance on the probability that the “parent company . . . [will] purchase the shares” if any are offered. The “parent company ... is not bound to take any new stock and is as much entitled as any other potential investor ... to have . . . [C]ompany earn a fair return.” See Salisbury Water Supply Co. v. Department of Pub. Util. 344 Mass. 716, 722-723. 5. The principles governing the appropriate rate of return on equity capital were stated generally in Federal Power Commn. v. Hope Natural Gas Co. 320 U. S. 591, 603, as related to the investor’s ‘ ‘ legitimate concern with the financial integrity of the [regulated] company.” From the investor’s standpoint there should “be enough revenue not only for operating expenses but also for . . . capital costs,” including debt service and dividends. 13 The two principal expert witnesses on the issue of the rate of equity return adopted differing methods of reaching their conclusions. McCarthy (called by Company) used a “comparable earnings” test (see e.g. Permian Basin Area Rate Cases, 390 U. S. 747, 806) viz., using “the rate of earnings on book value of common equity capital” of comparable companies as a basis of assigning a generally similar *304 rate to Company. 14 He then suggested as a comparable rate of return on common book equity for Company the range 12% to 12.5%, described by him as "below that of . . . seven of the eleven . . . companies and above that of the other four.” He tested the fairness of the comparison in various ways. 6. Kosh employed what he called the "discounted cash flow” (DCF) method, which the D.P.U. in its opinion described as assuming "that the price at which an investor purchases stock is based on his expectations as to the future yield in terms of dividends and sales price of the stock. By discounting the [expected] future yields at a rate which equalizes the present value of the future yields and the market price of the stock, the method arrives at a rate of return that the hypothetical investor is demanding on the given common stock issue. Calculation of this rate can also be obtained by adding the current dividend yield to the expected rate growth in dividends.” In applying this DCF formula, Kosh needed "an estimate of growth and a current dividend yield.” 15 Since Company’s "stock ... is not *305 traded,” he selected a number of gas distribution companies, which he regarded as similar to Company, for comparison (by computation of their cost of equity). 16 From his assumptions and on the basis of his computations from the figures for these companies deemed by him to be comparable, Kosh reached his conclusion that Company should have a 10.5% return on its equity investment. Company takes issue with Kosh’s methods or conclusions on various grounds outlined in the margin. 17 In the circumstances before us, we are not persuaded that D.P.U. should be required to rely upon any particular group of comparative figures employed to test (a) what would be the market price of a stock not traded, if it were to be traded, and (b) what return on such a stock would be required to permit its issuer to sell it to prudent investors. Such comparisons usually can be no more than general guides to be appraised by the D.P.U. in considering the fairness of rates, and by us in determining whether the rates allowed by the D.P.U. will result in confiscation, a matter properly for our concern. See New England Tel. & *306 Tel. Co. v. Department of Pub. Util. 331 Mass. 604, 612, 617. We recognize that Kosh’s assumptions and his use of selected and supposedly comparable data in connection with the complicated DCF formula 18 may well be open to challenge. The evidence as a whole, however, is sufficient to permit a decision on the issue of the proper cost of equity (a matter not susceptible of very precise determination). See Wannacomet Water Co. v. Department of Pub. Util. 346 Mass. 453, 465-471. On the present record, we are of opinion that Company has not thus far established the DJP.U.’s allowance of only a 10.75% cost of equity capital, by itself, to be below the range of permissible judgment and thus to be confiscatory. The present record, however, essentially ends with 1968, except for the limited additional evidence (already discussed, see fn. 3). Company’s 1969 report has not been fully analyzed in the record and briefs. The 1970 report should be on file with the D.P.U. by the time this opinion is .published. As in any event there must be reconsideration by the D.P.U. of the allowed rates on other aspects, it is appropriate that the adequacy of the 10.75% rate of equity return also be reexamined in the light of the 1969 and 1970 figures. This reexamination may be pertinent to reconsideration of the “attrition” problem next discussed. *307 Attrition. 7. Company contends that the D.P.U. improperly denied any allowance for “attrition” in Company’s rate of return. 19 The D.P.U. allowed Company’s request for a ruling that, where “no allowance is made by other means for demonstrated attrition of a utility’s economic ability to earn a fair return, a fair return must include an allowance in excess of the utility’s cost of capital.” The D.P.U. thus recognized the principle that an attrition allowance may be necessary. In its decision, the D.P.U. said, in part, “If the existence of attrition were demonstrated, proper regulation might require some adjustment reasonably related to the rate of attrition.” It found in the record, however, (a) no evidence of “relationship between the amount of attrition and the amount of additional revenues that would be produced by use of the year-end rate base” for which Company contends; and (b) no indication of attrition “for a significant period that would not be covered by adjustments to certain . . . expense accounts.” It concluded that no attrition allowance was justified “until it is demonstrated that the new [Company] policy of large additions to plant will not produce equivalent increases in income or reductions in expenses.” To support this conclusion there were no substantial subsidiary findings concerning the *308 past and expected relative growth in Company’s earnings, expenses, and plant, except as these may have been implied by the D.P.U.’s substantial acquiescence in testimony of a witness called by it. Cf. Westborough v. Department of Pub. Util. 358 Mass. 716, 717-718, and cases cited. Company rests its attrition contention on the actual figures with respect to its greatly increased expenditures for additional plant beginning in 1965 and those projected for the period 1969 to 1971. These were introduced through William J. Pruyn, a vice-president of Company, and, until September, 1969, its treasurer and chief accounting officer. Pruyn pointed out, as did another witness, that “when natural gas became available ... fix the mid-1950’s, the doubled heat content of natural gas, as compared with manufactured gas, virtually doubled the capacity of . . . ^Company's] distribution system” so that it was “able to live off this system with only modest capital additions while servicing substantial load growth for a number of years.” About 1965, “it became apparent that . . . load was growing at such a fast pace that . . . substantial capital investment in . . . ^Company’s] distribution system” would be required for efficiency and economy. This led to substantial net additions to plant, set out in the margin 20 . Analysis of this recent trend led Pruyn to predict for the period 1967 to 1970 the following disparities *309 among rates of growth of (a) revenues, (b) operating expense, and (c) net plant, which to him showed a distinct attrition or erosion of return: Growth Rates — 1967 to 1970 1967 Actual (000) 1970 Projected (000) Annual Average Rate of Growth Revenues $49,918 $61,483 7.2% Operating Exp. 45,322 56,788 7.8% Net Plant (12-31) 63,485 91,977 13.2% The witness concluded that the disparity was such that “if rates are made that will ‘allow’ . . . Company to earn an 8% return on 1968 average rate base . . . Company will Qn fact] earn only a 6.4% return on its average rate base in 1970.” To meet this erosion of return, Pruyn proposed the use of a rate base as of the end of the test year (see fn. 19, supra) as a “workable approximation” adjustment, counteracting attrition, which has. been “widely accepted.” 21 A witness called by the D.P.U. (Dennis R. Bolster, an associate of Kosh) testified that, from 1961 to 1967, Company’s “revenues, expenses, and plant investment . . . *310 tended to remain in balance, each growing approximately 4.0% per year.” The large increase of plant in 1968 he regarded as an “isolated aberration.” He recognized that, “in the past two or three years,” property taxes “have grown at the rate of 12.7% per year.” For this he made a special allowance in 1968 test year “operating expenses for known increased property taxes.” From his examination of the 1961-1967 figures, he concluded that, apart from the property taxes, there was “no evidence of attrition.” He felt that use of a year-end rate base “simply adds to the rate base an amount equal to one-half of the net plant additions during the test year” without more than coincidental relevance or relationship to any future “imbalance in investment, revenues, and expenses.” He suggested that one method of making allowance for attrition would be “continuous regulation” making rate adjustments (by discussion between the regulatory body and Company) with some frequency. On cross-examination, it was disclosed that Bolster had given slight weight, if any, to predictions (either Pruyn’s or his own) of continuing plant additions and disparate operating results for 1969 and 1970. The D.P.TJ. also, in effect, has given no weight to whatever erosion in contemplated return to Company w|s indicated by the evidence of actual and expected results for 1969 and 1970. In part, failure to do so may have been based upon the D.P.U.’s distrust of the feasibility of predicting (during the 1969 hearings) the results for 1969 and 1970. In part, it may have been reluctance to adjust the actual and adjusted results of the test year by the imprecise remedy (for attrition) of using a year-end rate base instead of an average test year rate base. The former difficulty no longer exists. The actual results for 1969 and 1970 are now known and Pruyn’s predictions for those years can now be tested accurately. If erosion is shown to have taken place, and a year-end rate base is deemed unsuitable, some other suitable form of compensatory attrition adjustment can be made. *311 The substantial evidence of attrition in recent years (i.e. since 1965) adduced by Company leads us to remand this issue (of attrition) to the D.P.U. for further consideration and for suitable adjustment of the rates (allowed by the decision of January 8, 1970) to reflect any attrition trend which may be confirmed by the 1969 22 and 1970 figures. This reexamination, as has already been indicated, should take place in connection with reconsideration of the proper cost of equity capital and should include subsidiary findings of the basic facts bearing upon whether attrition exists and is likely to continue for the future. Cost of Service. 8. Company contends that in various respects the D.P.U. has underestimated, with respect to 1968 (the “test year”), the cost to Company of rendering service and thus has understated Company’s revenue needs. Two of these contentions, relating to the amount of (1) interest deduction and (2) tax expense, depend upon the “cost of debt” issue, already discussed, and at further hearings before the D.P.U. must be reconsidered in connection with the “cost of debt” issue (see part 3 of this opinion, supra). The Transition to LNG Facilities for “Peak-Shaving.” 9. During 1968, Company put into operation a new “peak-shaving” liquefied natural gas (LNG) facility to supply supplemental gas (in addition to that normally drawn from pipelines) during cold winter days of peak demand. Theretofore, to meet this peak demand Company had been using an obsolescent manufactured gas plant. A decision to build the LNG plant was made in 1966. Although it began operation late in 1968, only minor quantities of liquefied gas *312 were used during the test year, 23 which (with respect to these peak-shaving facilities) was a transition year. The D.P.U. took the position that, in view of the LNG transition, adjustments to test year figures for rate base and for operating results were necessary to make 1968 figures reflect (and serve as an appropriate guide to) the earnings situation likely to exist in 1970 and thereafter when the new rates would be in effect. The D.P.U. rejected methods of adjustment of the “test year” proposed by Company and Bolster (fn. 23). Instead, it adopted a highly complicated adjustment method of its own in an effort to avoid possible distortion because of the transition. Company contends that the D.P.U. (in applying this complicated method) made several errors seriously adverse to Company’s position and greatly understating Company’s revenue needs. The problem arose from the circumstance “that the LNG plant and the manufacturing gas facilities were both in service during the . . . test year.” This situation probably was not going to continue long into 1969, although both plants were needed (as a precaution at least) to deal with peak loads in the winter of 1968-1969. The D.P.U. wished to reflect eventually the situation which would have existed if only the LNG method of peak shaving had been employed. Accordingly, the D.P.U. decided initially (1) “to eliminate [the LNG investment] from the test year rate base” and (2) to “relate . . . test year operating results to the actual plant which produced them,” by (a) separating from 1968 operating results the effects which the LNG plant had on gas costs, plant operation, and, maintenance expenses, and (b) reflecting in such results the additional expenses which would have been incurred if Company in 1968 had not been expecting to operate with LNG in 1969 and later years. *313 The result of these steps was to create a hypothetical no-LNG year. To give a more accurate indication of conditions likely in an anticipated all-LNG year, when the proposed rates would be in effect, a third step was necessary: viz., to reflect the savings in cost of service which LNG operations were expected to bring. There are probably reasonable objections to any method of adjustment of test year data. Probably also, no method suggested can serve as more than a general guide in determining the amount of appropriate adjustment. If the D.P.U. method is used, however, it must be employed fairly and in a manner which, so far as possible, does not overstate or understate Company’s revenue needs. We are of opinion that the D.P.U. adjustments of the test year data (with respect to “peak shaving”) have not been adequately explained (and shown to be supported by substantial evidence) in the following respects about which Company complains. A. Company contends that, in transforming 1968 into a hypothetical no-LNG year, the D.P.U. should have included at least $202,000 as a cost of carrying a hypothetical investment of $900,000 24 for an additional plant for manufacturing gas. Such an investment would have been necessary and should have been made in 1967 if the LNG plant had not been planned for use in time to meet the 1968-1969 winter peak demands, because the obsolescent manufactured gas plant would not have been adequate to meet those demands. Accordingly, Company says that, if the LNG plant is to be treated (as it was by the D.P.U.) as wholly removed from test year rate base, (a) there must be substituted (in the hypothetical no-LNG rate base) the cost of the necessary additional manufactured gas facilities, and (b) there should have been included in the expenses of the hypothetical *314 no-LNG year the carrying charges of that additional theoretical investment, if the theoretical year is adequately to reflect Company’s revenue needs. We are persuaded that, in applying its adjustments, the D.P.U. has not taken adequate account (as required by its own theory of adjustment) of the benefit accruing to Company during the test year from its investment in the LNG plant. The plant was available to meet the winter (1968-1969) demands which, as it happened, arose principally in the early months of 1969, after the test year, but (if weather had been more severe) could have occurred in November and December of 1968, as well as in January-March, 1970. B. Company contends that, in adjusting 1968 operating expenses to a no-LNG basis, the D.P.U. has inadequately reflected the savings (in 1966 and thereafter) in the expense of manufactured gas production and storage made possible only by the prospective availability of the new LNG facility, and thus properly to be taken into account as an expense in the hypothetical no-LNG year. The D.P.U. did recognize that Company made savings in maintenance and other expenses connected with the obsolescent manufactured gas facility because Company knew that the manufactured gas facility, soon to be supplanted by LNG, need not be maintained at the levels required if it had to be kept in operation for a substantial period. The D.P.U. gave effect to this saving, however, only in terms of the difference between such maintenance expenses for 1967 and those for 1968 (which were less than those of 1967 by about $163,000). Company contends that even the 1967 expenses (and possibly also the 1966 expenses) for this purpose were significantly below the level which would have been necessary if Company had not in 1966 planned on the use of the LNG plant at least as soon as late 1968. We think that there is sufficient evidence that these savings began earlier than 1967 to require that the D.P.U. reexamine this aspect of the case, make proper subsidiary findings on the issue, and modify the terms of its adjustment appropriately. C. Company contends that, in computing the savings to *315 be expected in an all-LNG hypothetical test year, the D.P.U. did not include (for purposes of comparison with the hypothetical no-LNG year) any return on the $3,794,000 of Company’s investment in manufactured gas facilities which had not been amortized (or depreciated) by the beginning of 1968. It would seem that this item (i.e. carrying any unamortized balance of a prudent investment in an obsolete facility after it has been supplanted by a modern facility) would be proper for consideration in some manner in comparing the results of a hypothetical all-LNG year with those of a hypothetical no-LNG year. Precisely what the D.P.U. has done with this item in its comparison is not clear from its decision or from the D.P.U.’s brief. The computation and comparison must be reexamined, and any discrepancy appropriately corrected, with a clear explanation of the relevant steps in the comparison. D. Company contends that the D.P.U. has overstated the savings in gas costs to be realized from an all-LNG peak-shaving operation (as compared with a no-LNG peak-shaving method) by computing the savings on the 1969-1970 level of gas sales rather than on the lower sales level of 1968, the test year. Bolster (fn. 23) calculated the savings for 1968 at $408,173 (the reduction of the cost of purchased natural gas of $596,673 less $188,500 for LNG purchases). On the other hand, the D.P.U. in its decision, states that it took its figures of saving (about $967,000) from Company’s exhibit B in another proceeding (D.P.U. 15,513). This exhibit, Company asserts, was based on 1969-1970 heating year estimated gas volume wrhich was significantly greater than the gas volume of 1968. The pertinent evidence from the earlier case, in which exhibit B was introduced, should be reexamined and put in evidence in the present case. If the gas cost savings shown by exhibit B are in fact based upon ail969-1970 gas sales volume higher than that of the test year, there should be a suitable adjustment of the costs of the theoretical all-LNG year for comparison with those of the theoretical no-LNG year, so that the LNG estimated savings will not be overstated. *316 Conclusions on Company’s Objections to the D.P.U. adjustments to test year results because of the “peak-shaving” transition. We conclude (1) that the record contains substantial support for each of Company’s objections (summarized above) to the D.P.U. adjustments of test year results (to avoid distortion because 1968 was a transition year), and (2) that the test year results (as thus adjusted by the D.P.U.) may substantially understate Company’s revenue requirements. These adjustment issues are to be reconsidered by the D.P.U. in the light of this opinion, after further hearings if necessary. A new statement of the adjusted test year results is to be made with (a) an adequate explanation of (i) the source of the basic figures, and (ii) the reasons for the adjustments affecting the transition problem here discussed; and (b) subsidiary findings based upon the evidence concerning such adjustments. Alleged Understatement op Amortization op Conversion Expense. 10. During 1968, Company charged $625,935 to expense as amortization of gas conversion expense incurred upon the introduction of natural gas (instead of manufactured gas) when there was a heavy expenditure (apparently in 1955-1961) for converting customer appliances. The D.P.U. by order had allowed Company to amortize over a. ten year period this aggregate sum (built up over several years). At the beginning of 1968, the unamortized book balance of this conversion expense was $1,036,950. At the end of 1968, the unamortized book balance was $411,015. There was evidence that Company expected to write off about $250,000 of this balance in 1969, and most of the remainder in 1970. The 1969 return shows that this was an accurate estimate of the 1969 write-off. The D.P.U. in its decision refused to include the 1968 charge ($625,935) as a 1968 expense for test year purposes. The D.P.U. took the position (a) that Company in fact had recovered the $625,935 in 1968 (not mentioning, however, *317 whether 1968 rates were inadequate) and (b) that the balance of $411,015 in fact would be recovered by Company in two years, with the consequence that this amortization charge would not long be an expense of Company while the new rates were in effect. Accordingly, the D.P.U. allowed as a test year expense only about one-fifth ($82,220) of the 1968 year-end unamortized balance ($411,015) on the theory (1) that this year-end balance appropriately should be written off over five years and (2) that to allow the full amount ($625,935) in the test year would distort test year results so that they would not represent conditions likely to exist when the new rates were in effect. Company contends that (a) the basic principle of the test year requires that the earnings results should be taken as they are, and (b) the fact that a particular expense is “extraordinary and non-recurring” is no reason for disallowing it and substituting for it a greatly reduced amount. To do so, Pruyn testified, “ignores the fact that other extraordinary . . . expenses are almost certain to be incurred in the future that will offset the cessation of” the amortization of a particular item of extraordinary expense. There is force to the D.P.U. view that the large write-off of $625,935 would have distorted 1968 results for test year purposes. This was a large non-recurrent expense and we think that its impact in the test year could be given some special treatment. It was reasonable to regard the unamortized balance at the beginning of 1968 as an item which, for test year purposes, should not be amortized in full measure in accordance with the earlier D.P.U. order. Extraordinary situations affecting test year results may be taken into account by reasonable adjustments. See e.g. Michigan-Wisconsin Pipe Line Co. v. Federal Power Commn. 263 F. 2d 553, 556-557 (6th Cir.). See also Central Maine Power Co. v. Public Util. Commn. 153 Maine, 228, 234-236. The D.P.U., however, concluded that the reasonable and proper diminution of the effect of the extraordinary expense would be to spread the unamortized 1968 year-end balance over five years, 1968 and the next four years. This runs *318 counter to usual accounting methods by which unamor-tized balances of such items, as of the beginning of a year, are written off (in that year and in later years) in accordance with a pre-determined schedule. When Company started 1968 it had a prudently incurred investment of $1,036,950 to amortize. If to do it (without distortion of test year results) required treating it as to be dealt with over five years, as the D.P.U. decided, the appropriate balance to amortize by one-fifth in 1968, to reflect 1968 results, was the balance at January 1, 1968 ($1,036,950), and not the year-end balance of $411,015. Conclusion. 11. The foregoing discussion indicates that the D.P.U. in its decision has closely restricted the rates allowed to Company, and in doing so has resorted to tests and estimates which (in respects noted) have tended to understate Company’s reasonable revenue requirements. It is now possible to correct the D.P.U.’s 1968 test year analysis in the light of Company’s 1970 financing and of its actual 1969 and 1970 operating results. Also some determination now may be made of what would have been the approximate impact of the new rates (as they may be adjusted in the light of this opinion) if they had been in effect in 1969 and 1970. To do so may reveal the extent, if any, to which they are inadequate. The issues remaining in dispute (or subject to further explanation and findings) are not numerous and should be susceptible of D.P.U. reconsideration within a brief period and prompt report to the county court. 12. The case is remanded to the county court. An order there is to be entered directing (a) further consideration by the D.P.U., in the light of this opinion, of its rate order and appropriate modification of that order, and (b) a further report by the D.P.U. to the county court. The stay heretofore granted is to remain in effect until further order of the county court, but without prejudice to the right of either party to request appropriate modification of the penal sum of the bond securing the stay. So ordered. Footnotes [1] Company contends that, to assure itself of a fair return on what it regards as its proper rate base, it actually needed $6,500,000 to $7,500,000 in additional revenues and that it sought only $3,493,600 because of its competitive situation. [2] An application for a stay of the D.P.U. order was originally denied (April 7, 1970) without prejudice to its renewal after Company’s brief was filed with the clerk of the full court. Thereafter (December 11, 1970) a stay was granted, after Company furnished a surety bond in the penal sum of $1,250,000. [3] This additional evidence included Company’s 1969 return to the D.P.U. filed about March 31, 1970, two orders of the D.P.U. authorizing specified financing by Company, and excerpts from certain testimony before the D.P.U. concerning that financing. See D.P.U. 16,470 and 16,470-A. [4] Company prior to 1964 had operated with a low (20%) debt ratio. In 1964 Company changed its financial policy and increased its debt ratio by cancelling 75,000 shares of common stock held by its parent association ana paying certain debt to the parent. Thereafter Company’s new capital has been in the form of debt obtained from sources other than the parent company. See G. L. c. 164, § 13 (as amended by St. 1967, c. 681, eliminating the 50% debt ratio limit existing under § 13, as theretofore appearing in St.. 1953, c. 85). [5] The new bonds increased Company’s debt ratio only to 56% because (as the D.P.U. was aware) limiting the new issue to $20,000,000 caused Company to abandon any “plan to reduce its capital stock” at the time of the new bond issue. [6] The D.P.U. contentions on this matter essentially are: (1) that such evidence is not “necessary” for the determination of the constitutional issue of confiscation under G. L. c. 25, § 5; (2) that to consider the additional, evidence, or to reopen the record to have it considered, would be to disregard certain Federal rate precedents; and (3) that such evidence cannot be considered because it has not been adduced before the D.P.U. at a rate hearing (as contrasted with a hearing on new financing). The third argument is rendered irrelevant because we direct the D.P.U. to verify the evidence, at further rate hearings if necessary, and, if verified, to take it into account thereafter in accordance with principles stated in this opinion. [7] The chronology of these proceedings is as follows: February IS, 1969 — Company’s rates filed. February 19, 1969 — Rates suspended until January 9, 1970. November 9 to December SO, 1969 — Hearings held. January 8, 1970, D.P.U. decision. March, 4,1970, D.P.U. amends its decision. March, 6, 1970 — Appeal to the county court. July S9, 1970 — Case reserved and reported. March 4, 1971 — Arguments before the full court. [8] The Justices went on to say (at p. 690), “ Otherwise, in view of the inevitable delays sometimes encountered in the decision of these cases, especially in times of rapidly changing costs, the court might sometimes be compelled to decide upon evidence already outdated and no longer applicable to the existing situation. Moreover, intervening experience may furnish the very best guide and should not be excluded from consideration, especially where violation of the Constitution is at stake.” The Justices also said (at p. 687), that the “court should have control over the evidence applicable to constitutional issues substantially equivalent to that which it has over evidence heard by a master appointed by it.” [9] Section 5 (as amended through St. 1956, c. 190) reads in part, as follows (emphasis supplied): “No evidence beyond that contained in the record shall be introduced before the court, except that in cases where issues of confiscation or of constitutional right are involved the court may order such additional evidence as it deems necessary for the determination of such issues to be taken before the commission and to be adduced at the hearing in such manner and upon such terms and conditions as to the court may seem proper. Whenever the court shall order additional evidence to be taken, the commission shall promptly hear and report such evidence to the court so that the proof may be brought as nearly as reasonably possible down to the date of its report thereof to the court. The commission may, after hearing such evidence, modify its findings as to facts and its original decision or orders by reason of the additional evidence so taken, and it shall file with the court such amended decision or orders and such modified or new findings. If the commission shall modify or amend its original decision or orders, the appealing party or any other party aggrieved by such modified or amended decision or order may file with the court, within such time as the court may allow, a specification of any errors of law claimed to have been made by the commission in such modified decision or orders, which specification of errors shall thereupon be considered by the court in addition to the errors of law asserted in the claim of appeal.” [10] If, during the pendency of this appeal, Company had been able to issue debt obligations at a substantially lower interest cost than had been predicted, parties other than Company could have brought that evidence to the attention of this court. [11] In its decision of March 25, 1970, the D.P.U. described the financing difficulties in part as follows: “Company is relatively unknown in the financial community and the current [March, 1970] competition for funds may place it at a disadvantage . . . where well known companies are seeking to place their securities at the same period of time.” The D.P.U. later referred to “the present market where the scarcity of available funds results in more sellers than buyers” and to “the limited amount of funds available for investment at reasonable rates.” It concluded that “the issuance of bonds at this time is consistent with the public interest and reasonably necessary in order to permit the Company to reduce its short-term indebtedness in anticipation of future borrowings to . . . finance [temporarily] its construction program. The issue of debt at this time is undoubtedly the most economical means of financing, which will result in a lower cost of money than the sale of other types of securities for . . . raising additional capital.” [12] There is nothing inherently unsound or unreasonable (see Re General Tel. Co. of Calif. 80 P. U. R. 3d 2, 23 [Calif. Pub. Util. Commn.]) about Company’s actual present debt ratio which justifies disregarding the firmly “imbedded” cost of the existing debt. See Phillips, Economics of Regulation, 282-283; Bonbright, Principles of Public Utility Rates, 240-246. See also Garfield and Lovejoy, Public Utility Economics, 124-125, 128-131. [13] Under standards laid down in the Hope Natural Gas Co. case, “the return to the equity owner should be commensurate with returns on investments in other enterprises having corresponding; risks. That return, moreover, should be sufficient to assure confidence in the financial integrity of the enterprise, so as to maintain its credit and to attract capital.” The text authorities, already cited, disclose a variety of different methods and guides for ascertaining, in cases like the present, the return necessary “to attract [equity] capital.” See Phillips, op. cit. 278, 284-288; Bonbright, op. cit. 246-276; Garfield and Lovejoy, op. cit. 125-128, 131-133; Priest, Principles of Public Utility Regulation, 218-226, 451-465; Leventhal, Vitality of the Comparable Earnings Standard for Regulation of Utilities in a Growth Economy, 74 Yale L. J. 989. [14] McCarthy listed “ all of the [eleven] natural gas operating . . . companies in the United States with revenues between $30 million and $70 million which have derived . . . 90% of their revenues from natural gas sales and which have common stock in public hands.” His contention that they were comparable to Boston Gas Company rests, in large part, on the following analysis. Average of 11 companies (000) Boston Gas Company (000) Total operating revenues, 1968 $48,406 $54,057 Operating revenues — average 1959-1968 — Total 37,000 45,328 Residential gas [1959-1968] 53.6% 72.2% Total Capitalization —• Dec. 31, 1968 84,942 63,029 Debt ratios — Dec. 31, 1968 Long Term Debt 50.3% 53.7% (6-30-69) The average rate of earnings of the eleven companies on average equity (1959-1968) was 13% (with a range from 10.4% to 15.9%) and for 1968 was 13.3% (with a range from 10.3% to 17.7%). [15] Kosh assumes, for purposes of his complicated formula (the derivation of which consists of a page of mathematical equations), that the market price of a stock will increase in proportion to any increase in earnings per share. This assumption in effect is that the price/earnings ratio of a stock will remain constant or fluctuate around some normal level. [16] Kosh (in choosing companies for comparison) “applied certain criteria to all . . . gas companies whose stock is traded,” viz. (1) 1968 revenue above 810,000,000; (2) revenue primarily from retail gas distribution; (3) Company subject to Statewide regulation on the basis of an original cost rate base. He eliminated four of fourteen companies (which met these criteria) whose ratio of market price to book value exceeded two times. This he suggested was done in order to “eliminate . . . companies whose securities may be subject to speculative pressure.” He then assumed that earnings are adequate if the market prices of the stocks are high enough above book value to permit the Company to realize net proceeds above book value (per share) from the sale of additional stock. He selected ten companies which showed an average growth of dividend per share (1957-1968) of 5.79% and of book value per share of 4.65%. He estimated this “growth experience” as “on average . . . representative of what informed investors can expect in the near term future.” [17] Company suggests (1) that Kosh eliminated from his comparison companies with “abnormally high” ratios of market price to book value, but did not disregard Providence Gas and Springfield Gas Light, whose earnings on Kosh’s own test “were so low as to indicate . . . that adequate earnings were not being allowed” to them, and also omitted, without adequate explanation, two companies (which Company says met Kosh’s criteria); _ (2) that, apart from these eliminations, Kosh’s own comparative computations on his formula would have resulted in an “ultimate cost of equity” of at least 12%; (3) that Kosh’s assumption that a price/earnings ratio will be constant is “patently false,” and (4) that Kosh disregarded entirely earnings growth rates in computing his trends. [18] For decisions of other commissions discussing the DCF formula (and somewhat similar theory discussed by Kosh), see e.g. Washington Util. & Transp. Commn. v. Pacific N. W. Bell Tel. Co. 81 P. U. R. 3d 371, 381; Re Potomac Elec. Power Co. 83 P. U. R. 3d 372, 393 (Md. Pub. Serv. Commn.); New England Tel. & Tel. Co. R. I. Pub. Util. Commn. Docket No. 1024 (1970). The DCF formula, even if theoretically sound, can produce no better results than are permitted by the comparative data and assumptions employed (as a matter of judgment) in its application. [19] The term “attrition” refers to “the tendency of the rate of return to diminish in a period of comparatively high construction costs” when new and “relatively expensive” plant is being added. “As the high cost plant comes into service, it tends to increase the applicable rate base at a more rapid pace than the resultant earnings; and the rate of return [i.e. actually received in relation to rate base] decreases accordingly.” See New England Tel. & Tel. Co. v. Department of Pub. Util. 331 Mass. 604, 622-623; Phillips, Economics of Regulation, 243-244; Priest, Principles of Public Utility Regulation, 203-206. See also Bonbright, Principles of Public Utility Rates, 150, n. 7. The problem is particularly acute in an inflationary period. Bonbright summarizes thus, “As to the rate base applicable to the test year, the usual procedure has been to accept some . . . average rate base in taking account of growth of plant investment during the very year under review. . . . £S]ome commissions that accept an actual-cost- measure of the rate base have conceded a year-end rate base, presumably as a crude means of allowing for the so-called attrition factor during a period of price inflation.” [20] The figures are shown by the following table: Year Gross Additions to Plant Retirement-Depreciation Millions of Dollars Net Additions to Plant 1965 4.6 2.4 2.2 1966 4.6 2.1 2.5 1967 5.1 2.2 2.9 1968 12.4 1.8 10.6 1969 10.2 * 7.2 * 3.0 * 1970 13.5 * 2.8 * 10.7 * [*] who!ly or partly projections at the time of the hearings. The bulge in capital expenditure in 1968 was based on the large expenditure on a new LNG plant. This expenditure, if spread over 1968 and 1969, would have resulted in a more even rate of net additions to plant. [21] For cases recognizing this imprecise method of taking account of attrition, see e.g. Lynchburg v. Chesapeake & Potomac Tel. Co. 200 Va. 706, 715-716; Re Florida Power Corp. 99 P. U. R. (N.S.) 129, 134-135 (Fla. RR. and Pub. Util. Commn.); Re Public Serv. Co. of Colo. 34 P. U. R. 3d 186, 216-217 (special attrition factors, Colo. Pub. Serv. Commn.); Re Chesapeake & Potomac Tel. Co. 56 P. U. R. 3d 91, 104-105 (Md. Pub. Serv. Commn.); Re Michigan Consol. Gas Co. 79 P. U. R. 3d 375, 379 (Mich. Pub. Serv. Commn.); Re Potomac Elec. Power Co. 83 P. U. R. 3d 113, 123-128 (D. C. Pub. Serv. Commn.). See also Miami v. Florida Pub. Serv. Commn. 208 So. 2d (Fla.) 249, 257-258. Other methods of taking account of attrition include a special allowance in rate of return. See New York v. Public Serv. Commn. 17 App. Div. 2d (N. Y.) 581, 584-585. Cf. e.g. Pennsylvania Pub. Util. Commn. v. York Water Co. 78 P. U. R. 3d 113, 128 (Pa. Pub. Util. Commn.); Re United Inter-Mountain Tel. Co. 79 P. U. R. 3d 499, 507-509 (Tenn. Pub. Serv. Commn.). [22] Company in its brief points out that the figures shown in Company’s 1969 annual return (pp. 11, 18) closely conformed with Pruyn’s predictions of operating gas revenues, operating gas expenses, and gross additions to plant. [23] Bolster (called by the D.P.U.) defined the test year as “a recent operating period during which revenues, expenses, and plant requirements are either generally in balance, in the sense that thepr] relationship ... is generally representative of the recent past, and therefore likely to be representative of the near term future, or can be so adjusted that such relationships are representative of the near term future.” [24] This figure appears to have been accepted by the D.P.U. in its computation of the savings to be expected from all LNG operations. Company contends that the figure of $202,000 is also an understatement of the appropriate carrying charge. The accuracy of this figure is to be reconsidered when the case is again before t.he D.P.U.