Bellermann v. Fitchburg Gas & Electric Light Co.

This case arises out of a major ice storm that struck areas of the northeastern United States in December, 2008 (Winter Storm 2008). The defendant, Fitchburg Gas and Electric Light Company (FG&E), is a public utility that provides electric service to customers in the municipalities of Fitchburg, Lunenburg, Townsend, and Ashby, which were among those affected by the storm. FG&E is one of the utilities owned by Unitil Corporation (Unitil). The plaintiffs are twelve residential and business customers of FG&E who lost power during Winter Storm 2008. They filed a suit in the Superior Court on behalf of themselves and those similarly situated, asserting claims of gross negligence and violation of G. L. c. 93A. Pursuant to G. L. c. 93A, §§ 9 (2) and 11, 2 and Mass. R. Civ. P. 23, as amended, 452 Mass. 1401 (2008) *45 (rule 23), 3 the plaintiffs moved to certify a class consisting of FG&E’s residential and business customers; their dependents, tenants, and employees; and other users of electricity who sustained damages as a result of FG&E’s inadequate preparation for and response to Winter Storm 2008. The parties also filed cross motions for partial summary judgment on the plaintiffs’ G. L. c. 93A claims. In their motion for partial summary judgment, the plaintiffs sought issue preclusive effect of certain findings made by the Department of Public Utilities (DPU) in two prior administrative adjudications related to FG&E’s conduct during Winter Storm 2008. See D.P.U. 11-01 (2011); D.P.U. 09-01-A (2009).

After a combined hearing on these motions, the judge issued two decisions. He denied the plaintiffs’ motion for class certification, and, while determining that the application of offensive issue preclusion was appropriate, he also denied the motions for summary judgment as to all but two claims that are not at issue here. The judge then reported his decision denying class certification to the Appeals Court pursuant to Mass. R. Civ. P. 64 (a), as amended, 423 Mass. 1403 (1996), and FG&E sought interlocutory review, pursuant to G. L. c. 231, § 118, of the judge’s decision as to issue preclusion. A single justice of the Appeals Court allowed the petition for interlocutory review, and the two appeals were consolidated in the Appeals Court. We allowed the plaintiffs’ application for direct appellate review. We conclude that the judge did not abuse his discretion in declining to certify *46 a class and in applying issue preclusion to facts found after evidentiary hearings at the DPU. 4

Background. We summarize the facts set forth in the judge’s decisions, supplemented by other undisputed facts in the record.

FG&E receives all of its electric power from four transmission supply lines owned by National Grid USA Service Company, Inc. (National Grid). Each of these lines ties into a substation in southwest Fitchburg. From this substation, FG&E’s network consists of sixty miles of subtransmission lines that feed into 680 miles of distribution lines. At the time of Winter Storm 2008, FG&E had approximately 28,500 customers.

1. Winter Storm 2008. Winter Storm 2008 was an ice storm that struck the northeastern United States, including FG&E’s service territory, on December 11 and 12, 2008. As a result of that storm, ice accumulated on utility poles and tree limbs in FG&E’s territory, causing limbs and whole trees to fall onto FG&E utility poles, electrical lines, and other electrical infrastructure. The storm also damaged the National Grid transmission lines that supply FG&E’s system with power. In total, Winter Storm 2008 resulted in power outages for over one million customers in New England, New York, and Pennsylvania, including one hundred per cent of FG&E’s customers.

2. FG&E’s preparedness and restoration efforts. In 1992, sixteen years before Winter Storm 2008, the DPU had directed FG&E and several other utilities to “implement or to maintain a system,,for reviewing emergency plans on an annual basis,” and had advised them that “it would be beneficial for each company to examine the other [cjompanies’ plans in order to consider the incorporation of useful changes.” D.P.U. 91-228, at 4 (1992). In 2002,. the DPU further advised FG&E to “consider the use of extreme weather condition forecasts with outages or contingencies simulated in the power flow model for plan[n]ing and designing [transmission and distribution] facilities.”

In 2006 and 2007, FG&E sent newsletters to its customers extolling its ability to respond to outages and asserting that “safety and service reliability are our first priorities.” FG&E also filed annual emergency restoration plans (ERPs) with the DPU; these plans set forth an overview of FG&E’s service restoration *47 processes and priorities, defined organizational and functional responsibilities, identified communications protocols, and described the framework required to restore power in the event of a major storm or other emergency. Other severe ice storms occurred in Unitil’s service area, but contrary to the recommendation in the DPU’s 1992 order, FG&E did not study the storm preparation and response practices of other Massachusetts utilities. FG&E’s ERP in effect at the time of Winter Storm 2008 did not address a storm as severe and widespread as that storm.

At the time of Winter Storm 2008, FG&E also had in effect a vegetation management policy providing tree-trimming cycles for each of the circuits in its system. By the end of 2006, however, FG&E’s vegetation management program had become underfunded, and FG&E had fallen behind on its tree-trimming schedule. Rather than increasing its vegetation management budget, FG&E responded by adjusting downward its clearance standards and vegetation management cycles.

After Winter Storm 2008 struck, FG&E began restoring power to its customers on a rolling basis. The process started early on December 12, once National Grid had repaired its transmission supply lines, when FG&E restored power to about 500 customers in downtown Fitchburg whose underground network system was largely insulated from the weather. To complete full restoration, however, FG&E had to repair 244 utility poles, 192,729 feet of wire conductor, and 170 transformers. FG&E did not employ a sufficient number of its own restoration crews to respond to Winter Storm 2008, and had only limited success in obtaining additional crews from other utilities. By December 20, more than 4,000 FG&E customers remained without power, and at the Commonwealth’s request, National Grid stepped in to complete the necessary restoration work. The last customers did not have power until December 25.

The plaintiffs experienced outages beginning on either December 10, 11, or 12, and persisting for periods ranging from four to twelve days. Although all were attributable to the storm, the immediate reasons for their loss of power differed. For example, some plaintiffs lost power due to damage to FG&E’s subtrans-mission lines, whereas others lost power when trees in their yards, for which FG&E does not bear responsibility, fell on their individual service lines. The plaintiffs’ affidavits set forth various damages caused by the outages, including lost income and business profits, burst pipes, dead plants and pet fish, spoiled food, *48 and other expenses incurred as a consequence of and in response to the outages.

3. FG&E’s public communications. FG&E’s ERP at the time of Winter Storm 2008 provided that one of its primary responsibilities during an emergency was “[t]o keep the public informed of the status of the restoration in a timely manner through direct contact with town and city officials and the news media.” According to FG&E, its principal method for communicating with the public during Winter Storm 2008 was through public service announcements (PSAs). FG&E issued between one and five PSAs daily from December 11 through December 24.

As of the morning of December 11, FG&E knew that the weather forecast was for “anywhere from 1/4 inch to more than 1 inch of ice,” that “the current thinking is that FG&E is in the worst position” of the Unitil subsidiaries, and that “1/4 inch of ice and a little wind would be problematic, so having a forecast of over 1 inch would likely result in an extended restoration period that could easily exceed one week.” Nevertheless, FG&E issued the following PSA later that day: “Most electrical outages are expected to be for relatively short periods of time, only. However, severe weather conditions can create substantial damage to the electrical system, and restoration can take an extended period of time.”

Within the first few hours of Winter Storm 2008, FG&E had information that its system had suffered extensive damage that affected more than just its subtransmission and distribution facilities. Accordingly, when it energized its subtransmission lines on December 14, it knew that this would not restore power to a large majority of its customers. During this time, FG&E also learned that it would not have access to repair crews that it previously had believed would be available. Nonetheless, from December 12 through December 15, FG&E’s PSAs stated that it would “take days” or “several days” to restore power. On December 16, FG&E’s PSAs began referring to “restoration of all primary circuits” by the end of the week, without explaining that this did not mean that all customers would have their power restored by that time. FG&E had information about the need for “extensive rebuilding of circuits,” but it did not communicate this fact until December 19.

FG&E’s PSAs also served as the source of information for its customer service representatives. From December 11 through December 25, FG&E’s call center received 164,136 calls, of which *49 only 32,327 reached a customer service representative. Some customers who did speak with representatives, including two of the plaintiffs, received false predictions as to when they could expect their power to be restored. Others who attempted to call did not speak to representatives but were connected to automated recordings that provided no information about restoration times.

4. DPU proceedings. On January 7, 2009, the DPU, pursuant to its regulatory authority under G. L. c. 164, §§ IE and 76, opened an investigation into FG&E’s preparation for and response to Winter Storm 2008. The purpose of the investigation was to determine whether FG&E had satisfied its public service obligation to provide safe and reliable service. 5 See D.P.U. 09-01-A, at 6-8, and authorities cited.

The DPU began its investigation by holding two public hearings and requesting written comments. The DPU accepted both sworn and unsworn statements, although it informed the public that it would not rely on unsworn statements in its decision-making process. See 220 Code Mass. Regs. § 1.10(1) (2008) (“The [DPU] shall follow the rules of evidence observed by courts when practicable .... All unsworn statements appearing in the record shall not be considered as evidence on which a decision may be based”). As summarized in its final decision, the DPU received a total of 243 oral comments and eighty written comments, describing a variety of problems regarding FG&E that customers experienced during the storm. See D.P.U. 09-01-A, at 9-17.

The DPU next conducted an adjudicatory proceeding pursuant to 220 Code Mass. Regs. §§ 1.01-1.15 (2008). The Attorney General intervened in the proceeding as authorized by G. L. c. 12, § 11E (a). On March 25, 2009, FG&E, represented by counsel, provided the DPU with prefiled testimony of various Unitil officials. It also submitted a self-assessment report that included *50 twenty-eight recommendations on how to improve its preparedness for and response to emergency events. The DPU held evidentiary hearings from May 11 to May 15, at which both the Attorney General and FG&E presented witness testimony. Both parties then filed initial briefs and reply briefs.

The Attorney General proposed a number of remedies to address FG&E’s alleged failures related to Winter Storm 2008. It urged the DPU to (1) require FG&E to adopt the twenty-eight improvements recommended in FG&E’s self-assessment report, (2) fine FG&E $4.6 million, (3) deny FG&E recovery of storm-related costs, and (4) reduce FG&E’s return on equity.

In a 215-page decision, the DPU found “numerous and systematic” deficiencies in how FG&E prepared for and responded to Winter Storm 2008. D.P.U. 09-01-A, at xiii. These deficiencies included inadequate planning and training for significant emergency events; inadequate preparation for Winter Storm 2008; unsatisfactory damage assessment after the storm; insufficient acquisition of external repair crews; various problems that prevented FG&E from restoring service in a timely manner; and a failure to provide the public with useful and accurate information, which “resulted in the inability of customers to plan appropriately for an extended outage.” See id. at 47, 60, 69, 83, 102, 121, 125. The DPU concluded that each of these deficiencies constituted a violation of FG&E’s obligation to provide safe and reliable service. See id. at 52, 60, 72, 83-84, 102, 121, 125. The DPU also found that FG&E had engaged in deficient tree-trimming practices, which likely contributed to outages experienced during the storm. See id. at xiii, 160.

To remedy these failures, the DPU ordered FG&E to undertake a comprehensive independent management audit at its own expense; to implement, with some modifications, the twenty-eight improvements in its self-assessment report; and to provide progress reports regarding the implementation of these improvements. See id. at 193-194, 208-214. While acknowledging that FG&E’s conduct might warrant monetary penalties, the DPU concluded that it lacked authority to impose the requested $4.6 million fine. See id. at ix, 185, 189. 6 The DPU decided to defer the questions of FG&E’s storm-related cost recovery and return on equity until *51 its next rate-setting proceeding. See id. at 196, 198-199. FG&E was notified properly of its right to appeal under G. L. c. 25, § 25, see 220 Code Mass. Regs. § 1.13, but did not file an appeal.

The DPU addressed FG&E’s cost recovery and return on equity in its next rate-setting decision, at which time FG&E sought to recover storm-related costs totaling $22,120,286. See D.P.U. 11-01, at 13-15. The DPU adopted the findings in its investigatory decision, including those pertaining to specific failures by FG&E. See id. at 16, 59, 66. 7 Relying on these findings, the DPU allowed recovery of costs that FG&E could not have avoided given the magnitude of Winter Storm 2008, but denied FG&E’s request with respect to $6,954,492 in costs incurred because of FG&E’s failures leading up to and during the storm. See id. at 13-14. The DPU also reduced FG&E’s return on equity in part to reflect its deficient performance. See id. at 14. FG&E again did not appeal from the DPU’s decision.

Discussion. 1. Class certification. With respect to both rule 23 and G. L. c. 93A, we review a grant or denial of class certification for an abuse of discretion. See Salvas v. Wal-Mart Stores, Inc., 452 Mass. 337, 361 (2008); Moelis v. Berkshire Life Ins. Co., 451 Mass. 483, 486 (2008). Under both sets of certification requirements, the plaintiffs “do not bear the burden of producing evidence sufficient to prove that the requirements [of class certifi *52 cation] have been met,” but need only provide “information sufficient to enable the motion judge to form a reasonable judgment” that the class meets the relevant requirements. 8 Weld v. Glaxo Wellcome Inc., 434 Mass. 81, 87 (2001). See Aspinall v. Philip Morris Cos., 442 Mass. 381, 391-392 (2004); Kwaak v. Pfizer, Inc., 71 Mass. App. Ct. 293, 297 (2008). General Laws c. 93A, however, affords a judge less discretion to deny class certification than does rule 23. See Salvas v. Wal-Mart Stores, Inc., supra at 370 n.66, citing Aspinall v. Philip Morris Cos., supra.

To support class certification under rule 23, plaintiffs must satisfy the four elements of rule 23 (a) and the two additional elements of rule 23 (b). Rule 23 (a) requires the plaintiffs to show that “(1) the class is sufficiently numerous to make joinder of all parties impracticable, (2) there are common questions of law and fact, (3) the claim of the named plaintiff representative is typical of the claims of the class, and (4) the named plaintiff will fairly and adequately represent the interests of the class.” Weld v. Glaxo Wellcome Inc., supra at 86. In addition, rule 23 (b) requires them to show “that common questions of law and fact predominate over individualized questions, and that the class action is superior to other available methods for fair and efficient adjudication of the controversy.” Weld v. Glaxo Wellcome Inc., supra.

Plaintiffs may bring a class action under G. L. c. 93A if they can show that they may seek relief for an unfair or deceptive act or practice under G. L. c. 93A, § 9 (2) or 11, that the act or practice “caused similar injury to numerous other persons similarly situated,” and that they would “adequately and fairly represent[ ] such other persons.” G. L. c. 93A, §§ 9 (2), 11. Where appropriate, “the public policy of the Commonwealth strongly favors G. L. c. 93A class actions.” Feeney v. Dell Inc., 454 Mass. 192, 200 (2009). In considering certification under G. L. c. 93A, a judge must bear in mind the “ ‘pressing need for an effective private remedy’ . . . and that ‘traditional technicalities are not to *53 be read into the statute in such a way as to impede the accomplishment of substantial justice. ’ ” Aspinall v. Philip Morris Cos., 442 Mass. at 391-392, quoting Fletcher v. Cape Cod Gas Co., 394 Mass. 595, 605-606 (1985).

Although the requirements of rule 23 (a) provide a “useful framework” for considering class certification under G. L. c. 93A, the similarity requirements of the rule do not equate with the requirement of G. L. c. 93A that the plaintiffs be “similarly situated” and have suffered a “similar injury” as members of the class they seek to represent. See Aspinall v. Philip Morris Cos., 442 Mass, at 391, quoting Fletcher v. Cape Cod Gas Co., 394 Mass, at 605. The class action provisions of G. L. c. 93A also have “a more mandatory tone” than does rule 23 in that they omit the predominance and superiority elements of rule 23 (b), see Fletcher v. Cape Cod Gas Co., supra at 605, quoting Baldassari v. Public Fin. Trust, 369 Mass. 33, 40 (1975), but a judge retains some discretion to consider these factors in determining whether putative class members are “similarly situated” and have suffered a “similar injury.” See Moelis v. Berkshire Life Ins. Co., 451 Mass, at 489-490; Fletcher v. Cape Cod Gas Co., supra at 605-606.

a. Chapter 93A. The Superior Court judge determined that the individual plaintiffs had proffered sufficient information to show unfair and deceptive conduct by FG&E against all its customers, but denied their motion for class certification based on his conclusion that they had not shown that such conduct caused “similar injury” to the putative class members. The judge did not abuse his discretion.

To pursue a class action under G. L. c. 93A, plaintiffs must show that the putative class members suffered “similar,” although not necessarily identical, injuries as a result of the defendant’s unfair or deceptive conduct. 9 G. L. c. 93A, §§ 9 (2), 11. See Salvas v. Wal-Mart Stores, Inc., 452 Mass. at 364, quoting Spear v. H.V. Greene Co., 246 Mass. 259, 266 (1923). As with G. L. c. 93A claims generally, a causal connection must exist between the unfair or deceptive conduct and the injury, see Hershenow v. Enterprise Rent-A-Car Co. of Boston, Inc., 445 Mass. 790, 797-800 (2006), and the injury must be a “separate, identifiable harm” that *54 is “distinct” from the unfair or deceptive conduct itself. Tyler v. Michaels Stores, Inc., 464 Mass. 492, 503 (2013). General Laws c. 93A does not permit class certification for the purpose of addressing limited common issues short of class-wide liability. See Fletcher v. Cape Cod Gas Co., 394 Mass, at 602. Accordingly, plaintiffs must show that they can establish causation of such similar injury on a class-wide basis. See Aspinall v. Philip Morris Cos., 442 Mass, at 397 n.19; Fletcher v. Cape Cod Gas Co., supra at 603-604.

The plaintiffs concede that Winter Storm 2008 would have caused widespread power outages without regard to any failures by FG&E, and they do not seek to certify a class on the basis of such loss of power. They press two other theories of injury instead: (1) prolonged power outages caused by FG&E’s failure to restore power more expeditiously, and (2) an inability to plan for the widespread and extended outages due to FG&E’s unfair and deceptive communications system. 10 We address these theories of injury in turn and conclude that neither one required the judge to certify a class.

i. Delay in restoration of power. The plaintiffs argue that regardless of the cause of any initial loss of power, FG&E’s numerous and systemic failures in planning for and responding to Winter Storm 2008 injured the putative class members by delaying restoration. This theory of injury requires the plaintiffs to prove that the class members suffered longer outages than they would have but for FG&E’s failures. As the judge concluded, this *55 theory of similar injury fails for the same reason that the plaintiffs cannot pursue a class action based on the initial loss of power: causation would have to be resolved “on an individual (or small group) basis.”

As reflected by the differing lengths of time that outages were experienced, FG&E customers lost electric power for various reasons, and distinct obstacles impeded restoration of their power. The plaintiffs have not indicated how long outages would have lasted in the absence of FG&E’s failures. Their expert testified that he made no attempt to estimate what a reasonable restoration period would have been, and the plaintiffs have not otherwise identified a model or method for determining how long outages would have lasted but for FG&E’s failures.

Even granting that FG&E’s alleged systemic failures may have had a general tendency to delay restoration efforts, relying on such a general tendency to prove causation with respect to each individual class member would involve impermissible speculation or generalization. See Salvas v. Wal-Mart Stores, Inc., 452 Mass. at 361, quoting Weld v. Glaxo Wellcome Inc., 434 Mass. at 85. Many customers, especially those who experienced shorter outages, may not have suffered prolonged outages due to FG&E’s conduct. If the plaintiffs were to prove FG&E’s systemic failures, it would still remain for each class member to show that those failures caused a prolonged outage on an individual basis. The nominal class action would “degenerate in practice into multiple law-suits separately tried.” Fletcher v. Cape Cod Gas Co., 394 Mass, at 604 n.8, quoting Advisory Committee Notes to 1966 Revision, Fed. R. Civ. P. 23(b)(3), 28 U.S.C. (1982).

The plaintiffs suggest that this inquiry would involve merely a question of damages. However, if customers did not suffer a longer outage than otherwise would have occurred, their claims must fail for lack of causation. See Hershenow v. Enterprise Rent-A-Car Co. of Boston, Inc., 445 Mass. at 800-801 (rejecting G. L. c. 93A claim where plaintiffs failed to establish that statutorily noncompliant contract terms made them worse off than they would have been had contract complied with statute); Lord v. Commercial Union Ins. Co., 60 Mass. App. Ct. 309, 317-318, 322-323 (2004) (defendant entitled to judgment on G. L. c. 93A claim where its unfair or deceptive act did not cause plaintiff’s injury). The judge did not abuse his discretion in denying class certification with respect to this theory of injury.

ii. Inability to plan. The plaintiffs’ second theory of injury does not depend on what caused either the initial loss of power or the *56 length of a particular outage. They argue that FG&E’s communications leading up to and during Winter Storm 2008 consisted of deceptions and omissions that failed to provide them with accurate and useful information regarding the expected scope and duration of outages. They contend that this failure caused all members of the putative class to suffer the similar injury of an inability to plan for the outages. The plaintiffs’ affidavits aver that had FG&E properly informed them, they would have prepared for the outages in various ways, including arranging alternate living accommodations for themselves, their relatives, and their pets; purchasing generators; draining water from pipes to keep them from bursting; obtaining firewood; and stocking less perishable food or storing food elsewhere. The plaintiffs also allege, pointing to statements made at the DPU public hearings, that the other putative class members likewise suffered damages due to their inability to plan. The judge did not abuse his discretion in denying class certification on this theory of injury.

Both in the Superior Court and in this appeal, the plaintiffs’ arguments have not been clear about whether their claim rests on an alleged system of actively false and misleading statements by FG&E, or whether the claim rests instead on FG&E’s alleged violation of an affirmative duty to provide its customers with useful information. These claims are distinct. To the extent the plaintiffs assert the former claim, we agree with the motion judge that, although the named plaintiffs may prove that FG&E made deceptive statements as to them, they cannot establish causation on a class-wide basis because not all class members were exposed to the same deceptive statements and because the announced time frames for restoration may have been accurate as to many class members. See Kwaak v. Pfizer, Inc., 71 Mass. App. Ct. at 300-301 (rejecting class certification of those seeking relief for deceptive advertising campaign, where not all class members encountered advertisements and variations existed in advertisements that proposed class members did encounter).

To the extent the plaintiffs assert the second claim, they fare only slightly better. The plaintiffs may be able to show that FG&E engaged in unfair or deceptive conduct with respect to all putative class members by violating an affirmative duty to provide useful information regarding the scope and duration of expected outages. FG&E’s ERP stated that one of its primary responsibilities during an emergency was “[t]o keep the public informed of the status of the restoration in a timely manner,” and the DPU con- *57 eluded that FG&E’s public service obligation encompassed a duty to provide the public with “accurate and useful information” during an emergency event, designed to enable customers to plan accordingly. See D.P.U. 09-01-A, at 122, 125. Cf. 940 Code Mass. Regs. § 19.05(1) (1998) (“It is an unfair or deceptive act or practice for a retail seller of electricity to fail to disclose material information about its products, services, or business, where such failure has the capacity or tendency to deceive or mislead a reasonable consumer, or has the effect of deceiving or misleading such a consumer, in any material respect”); Evans v. Lorillard Tobacco Co., 465 Mass. 411, 465 (2013) (suggesting that breach of voluntarily assumed duty to provide information may constitute violation of G. L. c. 93A). The DPU found that FG&E did not “provide accurate and useful information to the public regarding restoration times” and noted that this “failure in communicating the extent of damage resulted in the inability of customers to plan appropriately for an extended outage.” D.P.U. 09-01-A, at xi-xii, 125. As the plaintiffs argue, and the judge agreed for purposes of FG&E’s motion for summary judgment, although some of FG&E’s PSAs did predict that restoration would take “days” or “several days,” which arguably turned out to be accurate as to many customers, a fact finder could determine that these communications were too vague and general to satisfy a duty to inform. Cf. Aspinall v. Philip Morris Cos., 442 Mass, at 394 (“advertising need not be totally false in order to be deemed deceptive in the context of G. L. c. 93A”).

The judge was not required, however, to certify a class based on FG&E’s alleged violation of this affirmative duty, especially where the plaintiffs did not proffer a method for proving on a class-wide basis that FG&E’s failure to provide useful information actually interfered with the ability of the different putative class members to plan for the outages. 11

*58 As reflected by the variety of ways that the plaintiffs contend they would have responded to better information, this theory of injury would necessitate individualized inquiry regarding the counterfactual mental processes of each class member. Indeed, many class members — such as, perhaps, the approximately 500 customers in downtown Fitchburg who lost power for only a few hours or those customers who already possessed power generators — may not have planned any differently and therefore suffered no injury under this theory. Although individualized inquiries regarding affirmative defenses and, especially, calculation of damages do not preclude class certification on the question of liability, see Salvas v. Wal-Mart Stores, Inc., 452 Mass. at 367-368; Aspinall v. Philip Morris Cos., 442 Mass. at 402, a judge retains discretion to deny certification based on such considerations. See Moelis v. Berkshire Life Ins. Co., 451 Mass. at 490; Fletcher v. Cape Cod Gas Co., 394 Mass. at 603-604, 606-607. Of course, the judge remains free to alter his decision as the litigation proceeds. See Moelis v. Berkshire Life Ins. Co., supra at 491-492; Aspinall v. Philip Morris Cos., supra at 398 n.22, quoting School Comm. of Brockton v. Massachusetts Comm’n Against Discrimination, 423 Mass. 7, 14 n.12 (1996) (“decision as to class certification is not immutable”).

iii. Alternative classes and subclasses. The plaintiffs propose three alternative classes and subclasses that they contend the judge should have certified: (1) a subclass of approximately 17,000 FG&E customers who lost power for four days or more, (2) a class seeking only injunctive or declaratory relief, and (3) a subclass of approximately 4,000 customers who received estimated bills for December, 2008, that made no adjustment for the loss of power during that period. Where a natural alternative class or set of subclasses would address a judge’s concerns about certifying a class as initially proposed, the judge should redefine the original class or certify subclasses as appropriate. See Kwaak v. Pfizer, Inc., 71 Mass. App. Ct. at 302 n.8, quoting Richardson v. Byrd, 709 F.2d 1016, 1019 (5th Cir.), cert, denied sub nom. Dallas County Comm’rs Court v. Richardson, 464 U.S. 1009 (1983). Furthermore, a judge may certify a subclass represented by only some named plaintiffs, with other named plaintiffs asserting different claims, if it makes sense to litigate these various *59 claims in a single proceeding. See, e.g., McGonagle v. Home Depot U.S.A., Inc., 75 Mass. App. Ct. 593, 597 (2009) (judge certified two subclasses each represented by only one named plaintiff). Nonetheless, the judge did not abuse his discretion in declining to certify any alternative classes or subclasses here.

As to the first proposed alternative, narrowing the class to those who suffered outages of at least four days would not address the judge’s concerns because it would not eliminate the need for individual inquiries regarding whether the extended outage resulted from FG&E’s unfair or deceptive conduct and how the customers would have planned differently had they received better information. As to the second alternative, the judge noted that the responses of the Legislature and the DPU to FG&E’s conduct during Winter Storm 2008, see note 6, supra, may already have remedied the allegedly unfair or deceptive practices for which the plaintiffs seek equitable relief. The judge permitted the plaintiffs’ request for equitable relief to proceed on a nonclass basis, and if the plaintiffs succeed in proving the need for such relief, the judge may revisit his class certification decision at that time. See Moelis v. Berkshire Life Ins. Co., 451 Mass. at 491-492. Cf. Brantley v. Hampden Div. of the Probate & Family Court Dep’t, 457 Mass. 172, 184 n.15 (2010) (class certification properly denied where equitable remedies requested by named plaintiff would afford relief to those similarly situated notwithstanding absence of class). Finally, as to the third alternative, the judge acknowledged that an estimated-billing class might warrant certification but did not so certify because not all named plaintiffs had received estimated bills and the plaintiffs had requested such a class as just one of sixteen alternatives, without explaining how this class claim would fit together with the other claims brought by the named plaintiffs. Because the estimated-billing claim rests on conduct distinct from that underlying the claims that FG&E failed to restore power and failed to provide useful information, the judge properly could demand additional information regarding such a class and how it would comport with the larger litigation. See Kwaak v. Pfizer, Inc., 71 Mass. App. Ct. at 302 n.8.

b. Rule 23. The plaintiffs also seek to certify a class under rule 23 on a claim of gross negligence. This claim rests on FG&E’s asserted “deficient emergency preparedness, service restoration and communications system.” We understand the plaintiffs’ gross negligence claim to assert theories of injury analogous to those pressed under G. L. c. 93A, and therefore affirm the judge’s *60 denial of such a class for the same reasons. See Kwaak v. Pfizer, Inc., 71 Mass. App. Ct. at 298 (“a certification that fails under c. 93A would fail under the requirements of rule 23 as well”).

2. Issue preclusion. In its cross appeal, FG&E challenges the judge’s application of offensive issue preclusion, also known as offensive collateral estoppel, to factual findings made by the DPU. “The offensive use of collateral estoppel ‘occurs when a plaintiff seeks to prevent a defendant from litigating issues which the defendant has previously litigated unsuccessfully in an action against another party.’ ” Evans v. Lorillard Tobacco Co., 465 Mass. at 466, quoting Matter of Cohen, 435 Mass. 7, 15 (2001). Offensive issue preclusion “does not require mutuality of parties, so long as there is an identity of issues, a finding adverse to the party against whom it is being asserted, and a judgment by a court or tribunal of competent jurisdiction.” Pierce v. Morrison Mahoney LLP, 452 Mass. 718, 730 (2008), quoting Miles v. Aetna Cas. & Sur. Co., 412 Mass. 424, 427 (1992). Additionally, “the determination of the issues for which preclusion is sought must have been essential to the underlying judgment.” Matter of Brauer, 452 Mass. 56, 67 (2008).

Once a plaintiff establishes these initial requirements, the “central inquiry” becomes whether the defendant had a “full and fair opportunity to litigate the issue in the first action.” Pierce v. Morrison Mahoney LLP, 452 Mass, at 730, quoting Matter of Goldstone, 445 Mass. 551, 559 (2005). Offensive issue preclusion may apply to the findings of an administrative agency “so long as the tribunal rendering judgment ha[d] the legal authority to adjudicate the dispute.” Alba v. Raytheon Co., 441 Mass. 836, 841 (2004). See, e.g., Stowe v. Bologna, 415 Mass. 20, 22 (1993); Haran v. Board of Registration in Med., 398 Mass. 571, 578-579 (1986); Commonwealth v. Two Parcels of Land, 48 Mass. App. Ct. 693, 698 (2000). A judge has wide discretion in deciding whether the doctrine should apply in a particular case. Matter of Brauer, 452 Mass. at 67.

a. Identity and essentiality of issues. FG&E argues that the judge failed to identify the particular issues to be precluded and therefore did not determine whether they were essential to the DPU’s judgment and identical to issues presented here. FG&E points to the judge’s statement that “[i]f this case goes to trial, the trial judge will continue to have discretion to make the final decision as to what issues are already determined and what issues remain for trial.”

*61 Contrary to FG&E’s assertion, however, the judge identified a number of DPU findings subject to issue preclusion. He quoted the summary of the DPU’s investigatory findings contained in its rate-setting decision, see note 7, supra, among them that FG&E did not plan or train sufficiently for emergency weather events, that it did not prepare adequately for Winter Storm 2008, and that it did not communicate accurate and useful information to the public. The judge then treated these findings as settled in determining whether the plaintiffs could establish G. L. c. 93A violations as a matter of law.

These factual findings were essential to the DPU’s decision that FG&E’s various failures in preparing for and responding to Winter Storm 2008 constituted independent violations of its duty to provide safe and reliable service. See D.P.U. 09-01-A, at 52, 59, 71-72, 83-84, 121, 125. The DPU also imposed numerous requirements on FG&E designed to remedy each of its noted failures. See id. at 210-214. Accordingly, the DPU’s factual findings “b[ore] on the outcome of the case” and were essential to its judgment. See Jarosz v. Palmer, 436 Mass. 526, 533 (2002).

Likewise, the DPU’s factual findings are identical to issues here because the plaintiffs’ various G. L. c. 93A and gross negligence claims rest on the same conduct that the DPU found deficient. We reject FG&E’s contention that the two adjudications do not present identical issues because the plaintiffs assert claims under G. L. c. 93A that the DPU does not have authority to address. See D.P.U. 09-01-A, at 189. Issue preclusion may apply where the two adjudications involve the same subsidiary findings, even if they involve different ultimate claims. See Alba v. Raytheon Co., 441 Mass. at 843.

The DPU’s 215-page decision contains numerous subsidiary findings, and the plaintiffs asserted multiple theories of liability resting on different alleged failures by FG&E. We therefore understand the motion judge’s statement, that a trial judge has continued discretion as to issue preclusion, to be merely an acknowledgment of the inefficiency of conducting a full analysis of the preclusive effect of each subsidiary finding before it becomes clear what facts matter to the particular theories of liability that the plaintiffs assert at trial. The motion judge committed no error in deciding that the trial judge should retain discretion to make final decisions regarding which issues are precluded and which ones remain.

b. Fairness. FG&E also argues that, under the circumstances, application of issue preclusion would be unfair. “[F]airness is the *62 ‘decisive consideration’ in determining whether to apply offensive issue preclusion.” Pierce v. Morrison Mahoney LLP, 452 Mass. at 730, quoting Matter of Goldstone, 445 Mass. at 559. In making this determination, “courts generally ask whether (1) the party in whose favor the estoppel would operate could have joined the original action, (2) the party against whom it would operate had an adequate incentive to defend the original action vigorously, (3) ‘the judgment relied upon as a basis for the estoppel is itself inconsistent with one or more previous judgments in favor of the defendant,’ and (4) ‘the second action affords the defendant procedural opportunities unavailable in the first action that could readily cause a different result.’ ” Matter of Brauer, 452 Mass. at 70, quoting Haran v. Board of Registration in Med., 398 Mass. at 577-578. See Bar Counsel v. Board of Bar Overseers, 420 Mass. 6, 11-12 (1995), quoting Restatement (Second) of Judgments § 29 (1982) (listing eight circumstances for judge to consider when determining propriety of applying offensive issue preclusion). The judge enjoys “ ‘wide discretion in determining whether’ applying offensive collateral estoppel ‘would be fair to the defendant.’ ” Pierce v. Morrison Mahoney LLP, supra at 731, quoting Bar Counsel v. Board of Bar Overseers, supra at 11. The party facing preclusion bears the burden of proof on the question of fairness. See Bailey v. Metropolitan Prop. & Liab. Ins. Co., 24 Mass. App. Ct. 34, 37 (1987). FG&E does not contest the first and third factors; its appeal focuses on the second and fourth factors.

i. Incentive to litigate. FG&E argues that it did not have an adequate incentive to dispute its purported failures before the DPU. It points out that the remedies imposed by the DPU in its investigatory decision consisted largely of improvements that FG&E already had volunteered to undertake in its self-assessment report. FG&E argues also that affording preclusive effect to the DPU findings would undermine its cooperative relationship with the DPU and frustrate the public utility regulatory scheme by discouraging utilities from offering such voluntary submissions. See Bar Counsel v. Board of Bar Overseers, 420 Mass, at 11, quoting Restatement (Second) of Judgments § 29(1), (5) (in determining whether to apply issue preclusion, judge should consider whether “[tjreating the issue as conclusively determined would be incompatible with an applicable scheme of administering the remedies in the actions involved” and whether “[t]he prior determination may have been affected by relationships among the *63 parties to the first action that are not present in the subsequent action”). These arguments are unavailing.

First, FG&E understates its incentive to litigate before the DPU. During the investigatory proceeding, the Attorney General urged the DPU to deny FG&E recovery of storm-related costs, to reduce FG&E’s return on equity, and to impose a $4.6 million fine as a result of FG&E’s failures during Winter Storm 2008. Relying largely on the various failures found in its investigatory decision, the DPU in its rate-setting decision denied FG&E recovery of nearly $7 million in storm-related costs, and reduced FG&E’s return on equity. See D.P.U. 11-01, at 14, 23, 72-73, 374-375, 425-427. Furthermore, although the DPU ultimately concluded that it lacked authority to impose a fine, FG&E faced the possibility that the DPU would reach a different conclusion. Given the large financial stakes involved, FG&E had adequate incentive to litigate vigorously the facts found in the DPU decision. See Matter of Goldstone, 445 Mass. at 559-560.

Second, applying issue preclusion would not undermine the public utility regulatory scheme because the large financial stakes involved here already provided a significant incentive for FG&E to litigate the DPU action “to the hilt.” See Commissioner of the Dep’t of Employment & Training v. Dugan, 428 Mass. 138, 145 (1998). Moreover, partly in response to FG&E’s failures during Wnter Storm 2008, the Legislature authorized the DPU to impose fines of up to $20 million for violations of a utility’s emergency preparation and service restoration duties, an action the Legislature would not have taken had it believed that potential, liability of this magnitude would undermine the regulatory scheme. See G. L. c. 164, § 11

ii. Different procedural opportunities. As to the fourth fairness consideration, FG&E identifies three procedural distinctions between the two actions that, it contends, render issue preclusion inappropriate here: (1) the DPU considered evidence not admissible in the Superior Court proceeding; (2) different parties bore the burden of proof in the two actions; and (3) the DPU proceedings involved a more limited right of appeal. A party seeking to avoid issue preclusion must show that the procedural distinctions “could likely result in the issue being differently determined.” Matter of Goldstone, 445 Mass, at 561 n.7, quoting Restatement (Second) of Judgments § 29(2). We conclude that FG&E has not demonstrated that these distinctions affected the DPU findings at issue.

*64 First, FG&E emphasizes that the DPU considered extensive public comments from individuals not subject to cross-examina-tian. See D.P.U. 09-01-A, at 9-17. After reviewing the record, the judge concluded that the DPU did not rely on the public comments in making any findings that might have preclusive effect in this case. The DPU did reference certain public comments in discussing deficiencies in FG&E’s management of restoration crews and the performance of its call center. See id. at 101, 103, 106, 120. However, the DPU also relied on other evidence in finding these failures, and FG&E conceded that its crew logistics needed improvement and that its call center was overwhelmed during Winter Storm 2008. See id. at 101-102, 104-106, 112, 120-121. Accordingly, the public comments likely did not affect the DPU’s findings and therefore do not undermine the judge’s application of issue preclusion. See Commonwealth v. Two Parcels of Land, 48 Mass. App. Ct. at 699-700 (issue preclusion warranted notwithstanding agency’s consideration of hearsay evidence in first proceeding).

Second, FG&E points out that the burden of proof has shifted from FG&E in the DPU proceedings to the plaintiffs in this case. Compare Fitchburg Gas & Elec. Light Co. v. Department of Pub. Utils., 375 Mass. 571, 578-579 (1978) (once DPU challenges company’s decisions, company faces burden of proving that those decisions comply with valid DPU policies), with Cleary v. Cleary, 427 Mass. 286, 297 (1998) (plaintiff has burden of proving violation of G. L. c. 93A). Although a determination in a prior proceeding ordinarily has “no preclusive effect” where the burden has shifted away from the party facing preclusion, Jarosz v. Palmer, 436 Mass, at 532, a judge has discretion to grant preclusive effect to an issue if the burden of proof did not affect the outcome of the prior determination. See Matter of Goldstone, 445 Mass. at 563-564. Cf. Bar Counsel v. Board of Bar Overseers, 420 Mass. at 12 (noting that burden shifted between two proceedings but remanding for decision whether to apply issue preclusion).

The DPU did not recite the burden of proof in its investigatory decision, and the decision contains no language suggesting that the DPU’s factual findings rested on FG&E’s failure to carry its burden. See D.P.U. 09-01-A, at 58-60, 68-72, 81-84, 97-102, 119-128, 132-136, 143-147, 158-160. Furthermore, as reflected by the twenty-eight recommended improvements in its self-assessment report, FG&E largely did not dispute the DPU’s find *65 ings pertaining to FG&E’s deficient storm-related conduct. See id. at 199-204. Thus, the shift in the burden of proof did not foreclose the judge from applying issue preclusion. See Matter of Goldstone, 445 Mass. at 563-564 (notwithstanding shift in burden of proof, issue preclusion warranted because facts were not disputed in first adjudication).

Third, FG&E contends that it could not appeal the DPU’s factual findings because G. L. c. 25, § 5, limited its right of appeal to “matters of law.” To the contrary, however, FG&E could have challenged the DPU’s findings as “[unsupported by substantial evidence.” See G. L. c. 30A, § 14 (7) (e); Fitchburg Gas & Elec. Light Co. v. Department of Pub. Utils., 460 Mass. 800, 811 (2011). This limited right of appeal has not prevented us from affording preclusive effect to administrative findings. See, e.g., Stowe v. Bologna, 415 Mass. at 22; Brunson v. Wall, 405 Mass. 446, 451 (1989). Nor does FG&E’s decision not to appeal from the DPU’s adjudications render the application of issue preclusion improper. See Stowe v. Bologna, supra; Brunson v. Wall, supra; Conservation Comm’n of Falmouth v. Pacheco, 49 Mass. App. Ct. 737, 741-742 & n.5 (2000), and authorities cited.

The DPU conducted a five-day adjudicatory hearing at which FG&E was represented by competent counsel, and FG&E had a right to proffer evidence, subpoena witnesses, cross-examine witnesses under oath, present oral and written arguments, and appeal an adverse decision. See 220 Code Mass. Regs. §§ 1.06(6)(f), 1.10(9), 1.13. See also Martin v. Ring, 401 Mass. 59, 63-64 (1987) (emphasizing that precluded party “had ample opportunity in the prior adjudication to present evidence and to cross-examine witnesses” and that he could have appealed'from adverse administrative decision); Haran v. Board of Registration in Med., 398 Mass. at 578 (emphasizing that precluded party was represented by counsel at four-day administrative hearing). The judge’s application of issue preclusion was within the scope of his broad discretion.

Conclusion. We affirm the orders denying the plaintiffs’ motion for class certification, denying the plaintiffs’ motion for partial summary judgment, and denying in part FG&E’s motion for partial summary judgment.

So ordered.