State Ex Rel. Commissioner of Insurance v. North Carolina Rate Bureau

331 S.E.2d 124, 75 N.C. App. 201, 1985 N.C. App. LEXIS 3622

• Court: Court of Appeals of North Carolina
• Decided: June 18, 1985
• Docket: 8410INS744
• Status: Published

Opinion

WELLS, Judge.

The Rate Bureau, in its 108 page brief, brings forward forty-four assignments of error incorporated into eleven arguments based on 274 exceptions noted in the record. Because of the extensive number of errors alleged, and because of the complexity of the issues presented, only those arguments necessary to a determination of the material issues presented by this appeal will be addressed. We deem it unnecessary to discuss the Rate Bureau’s arguments numbered eight, nine, and eleven because these arguments present issues addressed in other assignments of error, or they involve findings of fact by the Commissioner that are unnecessary to the order. We ultimately find that the Commissioner’s order must be vacated and remanded.

I. Rate Bureau Jurisdiction

The Rate Bureau argues that the Commissioner erred by conditioning approval of an 11.7 percent rate increase for farm-owner insurance coverages subject to the Rate Bureau’s jurisdiction on a filing for a rate decrease for farmowner insurance coverages not subject to the Rate Bureau’s jurisdiction. We agree and vacate that part of the Commissioner’s order.

The North Carolina General Assembly created the North Carolina Rate Bureau and empowered it with the authority to promulgate rates for all insurance companies writing specified lines of insurance in this state. N.C. Gen. Stat. §§ 58-124.17 to -124.30 (1982 and Cum. Supp. 1983). Article 12B of Chapter 58 prescribes the lines of insurance subject to the Rate Bureau’s jurisdiction. Among the various insurance coverages subject to the Rate Bureau’s authority, the General Assembly provided that:

The Bureau shall have the duty and responsibility of promulgating and proposing rates for insurance against loss to residential real property with not more than four housing units located in this State and any contents thereof or valuable interest therein and other insurance coverages written in connection with the sale of such property insurance. . . . The Bureau shall have no jurisdiction over . . . farm buildings other than farm dwellings and their appurtenant structures [and] farm personal property . . .

G.S. § 58-124.17(3) (emphasis added). Insurance companies writing property and casualty lines of insurance which are not subject to the Rate Bureau’s jurisdiction are subject to a separate rate making scheme under Article 13C of Chapter 58. N.C. Gen. Stat. §§ 58-131.34 to -131.60 (1982). The farmowner insurance coverages specifically excluded from the Rate Bureau’s jurisdiction in Article 12B are explicitly included within Article 13C:

[T]his Article [13C] shall apply to insurance against loss to farm buildings (other than farm dwellings and their appurtenant structures) [and] farm personal property . . .

G.S. § 58-131.36(11).

The authority of the Commissioner to review, approve, modify, or disapprove insurance rates promulgated by the Rate Bureau is limited to that authority granted by the General Assembly. E.g., Comr. of Insurance v. Rate Bureau, 300 N.C. 381, 269 S.E. 2d 547, reh’g denied, 301 N.C. 107, 273 S.E. 2d 300 (1980); Comr. of Insurance v. Rate Bureau, 43 N.C. App. 715, 259 S.E. 2d 922 (1979), disc. rev. denied, 299 N.C. 735, 267 S.E. 2d 670 (1980). Since the Commissioner’s duties and responsibilities are fixed by the General Assembly, “he may act only to the extent and in the manner legislatively prescribed.” Comr. of Insurance v. Rate Bureau, 43 N.C. App. 715.

We conclude from the explicit language of G.S. §§ 58-124.17 (3) and -131.36(11) that the Commissioner did not have statutory authority to withhold approval of the 11.7 percent rate increase on the condition that ISO file for a rate decrease for Article 13C coverages. As applied to the farmowner insurance program, both statutes contemplate that coverages applicable to farm residences and appurtenant structures are subject to Article 12B and that the insurance rates for coverages applicable to the commercial operations and property of the farmowner are subject to Article 13C. The distinction drawn for the farmowner insurance program is consistent with the General Assembly’s intent to subject “essential” lines of insurance to Rate Bureau jurisdiction, while permitting Article 13C filing for most commercial lines of insurance, deemed “non-essential.” See Comr. of Insurance v. Rate Bureau, 300 N.C. 381.

The Commissioner relies on N.C. Gen. Stat. § 58-44.3 (1982) for the requisite authority to withhold approval of Article 12B rates. G.S. § 58-44.3 is a part of the anti-rebate statutes. N.C. Gen. Stat. §§ 58-44.3, -44.5, -54.4(8) (1982). These statutes prohibit an insurer or insurance agent from “discrimination” in setting rates for any person. They are obviously designed to prohibit an insurance agent or company from charging reduced or excessive insurance rates contrary to the established rating rules applicable to the risk, cf., Insurance Agency v. Leasing Corp., 26 N.C. App. 138, 215 S.E. 2d 162 (1975) (agent’s agreement to waive short rate cancellation), and are not applicable to rate making. Both Article 12B and Article 13C contain anti-discrimination provisions. G.S. § 58-124.19(1), which governs the Rate Bureau’s filings, specifically provides that rates promulgated by the Rate Bureau cannot be “unfairly discriminatory,” and this provision applies only to insurance coverages subject to the Rate Bureau’s jurisdiction. The Article 13C definition of “discrimination” is substantially different.

The provisions of the Commissioner’s order withholding implementation of the 11.7 percent increase must be vacated. N.C. Gen. Stat. § 58-9.6(b) (1982). Because that part of the Commissioner’s order that must be vacated is clearly separable from the balance of the order in which the Commissioner found an 11.7 percent increase justified, we consider the Rate Bureau’s remaining assignments of error addressed to the merits of the rate increase.

II. Modified Farm Projection Factor

The Rate Bureau contends that the Commissioner erred by disapproving the premium trend and modified farm projection factor contained in its filing. Specifically, the Commissioner, in findings of fact numbered 28 through 45 and conclusion of law number 1, found that the modified farm projection factor employed by the Rate Bureau was not actuarily sound and resulted in excessive rates.

The “modified farm projection factor” represents a mathematical calculation by which the effect of inflation on future losses and the effect of increases in premiums due to insureds increasing policy limits are accounted for during the period covered by the rate filing. By “trending,” the rate maker projects known losses and anticipated premiums into the future to reflect factors that will either increase or decrease losses to be paid or premiums to be collected. The farm projection factor consists of two separate components. First, future losses must anticipate inflation in repair or replacement costs, thereby requiring greater premiums. Second, policy owners increase the amount of their insurance coverage to meet rising replacement or repair costs, thereby generating additional premiums from which insurance companies can pay losses. Rates must be adjusted downward when premiums rise from increased coverage limits. The Rate Bureau’s formula was expressed as a percentage by dividing the trended losses, the numerator, by trended premium increases, the denominator.

The Rate Bureau’s filing was based on loss costs rising by 7.04 percent, projected for the next 20.5 months, and premiums increasing 6.55 percent, projected for the next 14.5 months. The modified farm projection factor netted a result that losses would increase 4 percent more than premiums from increased coverage limits. The Commissioner disapproved the premium trend factor used as the denominator in the modified farm projection factor, finding that the Rate Bureau had provided either inadequate or unreliable supporting data, and he adopted a modified farm projection factor of one (unity).

The appellate standards of review we must apply to the Commissioner’s order are found in the Administrative Procedures Act, particularly N.C. Gen. Stat. § 150A-51 (1983) and the provisions of G.S. § 58-9.6. Our supreme court, in applying these standards, has stated:

[I]t is for the administrative agency to determine the weight and sufficiency of the evidence and the credibility of the witnesses, to draw inferences from the facts, and to appraise conflicting and circumstantial evidence. ... It is not our function to substitute our judgment for that of the Commissioner when the evidence is conflicting. However, as also in dicated above, when evidence is conflicting, the standard for judicial review of administrative decisions in North Carolina is that of the ‘whole record’ test. . . . ‘The “whole record” test is not a tool of judicial intrusion; instead, it merely gives a reviewing court the capability to determine whether an administrative decision has a rational basis in the evidence.

Comr. of Insurance v. Rate Bureau, 300 N.C. 381 (citations omitted). 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.

While the Commissioner’s order must be based on material and substantial evidence in the record, the ultimate burden of proof to justify a rate adjustment and its amount is on the Rate Bureau. Comr. of Insurance v. Rate Bureau, 300 N.C. 381; Commissioner of Insurance v. Rate Bureau, 40 N.C. App. 85, 252 S.E. 2d 811, cert. denied, 297 N.C. 452, 256 S.E. 2d 810 (1979). The Commissioner, however, may not simply declare that the Rate Bureau has failed to meet its burden of proof. The Commissioner must show:

[Specifically . . . how the Bureau has not carried its burden of proof, and, if the Commissioner fails to do so by substantial evidence, the presumption of prima facie correctness given to an order of the Commissioner ... is rebutted.

Commissioner of Insurance v. Rate Bureau, 40 N.C. App. 85 (emphasis in original).

Applying these principles, we review the essential evidence surrounding the modified farm projection factor adopted by the Commissioner. The Rate Bureau’s evidence supporting its projection factor consisted of testimony of Charles Orlowicz, an ISO actuary employed by the Rate Bureau and qualified as an expert witness.

To determine the loss trend component of the farm projection factor, Orlowicz used the North Carolina Boeckh residential index, which measures construction cost changes for housing in North Carolina, and the modified Consumer Price Index, which measures cost changes for items other than construction costs. His ultimate conclusion was that losses had increased at a 7.04 percent annual rate. To determine the premium trend component of the farm projection factor, he used actual North Carolina farm-owner policy data to calculate percentage increases in premiums for the years 1980 and 1981, the only years for which actual farm-owner data was available. Premiums increased 7.8 percent from 1979 to 1980, and increased 5.8 percent from 1980 to 1981. For years prior to 1980 for which farmowner policy data was not available, he used homeowner premium trend data which he modified. The homeowner premium trend was modified by taking the farmowner premium trend for the years in which actual farmowner premium data was available and comparing that data with premium trends in homeowner insurance for the same periods. The difference in the percentage of growth between homeowner and farmowner premium trend was 65.5 percent, and he applied that differential to homeowner data from previous years.

Orlowicz stated that premium trends in homeowner and farmowner insurance were sufficiently analogous to permit the use of homeowner data to derive an actuarily sound farmowner trend. In Orlowicz’s opinion, homeowner premium trends would be greater than farmowner premium trends for several reasons. First, the years for which actual data was compared confirmed that the percentage of farmowner premium trend was less than homeowner premium trend. Second, a homeowner inflation guard endorsement was marketed which automatically increases coverage limits in relation to inflation. Approximately 75 percent of the homeowner policies written in North Carolina have the inflation guard endorsement. The inflation guard endorsement is not written in conjunction with farmowner policies in North Carolina. Third, he testified that the cost of new construction in the non-farm residential market would account for the higher premium trend factor for homeowner policies because there would be less new farm home construction. In his opinion, the Rate Bureau premium trend was conservative when compared to other methodologies employed to determine farmowner premium trend, and the method he employed would result in lower overall premium rates than if other methodologies had been employed.

Philipp Stern, a consulting actuary employed by the Department of Insurance and qualified as an expert witness, testified that the premium trend and modified farm projection factor em ployed by the Rate Bureau were actuarily unsound. He stated that the Rate Bureau had failed to provide underlying data from which the 1980 and 1981 farmowner premium trend could be properly verified. He classified the Rate Bureau’s contention that the farmowner premium trend was approximately two-thirds of the homeowner premium trend as “absurd.” First, there had been no evidence and there was no reason to assume that replacement or repair costs would be different for farmowner and homeowner losses. Second, he concluded that the homeowner inflation guard endorsement was an immaterial consideration. He based this conclusion on the fact that both the homeowner and farmowner policies “required” insureds to maintain coverage at 80 percent of the dwelling replacement cost in order to receive full replacement cost protection. Despite the fact that homeowners could purchase a policy endorsement to increase coverage limits due to inflation, Stern stated that farmowners, and their insurance agents, would have equal motivation to increase coverage limits to maintain replacement cost coverage. Third, Stern stated that the Rate Bureau had not provided sufficient documentation of the farm-owner premium trend for 1980 and 1981, based on actual North Carolina experience. He explained that the Rate Bureau had employed data based on the number of policy transactions. Stern had requested, and he alleged that the Rate Bureau had not provided, data on average farmowner premiums by risk exposure.

Based on Stern’s calculations from seven levels of risk exposure, he testified that the farmowner premium trend for 1979 to 1980 was between 8.2 percent and 8.7 percent, not the 7.8 percent calculated by the Rate Bureau. For 1980 to 1981, Stern calculated a farmowner premium trend factor of 2.3 percent to 2.5 percent, not the 5.8 percent calculated by the Rate Bureau. He noted that the Rate Bureau, using its policy transaction methodology, had reported a first quarter 1980 premium increase of 30.6 percent. When compared to all other quarters reported, Stern stated that the percentage increase was inordinate and not reliable. Stern determined that if the first quarter premium trend was incorrect that the ultimate result would be that the farm-owner premium trend would be 43.6 percent of homeowner as compared to Orlowicz’s determination of 65.5 percent.

Based on his review, it was Stern’s opinion that the Rate Bureau had failed to support its filing and recommended that the filing be disapproved on that basis. In the alternative, Stern proposed that the loss trend and premium trend be equated as being equal, resulting in a modified farm projection factor of one (unity).

From this evidence and from other conflicting evidence in the record that is too lengthy to be outlined here, the Commissioner found that the Rate Bureau had not met its burden of proof on the premium trend. The Commissioner then disapproved the Rate Bureau’s premium trend and adopted Stern’s proposal that the modified farm projection factor be given a value of one: in other words, premium trend and loss trend being held equal. The effect was to reduce the Rate Bureau’s general rate request.

The Rate Bureau vigorously attacks Stern’s actuarial methodology and conclusions contending that his findings were based on guesswork. Applying the whole record test, we hold that the Commissioner did not err in rejecting the Rate Bureau’s premium trend calculation. Stern’s expert evidence in this case was substantial. It afforded the Commissioner the necessary evi-dentiary foundation from which he could find that the percentage increase urged by the Rate Bureau was not properly justified. Stern demonstrated that the first quarter 1980 farmowner premium trend was inconsistent with all other quarters reported, and, contrary to Orlowicz’s testimony, would substantially impact the Rate Bureau’s premium trend. Most importantly, Stern testified that Orlowicz’s methodology in determining that the farm-owner premium trend as a percentage of the homeowner premium trend was inappropriate because it was based on fallacious assumptions. In this respect, the testimony of both Stern and Orlowicz was highly contradictory, both witnesses relying on subjective judgments, especially in determining the effect of the homeowner inflation guard policy endorsement, to derive a farm-owner trend from homeowner data. The weight and credibility of the conflicting evidence was for the Commissioner to decide. We cannot say, as we must in order to sustain the Rate Bureau’s assignments of error, that the Commissioner’s findings did not have a logical basis in the evidence presented.

While we find that there was substantial and material evidence from which the Commissioner could reject the Rate Bureau’s premium trend, we hold that the Commissioner erred in adopting Stern’s recommendation that the premium and loss trends be equated as equal. The Commissioner found as a fact that the Rate Bureau’s loss trend of 7.04 percent was justified. By adopting Stern’s recommendation that the premium and loss trends be equal, the Commissioner impliedly determined that the premium trend was 10.08 percent. Furthermore, Stern admitted that his recommendation was based on an assumption that the premium trend should equal the loss trend. There is no substantial or material evidence to support Stern’s assumption and his recommendation. We, therefore, reject the Commissioner’s modified farm projection factor of one.

On remand of this matter to the Commissioner, he may consider Stern’s testimony that the Rate Bureau’s premium trend was not supported. If the Commissioner should elect to reject the Rate Bureau’s premium trend in its entirety, the Commissioner must give due consideration to a rate adjustment for the Rate Bureau’s loss trend which he found to be justified. We note, however, that ample evidence in the record would permit the Commissioner to calculate a premium trend by modifying the Rate Bureau’s proposed premium trend based on Stern’s testimony that the premium trend should more closely parallel that of the homeowner premium trend. The record evidence of the relationship of farmowner and homeowner premium trend was highly contradictory, especially testimony on the effect of the homeowner inflation guard endorsement on the relationship, and the weight and credibility of the conflicting evidence is for the Commissioner’s determination.

III. Excess Multiplier

The Rate Bureau next contends that the Commissioner erred by disapproving its excess multiplier used in the filing. The Rate Bureau advances numerous arguments by which it contends that the Commissioner failed to base his order on substantial or material evidence in ordering an excess multiplier of 5 percent.

The “excess multiplier” is a computation which provides a premium against catastrophic losses. In the context of farmowner insurance, catastrophic losses would normally result from hurricanes, tornadoes, and severe windstorms. The excess multiplier defines the limits of losses that can normally be anticipated over a period of time, segregates those losses that would be attributable to catastrophic occurrences, and then spreads the losses over a number of years. The intended effect is to prevent rates from fluctuating excessively due to catastrophic losses in any one year.

The Rate Bureau requested a 10.6 percent excess multiplier. Orlowicz conducted a study of hurricanes which affected North Carolina from 1871 to 1983. The study revealed that 31 hurricanes entered North Carolina during twenty-three separate years, an average of one hurricane each 4.9 years. Next, Orlowicz determined that incurred losses which exceeded 80 percent of earned premiums in any given year would be classified excess losses. Because no excess losses had occurred since the farmowner policy was first written in 1962, he added incurred losses and earned premiums for the standard fire and extended coverage policy which was written for farm dwellings and buildings both prior to and following 1962. After adding farm fire and extended coverage statistics, he determined that since 1950, excess losses had occurred in three years; 1954, 1955 and 1960. By dividing normal losses in each of the years reviewed into excess losses in the three years determined to have such losses, he concluded that a 10.6 percent excess multiplier was justified.

Stern testified that the 10.6 percent excess multiplier employed by the Rate Bureau was not based on proper rating experience. He testified that the fire and extended coverage data used by the Rate Bureau was not proper rate making data because the excess multiplier should be based on the same loss experience as used to determine the rate levels. The Rate Bureau had not used fire and extended coverage data in the balance of the rate filing. Stern also stated that use of the 80 percent cut off to determine excess losses was not reasonable. The Rate Bureau had used the 80 percent demarcation because it generated the same number of excess loss years as did a 100 percent cut off applied to fire and extended coverage insurance. In Stern’s opinion, fixing an 80 percent demarcation for excess losses for farmowner insurance because it produced the same number of excess years as fire and extended coverage policies was not actuarily sound. Stern proposed a 100 percent demarcation because it was the same standard applied to fire and extended coverage policies. In addition, Stern testified that the Rate Bureau’s filing only used data from insurance companies reporting to ISO, and did not include data available from the American Association of Insurance Services (hereinafter AAIS) and the National Association of Independent Insurers (hereinafter NAII). Based on ISO, AAIS, and NAII data combined, and using a 100 percent demarcation, the excess multiplier would be 7 percent.

Stern ultimately proposed an excess multiplier of 5 percent. Because there were no excess losses, either at the 80 percent or 100 percent demarcation, based solely on farmowner policy data, Stern calculated his excess multiplier from certain assumptions. First, for the rating years of 1982 and 1983, he assumed that those years would have a normal loss ratio of approximately 65 percent. Second, he assumed that 1985 would have a catastrophic loss ratio of 150 percent. Using a 100 percent demarcation would result in an excess multiplier of 3.9 percent. To arrive at the 5 percent recommended, Stern took a mean of the 7 percent excess multiplier based on combined rating data with a 100 percent demarcation and the 3.9 percent excess multiplier he derived. The mean of 4.1 percent was rounded to 5 percent.

J. Robert Hunter, a member of the Casualty Actuarial Society and the American Academy of Actuaries, and qualified as an expert in property and casualty insurance with a specialty in profitability of property and casualty insurance companies, testified on behalf of the Insurance Department. Hunter testified that the Rate Bureau’s use of fire and extended coverage data in determining the excess multiplier was improper. Instead, he proposed that the 80 percent demarcation be retained, but applied the demarcation to loss ratios exceeding 80 percent for the years 1980 and 1981, not deemed excess years by the Rate Bureau. By including these years in his proposal, Hunter determined an excess multiplier slightly less than the excess multiplier proposed by Stern.

The Rate Bureau first contends that the Commissioner erred by disapproving their excess multiplier because only ISO data was used. In his order, the Commissioner found that the companies reporting to ISO only accounted for some 30 percent of the farmowner business in North Carolina. The Commissioner also found that the balance of farmowner business written in North Carolina was reported to the NAII and AAIS. The Rate Bureau argues that N.C. Gen. Stat. § 58424.20(c) (1982) does not require that data from all companies be used in its rate filings, the NAII data for some years was not compiled on a basis which permits consolidation with ISO data, and the AAIS had collected data only from 1973. The Rate Bureau also contends that there was no evidence suggesting that use of all ISO, NAII, and AAIS data would have produced a result different than that they proposed if each non-ISO organization had compiled data for the thirty-one year period reviewed and that use of all data available resulted in an excess multiplier similar to the one proposed in the filing.

N.C. Gen. Stat. § 58-124.19(2) (1982) provides that in developing insurance rates “consideration shall be given ... to the hazards of conflagration and catastrophe.” Beyond this language, the General Assembly has not given any guidance to the rating standards by which such an allowance is to be made. While we note that the Commissioner’s findings of fact in regard to the appropriate excess multiplier are vague and afford minimum guidance in determining the precise basis on which he reached his determination, it is apparent from his order that the Commissioner adopted Stern’s proposal. We must again apply the “whole record” test to determine the sufficiency of this aspect of the Commissioner’s order.

Initially, we agree with the Rate Bureau’s contention that they are not required by statute to base a rate filing on data from all insurance companies comprising their membership. G.S. § 58-124.20(c) requires the Rate Bureau to “maintain reasonable records ... of the experience of its members and of the data, statistics or information collected or used by it in connection with the rates, rating plans, rating systems, . . . surveys or inspections made or used by it.” While the Rate Bureau’s data, based on ISO’s statistical base, represented approximately 30 percent of the farmowner experience in North Carolina, there is nothing in the record which suggests that this data base, standing alone, would be insufficient to reach a determination of the excess multiplier. That question, however, is not the issue which must be resolved.

The Commissioner’s order indicates that he disapproved the Rate Bureau’s excess multiplier because it was based on ISO data alone while reliable data from the NAII and AAIS was available, and this rating experience dictated an excess multiplier less than that based only on ISO data. At the request of the Department of Insurance, the Rate Bureau combined ISO, NAII, and AAIS data available and applied an 80 percent demarcation to determine excess losses. This calculation indicated an excess multiplier of 9.7 percent. If Stern’s 100 percent demarcation was applied to the combined data, the excess multiplier would be 7 percent. Orlowicz testified that the combination of ISO data, compiled on a premium earned and loss incurred basis, with NAII data, compiled on a premium written and loss paid basis, would understate the excess multiplier. Stern’s expert testimony tends to show that the use of combined data provided a more adequate basis on which to base an excess multiplier. The conflict in expert opinions goes to the weight and credibility of the evidence. The Commissioner resolved the use of combined data against the Rate Bureau, and, based on the whole record, we find no error in the Commissioner relying on the rate making data beyond that compiled by ISO.

The Rate Bureau next advances several arguments in which they contend that the Commissioner erred in disapproving the Rate Bureau’s excess multiplier because fire and extended coverage premium and loss data was used in calculating their excess multiplier. The Commissioner found as a fact that the Rate Bureau had used, “in addition to the farmowners business of the ISO companies, the experience of other business pertaining to farm dwellings written under different kinds of policies which are not under rate review in this filing.”

First, the Rate Bureau argues that the Commissioner failed to give adequate notice of his objections to the use of fire and extended coverage data in the notice of public hearing required by G.S. § 58424.21(a). The statute provides:

At any time within 30 days from and after the date of any filing, the Commissioner may give written notice to the Bureau specifying in what respect and to what extent he contends such filing fails to comply with the requirements of this Article . . .

Id. Our supreme court, in Comr. of Insurance v. Rate Bureau, 300 N.C. 381, discussed the requirements of the statute. The issue before that court was whether the Commissioner knew in advance of the hearing whether the Rate Bureau had used unaudited rate making data. The supreme court found that the record on appeal in that case established that the Commissioner knew the data was unaudited, and held that the statute and fundamental fairness required that he give notice of the deficiency. The court continued:

We wish to emphasize the narrow holding in this portion of our opinion. The Commissioner correctly notes that it was clearly not the intent of the Legislature to prevent the Commissioner from disapproving a filing if matters coming to his attention during the course of a hearing would compel such disapproval. Obviously, matters relating to credibility or other factors might arise during the course of a hearing for which the Commissioner could not have provided notice prior to the hearing. What we hold here, and all that we hold here, is that when the Commissioner knows prior to the giving of public notice ‘in what respect and to what extent he contends such filing fails to comply with the requirements of [the] Article,’ then he must give the specifics in his notice of public hearing. Here, the Commissioner clearly failed to do this with respect to the reliability of the data.

Id.

The Rate Bureau’s farmowner filing stated that the “Excess Loss Procedure is based on North Carolina all class Monoline Farm [fire and extended coverage policies] and Farmowners underwriting experience for all available years, 1950 through 1981 with the earliest available Farmowners experience being reported in 1962.” At the hearing, Stern testified that the filing had given notice that the Rate Bureau had relied on fire and extended coverage data even though he had not realized the implications of its usage at that time. The record on appeal also discloses that the Commissioner knew at the issuance of the notice of public hearing that the Rate Bureau had used fire and extended coverage data. In a letter to the Rate Bureau dated on the same day as the notice of public hearing, the Commissioner requested additional information concerning the filing stating, “[a]lso, it [the excess multiplier] is based on the ‘all class monoline farm and farm-owners underwriting experience.’ ”

Clearly, the Commissioner knew that the Rate Bureau had used fire and extended coverage data prior to the notice of hearing. We hold that the statute and fundamental fairness required the Commissioner to give notice of the nature and extent of any alleged deficiency in the use of fire and extended coverage data. Our holding, however, does not mean that the Commissioner was required to adopt the Rate Bureau’s excess multiplier. We hold, and all that we hold is, that having failed to give notice that the use of farm fire and extended coverage data was deficient, the Commissioner was prohibited from disapproving the Rate Bureau’s excess multiplier solely on that basis. Comr. of Insurance v. Rate Bureau, 300 N.C. 474, 269 S.E. 2d 595 (1980).

The Rate Bureau also contends that the Commissioner erred by rejecting the 80 percent demarcation to separate normal losses from excess losses. The Commissioner, in his order, determined that a 100 percent demarcation was reasonable. The Rate Bureau’s argument raises three issues. First, they contend that the 80 percent cut off was fully justified in the record and this demarcation is used in every other state. Second, they argue that the Commissioner failed to find facts that justified the use of the 100 percent demarcation. Third, they attack the 5 percent multiplier adopted by the Commissioner in his order because it was not based on any data and was premised on erroneous assumptions.

We hold that the Commissioner did not err in disapproving the Rate Bureau’s excess multiplier demarcation of 80 percent. First, the Rate Bureau relies on the fact that the 80 percent demarcation is used in every other state. On such evidence, the Commissioner could have adopted the Rate Bureau’s excess multiplier, but such evidence does not prohibit the Commissioner from adopting a demarcation different from that used in other jurisdictions. See Comr. of Insurance v. Automobile Rate Office, 293 N.C. 365, 239 S.E. 2d 48 (1977) (approving automobile rate system based on motor vehicle point system rather than insurance point system typically used).

Second, the Rate Bureau contends that use of an 80 percent demarcation for farmowner policies was fully justified in the record because it produced a similar number of excess years when compared to fire and extended coverage policies. Orlowicz testified that the farmowner policy provided coverages in addition to the fire and extended coverage policy which further justified the use of an 80 percent demarcation. It was conceded, however, that the amount of losses due to catastrophic events in the three years determined excess years by the Rate Bureau could be related to rate inadequacy rather than catastrophic losses even though hurricanes entered North Carolina in each of the three years described as excess years.

Stern testified that since a 100 percent demarcation was used in fire and extended coverage policies, the same demarcation should be applied in farmowner insurance. In recommending a 5 percent multiplier, Stern based his proposal on combined ISO, NAII, and AAIS data applying a 100 percent demarcation. We conclude that from the conflicting evidence, the Commissioner could properly adopt a 100 percent demarcation.

We hold, however, that the Commissioner erred in adopting Stern’s recommendation of an excess multiplier of 5 percent. Stern developed his recommendation by assuming that 1983 and 1984 would be normal loss years. Then, he developed a hypothetical excess loss of 150 percent which produced an excess multiplier of 3.9 percent. He conceded that if the 1985 excess loss ratio was 200 percent his excess multiplier would double. No evidence in the record before us supports Stern’s assumptions since they are based on speculation rather than any underlying rate making data. While “[t]he language of G.S. 58-248 does not restrict the Commissioner’s consideration to the statistical data furnished by the Rate Office [now the Rate Bureau] and he may consider evidence from other sources if it is otherwise competent,” Comr. of Insurance v. Automobile Rate Office, 292 N.C. 1, 231 S.E. 2d 867 (1977), the Commissioner’s order of a 5 percent excess multiplier is based on evidence that is not material or substantial. We reject the Commissioner’s 5 percent excess multiplier.

On remand, the Commissioner must give due consideration to the fire and extended coverage data used by the Rate Bureau because he failed to place the Rate Bureau on notice that the use of this data was objectionable; however, he may also consider NAII and AAIS data. The Commissioner must also determine the proper demarcation used to identify excess losses. There is ample evidence in the record from which the Commissioner could conclude that the Rate Bureau’s excess loss demarcation is justified and ample evidence from which the Commissioner could modify the Rate Bureau’s request by the use of combined rating data and use of a 100 percent demarcation.

IV. Masonry Veneer and Three Year Policies

By its next argument, the Rate Bureau contends that the Commissioner erred in failing to make a rate adjustment for the reclassification of masonry veneer structures and the incidence of three year policies. We find that the Commissioner erred in failing to consider the Rate Bureau’s filing for rate changes based on masonry veneer classifications, but that the Commissioner did not err in failing to consider the effect on rates of writing three year farmowner policies.

G.S. § 58424.20(a) requires the Rate Bureau to:

[F]ile with the Commissioner copies of the rates, classification plans, rating plans and rating systems used by its members. Each filing shall become effective immediately on the date specified therein but not earlier than 90 days from the date such filing is received by the Commissioner.

In formulating and promulgating rates the General Assembly provided:

Risks may be grouped by classifications and lines of insurance for establishment of rates and base premiums. Classification rates may be modified to produce rates for individual risks in accordance with rating plans which establish standards for measuring variations in hazards or expense provisions or both.. . .

G.S. § 58-124.19(4). When the Commissioner receives a filing he has 30 days in which he “may give written notice to the Bureau spécifying in what respect and to what extent he contends such filing fails to comply with the requirements of this Article and fixing a date for hearing not less than 30 days from the date of mailing of such notice.” G.S. § 58424.21(a).

In its filing, the Rate Bureau clearly stated that masonry veneer farmowner dwellings previously rated as frame structures would be rated as masonry structures. The filing contained no further documentation or explanation for the change. In his notice of hearing, the Commissioner did not specify any deficiency in the requested classification change. Furthermore, the Commissioner’s letter of 1 November 1983 to the Rate Bureau never mentioned the brick veneer change. At the hearing, the Rate Bureau offered substantial evidence as to the requested classification change and its effect on premiums.

The Commissioner’s order mentioned the brick veneer change requested in only three instances. Even though the Commissioner never specifically states his findings of fact on the requested classification change, it is apparent that he disapproved the brick veneer classification. The Commissioner now argues that the proposed change was not properly before him at the hearing. We disagree.

The Rate Bureau’s filing specifically requested the brick veneer classification change even though supporting data was not provided. The controlling statutes do not require the Rate Bureau to provide justification of classification changes in the rate filing itself. G.S. § 58-124.20(a) merely requires that the Rate Bureau file “classification plans, rating plans and rating systems used by its members.” The Rate Bureau filed accordingly. Certainly, the Commissioner had ample authority to give notice to the Rate Bureau that it had not adequately documented the need for the classification. Under G.S. § 58-124.20, the Commissioner could have required the Rate Bureau to submit documentation supporting the requested change. And, since the Rate Bureau has the ultimate burden of proof at the hearing, the Rate Bureau’s failure to provide adequate documentation and subsequent failure to produce any substantial evidence justifying the reclassification at the hearing would have subjected that portion of the filing to disapproval. Having failed to give the Rate Bureau notice of the alleged deficiency, however, the Commissioner was precluded from raising the classification change as an issue at the hearing and was required to permit a rate adjustment on this basis because of the material and substantial evidence offered by the Rate Bureau. Comr. of Insurance v. Rate Bureau, 300 N.C. 381.

The statutory procedures applicable to the brick veneer change are equally applicable to the question of the three year policy term. At the time of the filing, farmowner rating rules required that most policies be written for a three year term, but the premium collected annually. The Rate Bureau had filed with the Commissioner, in a separate filing, a request to write farm-owner policies on a one year basis as is normal practice with homeowner insurance. The Rate Bureau based its farmowner fil ing on one year policies rather than three year policies in anticipation of the Commissioner’s approval of its separate filing. At the time of the rate hearing, the Rate Bureau’s request to write one year policies had been disapproved. Therefore, the Rate Bureau introduced evidence to revise premium trend and loss trend factors upwards to reflect three year policies.

The Commissioner’s order did not allow for additional premiums for the writing of three year policies. He argues that the rate increase needed to reflect the writing of three year policies was not before him at the hearing. We agree.

The legislative rating scheme clearly contemplates that the Rate Bureau must give the Commissioner sufficient notice of its rating plan in order that he may review the plan within 30 days and specifically determine the nature and extent of any deficiencies. Unlike the brick veneer change requested, the Rate Bureau’s filing does not note that disapproval of the separate filing to write three year policies would require increases in premium trends in the filing before the Commissioner. If the Commissioner had simply declined to hold a hearing on the farmowner filing, thereby permitting the Rate Bureau’s filing to become automatically effective, the rates proposed would have been based on the one year policy assumptions filed by the Rate Bureau. Because the Commissioner is required to give notice of all deficiencies within 30 days, and conduct a hearing within 30 days following his notice of hearing, it is only reasonable that the Rate Bureau’s filing give adequate notice of its basis in order that the Commissioner may properly review the filing and be afforded an opportunity to develop proper evidence.

Our holding should not be interpreted to mean that the Commissioner could not have considered the effect of his disapproval of the Rate Bureau’s request to write one year rather than three year policies. Our holding only provides that he was not required to do so. The proper procedure was for the Rate Bureau to base its proposed farmowner filing on the existing law requiring farm-owner policies to be written on a three year basis and to have proposed an amendment to the filing if the Commissioner had approved the separate filing. See Comr. of Insurance v. Rating Bureau, 291 N.C. 55, 229 S.E. 2d 268 (1976). Or, the Rate Bureau could have simply withdrawn its farmowner filing and made a subsequent filing reflecting the Commissioner’s disapproval of a one year farmowner policy. Id.

On remand of this case to the Commissioner, he must consider the Rate Bureau’s brick veneer classification change and order an adequate rate adjustment based on the record evidence. The Commissioner may, but is not required to, consider the effect of three year farmowner policies.

V. Profit and Contingencies

The Rate Bureau, based on ten assignments of error, argues that the Commissioner erred in setting a rate level intended to produce a 13.5 percent return on net worth. Based on twelve assignments of error, the Rate Bureau argues that the Commissioner erred by approving a minus 3 percent underwriting profit.

G.S. § 58-124.19 establishes the standards by which insurance rates must be set. It provides:

(1) Rates shall not be excessive, inadequate or unfairly discriminatory.

(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.

(3) In the case of fire insurance rates, as are subject to the rate-making authority of the Bureau, consideration may be given to the experience of such fire insurance business dur ing the most recent five-year period for which such experience is available.

(5) In the case of workers’ compensation insurance and employers’ liability insurance written in connection therewith, due consideration shall be given to the past and prospective effects of changes in compensation benefits and in legal and medical fees that are provided for in General Statutes Chapter 97.

Id. By adopting subsection (1), the General Assembly intended that the term “inadequate” should serve to protect the interest of the insurance companies, filing as if one company through the Rate Bureau, by ensuring that rates are sufficient to earn a reasonable profit, and the term “excessive” protects the interests of consumers by prohibiting insurance companies from earning unreasonable profits. Comr. of Insurance v. Rate Bureau, 54 N.C. App. 601, 284 S.E. 2d 339 (1981), appeal dismissed, 305 N.C. 298, 290 S.E. 2d 708 (1982). In subsection (2), the General Assembly mandated that the Commissioner consider the specific rating criteria listed in reaching an ultimate rate level that complies with the subsection (1) standards. Hunt v. Reinsurance Facility, 302 N.C. 274, 275 S.E. 2d 399 (1981).

The Commissioner, in findings of fact numbered 46 through 60, disapproved the Rate Bureau’s request for an underwriting profit of 6 percent. The Commissioner ordered an underwriting profit of minus 3.3 percent. While the 11.7 percent rate increase approved by the Commissioner would have produced a negative underwriting profit, it would produce a 13.5 percent rate of return on net worth which the Commissioner found appropriate.

The Rate Bureau contends that G.S. § 58-124.19(2) required the Commissioner to approve rates which provided a positive underwriting profit and margin for contingencies as a matter of law. We disagree.

G.S. § 58-124.19(2) 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 established on the statu tory rate criteria is not inadequate, excessive, or unfairly discriminatory. We agree with the Commissioner’s interpretation of the statute, that if income from investments on loss reserves, loss expense reserves, and unearned premium reserves is sufficient to produce an overall profit that is not inadequate, excessive, or unfairly discriminatory the rate level thus determined does not violate the statutory criteria.

Our conclusion that the Commissioner could properly order a rate level that would produce an underwriting loss while providing for an overall adequate profit is supported by previous decisions of our appellate courts. In Comr. of Insurance v. Rating Bureau, 292 N.C. 471, 234 S.E. 2d 720 (1977), our supreme court considered the requirements of the rating scheme provided by G.S. § 58-132.1, subsequently repealed and replaced by G.S. § 58-124.19. The former statute is similar to that currently used; the current statute being more specific as to the various rating standards. Justice Lake precisely and cogently assessed the Commissioner’s duty in applying the statutory rate standards:

[I]t was obviously not the intent of the Legislature 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. . . .

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. . . .

Comr. of Insurance v. Rating Bureau, 292 N.C. 471. We conclude, therefore, that the Commissioner is required to consider each statutory rating factor specified by the General Assembly, and, when having done so, the Commissioner may balance the various factors to reach an adequate rate level.

The interplay between investment income and underwriting profit was also considered in Comr. of Insurance v. Automobile Rate Office, 292 N.C. 1. In that case the Commissioner reduced the requested allowance for an underwriting profit from 5 percent to 2.7 percent of earned premiums, based on an allowance for investment income. The supreme court specifically held that the Commissioner, acting on material and substantial evidence, could reduce the underwriting profit by allowing for investment income. Even though the court decided that case under slightly different rating standards than currently effective, the statutory schemes are analogous. We find no reasoning in that case which would prevent the Commissioner from ordering a negative underwriting profit if investment income alone would produce a reasonable profit.

Furthermore, the Rate Bureau’s farmowner filing was based on an underwriting loss of minus 3.3 percent. The Rate Bureau’s filing indicated that a rate increase of 35.7 percent was needed to produce an adequate rate. The request, however, was lowered to 25 percent. Paul Mize, General Manager of the Rate Bureau, testified that the “rate change was capped at 25% at the direction of the Property Committee and the Governing Committee of the Rate Bureau in order to ameliorate swings in the rate level.” The undisputed evidence in the record confirms that by “capping” the rate request at 25 percent the Rate Bureau’s margin for underwriting profit was minus 3.3 percent.

The Commissioner was not required to approve an underwriting profit greater than that requested by the Rate Bureau. The logical conclusion of the Rate Bureau’s argument is that the Commissioner was required to approve an underwriting profit even though the Rate Bureau requested an underwriting loss. We considered a similar situation in Comr. of Insurance v. Rate Bureau, 54 N.C. App. 601, where the Commissioner disapproved a Rate Bureau filing which requested an underwriting loss because it produced inadequate rates. We held that the Commissioner erred in disapproving the rate request on that basis:

[W]e are satisfied that the limiting effect of the term ‘inadequate’ as it is used in the statute is that the Commissioner may not disapprove of such portions or parts of a filing as will result in rates which are inadequate to produce a fair and reasonable profit to the companies represented in the filing. . . .

An additional reason for rejecting the reasoning of the plaintiff in this case is the bizarre result reached by the order. Upon concluding that the filing must be rejected because the requested rates are ‘inadequate,’ plaintiff would leave in effect rates which are even more inadequate. We cannot believe that the legislature even contemplated such a result.

Id. (Citations omitted.) For all the reasons set forth above, we find that the Commissioner did not err in not approving an underwriting profit greater than that proposed by the Rate Bureau.

The Rate Bureau also argues that the Commissioner erred in setting a 13.5 percent return on net worth. Their position is that the testimony of their expert witness, Dr. Irving Plotkin, that a 19 percent rate of return was justified was the only material and substantial evidence in the record as to the needed return on net worth.

The Department of Insurance tendered two expert witnesses, Dr. John Wilson and J. Robert Hunter, who testified on the rate of return needed. The Commissioner’s order primarily relies on Dr. Wilson’s testimony. Dr. Wilson offered an opinion that a 13.5 percent rate of return was justified relying on a complex stock analysis model comparing common stock prices to book value. On the whole record, we find that Dr. Wilson and Hunter’s testimony was material and substantial.

The Rate Bureau advances numerous arguments designed to discredit the theoretical basis, methodology, and conclusions of Dr. Wilson’s analytical model. For example, the Rate Bureau notes that Dr. Wilson determined that a return on investment income of 8 percent was justified. Dr. Wilson’s actual calculations resulted in an investment income allowance of 7.7 percent which he rounded to 8 percent, and Dr. Wilson’s calculation also failed to deduct investment expenses in deriving investment income. Reducing investment income by investment expenses would result in a net investment income allowance of 7.3 percent rather than 8 percent. Hunter’s testimony, which the Commissioner relied upon to support Dr. Wilson’s findings, clearly indicates that his methodology to determine investment income relied, in part, on investment earnings from surplus which is specifically prohibited. Comr. of Insurance v. Rate Bureau, 300 N.C. 381.

The Commissioner’s findings of fact do not reflect his consideration of the contradictory evidence and the basis on which he determined the adequacy of a 13.5 percent rate of return. It ap pears that he adopted Dr. Wilson’s recommendation, but his findings of fact do not indicate whether he considered the undisputed testimony that Dr. Wilson had rounded certain results and failed to include investment expenses in his analysis. Furthermore, the Commissioner may not consider investment income from capital and surplus accounts, and his order does not reflect the Rate Bureau’s evidence that Hunter included income from these sources in his analysis. After a thorough review of all the evidence and the Commissioner’s findings of fact, we are unable to delineate the manner in which the Commissioner reached his determination of a 13.5 percent rate of return. His findings of fact on this issue are inadequate to afford an effective appellate review. We must, therefore, remand this issue to the Commissioner for meaningful findings of fact.

On remand of this matter, the Commissioner must make findings of fact which establish the basis of the rate of return on net worth adopted. The testimony of Hunter and Dr. Wilson is substantial and material evidence for the Commissioner’s consideration. The Commissioner must consider, however, the uncon-tradicted evidence that Dr. Wilson rounded his mathematical results and omitted investment expenses from his calculations. The Commissioner must consider the evidence tending to show that Hunter based part of his calculations on investment income from capital and surplus because investment income from these sources may not be considered in insurance rate making. In reaching his ultimate determination, the Commissioner must make findings of fact which clearly indicate the facts on which he bases his order, the resolution of conflicting evidence, and the consideration given to the material and substantial evidence that has been offered.

VI. Notice of Hearing

The Rate Bureau next contends that the Commissioner erred in admitting evidence and making findings of fact and conclusions of law with respect to matters of which the Rate Bureau was not given proper notice in the notice of public hearing. This argument centers on the Rate Bureau’s use of fire and extended coverage data, their failure to segregate wind losses in determining an excess multiplier, their failure to provide for a simultaneous effective date for rate decreases in non-Rate Bureau jurisdiction coverages, and the Commissioner’s failure to specify his contentions regarding the proper method of determining an appropriate rate level.

We have previously discussed the Rate Bureau’s arguments in relation to the excess multiplier and the simultaneous filing date for non-Rate Bureau coverages included in the farmowner policy package. No further discussion is needed here. As to the Rate Bureau’s contention that the Commissioner failed to provide the exact rating formula that he planned to employ, we find no error.

The Rate Bureau concedes that the Commissioner’s notice of hearing was sufficient to place them on notice that he found their profit determination deficient, but they argue that he failed to provide in his notice the manner in which profitability would be determined. G.S. § 58-124.21 only requires the Commissioner notify the Rate Bureau of the deficiencies of which the Commissioner knows after having reviewed the Rate Bureau’s filing within the 30 day period of review. See Comr. of Insurance v. Rate Bureau, 300 N.C. 381. There is no evidence in the record before us to indicate that the Commissioner knew the precise rating methodology that he would propose at the hearing before the notice of hearing was required. The Commissioner’s notice of hearing specifically provided that investment income had not been considered and that the Rate Bureau had failed to justify the 6 percent profit and contingency margin requested. We conclude that the Commissioner’s notice of hearing gave adequate notice of the alleged deficiencies subsequently raised at the hearing.

VII. Expert Witness

The Rate Bureau next contends that the Commissioner erred in accepting Hunter as an expert Property and Casualty Actuary for farmowner insurance. The Rate Bureau argues that Hunter has never testified in a farmowner rate hearing and that premium trend and excess multiplier calculations require actuarial experience beyond a general actuarial background and experience.

N.C. Gen. Stat. § 8-58.13 (1981) provides that:

If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education, may testify thereto in the form of an opinion.

The decision to qualify a witness as an expert is ordinarily within the exclusive province of the trial judge or hearing officer. E.g., Utilities Comm. v. Telephone Co., 281 N.C. 318, 189 S.E. 2d 705 (1972).

The record discloses that Hunter is a fellow of the Casualty Actuarial Society, and a member of the American Academy of Actuaries. Among numerous other qualifications, Hunter has served as chief actuary of the Federal Insurance Administration, published several scholarly works on investment income and insurance rate making, and testified in numerous rate hearings, including a homeowner policy hearing in North Carolina. He had reviewed one farmowner rate filing, but had never testified in a farmowner rate hearing.

The undisputed evidence in the record demonstrates that actuaries are not designated by lines of insurance: actuarial certification is normally as a property and casualty actuary. Hunter’s educational qualifications and experience as a property and casualty actuary qualified him as an expert witness. Furthermore, while the Rate Bureau contends that a farmowner rate filing requires specialized expertise, no evidence of this requirement was offered at the hearing and nothing in the Rate Bureau’s argument to this court supports their position. We conclude that the Commissioner did not err in qualifying Hunter as an expert witness.

Disposition

In reviewing an order of the Commissioner, N.C. Gen. Stat. § 58-9.6(b) empowers an appellate court to:

[AJffirm or reverse the decision of the Commissioner, declare the same null and void, or remand the case for further proceedings; or it may reverse or modify the decision if the substantial rights of the appellants have been prejudiced . . .

That part of the Commissioner’s order which requires the Rate Bureau to refile for a rate increase with non-Rate Bureau cover ages must be vacated because the Commissioner acted in excess of statutory authority.

The Commissioner, however, did approve part of the Rate Bureau’s requested rate increase subject to the improper requirement that rates for non-Rate Bureau must be filed simultaneously. There was material and substantial evidence that the entire rate increase requested by the Rate Bureau was not justified even though the Commissioner based his ultimate determination as to the rate justified on evidence that was not material or substantial. The record discloses substantial and material evidence from which the Commissioner could have based his decision, therefore, we vacate the Commissioner’s order and remand for further proceedings. We summarize the deficiencies in the order to guide the Commissioner’s consideration on remand.

As to the premium trend portion of the farm projection factor, there was sufficient evidence to support the Commissioner’s finding that the trend determined by the Rate Bureau was not fully justified. Stern’s testimony would permit the Commissioner to find that the farmowner premium trend should more closely parallel the homeowner premium trend. If the Commissioner rejects the Rate Bureau’s premium trend component of the modified farm projection factor, the Rate Bureau’s undisputed evidence of loss trend, found actuarily sound by the Commissioner, requires that he give due consideration to that factor.

In determining an excess multiplier, the Commissioner may properly consider combining ISO, NAII, and AAIS data. He may also consider the use of a 100 percent demarcation to determine excess losses. The Commissioner, having failed to give adequate notice to the Rate Bureau of any deficiency in the use of fire and extended coverage policy data, must consider that evidence. We reiterate, however, the Commissioner is not required to reach the same conclusions from this evidence as reached by the Rate Bureau.

The Commissioner must consider the Rate Bureau’s request for the masonry veneer classification change, and we note that the record evidence supporting this classification change is uncon-tradicted. The Commissioner may, but is not required to, consider the Rate Bureau’s evidence of the effect of writing three year policies rather than one year farmowner policies.

The Commissioner must consider each rating factor arising on the evidence which is enumerated in G.S. § 58-124.19. He may balance these factors so long as he reaches a rate that is neither inadequate, excessive, nor unfairly discriminatory. If the overall rate level is adequate, he may approve a negative underwriting profit.

Dr. Wilson’s testimony as to investment income from premium, loss, and loss expense reserves is substantial and material evidence which the Commissioner could properly consider. The Commissioner must, however, consider the undisputed evidence in the record that Dr. Wilson rounded certain data results and failed to consider investment expenses in his calculations. The Commissioner may not consider evidence of investment income from capital or surplus.

On remand, the Commissioner should be specific in his findings of fact. The order before us merely recites witness testimony in substantial parts, it is vague as to the manner in which the Commissioner reached mathematical results, and it does not reflect the weight that each rating factor was given in reaching his ultimate determination.

Finally, the Commissioner may issue a new order in this case without further hearings as we conclude that the record is sufficient to permit a proper order. We note, however, that the Commissioner who issued the order before us is not the present Commissioner. In fairness to the present Commissioner, who did not hear the evidence in this case, we grant the authority to conduct further hearings in this matter, if he deems it appropriate.

In the face of the Commissioner’s disapproval of their rate increase, the Rate Bureau elected to continue to use the rate requested during the pendency of this appeal and placing the amount of the rates disapproved in escrow in accordance with G.S. § 58-124.22(b). The Rate Bureau is hereby ordered to continue such procedure until a final determination of this matter is reached.

After carefully reviewing the entire record before us, and based on the holdings we have reached in each of the arguments presented, the order of the Commissioner dated 26 January 1984 is hereby

Reversed in part, vacated in part, and remanded.

Judges WHICHARD and Becton concur.