Hennessey, C.J.

This case was commenced, pursuant to G. L. c. 175A, § 5A, by the filing in the Supreme Judicial Court for the county of Suffolk of a complaint for review of the decision of the Commissioner of Insurance (Commissioner) establishing medical malpractice insurance rates for physicians *44 and surgeons, effective July 1, 1983. The case was reserved and reported to the full court without decision by a single justice, upon the record of the hearing before the Commissioner.

The complaint 1 was brought by The Medical Malpractice Joint Underwriting Association of Massachusetts (JUA), a nonprofit association established pursuant to St. 1975, c. 362, § 6, and to a plan of operation promulgated by the Commissioner. The JUA is authorized to issue policies of medical malpractice insurance to licensed physicians and hospitals in Massachusetts at rates fixed and established by the Commissioner under G. L. c. 175A, § 5A. The single justice allowed the motion of the Massachusetts Medical Society (MMS), a participant in the rate hearings, to intervene in the appeal.

The relevant facts are as follows. On April 1, 1983, the JUA filed a recommendation for a 162.7% increase in medical malpractice insurance rates for physicians for the period July 1, 1983, to June 30, 1984. On June 2,1983, the Commissioner published a notice of hearing on that request. The hearing, presided over by the Commissioner, was held on six days between July 13 and August 15, 1983. The JUA, the MMS, the Independent Insurance Agents of Massachusetts, Inc. (HAM), and the State Rating Bureau (SRB) participated.

On May 18, 1984, 2 the Commissioner issued his decision ordering a 42% average increase in rates for physicians for the period beginning July 1, 1983. Both the MMS and SRB had recommended that any rate increase be capped at 50%, although their advisory filings indicated that increases of 90.7% (SRB) or 62% (MMS) were actuarially indicated.

*45 On June 7, 1984, the JUA filed an appeal of the Commissioner’s decision alleging that the rates established are not “adequate, just, [and] reasonable” as required by G. L. c. 175A, § 5A, 3 and are not “actuarially sound” and “calculated to be self-supporting” as required by St. 1975, c. 362, § 6. 4 Specifically the JUA claims that the Commissioner erred in: (1) imposing an especially high standard of proof on proponents of a rate increase, (2) ruling that G. L. c. 175A, § 5A, does not require the rates to be self-sustaining in each rating period, and (3) establishing an average rate increase which is lower than the recommendation of any expert actuarial witness and which is unsupported by substantial evidence in the record. We agree that the Commissioner misinterpreted the applicable statutes and, as a result, reached a decision that is not supported by substantial evidence in the record. We therefore reverse and remand the case to the Commissioner for proceedings consistent with this opinion.

1. Standard of Proof.

The Commissioner applied to these proceedings the “evidentiary standard” set forth in a decision of his predecessor establishing medical malpractice rates for 1976-1977. See 211 Code Mass. Regs. § 61.00 (1978). He quoted from that earlier decision in relevant part: “Balancing the costs of an overestimate [of rates] against those of an underestimate, we feel compelled *46 to place a substantial evidentiary burden on those who would seek higher rates. We will not reject arguments for increases on this basis, but we will require that they be justified better than we might in a more conventional rate case.” The JUA contends that the Commissioner erred by imposing an especially high evidentiary burden on proponents of a malpractice insurance rate increase and that his error materially contributed to the establishment of an inadequate rate. We agree.

At the outset, we note that the controlling statutes do not subject proponents of rate increases to a higher standard of proof than is required of those seeking lower rates. General Laws c. 175A, § 5A, states that the Commissioner shall fix and establish “adequate, just, reasonable and nondiscriminatory premium charges . . . .” In St. 1975, c. 362, § 6, the Legislature further provides that “[t]he resultant premium rates shall be on an actuarially sound basis and shall be calculated to be self-supporting.” The Legislature has directed the Commissioner to set rates that meet these standards, regardless of whether such rates result in an increase or decrease of premium charges.

In the absence of a statutory provision to the contrary, we have held that “[p]roof by a preponderance of the evidence is the standard generally applicable to administrative proceedings.” Craven v. State Ethics Comm’n, 390 Mass. 191, 200 (1983). A higher standard of proof is required only “in a very limited number of cases where ‘particularly important individual interests or rights are at stake.’ ” Id., quoting Herman & MacLean v. Huddleston, 459 U.S. 375, 389 (1983). See Guardianship of Roe, 383 Mass. 415,423 (1981) (higher standard of proof required in civil case only when person may receive stigma comparable to criminal conviction and faces loss of liberty). We have noted that the “imposition of even severe civil sanctions that do not implicate such interests has been permitted after proof by a preponderance of the evidence.” Craven, supra at 200, quoting Herman & MacLean, supra at 389-390. According to these principles, it is clear that the monetary interests at stake in these proceedings do not merit *47 the imposition of a heightened evidentiary standard. Furthermore, it is unclear from the Commissioner’s decision just what standard he did apply. We have indicated that the adoption of an intermediate standard of proof, such as the “clear and convincing” standard, too often serves “as the functional equivalent for the more familiar ‘reasonable doubt’ standard.” Guardianship of Roe, supra at 424, quoting Custody of a Minor (No. 1), 377 Mass. 876, 885 (1979). See Stone v. Essex County Newspapers, Inc., 367 Mass. 849, 877-878 (1975) (Quirico, J., dissenting). At least once in his decision the Commissioner refers to the applicable standard as “more stringent” than the “compelling evidence standard.”

The Commissioner and MMS contend that it is routine practice to assign the burden of proof to the proponent of an insurance rate increase and that the Commissioner only required a “somewhat higher” “weight of that burden” in setting malpractice insurance rates. To support their contention, they cite cases arising under statutes regulating workers’ compensation insurance rates, G. L. c. 152, § 52, noncompulsory motor vehicle insurance rates, G. L. c. 175A, § 7, insurer contracts for medical and surgical services, G. L. c. 176B, § 4, and public utility rates, G. L. c. 164, § 94. However, such cases are inapposite for two reasons. First, the Commissioner did not simply allocate the burden of persuasion to the JUA as the proponent of a rate increase. He required an especially high degree of proof in support of the JUA’s recommendation. None of the cases cited requires proponents of rate increases to support their proposals with more than a preponderance of the evidence. More importantly, the cases relied upon arise under statutes authorizing the Commissioner only to approve or disapprove rate filings rather than to fix rates himself. In contrast, G. L. c. 175A, § 5A, requires the Commissioner to investigate, fix, and establish medical malpractice insurance rates in conformity with statutory requirements. Cognizant of this difference in the Commissioner’s statutory duties, we have placed the burden of proof on insurer-petitioners where they are responsible for filing a proposed rate increase. Liberty Mut. Ins. Co. v. Commissioner of Ins., 366 Mass. 35, 42 (1974). *48 Massachusetts Medical Serv. v. Commissioner of Ins., 346 Mass. 346, 348 (1963). Although the JUA was the first to request a malpractice insurance rate increase, it was the Commissioner who formally initiated the rate making proceedings by issuing notice of the hearing. In light of his duties under G. L. c. 175A, § 5A, we think the essential principle is that the record must support the Commissioner’s decision.

Despite the absence of any statutory provision which varies the standard of proof required in medical malpractice insurance rate hearings, the Commissioner argues that his view comports with legislative intent as evidenced by the design of St. 1975, c. 362, § 6. Incorporating the reasoning of a predecessor Commissioner’s 1976-1977 decision, he cites the statute’s “asymmetry of treatment between rate surpluses and rate deficits” and concludes that the remedy provided for premium undercharges is more explicit, more effective, and imposes lower social costs. Therefore, he concludes, the Legislature implicitly expressed a preference for erring on the side of setting inadequate rates and accordingly intended the use of “an asymmetrical burden of proof.” There are several flaws in this reasoning. At the time St. 1975, c. 362, § 6, was enacted — and at the time of the Commissioner’s 1976-1977 decision construing that statute — the JUA was to recoup deficits by either an assessment on policyholders or a prospective increase in rates. If, on the other hand, malpractice insurance rates resulted in any profit to the JUA, the surplus could either be refunded to policyholders in the form of a dividend or added to the JUA’s loss reserves. In the 1976-1977 decision the Commissioner reasoned that, given the imprecision of medical malpractice insurance rate setting and the likelihood that the rates set would be imperfect, the availability of two mandatory methods of recouping deficits meant that the JUA was better protected against inadequate rates than policyholders were protected against excessive rates. In 1980, the Legislature amended the statute by eliminating the assessment on policyholders as a means of recouping deficits, leaving only the possibility of prospective rate increases. St. 1980, c. 333. As a result of this legislative change, the alleged asymmetry in the statute has been eliminated.

*49 The Commissioner maintains that requiring a highter standard of proof for rate increases remains warranted by policy considerations. He contends that a rate increase of the magnitude proposed by the JUA would “cause serious dislocations in the health care market” which the Legislature intended to minimize by enacting St. 1975, c. 362, § 6. He speculates that, in the event rates prove excessive and a refund is given, physicians could pocket the refund dividends and reap a windfall while patients, who bore the cost of the high premiums, would receive nothing. The Commissioner further justifies his position by noting that, as the only underwriter of medical malpractice insurance, the JUA is not in danger of losing policyholders to private carriers should it incur a deficit and be forced to impose a prospective rate increase. Finally, he claims that growing JUA reserves will cushion the impact of inadequate rates by allowing the insurer to operate until losses are recouped.

The JUA has raised strong counter arguments to the Commissioner ’ s.rationale . 5 More importantly, however, we have stated with regard to the Commissioner’s analogous duty under G. L. c. 175, § 113B, to set “adequate, just, reasonable and nondiscriminatory premium charges” for motor vehicle liability insurance, that he “shall be bound in good conscience to consider the evidence, to be guided by that alone, and to reach his conclusion uninfluenced by extraneous considerations which in other fields might have play in determining purely executive action.” American Employers’ Ins. Co. v. Commissioner of Ins., 298 Mass. 161, 167-168 (1937), quoting Morgan *50 v. United States, 298 U.S. 468, 480 (1936). The Commissioner’s arguments do not demonstrate any legislative preference for erring on the side of inadequate medical malpractice insurance rates. Nor do we think that the statutes allow him the discretion to apply extraordinary standards of proof with that “extraneous consideration” in mind. Although deference may be due to longstanding administrative construction, it is ultimately our responsibility to interpret the applicable statutes. Johnson v. Martignetti, 374 Mass. 784, 790 (1978). In doing so, we look to the plain language of the statutes for our principal insight into legislative purpose. Hoffman v. Howmedica, Inc., 373 Mass. 32, 37 (1977).

While we appreciate the unusual complexity of malpractice insurance rate setting and realize that the Commissioner must exercise independent judgment in weighing widely varying predictive data, he cannot attempt to mitigate potential errors by skewing the process against a rate increase. The legislative mandate is clear. After a fair and even-handed appraisal of all the evidence, the Commissioner is to exercise impartial judgment in seeking to establish adequate, just, actuarially sound, and self-supporting rates.

In at least one area, the Commissioner’s decision suggests the impact of his error regarding the applicable standard of proof. Among the most contested of the rate components was the projection of over-all claim frequency for 1983-1984. The JUA proposed the adoption of an annual frequency trend, based upon past JUA experience. While conceding that claim frequency has been rising in the recent past, the Commissioner found insufficient evidence that this trend would continue. Instead he allowed a one-time increase of 15% in the base frequency rate to reflect the potential effect of our decision in Harnish v. Children’s Hosp. Medical Center, 387 Mass. 152 (1982) (recognizing a cause of action for failure to disclose information sufficient to allow the patient to give informed consent). In rejecting the JUA’s proposal, the Commissioner stated, “I would be justified in finding the evidence as a whole insufficient to dislodge the null hypothesis [i.e., presumption that no frequency adjustment should be made] in light of the *51 compelling evidence standard applied generally to frequency adjustments and, in particular, the even more stringent evidentiary standard applied in this proceeding” (emphasis added). Whether or not the JUA proposal on claim frequency was justified, it cannot properly be subjected to this “more stringent evidentiary standard.”

The Commissioner’s error with respect to the standard of proof applied in the proceedings is sufficient reason to reverse and remand his decision. However, because the other issues raised by the JUA’s appeal are likely to reappear in future proceedings, we will consider their merits as well.

2. Actuarially Sound Rates.

The JUA also argues that the Commissioner failed to comply with his statutory obligation to set “actuarially sound” and “self-supporting rates” for each rating period. According to G. L. c. 175A, § 5A, “[t]he commissioner shall, annually on or before December first, after due hearing and investigation, fix and establish fair and reasonable classification of risks and adequate, just, reasonable and nondiscriminatory premium charges on claims made and occurrence basis to be used and charged by companies in connection with the issue or execution of medical malpractice insurance for the ensuing calendar year or any part thereof.” In addition the Legislature has required that the rates for medical malpractice insurance be established on “an actuarially sound basis and. . . calculated to be self-supporting.” St. 1975, c. 362, § 6. Despite this legislative mandate, the Commissioner has indicated that, in his judgment, “the statute does not require the rates to be self-sustaining in each annual rating period.” While concluding that an average rate increase of 42% was actuarially necessary, he states that he would be inclined to place an artificial cap on increases, as proposed by the MMS and SRB, “were the actuarially indicated rates in my judgment significantly higher than I have found.” We disagree with the Commissioner’s interpretation of the statutes. A fair reading of the statutes in their entireties, including the recoupment provisions of c. 362, § 6, leads us to conclude that the Commissioner is required to set actuarially sound rates on an annual basis.

*52 The language of G. L. c. 175A, § 5A, is based upon and is virtually identical to the provisions of G. L. c. 175, § 113B, authorizing the Commissioner to fix rates for compulsory motor vehicle insurance. Consequently the decisions of this court interpreting and applying the standards of § 113B are equally applicable here. Commonwealth v. Hartnett, 3 Gray 450, 451 (1855) (“[W]hen the . . . legislature, in a later statute, use the terms of an earlier one which has received a judicial construction, that construction is to be given to the later statute”). In construing § 113B we have held that rates must be adequate for each rating period. Insurance Rating Bd. v. Commissioner of Ins., 359 Mass. 111, 115 (1971). We stated that “[rjates are not adequate, fair and reasonable if a large aggregate loss for the preceding year and a probable greater loss for the year in issue are ignored and the preceding year’s rates simply renewed.” Id.

With respect to medical malpractice insurance, the JUA’s rates “shall be based upon the association’s loss and expense experience, and investment income from unearned premium and loss reserves together with such other information based upon such experience as the commissioner may deem appropriate.” The statute creating the JUA provides that the association’s purpose is “to provide medical malpractice insurance on a self-supporting basis.” Any profits earned by the JUA are to be added to the association’s reserves to meet future identifiable losses or are to be returned to policyholders as a dividend. We have already discussed the mechanism by which the JUA is to recoup deficits caused by inadequate premium rates. By providing that rates are to be based on the JUA’s “loss and expense experience” and by setting out specific recoupment procedures for inadequate or excessive rates, the Legislature has demonstrated that it intended malpractice rates to be self-supporting on an annual basis.

Although the Commissioner contends that the rate he set was calculated to make the JUA self-supporting for 1983-1984, he has continued to argue that the statutes do not compel him to do so. He relies upon rate setting decisions in the automobile insurance industry for the proposition that he has discretion to *53 spread the impact of rate changes over more than one year 6 and to engage in “normative ratemaking,” i.e., incorporating cost control into the rate making process. First, it is important to note that in setting motor vehicle insurance rates the Commissioner is not confronted with the specific requirements of St. 1975, c. 362, § 6. However, even if we assume identical rate setting duties for both medical malpractice and motor vehicle insurance, the Commissioner’s reliance on cases reviewing the latter rates is misplaced. In Massachusetts Auto. Rating & Accident Prevention Bureau v. Commissioner of Ins., 381 Mass. 592, 602-603 (1980), and Massachusetts Auto. Rating & Accident Prevention Bureau v. Commissioner of Ins., 384 Mass. 333, 337-338 (1981), we affirmed the Commissioner’s adoption of an allowance for company operating expenses that was lower than the average expenses of all insurers. Because of the wide variation in the expenses of competing automobile insurance companies, we recognized that the Commissioner had “considerable discretion to exclude the highest costs from consideration, even though he cannot prove how they can be reduced.” 384 Mass, at 337-338, quoting 381 Mass, at 603. Here, however, there can be no evidence of such variation in operating expenses because only the JUA is offering medical malpractice insurance in the Commonwealth. Nor does the record provide any evidence that the JUA’s expenses are excessive.

In Attorney Gen. v. Commissioner of Ins., 370 Mass. 791, 804 (1976), we agreed in principle “that regulation should provide cost control as well as cost observation.” There the parties disagreed over the use of internal versus external (i.e., Consumer Price Index) data in projecting automobile claim costs. The Commissioner proposed to examine both internal and external data and employ the lower estimate. We expressed our “doubt that the Commissioner’s rule of preference would always be appropriate. For if the cost changes reflected in the *54 internal data are not subject to control by the insurer, then recognizing only the lower costs shown in an external index would go beyond subjecting a regulated industry to the rigors of simulated competition; it would prevent legitimate price changes based on real and uncontrollable changes in costs.” Id. at 805. In this case there is no evidence that the JUA’s claim costs are excessive due to the association’s failure to exercise internal controls. Nor is there evidence that an external index, like the Consumer Price Index, would validly indicate a trend in the level of medical malpractice claim costs. Without such specific findings, there is no sound basis for the Commissioner to engage in normative rate making.

These arguments and others 7 by the Commissioner indicate that, although he claims to have set “actuarially sound” and “self-supporting rates, ” he was influenced by factors extraneous to and even inconsistent with that obligation. We cannot assume that such erroneous considerations played no role in the selection of a rate lower than that recommended as actuarially necessary by any of the expert witnesses. Accordingly, upon remand, the Commissioner is to set a rate that is actuarially indicated and calculated to be self-supporting for the 1983-1984 period.

3. Substantial Evidence.

Finally, the JUA alleges that the rates established by the Commissioner, which were significantly lower than the recommendations of any expert actuarial witness, were not supported by substantial evidence in the record. We agree.

We have held that motor vehicle insurance rates set under the analogous provisions of G. L. c. 175, § 113B, must be supported by reasonable evidence, Massachusetts Auto. Rating & Accident Prevention Bureau v. Commissioner of Ins., 389 Mass. 824, 827-828 (1983), and that this standard is indistinguishable from the substantial evidence standard contained in G. L. c. 30A, § 1 (6), Workers’ Compensation Rating & *55 Inspection Bureau v. Commissioner of Ins., 391 Mass. 238, 244-245 (1984). Substantial evidence is “such evidence as a reasonable mind might accept as adequate to support a conclusion.” New Boston Garden Corp. v. Assessors of Boston, 383 Mass. 456, 466 (1981), quoting from G. L. c. 30A, § 1 (6). Under this standard, “we are not required to affirm . . . merely on a finding that the record contains evidence from which a rational mind might draw the desired inference.” Id. Rather, our determination of the substantiality of the evidence must be made “upon consideration of the entire record,” id., quoting G. L. c. 30A, § 14 (8), including “whatever in the record fairly detracts from its weight.” Id., quoting Universal Camera Corp. v. NLRB, 340 U.S. 474, 488 (1951). See L. L. Jaffe, Judicial Control of Administrative Action 600-602 (1965).

The JUA contends that the Commissioner’s 42% rate increase is lacking in evidentiary support because it is 20% lower than the recommendation of any of the expert actuarial witnesses. Their recommended increases were 62% (MMS), 90.7% (SRB), and 162.7% (JUA). Because the Commissioner’s decision fell significantly below this range, the JUA argues that the rates he set were not actuarially justified. To support its position, the JUA cites Massachusetts Auto. Rating & Accident Prevention Bureau v. Commissioner of Ins., 362 Mass. 43, 46 (1972), where we reversed the Commissioner’s proposed rates for “increased limits” motor vehicle insurance as “contrary to the recommendations of all experts who testified at the hearing.” This case, however, is unlike the 1972 motor vehicle insurance rate setting case. There the Commissioner wholly rejected the recommendations of the expert witnesses with respect to adjusting the rates to reflect the impact of the recently enacted “no-fault law,” St. 1970, c. 670, § 6. Here the Commissioner combined different rate components proposed by various expert witnesses to arrive at a lower insurance rate than any of these experts proposed. 8 The Commissioner claims *56 that each of these separate components is an independent variable supported by substantial actuarial evidence in the record.* ***** 9 See Massachusetts Auto. Rating & Accident Prevention Bureau v. Commissioner of Ins., 381 Mass. 592, 609 (1980) (separate rate factors should be computed consistently with other assumptions made).

We cannot agree with the JUA that the Commissioner’s method is by itself sufficient proof that his decision was not based on substantial evidence in the record. In Boston Edison Co. v. Department of Pub. Utils., 375 Mass. 1, 12 (1978), we approved the department’s decision fixing a cost of capital for Boston Edison which was not based on the testimony of any one witness but was “a reasoned composite of all the evidence,” based upon an evaluation of “the strengths and weaknesses of each witness’s testimony.” Similarly, in Attorney Gen. v. Commissioner of Ins., 370 Mass. 791, 810 (1976), we refused to overturn the Commissioner’s choice of a trend and projection factor which was the average of the estimates of the hearing participants. In Appellate Tax Board cases we have long recognized that “[t]he board was not required to believe the testimony of any particular witness but it could accept such portions of the evidence as appeared to have the more convincing weight.” Assessors of Quincy v. Boston Con-sol. Gas Co., 309 Mass. 60, 72-73 (1941). See General Elec. Co. v. Assessors of Lynn, 393 Mass. 591, 605 (1984). North *57 American Philips Lighting Corp. v. Assessors of Lynn, 392 Mass. 296, 300 (1984). We see no reason, especially given the nature and complexity of malpractice insurance rate setting, that the Commissioner should not be allowed comparable discretion in exercising his independent judgment as long as his findings are sufficiently supported by the evidence. It is the latter safeguard that we find lacking in the Commissioner’s decision.

In each of the cases cited above, the decision we affirmed fell within the range of the expert testimony, even though it was not proposed by any single witness. Here the Commissioner’s rate increase was significantly lower than all the actuarial recommendations. In these circumstances we have noted that greater specificity in administrative findings and conclusions may be required in order to determine whether the substantial evidence standard has been met. New Boston Garden Corp. v. Assessors of Boston, 383 Mass. 456, 467 (1981). The purpose of our review is “to limit the opportunity for transmuting a preconception into judgment by picking and choosing what will support that preconception and willfully ignoring whatever weighs against it.” Id. at 474 n.ll, quoting Jaffe, supra at 607. In light of the specific requirement in St. 1975, c. 362, § 6, that the JUA’s rates be “actuarially sound” and “self-supporting,” we think that a decision setting rates below the level recommended by every actuarial expert — which nonetheless relies upon their methodologies — must clearly demonstrate that each selected rate component is truly independent and is supported by substantial evidence. On the record before us, we do not find sufficient support for the Commissioner’s decision.

For all of the foregoing reasons, we reverse and remand the case to the Commissioner for reconsideration in accordance with our interpretation of the controlling statutes. In reviewing this matter, the Commissioner should take into account more recent JUA experience in order to bring the advantages of hindsight to bear on the rate setting process. Massachusetts Auto. Rating & Accident Prevention Bureau v. Commissioner of Ins., 362 Mass. 43, 47 (1972). See New England Tel. & *58 Tel. Co. v. Department of Pub. Utils., 360 Mass. 443, 494 (1971); Opinion of the Justices, 328 Mass. 679, 687 (1952).

4. Conclusion.

Judgment is to be entered in the county court reversing the decision of the Commissioner setting rates for medical malpractice insurance for July, 1983-June, 1984, and remanding the case for further proceedings in accordance with this opinion.

So ordered.

Notes

Originally, a second appeal was filed in the county court by the Independent Insurance Agents of Massachusetts, Inc. (HAM), which participated in the hearings before the Commissioner. Upon the joint motion of the parties, the two cases were consolidated for the purpose of appeal. Following the reservation and report by a single justice, HAM withdrew its appeal.

On April 12, 1984, the JUA filed a complaint in the nature of mandamus in the Supreme Judicial Court for the county of Suffolk to compel the Commissioner to issue his decision. On April 18, 1984, the JUA and the Commissioner stipulated that a decision would be made by May 18,1984.

General Laws c. 175A, § 5A, provides in relevant part: “The commissioner shall, annually on or before December first, after due hearing and investigation, fix and establish fair and reasonable classification of risks and adequate, just, reasonable and nondiscriminatory premium charges on claims made and occurrence basis to be used and charged by companies in connection with the issue or execution of medical malpractice insurance for the ensuing calendar year or any part thereof.”

Statute 1975, c. 362, § 6, as amended by St. 1980, c. 333, provides in relevant part: “Effective after the initial year of operation rates, rating plans and any provision for recoupment through policyholder assessment or premium rate increase, shall be based upon the association’s loss and expense experience, and investment income from unearned premium and loss reserves together with such other information based upon such experience as the commissioner may deem appropriate. The resultant premium rates shall be on an actuarially sound basis and shall be calculated to be self-supporting.”

The JUA argues, with good sense, that unfairness to both physicians and patients may result from the establishment of either inadequate or excessive rates. While a refund to policyholders may never reach patients, a prospective rate increase will be assessed directly against future generations of physicians and indirectly against patients who have derived no benefit from the former inadequate rates.

The JUA also disputes the Commissioner’s reference to its loss reserves as cushioning the effect of inadequate rates. As a nonprofit association, the JUA maintains reserves in order to pay known and anticipated claims. These reserves do not represent a surplus or profit that allows the JUA to operate on inadequate premiums.

It is unclear what authority the Commissioner relies upon for this proposition. The motor vehicle rate cases he cites, which are discussed above, are relevant only to his argument concerning normative rate making.

Other arguments which could undermine the selection of an actuarially justified and self-supporting rate are discussed supra. They include construing the statutes as expressing legislative preference for erring on the side of inadequate rates and relying upon the JUA’s monopolistic status and loss reserves to cushion the impact of such inadequate rates.

For the purposes of this appeal, the relevant components are the frequency with which malpractice claims will occur in the future and the *56 “severity” or estimated average cost of claims in the projected period. The frequency component takes into account the ratio of claims closed with indemnity payment to total claims (CWIP) as well as the over-all frequency of reported claims. The Commissioner adopted the MMS approach on CWTP, a combination of the MMS and SRB positions on absolute frequency, and the SRB position with regard to severity of claims.

The JUA contends that the frequency and severity components are necessarily and mathematically interrelated and that by combining allegedly inconsistent estimates of these components from the testimony of different witnesses the Commissioner arrives at a decision which is logically inconsistent and actuarially incorrect. Without deciding the merits of this contention, we note that the Commissioner’s findings are insufficient to establish that each of the components he employed was in fact independent and supported by substantial evidence.