On 1 February 1994, the North Carolina Rate Bureau (Bureau) filed a general request for increased rates for private passenger automobiles and motorcycles. The rate increase requested an increase of 10.8% for automobile rates and 22.4% for motorcycle rates. The Commissioner held a comprehensive hearing during the summer of 1994. The filing request was more than 1,500 pages in length; there were an additional 800 pages of responses by the Bureau to the Commissioner’s requests for data to explain the filing; the hearing transcript is more than 3,500 pages in length and the evidence included more than 120 exhibits. The Commissioner’s lengthy order of more than 500 pages, including calculations and exhibits, disapproved the Bureau’s filing and ordered rate changes reducing rates for automobiles by 13.8% and increasing rates for motorcycles by 10.2%. The Bureau appealed from this order and brought forward 13 *678assignments of error based on more than 40 pages of exceptions to various findings of fact, conclusions of law and exhibits.
I. STANDARDS OF REVIEW
A. Appellate Court Review
[1] In reviewing orders of the Insurance Commissioner, the test is whether the Commissioner’s conclusions of law are supported by material and substantial evidence in light of the whole record. State ex rel. Comr. of Insurance v. N.C. Rate Bureau, 75 N.C. App. 201, 208, 331 S.E.2d 124, 131, disc. review denied, 314 N.C. 547, 335 S.E.2d 319 (1985). “The whole record test requires the reviewing court to consider the record evidence supporting the Commissioner’s order, to also consider the record evidence contradicting the Commissioner’s findings, and to determine if the Commissioner’s decision had a rational basis in the material and substantial evidence offered.” Id. Substantial evidence is “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Comr. of Insurance v. Automobile Rate Office, 287 N.C. 192, 205, 214 S.E.2d 98, 106 (1975). It is “more than a scintilla or a permissible inference.” Id. (quoting Utilities Commission v. Trucking Company, 223 N.C. 687, 690, 28 S.E.2d 201, 203 (1943)).
While this Court employs the “whole record” test in reviewing the Commissioner’s orders, “it is not our function to substitute our judgment for that of the Commissioner when the evidence is conflicting.” State ex rel. Comr. of Insurance v. N.C. Rate Bureau, 96 N.C. App. 220, 221, 385 S.E.2d 510, 511 (1989). The weight and sufficiency of the evidence as well as the credibility of the witnesses are determined by the Commissioner. Id.
B. Review by the Insurance Commissioner
[2] An order or decision of the Insurance Commissioner regarding premium rates is presumed to be correct if it is supported by substantial evidence. N.C. Gen. Stat. § 58-2-80. “Upon any appeal, the rates fixed or any rule, regulation, finding, determination, or order made by the Commissioner under the provisions of Articles 1 through 64 of this Chapter shall be prima facie correct.” N.C. Gen. Stat. § 58-2-90(e).
The Commissioner’s order regarding a rate filing must comply with the standards set forth in N.C. Gen. Stat. § 58-36-10:
(1) Rates shall not be excessive, inadequate or unfairly discriminatory.
*679(2) Due consideration shall be given to actual loss and expense experience within this State for the most recent three-year period for which such information is available; to prospective loss and expense experience within this State; to the hazards of conflagration and catastrophe; to a reasonable margin for underwriting profit and to contingencies; to dividends, savings, or unabsorbed premium deposits allowed or returned by insurers to their policyholders, members, or subscribers; to investment income earned or realized by insurers from their unearned premium, loss, and loss expense reserve funds generated from business within this State; to past and prospective expenses specially applicable to this State; and to all other relevant factors within this State: Provided, however, that countrywide expense and loss experience and other countrywide data may be considered only where credible North Carolina experience or data is not available.
N.C. Gen. Stat. § 58-36-70(d) regarding rate filings and hearings for motor vehicle insurance states, in part:
If the Commissioner after the hearing finds that the filing does not comply with the provisions of this Article, he may issue an order disapproving the filing, determining in what respect the filing is improper, and specifying the appropriate rate level or levels that may be used by the members of the Bureau instead of the rate level or levels proposed by the Bureau filing, unless there has not been data admitted into evidence in the hearing that is sufficiently credible for arriving at the appropriate rate level or levels.
“In reaching his ultimate determination, the Commissioner must make findings which clearly and specifically indicate the facts on which he bases his order, the resolution of conflicting evidence, and the consideration he has given to the material and substantial evidence that has been offered.” State ex rel. Comr. of Insurance v. N.C. Rate Bureau, 95 N.C. App. 157, 159, 381 S.E.2d 801, 803 (1989). This requires the Commissioner to be mathematically specific as to his findings of fact. Comr. of Insurance v. Rate Bureau, 300 N.C. 381, 456, 269 S.E.2d 547, 592, reh’g denied, 301 N.C. 107, 273 S.E.2d 300 (1980).
II. DIVIDENDS AND DEVIATIONS
[3] The Bureau contends the Commissioner exceeded his statutory authority and entered an order which is unsupported by material and *680substantial evidence when he ignored the requirements set forth in G.S. 58-36-10 by failing to give “due consideration” to dividends and deviations in ruling on this rate request. Particularly, the Bureau argues the Commissioner determined the aggregate losses, expenses and an appropriate profit; he then calculated and used underwriting profit provisions without any adjustment in the ratemaking formula for dividends and deviations. In so doing, the Bureau contends the Commissioner “camouflage [d] his continuing refusal to adhere to the requirement of the law that the rates reflect the effects of dividends and deviations” by devoting almost half of his order to an examination of the issue of dividends and deviations, but ultimately concluding that our current system of ratemaking already includes, within the average rate, a provision for dividends and deviations.
The Bureau argues this conclusion is erroneous and will result in rates which will not generate sufficient premium to provide for a reasonable profit for all automobile insurance. Because deviations and dividends reduce the cost of insurance to policyholders, the Bureau argues they should be treated as an expense item as opposed to profit. The Commissioner rejected the Bureau’s treatment of deviations and dividends as a reduction in premium (an expense) and he simply adjusted the expected premium back to the amount which would be collected if there were no deviations without making an adjustment for dividends and deviations in his rate calculations. With this adjustment, the Bureau contends the targeted profit is only generated “if one assumes that the premiums not charged (i.e. the amount deviated) are somehow collected by the companies and the premiums returned to policyholders (i.e. dividends) are somehow retained by the companies.”
The Bureau argues the Commissioner’s reasons for ignoring deviations and dividends are baseless and irrelevant. The Bureau notes the Commissioner’s concern that dividends and deviations are voluntary and discretionary has already been settled by our Court, which has held the discretionary nature of dividends and deviations is not a basis for the Commissioner to ignore them in developing the rate level. State ex rel. Comr. of Ins. v. N.C. Rate Bureau, 97 N.C. App. 644, 646, 389 S.E.2d 574, 575, disc. review denied, 326 N.C. 804, 393 S.E.2d 905 (1990). Furthermore, the Bureau points out the Commissioner’s finding that deviations and dividends are unfairly discriminatory ignores the fact that our General Assembly “created a system of ratemaking that is by design ‘discriminatory’ ” because the formula is based on the collective experience of all insured motorists, *681pooling drivers with both good and bad driving records. The Commissioner’s charges that dividends and deviations (1) are not competitive tools, (2) they lead to spiraling manual rates, and that (3) they are payable from surplus are not supported by material or substantial evidence, according to the Bureau. Finally, the Bureau contends the Commissioner’s last two arguments are erroneous: (1) ratemaking must assume that every company charges the manual rate and (2) there is an actual margin in the rates for dividends and deviations. The Bureau concludes by arguing the Commissioner’s failure to explicitly recognize deviations and dividends is in excess of his statutory authority, his reasoning is flawed and irrelevant, and his implicit provision for deviations and dividends is unsupported by material and substantial evidence.
The Commissioner contends his order is the product of a thorough consideration of dividends and deviations and that the concept of “due consideration” required by G.S. 58-36-10 does not necessarily mean that an adjustment to the rates must be made to reflect the effects of dividends and deviations. He argues there was substantial evidence that the Bureau’s proposed rate was excessive and that the Bureau formula results in the double counting of dividends and deviations which, in turn, leads to spiraling manual rates. Left uncorrected, this situation leads to excessive and unfairly discriminatory rates. Consequently, the Commissioner contends he remedied the inequities by: (1) using the manual rates actually in effect in North Carolina; (2) correcting the Bureau’s mathematical error in its ratemaking formula; (3) determining the appropriate amount of dividends and deviations based on the corrections; and (4) deriving a profit provision which included the appropriate dividends and deviations.
We agree with the Commissioner that “due consideration” does not mandate that a numerical adjustment to the rates must be made to reflect the effects of dividends and deviations. In State ex rel. Comr. of Ins. v. N.C. Rate Bureau, 75 N.C. App. at 224-25, 331 S.E.2d at 141, this Court addressed the meaning of “due consideration” in terms of the underwriting profit rating factor. Our Court said:
G.S. § 58-124.19(2) [now G.S. 58-36-10] only requires that the Commissioner give “due consideration” to the enumerated rating criteria, including allowance for an underwriting profit. Nothing in the language of the statute requires that the Commissioner provide for an underwriting profit so long as the rate level estab*682lished on the statutory rate criteria is not inadequate, excessive, or unfairly discriminatory.
Id. The Court quoted our Supreme Court as saying the General Assembly never intended “to make any one, or all, of these matters [statutory rating standards] conclusive. . . . The weight to be given the respective factors is for the Commissioner to determine in the exercise of his sound discretion and expertise. ...” Id. at 225, 331 S.E.2d at 141 (quoting Comr. of Insurance v. Rating Bureau, 292 N.C. 471). Like underwriting profit, dividends and deviations are statutory rating factors to which the Commissioner must give “due consideration,” but the Commissioner must then weigh the various statutory rate factors to achieve an adequate rate level and to ensure the proposed rate will leave the insurers with a fair and reasonable profit. Id.
After a careful review of the record, we find there is substantial support for a number of the Commissioner’s concerns and his rejection of the Bureau’s treatment of dividends and deviations. At the hearing, there was conflicting expert testimony as to where to reflect dividends and deviations — in the expenses or in calculating the underwriting profit. While two of the Bureau’s witnesses testified these factors are expenses, a number of experts testifying for the Department of Insurance (Department) stated the appropriate place to reflect dividends and deviations is in the margin for underwriting profits. There was expert testimony that the Bureau used neither of the two accepted ratemaking formulas (Loss Ratio and Pure Premium Methods) and consequently, the results lead to inflated levels of dividends and deviations. This testimony was illustrated by comparing the two accepted formulas and showing they always produced the same result. The expert then compared these formulas to the Bureau’s method and demonstrated the Bureau’s formula resulted in double counting, excessive and inflated rates, and was unfairly discriminatory as it ultimately led to spiraling rates. Department witnesses echoed the testimony as to the double counting and spiraling effect of the Bureau’s formula. Relying on this testimony, “the Commissioner corrected the Bureau’s mathematical and actuarial error and used the manual rates in effect in North Carolina without the improper reduction for deviations.”
After making the appropriate corrections, the Commissioner then provided for the appropriate amount for dividends and deviations in the profit provision by calculating the amount that is provided in his *683prospective rate level based on evidence in the Record on existing levels. The Commissioner made the following findings, in his order:
74. The testimony and evidence summarized in Exhibits C through H, attached, together with the matters and things set forth in the exhibits referred to in this Section are found to be substantial, credible, convincing, true and supportive of the Findings of Facts and Conclusions of Law set forth in this Order and collectively such evidence and testimony are hereby adopted as additional Findings of Facts and they are incorporated herein as fully as if set forth verbatim in the main body of this Order.
75. The evidence for the Department convincingly and repeatedly demonstrates that the average rate includes within it a provision for dividends and deviations.
76. DOI-44 shows the effects of the Bureau method and corroborates the testimony of the Bureau expert Michael Miller and the Department experts that manual rates based on average costs do, in fact, provide a margin for insurers to deviate and/or pay dividends. Furthermore, it demonstrates that for the prior decade, there have been deviations and dividends in North Carolina in excess of total savings and shows the actual dividends and deviations in dollars from 1983 through 1992.
77. Using the historical results in the evidence supplied by the Bureau and used in the Jordan model, it appears that a reasonable margin has been included in prior rates for the accumulation of surplus for the payment of dividends and deviations even without an extra explicit expense load provision as set forth in this filing. These margins are set forth below.
78. These margins were provided by an average manual premium. The Commissioner finds and concludes that any margin in excess of the margin provided for in the average manual premium is unreasonable and produces rates that are excessive and unfairly discriminatory.
79. Based on the foregoing, the Commissioner finds that profit provisions of -3.75 for liability and +1.75 for physical damage will provide 4.96% of manual premiums, or $90 million, that may be dividended and deviated as a savings to insureds, assuming the same book of business. DOI-44, p. 4; Jordan Transcript, pp. 1826-1828.
*68480. The 4.96% of premium or approximately $90 million provided in the manual rate for policyholder dividends and deviations by the Bureau member companies is reasonable, adequate and is provided in the rates which are adopted and approved hereinafter by this Order and which are not inadequate, excessive or unfairly discriminatory.
81. Dividends and deviations in excess of the 4.96% of premium or approximately $90 million may occur, as in the past. If so, the excess may come from companies which are prepared to accept, on an individual basis, less than the average profit provided in the manual rate, from accumulated surplus, from lower expenses, from an excessive rate level implemented by the Bureau or from sources which are not within the jurisdiction of the Commissioner.
82. This 4.96% of the premiums will become retained earnings, i.e. profit, if it is not dividended or deviated. Including more than 4.96% of premium for dividends and deviations in the rate calculation will cause rates to spiral and become excessive and unfairly discriminatory.
While there is substantial evidence to support the majority of the Commissioner’s findings regarding dividends and deviations, we are unable to determine from the record exactly how the Commissioner selected the figure of 4.96% as the amount which may be used for dividends and deviations. Department exhibit DOI-44 appears to be the basis for the Commissioner’s figure; however, we agree with the Bureau that the estimates shown in the chart for the years 1993, 1994 and 1995 are not supported by the evidence and there are no findings explaining the Commissioner’s estimates, particularly how the Commissioner chose the 4.96% figure. Consequently, we remand to allow the Commissioner to make specific findings that clearly show the facts upon which he based his decision that the rate contains a 4.96% margin for dividends and deviations. See State ex rel. Comr. of Ins. v. N.C. Rate Bureau, 97 N.C. App. at 647, 389 S.E.2d at 576 (remanding the issue of underwriting profit provisions due to insufficient findings in the Commissioner’s order).
III. INVESTMENT INCOME FROM CAPITAL AND SURPLUS FUNDS
[4] We agree with the Bureau’s next contention that the Commissioner erred as a matter of law in considering investment income from capital and surplus in his ratemaking calculations.
*685In Comr. of Insurance v. Rate Bureau, 300 N.C. 381, 269 S.E.2d 547, our Supreme Court examined the issue of income on invested capital. The Court said:
B. The Majority Rule
We also find our view consistent with that prevailing in other jurisdictions. In 2 Couch, Insurance Law § 21:38 at 494 (Anderson ed. 1959) it is said:
“In determining whether an insurer has made a reasonable profit, the amount of business done rather than its capital should be considered, and profits should be determined by subtracting losses and expenses from the total premiums actually received, to the exclusion of profit on capital and surplus, and excess commissions paid to agents but considering interest on unearned premiums and related elements.
Id. at 444, 269 S.E.2d at 586 (alteration in original). After summarizing the issue, the Court concluded, “prior decisions in this State have sustained the view that investment income from unearned premiums and loss reserve funds are appropriately considered in a ratemaking hearing. . . . Neither prior cases nor statutes, however, have permitted consideration of invested income from investment capital.” Id. at 446, 269 S.E.2d at 587. In subsequent cases, this Court has clearly followed the Supreme Court’s directive. In State ex rel. Comr. of Insurance v. N.C. Rate Bureau, 75 N.C. App. at 228, 331 S.E.2d at 142 our Court said, “the Commissioner may not consider investment income from capital and surplus accounts . ...” We remanded the issue of the Commissioner’s selection of underwriting profit provisions for more findings showing how the Commissioner’s figure was made without consideration of “investment income from capital and surplus” in State ex rel. Comr. of Ins. v. N.C. Rate Bureau, 97 N.C. App. at 647, 389 S.E.2d at 576.
In this order, the Commissioner set forth the methodology by which he calculated the underwriting profit and contingency factor. He discussed the various methods used by several department and Bureau expert witnesses and then found, “it is appropriate to use the formula of the Bureau and O’Neil [a department expert] to calculate the target underwriting profit and contingency factor in this proceeding, with due consideration and appropriate adjustments to each factor in the calculation.” This formula included a line item and calculation for “Income from Capital and Surplus.” The Bureau witness *686whose testimony the Commissioner relied upon in calculating this figure stated:
[I]n addition to including investment income from loss, expense and unearned premium reserves, it includes installment payment income and realized and unrealized capital gains. It also includes both investment income and capital gains on stockholder-supplied funds, i.e. capital and surplus. Thus, it is a total return and not just a return on insurance operations.
The Order also states that in addition to the Bureau witness, the Commissioner relied upon the figures presented by department witness, O’Neil. In prefiled testimony, O’Neil stated she reflected investment income in her calculations by “us[ing] the Rate Bureau’s calculation procedure. . . . My results differed slightly from the Rate Bureau’s because my expected loss ratio . . . differs from the Rate Bureau’s value because of differences in other underlying assumptions such as the treatment of dividends . . . .”
This order is remanded for recalculation of the underwriting profit provisions. The formula used must exclude investment income earned on capital and surplus.
IV. UNDERWRITING PROFIT PROVISIONS
The Bureau next argues the Commissioner erred in reducing the filed underwriting profit provisions from .8% to -3.75% for liability coverage and from 5.4% to 1.75% for physical damage. According to the Bureau, the Commissioner reached these figures by (1) accepting the 13.9% filed target return on net worth, but then improperly converting this to a return on statutory surplus; (2) improperly assuming a hypothetical capital structure by adopting a “normative” premium-to-surplus ratio rather than the existing ratio in North Carolina; and (3) improperly finding that the premiums used for prepaid expenses and agents’ balances remain available for investment.
A. Rate of Return
[5] Essentially, the Bureau argues part of the methodology employed by the Commissioner in determining the underwriting profit provisions was faulty because the Commissioner used the more conservative accounting system known as SAP (statutory accounting principles) as opposed to the GAAP system (generally accepted accounting principles). The Bureau contends SAP, established by the National Association of Insurance Commissioners (NAIC), is inap*687propriate because its purpose is to measure the liquidation value of a company and it does not include all of a company’s assets in its calculations. By contrast, the GAAP system measures the financial condition of a company as an ongoing concern, not its liquidation value. According to the Bureau, the more conservative SAP approach to measuring assets produced a chain of reactions: an understatement of the value of the aggregate insurance company and a smaller base upon which to apply the return, and ultimately resulted in a lower rate of return. Additionally, the Bureau contends the Commissioner’s calculations are not supported by material and substantial evidence.
The Bureau has not cited any authority and we find nothing in the cases or statutes which prescribe the system the Commissioner must use, either SAP or GAAP, in calculating these profit provisions. In Comr. of Insurance v. Rating Bureau, 292 N.C. 471, 489, 234 S.E.2d 720, 730 (1977), our Supreme Court said:
The ultimate question for the Commissioner’s determination is whether the proposed rates will, after provision for reasonably anticipated losses and operating expenses, leave for the insurers ... a fair and reasonable profit and no more. The purpose of the entire statutory plan is to provide for the public, at reasonable cost, insurance in financially responsible companies. The public interest extends as truly to the financial responsibility of the insurer as it does to the reasonable cost of the insurance to the insured, and vice versa, (citations omitted) (emphasis added).
The Commissioner is considered an expert in the field of insurance and his reliance on various “methods of analysis of the profit to which the insurance companies are entitled lies entirely within his discretion.” State ex rel Comr. of Insurance v. N.C. Rate Bureau, 96 N.C. App. at 223, 385 S.E.2d at 512. The rates the Commissioner determines in his order are prima facie correct so long as there is substantial and material evidence to support the Commissioner’s findings. G.S. 58-2-80; G.S. 58-2-90(e).
We find there is substantial and material evidence to support the Commissioner’s use of SAP in calculating the profit provisions. Not only was there expert testimony that SAP was the appropriate method, but as the Commissioner pointed out in his order, even our statutes refer to the accounting practices set forth by the NAIC (i.e. SAP system) in requiring insurance companies to evaluate and make regular reports of their financial positions. N.C. Gen. Stat. § 58-2-165. Additionally, the Commissioner reasons that since SAP represents *688that level of financial commitment an insurance company is legally required to make to its policyholders, it is a logical foundation upon which to base a rate of return in determining “a fair and reasonable profit and no more.” Comr. of Insurance v. Rating Bureau, 292 N.C. at 489, 234 S.E.2d at 730. “As we do not find error in the Commissioner’s judgment we cannot replace our judgment for his.” State ex rel Comr. of Insurance v. N.C. Rate Bureau, 96 N.C. App. at 223, 385 S.E.2d at 512.
We turn now to the Commissioner’s calculations and ultimate selection of 13.67% total rate of return as a percentage of surplus (SAP). The Bureau vigorously argues the Commissioner’s - calculations and findings, to the extent they portend to be based on the Bureau calculations, are “contrived in order to give the appearance of ‘comparable’ [Bureau] results.” While there may be some misleading language in the Commissioner’s findings as to comparisons with Bureau figures, including findings of Fact 47 and 48, we ultimately find there is material and substantial evidence to support the Commissioner’s calculations.
The Commissioner’s order included the following:
SELECTION OF RATE OF RETURN
44. The historical rate of return for the property and casualty insurance industry during the 1981 to 1990 period was 9.6% of GAAP and 9.4% of SAP. DOI-39.
45. Several recommendations as to the appropriate rate of return were advanced in the testimony. See Exhibit A, Section C, pp. 30, 31, line (12).
a. Plotkin testified that a broad spectrum of U.S. industries has historically achieved average returns on net worth (GAAP) in the range of 12% to 15%. He opined that 13.9% (GAAP) was not excessive. RB-15, Plotkin Prefiled Testimony, p. 15.
b. Vander Weide opined that the cost of equity capital for the average company writing private passenger automobile insurance in North Carolina is 13.0% to 15.25% (GAAP), or a fair rate of return of 15.0% to 17.25% on GAAP equity. RB-17, Vander Weide Prefiled Testimony, pp. 4, 17-18.
c. Cohn testified that the appropriate rate of return is on operations (not a total rate of return) based on risk premium, and the required risk premium is 5.2% of surplus. DOI-4, Cohn *689 Prefiled Testimony, pp. 8-16; Exhibit A, Section C, pp. 30-31, line (16). Cohn did not calculate the cost of capital, but would estimate it to be 10%. Cohn Transcript, p. 925.
d. Schwartz testified that the required rate of return on operations is 3.6% of premium (3.8% for liability and 3.2% for physical damage). Exhibit A, Section C, pp. 30-31, line (18). He calculated that the recommended 3.6% results in an overall post-tax rate of return in relation to surplus of 17.3%, which may be overly generous. DOI-5, Schwartz Prefiled Testimony, pp. 12-15.
e. Wilson concluded that a return of 10% on surplus (SAP) is appropriate based on the current cost of equity capital for private passenger automobile insurance companies. DOI-3, Wilson Prefiled Testimony, pp. 5, 44; Exhibit A, Section C, pp. 30-31, line (12).
f. O’Neil concluded that an overall 11% post-tax rate of return (SAP) was appropriate, noting that the reasonable expected rate of return for all industries would be in the range of 10% to 15% (GAAP). DOI-6, O’Neil Prefiled Testimony, pp. 47-49; Exhibit A, Section C, pp. 30-31, line (12).
46. The rates of return recommended by each of the witnesses were derived by reference, in varying degrees, to rates of return in industries other than the North Carolina private passenger automobile insurance industry.
47. The rate of return used by the Bureau in its calculation is a 13.9% return on net worth (GAAP) (see RBI, L-461 and L-465), which converts to a return on surplus of 13.66% for liability and 13.61% for physical damage. Exhibit A, Section C, pp. 32 and 33.
48. In light of all the evidence, the Commissioner selects a 13.67% total rate of return as a percentage of surplus (SAP) for both liability and physical damage coverages. This is the return derived from the adjusted Bureau calculation. Exhibit A, Section C, pp. 32 and 33. Based on the evidence in this case, a 13.67% return leads to a reasonable margin for underwriting profit and to contingencies for the average carrier, and no more. In Re Filing by Fire Ins. Rating Bureau. 215 [sic] N.C. 15, 32-33 (1969); Commissioner of Insurance v. Rate Bureau. 300 N.C. 381, 448-449 (1980). Rates of return of 13.67% for both liability and physical damage are appropriate and do not lead to rates that are excessive, inadequate or unfairly discriminatory. *69049. Therefore, 13.67% is entered into line (12) of Exhibit A, Section C, pp. 30 and 31.
After considering the wide variety of recommendations, the Commissioner used the Bureau’s profit components related to surplus and calculated a return on surplus of 13.66% by multiplying the return on premium (the Bureau's figure of 9.26%) by 1.475, the premium-to-surplus ratio for 1992, the latest single year of data available for this filing. We have already discussed the evidence which allows the Commissioner to use a return on surplus (SAP) as one of the bases for deriving profit. The Commissioner’s choice of the premium-to-surplus ratio for 1992 is also supportable. It represents data from the most current year; Bureau witnesses testified the 1992 data was credible; and the ratio was used in department witness O’Neil’s calculations.
B. Premium-to-surplus ratio:
[6] The Bureau argues “the Commissioner erred as a matter of law in adopting a hypothetical ‘normative’ premium-to-surplus ratio rather than the actual ratio.” This selection, according to the Bureau, further reduced the filed underwriting profit provisions. While the Bureau contends this hypothetical ratio is “error as a matter of law,” the only case cited for this proposition is Comr. of Insurance v. Rate Bureau, 300 N.C. at 450-51, 269 S.E.2d at 589-90. We find this case distinguishable.
In Comr. of Insurance v. Rate Bureau, 300 N.C. 381, 269 S.E.2d 547, the Court’s discussion of the use of a hypothetical rate of return was in the context of the Commissioner’s decision to use the “capital asset pricing model” for determining the underwriting profit margin. The Court stated:
[T]he Commissioner’s requirement for the use of a hypothetical “risk free” rate of return would clearly violate the intent of our Legislature in authorizing insurance companies operating in North Carolina to invest in certain securities. G.S. 58-79.1 specifically requires casualty insurance companies to invest reserve funds in one or more of ten different categories of investments.
It is inconceivable to us that our Legislature intended that insurance companies invest their funds in certain designated securities and then require that those companies’ underwriting profits shall be computed on the hypothetical assumption that they were *691invested in something else. Such an interpretation would, as appellants suggest, “make a mockery of the statute.”
Id. at 450-51, 269 S.E.2d at 589-90. The statute at issue, G.S. 58-79.1 has since been repealed.
In this case, we find there was substantial evidence to support the Commissioner’s selection of a 2 to 1 premium-to-surplus ratio. The 2 to 1 ratio is a traditional standard for the premium-to-surplus ratio and several expert witnesses used this 2 to 1 ratio in their calculations. Additionally, there was testimony that it is more appropriate to use a normative ratio than an historical one when determining rates on a prospective basis. We agree with the Commissioner there is no evidence of error as a matter of law; there is neither a statutory mandate for a premium-to-surplus ratio nor anything to preclude the Commissioner’s use of a hypothetical normative premium-to-surplus ratio as opposed to the actual ratio so long as there is substantial evidence to support the Commissioner’s selection.
C. Treatment of Reserves Subject to Investment
[7] The Bureau’s final contention under the underwriting profit provision argument is that the Commissioner erred by improperly finding that the premiums used for prepaid expenses and agents’ balances remain available for investment. We disagree.
G.S. 58-36-10 states, in part, “[d]ue consideration shall be given to . . . investment income earned or realized by insurers from their unearned premium, loss, and loss expense reserve funds generated from business within this State ....” Section F of the Commissioner’s order examined the issue of investment income from unearned premium, loss, and loss expense reserve funds. In this section, the Commissioner clearly defined the factors involved in considering investment income; selected a reasonable rate of return (7%) on investments; and carefully explained why he concluded the Bureau’s amount of reserves subject to investment was incorrect.
The Commissioner found the Bureau had excluded from the calculations of the reserves available for investment prepaid expenses and agents’ balances. Reasoning that decisions as to how to handle agents’ balances and to prepay expenses are discretionary and outside the control of policyholders, the Commissioner stated, “it is appropriate to allow the interest earned on the entire amount of these funds to accrue to the benefit of policyholders.” This conclusion was supported by the evidence in the record. Expert witness John Wilson *692testified that by reducing the reserve by prepaid expenses and agents’ balances, the Bureau “is saying that the working capital requirements of the insurance company — in particular things like agents’ balances and other important capital requirements, any type of accounts receivable — should be funded by the premium reserve.” Wilson further stated, “[t]hat is not the way in which business operates. ... To make a deduction from policyholder-provided funds rather than owner-supplied funds, for the working capital requirements of the enterprise is incorrect accounting for purposes of ratemaking.” The Commissioner also observed that none of the statutes authorize the Bureau to reduce investment reserves for prepaid expenses and agents’ balances and that the annual statement includes “the entire unearned premium reserve and loss and loss adjustment expense reserves are allocated in their entirety to the benefit of policyholders.”
We find there is substantial evidence in the record to support the Commissioner’s decision as to the Calculation of investment income from unearned premium, loss, and loss expense reserve funds.
Y. GENERAL AND OTHER ACQUISITION EXPENSES
[8] General and other acquisition expenses for liability coverage is figured from information which combines the automobile expense experience from the voluntary market and the Reinsurance Facility (facility market). To calculate the proper expense provision for a filing, this combined expense data is allocated between the voluntary and facility markets. In the past, the Commissioner has approved filings where the expense dollars are allocated to each market based on the number of policies (called exposures) in each market. In this filing, the Commissioner concluded the best method for determining general and other acquisition expenses was by allocating expenses between the voluntary and facility markets by premium volume as opposed to exposures. The Bureau contends the Commissioner erred in accepting this method because it is not supported by substantial or material evidence. We disagree.
As we have already stated, the Commissioner weighs the sufficiency of the evidence as well as the credibility of the witnesses. State ex rel. Comr. of Insurance v. N.C. Rate Bureau, 96 N.C. App. at 221, 385 S.E.2d at 511. Substantial evidence is “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Comr. of Insurance v. Automobile Rate Office, 287 N.C. at 205, 214 S.E.2d at 106. Additionally, we note the Commissioner is *693free to consider and adopt a new method of allocation. As pointed out in the Commissioner’s brief, our Supreme Court has stated, “[i]t is not a proper ground for the rejection of such evidence that such projection . . . has never before been used in the rate making process. The statute does not contemplate that procedures and methods for determining replacement costs for the future shall be frozen.” In re Filing by Fire Ins. Rating Bureau, 275 N.C. 15, 36, 165 S.E.2d 207, 222 (1969). “[T]he head of an administrative agency in the executive branch of our government clearly has the power, provided he follows legal means, to chart new courses in discharging the functions of his office.” Comr. of Insurance v. Rate Bureau, 300 N.C. 485, 491, 269 S.E.2d 602, 606 (1980).
In choosing to change the method by which general and other acquisition expenses are calculated, the Commissioner had before him expert testimony from department witness Allan I. Schwartz, an actuarial consultant who had conducted an independent analysis of the issue. Schwartz testified, “the key issue in this item is how these expenses are distributed between the voluntary and involuntary [North Carolina Reinsurance Facility (NCRF)] markets. There are two main methods used to allocate these types of expenses. One would be on the basis of premiums, the other would be on the basis of exposures.” Schwartz carefully explained the differences between allotment by exposure as opposed to allocating by premium. He reasoned that some expenses like advertising and a portion of boards, bureaus and surveys are already allocated on the basis of premiums and he stated, “ [i]t is clear that a significant portion of other acquisition plus general expenses are more closely related to premiums than to exposures.” Other evidence presented by Schwartz included a study performed by an industry-sponsored organization (AIPSO) responsible for the administration of automobile insurance residual markets in most jurisdictions. The results of this study showed the expense ratios are higher for the residual market than for the voluntary market whereas the Bureau method results in a much higher expense ratio for the voluntary market than the residual market. Schwartz was asked why insurance companies would take “contradictory positions through two of their agents (i.e. AIPSO and NCRB [Bureau]) with regard to the relationship of expenses to premiums for the residual market in comparison to the voluntary market.” He responded:
As it turns out, in most states other than North Carolina, residual market rates are more heavily regulated than the voluntary mar*694ket rates. Hence, in those jurisdictions a financial incentive exists for insurance companies to try and push as many expense dollars as possible into the residual market rate level.
In North Carolina, however, residual market rates are less regulated than the voluntary market rates. Hence, there is a financial incentive in North Carolina for insurance companies to try and push as many expense dollars as possible into the voluntary market rate level.
Hence, what at first may seem to be illogical and contradictory positions by insurance companies is simply consistent with an attempt by insurance companies to try and load as many expense dollars as possible into the rate level that is more closely regulated.
Schwartz went on to recommend to the Commissioner “that a reasonable procedure would be to allocate other acquisition and general expenses between the voluntary and residual markets based upon premiums.” We conclude there was substantial evidence before the Commissioner to support his method of calculating general and other acquisition expenses by allocating expense dollars based on premium volume as opposed to exposures. Therefore, we overrule this assignment of error.
VI. CURRENT COST AND EXPENSE TREND PROVISIONS
[9] The Bureau next argues the Commissioner erred in disapproving the filed current cost and expense trend provisions and in ordering rates based on inadequate current cost and expense trend provisions.
As we outlined at the beginning of this opinion, G.S. 58-36-70(d) states, in part:
If the Commissioner finds that a filing complies with the provisions of this Article, either after the hearing or at any other time after the filing has been properly made, he may issue an order approving the filing. If the Commissioner after the hearing finds that the filing does not comply with the provisions of this Article, he may issue an order disapproving the filing, determining in what respect the filing is improper, and specifying the appropriate rate level or levels that may be used by the members of the Bureau instead of the rate level or levels proposed by the Bureau filing, unless there has not been data admitted into evidence in *695the hearing that is sufficiently credible for arriving at the appropriate rate level or levels.
“In reaching his ultimate determination, the Commissioner must make findings which clearly and specifically indicate the facts on which he bases his order, the resolution of conflicting evidence, and the consideration he has given to the material and substantial evidence that has been offered.” State ex rel. Comr. of Ins. v. N.C. Rate Bureau, 95 N.C. App. at 159, 381 S.E.2d at 803. “[T]he present statute requires the Commissioner to be mathematically specific in rejecting proposed rate increases and future orders should specify ‘wherein and to what extent’ the proposed filings are deemed improper.” Comr. of Insurance v. Rate Bureau, 300 N.C. at 456, 269 S.E.2d at 592-93 (emphasis added).
After explaining the necessity of prospective loss and experience (trending) in insurance ratemaking cases, the Commissioner noted, “[t]he evidence in this case was conflicting concerning trends that will most accurately predict the prospective loss and expense experience . . . .” He then made detailed and specific findings based on department witness O’Neil’s “extensive analysis” on the issue of current cost and expense trend provisions, occasionally making general references to inconsistencies between the Bureau and O’Neil findings. The Commissioner ultimately chose to use O’Neil’s trend selections in the calculation of the rate level change. As to the Bureau’s trends, the Commissioner made the following findings of fact:
4. The Bureau trends were all selected by its committees. However, no testimony was offered by any member of these committees to explain the reasons for the selection of exponential curves as has been automatically made by the Bureau in all auto filings for many years. The mechanical process of calculating the pure premium and cost trend was briefly explained by the Bureau witness Woods, without independent evaluation, on pages 25 and 26 of RB-14. Woods was not a member of the Bureau committees.
. . . [findings 5-10 discussed O’Neil’s research and conclusions]
11. The trends selected by the Bureau will result in rates which are excessive in violation of the law of this state.
12. The trends selected by the Bureau committees and briefly explained by Woods were not determined upon the same thoughtful analysis as those derived by O’Neil and, thus, are found to be less credible, reasonable and reliable, and are rejected.
*69613. For the reasons set forth above, the trend selections derived by O’Neil are accepted as credible and reliable for use in the calculation of the rate level change in this proceeding, while the trends selected by the Bureau are found to lack credibility and are rejected.
O’Neil’s analysis and findings are supported by the evidence in the record; however, the Commissioner’s statements regarding the Bureau’s evidence are conclusory and unsupported by specific evidence. While “[tjhere is no burden upon the Commissioner to disprove the filing,” Comr. of Insurance v. Rate Bureau, 300 N.C. at 455, 269 S.E.2d at 592, the statutes do compel the Commissioner to be specific in rejecting rate increases by stating “ ‘wherein and to what extent’ the proposed filings are deemed improper.” Id. at 456, 269 S.E.2d at 592-93. The Commissioner’s recognition of conflicting evidence, but his failure to resolve the conflicts in precise detail along with his failure to specifically show he has given consideration to the material and substantial evidence the Bureau offered before rejecting the Bureau in favor of O’Neil’s evidence require us to remand this issue to the Commissioner for more specific findings as to the Bureau’s evidence.
VII. FILING DATE ADJUSTMENT
[10] The Bureau argues the Commissioner erred in (1) disapproving the filing date adjustment it was legally authorized to include in the proposed filing and (2) improperly calculating and ordering into effect his own adjustment, which resulted in further decreasing the overall ordered rate change. According to the Bureau, the Commissioner exceeded his statutory authority by reviewing the filing date adjustment because only the Bureau was given the power to make an adjustment to compensate for the changed filing date.
In 1993, the General Assembly changed the annual filing date for automobile insurance rate filings from 1 July to 1 February. Since this change prevented the Bureau from filing for a rate change in 1993, the General Assembly provided:
With respect to the nonfleet private passenger motor vehicle insurance rate filing made on or before February 1, 1994, the Bureau may file an additional factor for an additional rate increase or decrease to compensate for the changing of the filing rate (sic) from July 1 to February 1 as provided in Section 10 of this act.
*6971993 N.C. Sess. Laws 409, § 11 (emphasis added). Pursuant to this authority, the Bureau included an adjustment which “showed a need for an increase of 8.6% without the filing date adjustment. When combined with the filing date adjustment, the overall filed change was 10.7%.” The Commissioner, using the Bureau’s methodology, calculated and ordered into effect his own adjustment which resulted in a further decrease of the overall rate change.
While the legislation allows the Bureau an additional filing factor, we are not convinced the General Assembly intended that the ultimate decision on the amount of the additional rate increase or decrease be left solely in the hands of the Bureau. Indeed, the last line of this legislation states, “as provided in Section 10 of this act.” G.S. 58-36-10 is the listing of factors to which the Commissioner must give “due consideration” in insurance ratemaking cases. The filing date adjustment was part of the overall proposed filing and as such, the Commissioner was vested with the statutory authority to disapprove of any provision in the filing and to specify the appropriate rate to be used. G.S. 58-36-70(d).
VIII. CONCLUSION
We have reviewed the parties’ remaining assignments of error and find them to be without merit.
Affirmed in part; vacated in part; and remanded.