A single justice consolidated, and then reserved and reported, complaints by the Automobile Insurers Bureau of Massachusetts (AIB), and by the Alliance of American Insurers and three insurance companies (collectively, Alliance), for judicial review of a decision by the Commissioner of Insurance (commissioner) which established private passenger automobile insurance rates for 1999. The AIB and the Alliance argue that the commissioner exceeded her authority in seeking compliance with a discovery request and then improperly sanctioned them for alleged noncompliance. The AIB alone challenges three aspects of the commissioner’s decision concerning the calculation of the loss pure premium factor. These matters involve the commissioner’s handling of information concerning seat belt usage and the payment of sales tax on paint and materials, and the commissioner’s decision to use a so-called “dummy” variable in calculating the loss pure premium trend factors for rate year 1999. The issues were litigated as part of the main rate case. We conclude that the commissioner’s decision should be upheld. We take up first the rate sanction issue and thereafter the issues pertaining to the loss pure premium factor.
1. One component of the main rate case concerned calculation of a rate provision for expenses incurred by insurance companies. The expense provision includes general expenses, miscellaneous taxes, licenses and fees, and “other” acquisition expenses, including employees’ compensation. Since at least 1986, the method used to calculate this expense provision has included the calculation of an industry-wide company expense trend designed to project the change, into the rate period, in the companies’ expenses.
In its advisory filing for 1999, the AIB proposed a five per cent annual trend, which included, among other items, a 6.5 per cent increase in employee salaries and benefits. During the evidentiary hearings in the rate case, the Attorney General formally sought, in a discovery request, that the AIB provide certain specified information relating to employee compensation *287for each insurance company whose share of the private passenger automobile market was greater than .5 per cent. The AIB responded that it did not have the requested data. The Attorney General filed a motion to compel discovery with the presiding officers at the evidentiary hearing, seeking an order directing the AIB to respond to the discovery request. After hearing argument, the presiding officers issued an order directing the AIB, and affected individual insurance companies represented by the AIB, to produce the requested data in such a manner that individual companies and their individual employees were not identified.
After various appeals and motions with which we need not be concerned, the AIB began submitting responses to the discovery request. Dissatisfied with the responses, the Attorney General filed a motion for sanctions, arguing that the ABB’s and the companies’ compliance was inadequate and nonresponsive. The AIB opposed the motion, and the presiding officers heard argument and took the matter under advisement. In her rate decision, based on conclusions reached by the presiding officers, the commissioner determined that the AIB and the insurance companies had not fully complied with the discovery request, and that a negative inference should be drawn in calculating the company expense trend because of the noncompliance.
The AIB and the Alliance have challenged the commissioner’s determination, contending that she (a) lacked authority to issue the discovery order and specifically lacked authority to issue it against the individual insurance companies; (b) failed to make findings regarding whether the AIB and the companies had substantially complied with the discovery order; and (c) lacked authority to impose an industry-wide sanction.
(a) The discovery order. The commissioner is mandated by statute to fix and establish industry-wide private passenger automobile insurance rates through rate setting procedures, and hearings. G. L. c. 175, § 113B. G. L. c. 175E, § 5. See Massachusetts Auto. Rating & Acc. Prevention Bur. v. Commissioner of Ins., 381 Mass. 592, 603 (1980). In furtherance of this obligation, the commissioner has the authority to designate a “Presiding Officer” or officers to conduct evidentiary and other proceedings to obtain data and information pertinent to the rate setting process. See 211 Code Mass. Regs. § 77.02 (1997).
As might be expected, discovery is allowed in the rate setting *288proceedings to assist the parties in their need to obtain information from other parties. See 211 Code Mass. Regs. § 77.06 (1997) (“A party may request another party to produce or make available documents or tangible things, not privileged and not previously supplied, which are in the possession, custody or control of the party to whom the request is made”). The term “[p]arty” is defined as “[a]n insurer submitting an advisory filing, statutory intervenors, and interveners.” 211 Code Mass. Regs. § 77.02. The term “insurer” is defined as “[a]n insurance company authorized to write motor vehicle insurance in Massachusetts or a licensed rating organization authorized to act on its behalf.” Id. If a discovery request is not honored, the aggrieved party may file a motion to compel discovery with the presiding officer. See 211 Code Mass. Regs. §§ 77.06(3), 77.07(5) (1997).
The statute, and the implementing regulations, set forth above, expressly permit the Attorney General to obtain the information sought in his discovery request in the evidentiary hearings in the rate case. The Attorney General is a statutory intervener pursuant to 211 Code Mass. Regs. § 77.02, and G. L. c. 12, § 11F. The AIB is a licensed rating organization pursuant to G. L. c. 175A, §§ 8-11. Both submitted advisory filings in the rate case. Under the regulations, both are “parties” to these proceedings and therefore have the right to make discovery requests of one another.2
The AIB and the Alliance argue that the presiding officers’ discovery order could not extend to individual insurance companies because they were not “parties” to the proceeding. In her 1999 decision, the commissioner interpreted the govem-*289ing regulations defining the terms “party” and “insurer,”3 and she concluded that the insurance companies were parties to the proceedings, notwithstanding the fact that they had designated the AIB to act on their collective behalf. The commissioner also concluded that the insurance companies, through AIB’s representation, had an adequate opportunity to be heard on the discovery order.4
Common sense dictates that proper compliance with the discovery order required full cooperation by the companies. It was the companies’ requested rate increase that was at issue. The commissioner had the authority to interpret the regulations to impose the discovery order on the companies because they were parties and there was good cause to do so. We accept the commissioner’s determination. See Hurst v. State Ballot Law Comm’n, 428 Mass. 116, 120 (1998).
(b) Compliance. The AIB and the Alliance argue that the commissioner was required to make findings regarding whether the AIB and the insurance companies substantially complied with the discovery order. In “making her decisions, the commissioner is not required to make findings on every issue so long as her findings ‘indicate the over-all basis for [her] decision.’ ” Automobile Insurers Bur. of Mass. v. Commissioner of Ins., 415 Mass. 455, 457 (1993), quoting Aetna Cas. & Sur. Co. v. Commissioner of Ins., 408 Mass. 363, 374, 378 (1990). The commissioner concluded that the AIB and the insurance companies had not complied with the discovery order, and she listed specific findings of noncompliance.5 The commissioner’s decision as to noncompliance is adequately supported.
*290The AIB also argues that the commissioner failed to recognize any justification for the imperfect compliance and to account for whether the Attorney General was prejudiced by the lack of compliance. The commissioner was not persuaded that she needed to make determinations on these issues. We are not persuaded either. Lack of full compliance was established. The over-all basis for the commissioner’s decision to apply a negative inference is clear from her decision and supports a conclusion that the violation of the discovery order impeded the informed resolution of a major factor in the rate calculation process.
(c) The sanction. The AIB requested a five per cent expense trend (which included, among other items, a 6.5 per cent increase in employee salaries and benefits). The Attorney General challenged the AIB’s request. The commissioner set the expense trend at 2.7 per cent based on evidence and information in the main rate case, and she arrived at her figure without considering the failure of substantial compliance on the part of the insurance companies with the discovery order. The AIB and the Alliance do not .dispute the commissioner’s selection of a 2.7 per cent expense trend as a determination within her authority. They do challenge the commissioner’s additional decision that the trend should be reduced by twenty-five per cent as a sanction for violation of the discovery order. The basis for the sanction was a negative inference drawn by the commissioner as a result of the companies’ failure to produce pertinent data in response to the discovery order which might have permitted a more extensive and complete analysis of their expenses.
We reject the arguments made by the AIB and the Alliance that the commissioner lacked authority to impose the sanction. They argue that, if the commissioner was to seek enforcement and the imposition of a sanction, she could do so only under G. L. c. 175, § 8A, and G. L. c. 175A, §§ 16 and 18. Section 8A, in conjunction with G. L. c. 233, § 10, authorizes the commissioner to summons witnesses and gather information and to enforce any summons by means of an action to compel compliance in the Superior Court or the Supreme Judicial Court for Suffolk County. See generally Commissioner of Ins. v. First Nat’l Bank, 352 Mass. 74, 81-82 (1967); Massachusetts Bonding & Ins. Co. v. Commissioner of Ins., 329 Mass. 265, 279 (1952). Sections 16 and 18 prohibit the wilful withholding of information and provide for a penalty of $500 for each wilful violation.
*291The commissioner, of course, could have proceeded under these statutes. She was not restricted, however, to the procedures and remedies established by them. The specific authority to act in a particular way does not, in the absence of explicit direction, bar other action that is otherwise consistent with statutory authority. Massachusetts Elec. Co. v. Department of Pub. Utils., 419 Mass. 239, 246 (1994). We doubt that the Legislature intended that every discovery dispute in a rate setting case should reach the Superior Court or the county court with the resulting delay that collateral litigation might cause to the efficient determination of a rate. There is precedent for what the commissioner did in a related area. In appropriate circumstances, a fact finder in a civil dispute may draw a negative inference from the failure of the party with the burden of proof to call a witness or produce information within the party’s control which would shed light on the party’s position on a material issue. See Shafnacker v. Raymond James & Assocs., 425 Mass. 724, 735-736 (1997); Labor Relations Comm’n v. Fall River Educators’ Ass’n, 382 Mass. 465, 471-472 (1981). We see no reason why the commissioner, who is charged with setting a fair, industry-wide rate, could not draw an adverse inference. The AIB and the Alliance had the burden of producing evidence in the main rate case to support the rate they sought. See Automobile Insurers Bur. of Mass. v. Commissioner of Ins., supra at 463 n.11; Aetna Cas. & Sur. Co. v. Commissioner of Ins., supra at 373. The inference was properly based on their failure to produce complete information on a subject peculiarly within their knowledge and responsibility.
We also accept the commissioner’s imposition of the twenty-five per cent sanction as the product of the negative inference. By the AIB’s own calculation, of the general and other acquisition expenses that the proposed expense trend sought to capture, over fifty per cent was directly attributable to “salaries” and “employee relations,” and over seventy-three per cent was derived from data concerning “average weekly earnings.” Because of the magnitude of the proposed trend, and the large proportion of the increase that would arise from higher salaries and benefits, the Attorney General properly sought to determine the reasonableness of the AIB’s proposal through the discovery request to ascertain the historical trend in executive salaries to determine whether that trend was a proxy for company salaries generally. Because of the AIB’s and the companies’ non*292compliance, the Attorney General was unable to make this determination, and the AJB’s data stood virtually unchallenged on the record before the commissioner. The commissioner carefully weighed the issue and, on all the information before her, acted reasonably in setting the sanction at twenty-five per cent.
Finally, we reject the argument by the AIB and the Alliance that the commissioner improperly applied the sanction to all insurance companies, including companies that either complied with the discovery order or were not required to comply because their market share was less than .5 per cent. The commissioner is mandated by statute with fixing and establishing an industry-wide rate. She was required, therefore, to determine an expense trend that applied to the entire private passenger automobile insurance industry. The commissioner, consistent with her statutory obligation, could not effectively set differing rates proportionate to the degree of compliance with the discovery order or the fact that smaller insurers were exempted from compliance. The commissioner acted properly because the industry as a whole had failed to comply fully with the discovery order concerning a critical factor in the rate setting process.
2. We turn now to the issues concerning the loss pure premium factor.
(a) Seat belt usage. One aspect of the commissioner’s decision with respect to the calculation of the loss pure premium concerned seat belt usage. Chapter 387 of St. 1993, entitled “An Act relative to the use of seat belts in motor vehicles” (Act), expressly ties the issue of seat belt usage to the commissioner’s fixing and establishing of automobile insurance rates. The Act provides, in relevant part, that the commissioner be required to reduce bodily injury premiums by a minimum five per cent after the law had been in effect for one year, if the observed seat belt use rate in Massachusetts among all occupants equaled or exceeded fifty per cent. St. 1993, c. 387, § 7. The Act further provides that “[ajnnual safety belt survey results shall be a criterion in all future regulatory actions regarding bodily injury premiums,” and that if the seat belt usage rate in Massachusetts “exceeds the national average, additional reductions in bodily injury premiums shall take effect.” Id. The annual safety belt survey relied on by the parties and the commissioner was conducted by the Boston University School of Public Health for the Governor’s highway safety bureau (B.U. Survey).
*293In the main rate case, no party requested an adjustment to the loss pure premiums to reflect the impact of the Act. Based on the B.U. Survey, which was admitted in evidence,6 the seat belt usage rate in the Commonwealth did not exceed the national average. In her decision, the commissioner determined that the conditions that would require an adjustment based on seat belt usage were not present and made no adjustment.
This decision is consistent with the commissioner’s decisions in the prior four years. In 1995, one year after the law took effect, seat belt usage was at fifty-two per cent, and the commissioner made the required downward adjustment in her 1995 and 1996 rate decisions. In both the 1997 and 1998 rate cases, no party requested an adjustment to the rates to reflect the impact of the Act and, after considering the B.U. Survey, the commissioner determined that the conditions that would require an adjustment based on seat belt usage were not present, and she made no adjustment.
The AIB argues that the commissioner’s decision was not based on substantial evidence, and that the commissioner failed to make findings on any decrease in seat belt usage. We reject these arguments. It was reasonable for the commissioner to conclude that, based on the B.U. Survey, which only showed a slight decrease in seat belt usage, and the recommendations of the parties, no adjustment should be made. All the commissioner was required to do under the Act was consider the survey results as a “criterion.” St. 1993, c. 387, § 7. She did this.7 The Act does not constrain her judgment in how to use the results, and, on the information presented, it did not obligate her to make an adjustment.
(b) Sales tax on paint and other materials. A second aspect of the commissioner’s decision with respect to the calculation of the loss pure premium concerned the payment of sales tax on *294paint and other materials. On July 7, 1998, the automobile damage appraiser licensing board (board)8 issued an advisory ruling concerning the application of the sales tax to paint and other materials that are “transferred to the vehicle” in the course of automobile body repairs. The ruling stated that the entire amount of the cost of such items are taxable to the vehicle owner and, therefore, to the insurance company paying for repairs. The AIB sought an upward adjustment to the loss pure premium, on the ground that these sales taxes represent new and increased “losses” for which insurers would now be liable. The commissioner did not adjust the rates as a result of the board’s ruling. The AIB claims that the commissioner’s rejection of the AIB’s adjustment was wrong because the commissioner failed to make adequate findings to support this decision and, further, relied on erroneous and irrelevant statements made by the State Rating Bureau (SRB). We reject these contentions.
As the proponents of the adjustment to the loss pure premium, the AIB bore the burden of demonstrating its appropriateness. See Automobile Insurers Bur. of Mass. v. Commissioner of Ins., supra at 463 n.11; Aetna Cas. & Sur. Co. v. Commissioner of Ins., supra at 373; Massachusetts Med. Serv. v. Commissioner of Ins., 346 Mass. 346, 348 (1963). Although there was no dispute that, in the future, all sales taxes for paint and materials will be passed on to the insurance companies through the vehicle owners, the AIB still bore the burden of demonstrating that the board ruling would change the way this tax had been allocated in past years.9 It was this evidentiary burden that the AIB failed to carry. Based on the record before her, including the AIB’s filing and its one witness, the commissioner concluded that there should be no sales tax adjustment to loss pure premium. The basis for the commissioner’s decision is clear on the record. The AIB, which bore the burden of persuasion on the adjustment, failed to meet its burden.
(c) The “dummy ” variable. A third aspect of the commissioner’s decision with respect to the calculation of the loss pure premium concerned the substitution of a “dummy” variable for the actual 1996 loss experience data. A “dummy” variable is, in *295essence, a methodological adjustment employed to eliminate or temper the effect of unusual or aberrant data. When a portion of relevant data is aberrant and not expected to recur, a “dummy” variable is substituted for the unusual data.
In this case, the commissioner was calculating the loss pure premium trend in order to bring data from the most recent experience period (1997) to the level of the current rate year (1999). The purpose of calculating the trend is to project future losses from past loss experiences. This type of projection is necessary because the level of loss pure premiums is expected to change between the experience period and the period for which rates are being made.
To calculate this trend, the commissioner adopted the AIB’s proposed methodology of using a linear regression of the six most recent years of pure premium data. This methodology required using loss experience data for 1992 through 1997. Normally, the purpose of using linear regressions is to even out variations in data. The commissioner determined, however, that the 1996 loss data was unusual and, if it was included, it would result in a projection for 1999 that was excessive. For this reason, the commissioner chose to substitute a “dummy” variable for the 1996 loss data.
The AIB argues that the commissioner’s decision not to adopt the AIB’s recommendation and to use a “dummy” variable in place of the 1996 data point was erroneous. Further, the AIB contends that the commissioner’s decision violates the amendment to G. L. c. 175, § 113B, second par., inserted by St. 1988, c. 273, § 35.10 We reject these arguments.
As stated above, when reviewing the commissioner’s decision regarding methodology, we will consider whether her choices had reasonable support in the evidence, i.e., whether there is substantial evidence supporting the decision. See Automobile Insurers Bur. of Mass. v. Commissioner of Ins., supra at *296457; Aetna Cas. & Sur. Co. v. Commissioner of Ins., supra at 378. We shall not consider the evidence de nova and will “give due weight to the commissioner’s experience, technical competence, specialized knowledge, and discretionary authority.” Id. Deference to the commissioner’s expertise and discretion is particularly appropriate when reviewing her choice of methodology. Id. at 378-379. We have upheld methodological adjustments, similar to a “dummy” variable, in various rate-making contexts. See id. at 377-379 (commissioner’s decision to exclude data from time period characterized by unusually large increase in pure premium supported by reasonable evidence in record). See also Boston Gas Co. v. Department of Pub. Utils., 368 Mass. 780, 796-797 (1975) (department’s conclusion that weather adjustment from test year to a “normal” year appropriate in rate making context). In concluding that the use of a “dummy” variable was appropriate in the calculation of the loss pure premium, the commissioner relied on filings by the AIB, the SRB, and the Attorney General. In its various filings, the AIB had recommended that the “dummy” variable not be used in place of the 1996 loss data. The AIB contended that the use of such a variable was the mathematical equivalent of excluding the value and presented evidence in support of its position. The SRB and the Attorney General did not agree with the AIB’s recommendation and advised the commissioner that the 1996 data was unusual. The SRB recommended the use of a “dummy” variable in place of this data.11 The Attorney General recommended excluding the 1996 data altogether. Both presented evidence in support of their positions.
The commissioner possessed discretion to use a “dummy” variable for the actual 1996 loss data, and her decision to do so was reasonably supported.12 In reaching her decision, the commissioner discussed and considered the advisory filings of the AIB, the SRB, and the Attorney General. The commissioner devoted approximately eleven pages to discussion and analysis of the loss pure premium trend factors generally, and the 1996 *297data specifically. This portion of the decision included references to the evidence that the commissioner relied on, as well as evidence that she found unpersüasive. Further, the commissioner stated reasons why she found certain evidence persuasive or unpersuasive. The decision discloses an understanding of the factual record and the pertinent actuarial analyses and is sufficient on the point.
3. The case is remanded to the county court where a judgment is to be entered upholding the commissioner’s decision.
So ordered.
The Automobile Insurers Bureau of Massachusetts (AIB) argues that the relevance and necessity of the information received in the discovery request were never established. We disagree. The regulations on discovery are silent on the issue of relevance and necessity. As previously stated, the regulations allow a party to request documents or tangible things from another party. See 211 Code Mass. Regs. § 77.06(1) (1997). The regulations also provide that other forms of discovery “may be allowed at the discretion of the Presiding Officer upon a showing of good cause.” Id. In her 1999 decision, the commissioner specifically concluded that the Attorney General “demonstrated good cause” and that “[i]t is reasonable to assume that the production of the requested salary information might have led to relevant evidence.” The commissioner based this decision on the insurers’ request of an over-all 15.5 per cent rate increase, including expenses totaling over 300 million dollars, of which fifty per cent constituted salaries and employee relations. The commissioner’s conclusion was correct. See Automobile Insurers Bur. of Mass. v. Commissioner of Ins., 415 Mass. 455, 457 (1993).
Title 211 Code Mass. Regs. § 77.02 (1997).
The Alliance of American Insurers (Alliance) argues that the discovery order and the sanction violated the insurance companies’ due process rights because they did not have the opportunity to be heard. The insurance companies had the opportunity to be heard and were in fact heard in this proceeding through the AIB, a licensed rating organization that the insurance companies specifically designated to represent them in the course of the rate-setting proceedings.
Specifically the commissioner stated: “The AIB does not dispute that: (1) fourteen companies did not provide total compensation; (2) only three companies provided total compensation for all five years; (3) fourteen companies did not provide allocation factors from Massachusetts private passenger automobile insurance; (4) ten companies did not allocate salaries to lines 15, 18, and 22 of the expense call; (5) five companies failed to provide five years of data; and (6) three companies failed to provide unaggregated compensation of ten employees.”
The AIB made its advisory filing before the B.U. Survey was admitted in evidence. In its advisory filing, the AIB recommended that no adjustment be made to loss pure premiums based on seat belt usage. The B.U. Survey was later admitted in evidence and showed a slight decline in usage from fifty-three per cent to fifty-one per cent.
The AIB also contends that the commissioner rejected the seat belt model; violated the Act that requires usage to be a criterion; and deviated from prior rulings without any explanation. These contentions are without merit. The commissioner admitted the B.U. Survey in evidence and specifically referred to it in her decision. The commissioner’s ruling is consistent with her rulings in prior years.
The board is the agency within the division of insurance that licenses and regulates automobile damage appraisers. G. L. c. 26, § 8G.
The AIB’s critical witness on this point, its associate actuary, testified that he did not know whether the board’s 1998 ruling represented a change in industry practices.
Specifically, the AIB contends that, contrary to her statutory mandate, the commissioner’s decision does not make certain express findings; does not justify certain differences between the trend and projection factors produced by the commissioner’s decision and those implicit in the relevant data; and fails to resolve certain contested issues. These contentions are without merit. We note, for purposes of the discussion that follows, that we interpret c. 273, § 35, in the context of the standard of review for § 113B. See Aetna Cas. & Sur. Co. v. Commissioner of Ins., 408 Mass. 363, 378-379 (1990) (apart from § 35, decisions of commissioner regarding methodology are subject to judicial review under G. L. c. 175, § 113B).
The SRB advised the use of a “dummy” variable for property coverages only. The commissioner decided to use such a variable for all coverages.
The AIB argued that the commissioner erred by failing to determine the exact effect snowfall had on the 1996 data. The commissioner was not required to reach this issue in deciding whether the 1996 data was unusual. See Aetna Cas. & Sur. Co. v. Commissioner of Ins., supra at 378 (commissioner need not make findings on every issue so long as over-all basis for decision allows for judicial review).