OPINION
MURRAY, Justice.
On April 4, 1980, Valley Gas Company (the company), pursuant to the terms of G.L.1956 (1977 Reenactment) § 39-3-11, as amended by P.L.1977, eh. 236, § 2, filed with the Public Utilities Commission (the commission) a schedule of rates and charges which was to have become effective on May 5, 1980. The proposed schedule would have increased the company’s revenue by approximately $2,621,454. The commission, acting under authority of § 39-3-11, suspended the implementation of the proposed schedule on two separate occasions. The commission’s first order, issued on May 3, 1980, suspended the effective date for implementation of the rates for five months; the commission’s second order, issued on October 2,1980, suspended the effective date for implementation of the rates for an additional three months.
A number of hearings were held between September 23,1980, and December 23,1980. Appearances were entered by the company, the Division of Public Utilities and Carriers (the division), the Coalition for Consumer Justice, and the Rhode Island Federation of Chambers of Commerce. Expert witnesses were presented by both the company and the division.
On January 2, 1981, the commission issued a seventy-two-page report and order. The commission rejected the proposed rate schedule filed by the company and directed it to file a new schedule that would recover an additional $1,107,879 in annual revenues. Thereafter, on January 14, 1981, the commission issued another order that reduced the increase in annual revenues to $996,398. The subsequent order was needed to correct a computational error. The revised rates were to go into effect as of January 2,1981.
On January 9, 1981, the company, acting under authority of G.L.1956 (1977 Reenactment) § 39-5-1, filed a petition for writ of certiorari with this court. We issued the writ. Counsel for the company and counsel for the administrator of the division have appeared before us and presented argument. Since the company’s petition raises issues involving depreciation and its treatment by the commission, initially we shall set forth the relevant facts necessary for the disposition of the depreciation issues.
For a full understanding of the depreciation issues presented in the instant case, it is essential that we describe the developments that occurred the last time the company filed for a general rate increase. At that time, a proposed schedule of rate increases was filed on July 1, 1977. The company had employed Ebasco Services Incorporated (Ebasco) to conduct a depreciation study. The study was supervised by Julius Breitling (Breitling), an engineer with Ebasco. He made certain recommendations both regarding the composite depreciation rate to be applied to the company’s plant in service and regarding the treatment of the discrepancy that existed between the depreciation reserve carried on the company’s books and the newly calculated theoretical reserve. The commission accepted Ebasco’s suggestion regarding the depreciation rate to be applied but refused to adopt Ebasco’s suggestion regarding the deficiency.
On certiorari, this court sustained the commission’s decision on the depreciation issue.
Valley Gas Co. v. Burke,
R.I., 406 A.2d 366 (1979)
(Valley Gas I.).
In the instant case, the company again retained Ebasco to conduct a depreciation study of its plant in service as of August 31, 1979, and to recommend depreciation rates to be adopted as of such date. The study was to update the earlier study made of the year ending December 31, 1975.
Breitling also supervised the study and testified before the commission in support of the proposed rates. A two-page schedule of annual depreciation rates prepared by Breitling was entered into evidence. This exhibit detailed the average service lives, mortality dispersion, and net salvage ratios for the various classes of property used by the company to provide service to its customers. The exhibit also contained findings regarding the annual depreciation accrual amount and the annual depreciation rate for each functional group of property upon which the company accrues depreciation. Ebasco’s study concluded that the company should adopt a new composite depreciation rate of 2.933 percent on total depreciable plant in service as of August 31, 1979. According to Breitling’s exhibit, the composite depreciation rate of 2.933 percent produced a total annual depreciation expense of $738,964 for the test year ending August 31, 1979.
It is not essential for our purposes that we detail all of the various analyses Brei-
tling employed in reaching his conclusions. Suffice it to say that in determining the average service lives, the curve types, and the net salvage for each plant account, Breitling considered the company’s past retirement experience, the company’s present and anticipated-future systems requirements, and the experience of the industry in general with regard to various classes of property. He also used three different statistical methods to analyze the company’s previous experience with plant accounts: the actuarial method of life analysis, the geometric mean method, and the simulated plant record analysis-balances method. He used the actuarial method to analyze all but three accounts: “Account 382 — Meter Installations,” “Account 383— House Regulators,” and “Account 384— House Regulator Installations.” The actuarial method uses statistical methods developed in the life insurance industry and applies them to industrial properties. The three remaining plant accounts were analyzed through the use of the two other above-mentioned statistical methods.
In regard to net salvage, Breitling recommended net salvage ratios that are expressed as a ratio of original cost. He reached his conclusions after studying the company’s property-retirement experience from 1971 through 1979. He related the cost of removal and salvage to the original cost of the property that had been retired. The significance of the ratios he suggested are described as follows: a ratio of one to zero means that the cost of removal for property retired between 1971 and 1979 and the salvage value of such property are equal. A ratio higher than one to zero means that the cost of removal is greater than the salvage value — this occurrence is known as “negative net salvage” whereas a ratio that is less than one to zero means that the cost of removal is less than the salvage value — this occurrence is known as “positive net salvage.”
The composite depreciation rates recommended by Breitling were based upon a remaining-life method of depreciation. Under this method, the company has attempted to recover its unrecovered investment in plant in service adjusted for salvage over the average remaining life of the property. The method proposed did not require the computation of a theoretical reserve.
As he was being cross-examined, Brei-tling was asked to compare the method he was employing in the present case with the method he had proposed in the previous case. He stated that both methods were remaining-life methods, although the method recommended in the previous case was slightly more theoretically correct.
The method of depreciation accounting proposed in the last general rate case applied average life rates to all investments and it also amortized any difference between theoretical reserve and book reserve over the remaining life of the property. Breitling gave two reasons for his not utilizing that method of developing rates in the present case. His initial reason was that since the commission had not accepted his proposal in the earlier rate filing and it had accepted the methodology he was currently proposing in a recent filing by another utility, and that as an expert he must take the commission’s previously held positions into account when preparing depreciation rates to be used by his client. On re-direct examination, Breitling explained that he also felt that the commission’s deletion from the rate base of the difference between theoretical reserve and book reserve in the last general rate case was harmful to the company and its investors and that his current
remaining-life proposal would recoup such deficiency and alleviate the detrimental effect of the prior decision.
Evidently the commission was interested in obtaining new calculations based upon the method of depreciation accounting they had rejected in the last general rate case, because they requested that Ebasco prepare calculations using such method for the test year ending August 31, 1979. The calculations were supplied by Ebasco at a later time, and they were admitted into evidence as the commission’s exhibit despite the objection of the company’s attorney.
The division presented G. Raymond Armknecht (Armknecht), a senior rate consultant with the firm of Hess & Lim, as an expert witness. As part of his testimony, he reviewed the depreciation study presented by Breitling. He made no independent depreciation study, and he accepted Brei-tling’s findings concerning the lives to be assigned to the individual plant accounts as well as the curves proposed by Breitling. Although he also accepted the remaining-life method of depreciation recommended in this case and did not propose an alternative approach, he stated that his position might have been different had the commission not accepted the remaining-life approach in the last general rate case.
Armknecht did, however, make a few adjustments to Breitling’s calculations. He rejected Breitling’s proposed salvage factor for “Account 380 — Services.” Armknecht reasoned that because of the unprecedentedly high rates of inflation of the previous few years, it was conjectural to predict future salvage ratios based upon a study that compares the cost of removal for the years 1970 through 1979 with the original cost of the plant purchased years earlier. In place of recovering the cost of removal through the depreciation charges, Armk-necht proposed that the company account for the cost of removal as a current expense. Armknecht stated that this proposed treatment of removal costs was a generally accepted accounting principle for public utilities. He did not, however, have information available to identify other public-utility commissions that had in fact adopted this method. He also stated that he was merely suggesting that the method should be employed on an interim basis until the problem could be further examined. The actual expense allowance for the cost of removal proposed by Armknecht was $21,927, which represented the average cost of removal recorded on the books during the last five years. As a result of the $21,927 cost-of-removal allowance, Armk-necht directed his associate to deduct $78,-458 from the company’s claimed depreciation expense of $100,385.
Armknecht also increased the company’s book reserve by $1,013,757; this figure represented the difference between the book reserve and the theoretical reserve computed in the last rate case. In the previous case, the commission had taken the same action of increasing the company’s book reserve; but the company, in proposing a depreciation expense in the present case, had not maintained the adjustment. Since the method of depreciation accounting proposed by Breitling in the present case evidently would recoup such deficiency, Armk-necht also reduced the depreciation expense in the company’s calculations by $28,777 to eliminate the amortization of the difference between the book reserve and the theoretical reserve. Armknecht disagreed with the company’s position that it should be allowed to earn a return on the reserve deficiency. In his prepared testimony, Armknecht stated that he did not believe that the level of return on equity the company has received in previous years “is the proper test for [the] determination of the disposition of depreciation reserve deficiencies.” In addition, he believed that the methods used by the company in computing the equity returns were “suspect, if not outright invalid * * Armknecht’s position is that the company has earned sufficient revenues to recover its costs including book-depreciation expense. He asserted that if a company wants to switch to a remaining-life method of depreciation, then it should record the difference between book reserve and theoretical reserve at the time the switch in methods takes place. In conclusion, he
found no evidence that would cause him to recommend to the commission that the reserve deficiency be returned to the rate base so that the company’s investors could earn a return on such amount.
During rebuttal, the company presented Jay H. Price, Jr. (Price), a certified public accountant and partner of Arthur Andersen & Company, specifically to counter the adjustments made to the depreciation expense by Armknecht. In response to Armknecht’s proposal that the cost of removal be charged as a current expense item rather than be recovered through depreciation charges, Price had several observations. Initially, he noted that the cost-of-removal factor that Breitling had proposed was significantly less than the cost that the company had actually experienced in the recent past. This lower figure was given despite the fact that recent economic trends gave no indication of easing the pressures that had caused the amount to rise originally. In specific reference to the Armknecht proposal, Price stated it was not in accordance with generally accepted accounting practices for the regulatory industry. According to Price, the infirmities in Armknecht’s proposal create a situation in which the cost of removal is not charged to those customers who have received service from the related property but to subsequent customers who may refuse to pay for the removal of property with which they have had no prior connection. Price was also of the opinion that the net cost-of-removal amount of $21,927, representing the average cost of removal for the last five years and the amount Armknecht included as a charge to a current expense account, was unreasonably low. He based his opinion upon the fact that the actual amount charged against the reserve for depreciation for the 1979 test year was $43,332, and the fact that removal costs, like most other costs, could be expected to increase because of the inflationary pressures in the economy.
Price also responded to Armknecht’s proposed reduction from the rate base of $1,013,757 (representing the difference between the depreciation reserve recorded on the books at the time of the last general rate case and the calculated theoretical reserve) and Armknecht’s concurrent proposal to reduce depreciation expense by $28,777 (representing the amortization of the $1,013,757 reserve deficiency over the remaining life of the property). Price observed that the company’s depreciation rates have been prescribed by the commission for an appreciable period. Price also stated that, relying upon information provided by the company management, he was able to ascertain that almost all of the $1,013,757 deficiency must have been caused by economic conditions occurring after the company’s first depreciation study, whose test year ended December 31, 1965. In addition, he remarked that since the company had clearly demonstrated that it had not earned the level of return authorized by the commission after 1965, no basis existed for calling the return requirements inflated. Price also explained that the company’s return on equity since 1962 has been well below the industry average.
Price disagreed strongly with Armk-necht’s contention that a theoretical reserve should be recorded on the books when a company changes to a remaining-life method of depreciation. In fact, he could only think of possibly one narrow exception in which the recording of a theoretical reserve would be appropriate: when a company is both negligent in making depreciation studies or in requesting depreciation rates and when such company has earned a rate of return in excess of that authorized by a regulatory body. In Price’s opinion neither factor was present in this case.
Price further testified that in his opinion sufficient evidence existed to support the recovery of the deficiency. In particular, he mentioned the company exhibit that showed the company as not having recovered its cost-of-equity capital in the past, and he also noted the fact that almost all of the deficiency had arisen since the 1966 study. Also, in Price’s opinion, the remaining-life proposal for depreciation accounting made by the company would adequately and fully recover the undepreciated cost of plant over its remaining life. He explained
that if the depreciation expense is ascertained by making a best estimate of the amount needed to recover undepreciated plant over its remaining life, then no reserve deficiency will exist.
In the report and order of the commission the current-expense proposal for the cost of removal made by the division was rejected. The basis for this decision was the belief that the proposal did not equitably allocate the cost of removal to those customers who had in fact received service from the related property. The commission allowed an actual depreciation expense of $641,957; this amount was based upon the calculations that it requested Breitling to provide using the method of depreciation accounting he had employed in the last general rate case.
In regard to the $1,013,757 figure that represents the differential between the company’s book reserve and its theoretical reserve calculated in the last rate case, the commission specifically found that the company’s earnings were insufficient to recover such amount, especially in the last few years when “the bulk of the deficiency appealed] to have arisen.” The commission allowed the company to amortize the deficiency at a rate of $30,000 per year. The commission refused, however, to allow the company to earn a return on the remaining unamortized portion of the deficiency and consequently removed that portion of the deficiency from the rate base.
In sum, the commission allowed the company a depreciation expense of $641,957, plus an additional $30,000, which represented the amount allowed for amortization of the deficiency. The total depreciation expense of $671,957 placed in the cost of service was less than the $739,722 requested by the company.
In support of its petition, the company essentially raises two major arguments. The company contends that the commission acted arbitrarily, unreasonably, and unlawfully in reducing the company’s rate base by $1,013,757, which represented the difference between the book reserve for depreciation and the theoretical-reserve requirement at the time of the last general rate case. In addition, the company argues that the commission acted unreasonably and unlawfully and exceeded its authority in allowing the company an annual depreciation expense of $671,957 rather than the $739,-722 amount requested by the company. Each argument will be considered in a separate part of the opinion, and additional facts will be developed as needed in the resolution of the issues.
As a preliminary matter, we set forth the framework we follow in reviewing the commission’s report and order. In a utility rate case, we are required “to determine whether the commission’s [report] and order are lawful and reasonable * *
Rhode Island Consumers’ Council v. Smith,
111 R.I. 271, 277, 302 A.2d 757, 762 (1973);
see also New England Telephone & Telegraph Co. v. Public Utilities Commission,
446 A.2d 1376, at 1380-1381 (R.I., 1982). We concern ourselves with neither independent factfinding nor the weighing of evidence.
Valley Gas I,
406 A.2d at 369. Factfinding is the province of the commission under § 39-5-3. The commission’s findings of fact, if supported by substantial evidence, are generally “unassailable on review.”
Providence Gas Co. v. Burke,
R.I., 419 A.2d 263, 268 (1980). Since we do not search the record for evidence that supports the commission’s decision or speculate as to the true reasons for its decision, it is mandatory to our appellate function that the commission clearly set forth the findings and evidentiary facts upon which its decision rests.
Bristol County Water Co. v. Public Utilities Commission,
117 R.I. 89, 101-02, 363 A.2d 444, 451 (1976);
New England Telephone & Telegraph Co. v. Public
Utilities Commission,
116 R.I. 356, 363, 358 A.2d 1, 7 (1976). Of course, the commission will be given the benefit of factual determinations we can reasonably infer from its language and actions.
Bristol County Water Co. v. Public Utilities Commission,
117 R.I. at 102-63, 363 A.2d at 451.
I
We initially concern ourselves with the company’s contention that the commission acted arbitrarily and unreasonably in not allowing the company’s investors to earn a return on the unamortized portion of the deficiency while it acknowledged that the deficiency had not been recouped through earnings.
At the outset we shall set forth certain principles relating to the issue before us. “Depreciation * * * is the process of continued loss to property resulting from factors such as wear and tear and technological obsolescence which eventually leads to the retirement of that property.”
Narragansett Electric Co. v. Harsch,
117 R.I. 395, 411, 368 A.2d 1194, 1204 (1977). More generally, it is a property’s “decline in, or loss of, value[.]” Bonbright,
Principles of Pubiic Utility Rates,
194 (1961). In a public-utility rate case, depreciation is seen as an annual allowance to an expense account, whereas accrued depreciation is seen as a reserve that is deducted from a company’s total investment in a plant in service in determining a company’s rate base. It is the latter treatment — the negative term of the rate base — that we shall deal with presently.
The rate base of a public utility is that “quantum of invested capital” on which the company’s investors are entitled to earn a fair rate of return. Bonbright, supra, at 149-50. The failure of a regulatory agency to allow a public utility to earn a fair rate of return is a confiscation of the property of those who have invested in such utility and a violation of the Fourteenth Amendment to the United States Constitution.
See Federal Power Commission v. Natural Gas Pipeline Co. of America,
315 U.S. 575, 590, 62 S.Ct. 736, 745, 86 L.Ed. 1037, 1052 (1942);
New England Telephone & Telegraph Co. v. Public Utilities Commission,
116 R.I. 356, 362-64, 358 A.2d 1, 7-8 (1976).
In the last rate case the commission did not allow the company to recover the $1,013,757 deficiency, nor did it allow the company to earn a return on such amount. We sustained the commission’s decision solely on evidentiary grounds.
Valley Gas I,
406 A.2d at 370. We found that the record did not contain sufficient evidence to support the company’s position that the deficiency had not been recovered.
Id.
We did, however, recognize that “the company’s entitlement to collect the deficiency” was not at issue before us.
Id.
In the present rate filing, the company clearly and definitely set out to cure the evidentiary defects of the previous case. The company presented a nineteen-year study showing that it had not earned in any year either the rate of return authorized by the commission or the average yearly earnings for the industry. In contrast, the study presented by the company in the previous case covered only a five-year period. The commission, in a significant and crucial finding, acknowledged “that the [c]ompa-ny’s earnings were not adequate to recoup the deficiency during the recent years in which the bulk of the deficiency appear[ed] to have arisen.”
The commission, in accordance with the above-stated finding, allowed the company to recover such deficiency by way of a $30,000 per year amortization allowance that was to be added to the depreciation expense. The commission, however, deducted the unamortized portion of the deficiency from the rate base, thereby refusing to allow the company’s investors to earn a return on such amount. Evidently, this action was caused in part by the commission’s interpretation of our ruling in
Narragansett Electric Co. v. Kennelly,
88 R.I. 56, 143 A.2d 709 (1958), and in part upon “the [cjommission’s own conviction that [a] company [is] not entitled to a return on plant costs they should have depreciated in the past.” Ac
cording to the report and order, the commission has consistently taken this position in the past.
We are of the opinion that the commission’s position is not in accordance with our holding in
Providence Gas Co. v. Burke,
R.I., 419 A.2d 263 (1980), and is therefore erroneous.
Initially, we point out that ample evidence exists in the record to support the commission’s finding that the company’s earnings have been insufficient to recoup the deficiency. The nineteen-year study presented by the company as well as Price’s testimony concerning the company’s earnings clearly established that the company had sustained its burden of proving that the deficiency had not been recovered.
See Valley Gas I,
406 A.2d at 370.
We also observe that the record is devoid of evidence that indicates that any of the company’s property subject to depreciation is not used and useful in providing service to customers. This distinguishes the situation in
NEPCO Municipal Rate Committee v. Federal Energy Regulatory Commission,
668 F.2d 1327 (D.C.Cir.1981), a case relied upon by the division, from the present situation. In
NEPCO,
the utility had abandoned a planned construction project to build an electric generating facility and had attempted to include the amount it had expended in preparation for the project in its rate base. The Court of Appeals, relying in part upon the general rule that only property used and useful may be included in a company’s rate base, failed to overturn a Federal Energy Regulatory Commission refusal to allow a return on the amount of the expenditures.
Id.
at 1335. Also, in
Providence Gas Co. v. Burke,
419 A.2d at 268, the division attempted to argue that, relying upon
Narragansett Electric Co. v. Kennelly, supra,
it found the utility’s property not used and useful in supplying service to customers should be excluded from the rate base. We found “[n]ot a shred of evidence [to be] present in this record” indicating that the differential between book reserve and theoretical reserve “represent[ed] property not actually used and useful” in delivering service to company customers.
Providence Gas Co. v. Burke,
419 A.2d at 268.
In the case at bar it is also clear that the company has not been negligent in its depreciation accounting practices.
See id.
Initially, we observe that depreciation accounting is a mere approximation and rarely, if ever, will the depreciation on the books equal the depreciation experienced. Also, it can not be doubted that the company witness was absolutely correct in stating that the cost of removal of the company’s plant had caused the bulk of the deficiency to occur. The company management and the commission could not have been expected to foresee many years ago that inflationary pressures in the economy would cause these removal costs to rise so drastically. More importantly, the company’s four depreciation studies undertaken in the years 1966, 1969, 1975, and 1979 give rise to a strong inference of prudence on the part of the company management. This prudence is especially highlighted by the fact that in the recent years, when the company has experienced increases in the costs of removal, it has undertaken depreciation studies more frequently.
In conclusion, for the reasons just set forth we are of the opinion that the commission’s deduction from the company’s rate base of the $1,013,757 differential was an error as a matter of law in that it represents the penalization of the company for the “lack of precise clairvoyance” we denounced in
Providence Gas Co. v. Burke,
419 A.2d at 268.
II
We next deal with the company’s other major contention that the commission erred by adopting an annual depreciation expense of $671,957 rather than the $739,722 amount proposed by the company. The company bases this argument upon its view that the commission relied on inadmissible evidence in determining the depreciation expense and that the decision is not supported by probative evidence and is inherently contradictory.
As mentioned earlier, the commission, while cross-examining Breitling, asked him to compare the method of depreciation accounting he employed in the present case with the method he had proposed in the previous case. After explaining that both were essentially remaining-life methods, he stated that the previous method was slightly more theoretically correct. At this point, the commission asked Breitling if he could prepare calculations for use in the present case by using the methodology employed in the previous case. Breitling answered in the affirmative but stated that he would need additional time to supply the data. Later the calculations were admitted into evidence as the commission’s own exhibit. This exhibit became the basis for the commission’s decision to adopt a $671,957
annual depreciation expense.
Initially, we observe that the Public Utilities Commission is not bound by technical rules of evidence.
Providence Gas Co. v. Burman,
119 R.I. 78, 105, 376 A.2d 687, 701 (1977); G.L.1956 (1977 Reenactment) § 39-1-11, as amended by P.L.1979, ch. 95, § 2;
accord, New England Telephone & Telegraph Co. v. State,
113 N.H. 92, 101-02, 302 A.2d 814, 821 (1973). This does not mean that when the agency acts in a quasi-judicial capacity, all such rules are freely discarded. In addition, we have long taken the position that the Public Utilities Commission is “free to pick and choose even between inconsistent portions” of an expert’s testimony.
Rhode Island Consumers’ Council v. Smith,
111 R.I. at 295-96, 302 A.2d at 772.
In reviewing the commission’s order as it relates to the annual depreciation expense, we are of the opinion that the commission relied upon admissible evidence in making its determination and that its decision is supported by probative evidence.
Initially, we note that the record contains no indication of the commission’s acting as anything other than the neutral and detached decision maker envisioned by the Legislature when enacting § 39-1-11.
See Blackstone Valley Chamber of Commerce v. Public Utilities Commission,
R.I., 396 A.2d 102-04 (1979);
Narragansett Electric Co. v. Harsch,
117 R.I. at 403-05, 368 A.2d at 1200-01. The commission has not, as the counsel for the company asserted at oral argument, assumed the role of advocate by requesting that the company’s witness quantify the admission he had made under cross-examination. The request was a proper subject of cross-examination. After the results were quantified and presented, the commission was entitled to accept and rely upon such calculations even though they differed from those initially proposed by the company’s expert. Choosing between inconsistent portions of an expert’s testimony is the prerogative of the commission.
See Rhode Island Consumers’ Council v. Smith,
111 R.I. at 295-96, 302 A.2d at 772. In conclusion, we find nothing inherently wrong with the commission’s requiring an expert witness to quantify the admissions he has made under cross-examination and with the commission’s relying upon such calculations in modifying the depreciation expense. The commission, therefore, did not err in adopting a $671,957 annual depreciation expense.
The company’s petition for certiorari is sustained as it relates to the earning of a return on the $1,013,757 differential between the book reserve and the theoretical reserve calculated at the time of the last rate filing. The company’s petition for cer-tiorari is denied as it relates to the commission’s calculation of a $671,957 depreciation
expense. The papers in this case are ordered returned to the commission with our decision endorsed thereon and with the direction that the commission adjust the tariff rates to reflect the $1,013,757 addition to the company’s rate base.