Federal Court of Australia

Strand Fitness Pty Ltd v QBE Insurance (Australia) Ltd [2026] FCAFC 101

ORDERS

THE COURT ORDERS THAT:

1.    The applicants be granted leave to appeal from the orders of the primary judge made on 5 December 2024 in proceedings NSD637/2021 and NSD638/2021.

2.    The appeals be dismissed with costs.

Note:    Entry of orders is dealt with in Rule 39.32 of the Federal Court Rules 2011.

REASONS FOR JUDGMENT

THE COURT:

INTRODUCTION

1    We refer to the applicants for leave to appeal, who are also the applicants in the class actions, as the applicants, and the respondent insurers (QBE Insurance (Australia) Ltd and The Underwriting Members of Syndicate 2003 at Lloyd’s) as, respectively, QBE and Lloyd’s (or, when referring to both of them, the insurers).

2    The primary judge determined to declass four class actions that were being case-managed together: Cody Gemtec Retail Pty v Underwriting Members of Syndicate 2003 at Lloyd’s (Declassing Applications) [2024] FCA 1098 (PJ). Only the class actions against QBE and Lloyd’s are the subject of the appeals; the applicants in the two other class actions, against two other insurers — Insurance Australia Ltd (IAL) and The Hollard Insurance Company Pty Ltd (Hollard) — did not seek leave to appeal against the declassing orders made in those two proceedings.

3    The primary judge determined the declassing applications by reference to the issues identified in a document headed “Proposed List of Common Issues and Issues of Commonality to be determined at an initial hearing”, a copy of which was annexed to his Honour’s reasons (the Issues Table). That document covered all four class actions, and contained some sections that were only directed to policy wordings of particular kinds. Although framed as a list of “issues”, the parties to the appeals referred to the numbered items in that document as “questions”. We do likewise in these reasons, while retaining the language of “issues” when referring to the primary judge’s reasons, which used that terminology. Nothing turns on which word is used.

4    Before the primary judge, Mr Edwards KC of counsel appeared for the applicants in the class actions against IAL and Hollard, and Mr Mitchell of counsel appeared for the applicants in the class actions against QBE and Lloyd’s. The primary judge determined the declassing applications by focusing on issues 19–23, and did so on the basis that the applicants had accepted that those issues were the “high watermark of the issues of commonality requiring determination” (PJ [99]). However, reference to the transcript of argument before the primary judge shows that that concession was not made by the applicants in the QBE and Lloyd’s class actions; it was only made by Mr Edwards KC for the applicants in the IAL and Hollard class actions. QBE and Lloyd’s did not submit on the appeal that the applicants were wrong to object that the concession had not been made by them.

5    If the applicants had also focused on questions 19–23 in the conduct of the appeal, this may not have mattered. But their argument focused on entirely different questions: being questions 6(a), 11, 17 and (in the QBE proceedings only) 25, 36 and 40. By reason of this matter, we consider that the primary judge’s exercise of discretion did miscarry, and the Full Court must re-exercise the discretion, addressing the questions upon which the applicants focused on appeal (questions 6(a), 11, 17, 25, 36 and 40). Having addressed those questions, our reasons for concluding that the proceedings should be declassed largely correspond with the reasons of the primary judge for reaching the same conclusions (PJ [128]–[135]), albeit his Honour’s consideration of the exercise of the discretion followed his analysis which, as we have already noted, focused on the questions identified by counsel for IAL and Hollard as the high watermark of their case.

6    It is convenient to note, at this point, that although question 10 was among the group of questions by reference to which the applicants advanced their appeals, it was confirmed in oral submissions that the applicants did not rely on question 10 in support of their appeals, and the question was only relevant to determination of the notices of contention, if those notices were reached. Question 10 asked whether an insured had to have lodged a claim with the insurer before the insurer could be taken to have “withheld payment” within the meaning of s 57 of the Insurance Contracts Act 1984 (Cth), or whether it was sufficient that the insurer publicised its position that its policies did not respond to losses of a particular kind.

7    As will be seen, we address the issue of whether the class actions should be declassed by reference to the questions raised by the applicants, and to which their submissions were addressed. While acknowledging that those questions were put forward as “examples”, we proceed by reference to those questions for two reasons. First, the applicants did not make submissions on any other questions. Secondly, if the continued utility of the proceedings as group proceedings cannot be exposed by the very questions put forward by the applicants on that key issue in prosecuting their appeals, the prospect that they have overlooked a winning point lurking in another question is remote.

CLASS ACTIONS AND THE POWER TO DECLASS

8    The power to declass a class action is conferred by s 33N of the Federal Court of Australia Act 1976 (Cth) (the Act). Section 33N(1) provides:

33N     Order that proceeding not continue as representative proceeding where costs excessive etc.

(1)    The Court may, on application by the respondent or of its own motion, order that a proceeding no longer continue under this Part where it is satisfied that it is in the interests of justice to do so because:

(a)    the costs that would be incurred if the proceeding were to be continued as a representative proceeding are likely to exceed the costs that would be incurred if each group member conducted a separate proceeding; or

(b)    all the relief sought can be obtained by means of a proceeding other than a representative proceeding under this Part; or

(c)    the representative proceeding will not provide an efficient and effective means of dealing with the claims of group members; or

(d)    it is otherwise inappropriate that the claims be pursued by means of a representative proceeding.

9    Subsections 33N(1)(a)–(d) set out four independent — but potentially overlapping — reasons why the Court may conclude that it is in the interests of justice to declass a proceeding. Those matters are distinct from, and are not to be conflated with, the criteria specified by s 33C(1), for the commencement of a class action (the conflation of the enquiries under ss 33C and 33N featured in the successful appeal in Bright v Femcare Ltd [2002] FCAFC 243; (2002) 195 ALR 574 (Bright v Femcare)). In order for a class action to be validly commenced under s 33C(1), three cumulative criteria must be satisfied: there must be seven or more persons who have “claims against the same person”; the “claims” of all those persons must be “in respect of, or arise out of, the same, similar or related circumstances”; and the “claims of all those persons” must “give rise to a substantial common issue of law or fact” (emphasis added). As was explained in Wong v Silkfield Pty Ltd (1999) 199 CLR 255; [1999] HCA 48 at [28] (Gleeson CJ, McHugh, Gummow, Kirby and Callinan JJ), the word “substantial” in s 33C(1) does not indicate that which is “large” or “of special significance” or would have a “major impact” on the litigation but is, instead, directed to issues which are “real or of substance”.

10    We pause to note the distinct criteria for commencing a class action as the argument — particularly the submissions of QBE — at times tended to suggest that unless it were the case that some of the questions in the Issues Table would apply to all group members, the class actions ought to be declassed. Through the course of argument, it was clarified that QBE did not submit that the class actions were not validly commenced under s 33C, but pointed to the lack of a common question in the Issues Table arising for each and every group member as a matter that supported the exercise of a discretion to declass the proceedings under s 33N. We explain some background to the Issues Table at paragraphs 23–30 below.

11    The terminology of “common issues” and “issues of commonality” can be deployed to refer, respectively, to issues that are common to the claims of all group members, and to issues that are common to the claims of some group members. This is, as we apprehended it, the way in which the primary judge used that terminology. This distinction reflects the difference between the gateway requirements of s 33C, and other provisions of Pt IVA, which recognise that properly constituted class actions will often still involve group members whose claims will require determination of issues that are common only to a sub-set of group members: see, in particular, s 33Q, which allows the Court to make directions where it appears that determination of the issues common to all group members will not finally determine the claims of all group members, and provides for such directions to include directions establishing a sub-group with its own representative party where there are issues common to the claims of those group members. The class action regime also recognises that properly constituted class actions may also involve the determination of issues that are specific only to the claims of an individual group member: see s 33R.

12    We use the terminology of “common issues” and “issues of commonality” in this manner, but note that the expression “common issues” has been used — more commonly the further back one travels in the record of class action judgments — and continues to be used in a way that goes beyond issues that are common to all group members, and extends to what we, and the primary judge, refer to as “issues of commonality”. The broader meaning of the expression “common issues” is to be recalled when considering observations made in the course of declassing judgments.

13    At a practical level, class actions are typically conducted in this Court on the basis that there will be an initial trial that determines the whole of the representative parties’ claims, but it should not be thought that an initial trial is necessarily limited to the issues that must be tried in order for the Court to rule on the representative parties’ claims. What else may be included in orders identifying what will be dealt with at an initial trial will vary according to the circumstances, and is a matter of case management for the docket judge. For example, in Dillon v RBS Group (Australia) Pty Ltd (2017) 252 FCR 150; [2017] FCA 896 (Dillon), Lee J had made orders that the matters that would be determined at the initial trial included not only the whole of the applicants’ claims, the whole of the claim of a sample group member, specified questions of fact or law common to the claims of the applicants and group members, but also “any other questions of substantial commonality” identified in a schedule to the orders: at [12]. Similarly, in Gill v Ethicon Sàrl (No 3) [2019] FCA 587; (2019) 369 ALR 175, Lee J explained the importance of “Merck” orders in determining the issues to be decided at the initial trial, observing that (at [11]):

Nowadays a “Merck order” is made, almost always wholly or largely by consent, which identifies that at an initial trial the whole of the claim of the applicant or some other group representatives are determined, together with a list of questions identified in a schedule to the order. These questions reflect common issues or issues of commonality the answers to which, following judgment, can be the subject of s 33ZB orders (thus identifying the metes and bounds of the statutory estoppel).

14    While issues of commonality that do not arise for determination on the applicants’ claims may usefully be addressed through identification of sample group members whose claims are also accelerated for determination at the initial trial (see, eg, Stack v AMP Financial Planning Pty Ltd (No 2) [2021] FCA 1479; (2021) 401 ALR 113 at [224]–[227] (Beach J)), it is implicit in the approach taken in Dillon that issues of commonality may be addressed other than through a sample group member (although caution is required to avoid straying into hypothetical issues, as was raised by Lee J in Lloyd v Belconnen Lakeview Pty Ltd [2019] FCA 2177; (2019) 377 ALR 234 at [377]; see also the observations of Beech-Jones J (then a judge of the Supreme Court of New South Wales) in Rodriguez & Sons Pty Ltd v Queensland Bulk Water Supply Authority t/as Seqwater (No 5) [2015] NSWSC 1771 (Rodriguez) at [36]).

15    Determining issues of commonality may have real utility in disposing of the claims of the group even though they may not necessarily arise on the representative parties’ claims. One example, given by Lee J in Dillon at [67] in explaining why it may not be necessary to accelerate the claims of one or more other group members in order for an initial trial to determine issues not arising on the claims of the representative party, is where a named applicant bought shares early in the relevant period covered by a shareholder class action, and there may be utility in determining wrongdoing outside that early period where the issue arises for other group members (see also Dillon at [75]). Utility was also the touchstone of the approach of Beech-‍Jones J in Rodriguez at [14]–[20].

16    The point that an initial trial may include determination of issues of commonality that do not directly arise on the applicants’ claims is important, and we have dwelled on it because aspects of the insurers’ submissions appeared to assume that an initial trial could only address issues arising directly from the applicants’ own claims. For the reasons we have explained, that is not an accurate conceptual starting point.

17    Nevertheless, the ambit of the representative applicants’ claims, what issues still fall to be determined following the test cases, and the utility of the answers to the identified questions in the Issues Table, were the focal points of the appeals. That is, the appeals were conducted on a basis that did not seek to support the continued utility of the representative proceedings on the basis that an initial trial could determine issues of commonality even though they did not arise on the representative applicants’ cases. On the contrary (but subject to the clarification in oral submissions referred to in paragraph 86 below), the applicants’ position was that the questions addressed in argument did arise on one or more of the representative applicants’ cases, and that “principles” would “emerge” in addressing the applicants’ claims that would “at least provide guidance” for group member claims.

18    The insurers’ responses in respect of the questions upon which the applicants conducted the appeals was to argue (variously) that the questions either did not arise on the applicants’ cases or, if they did, had been determined by the test cases and/or involved individual issues so that addressing the applicants’ cases would be of little to no utility in determining the claims of group members. The applicants’ responses to this challenge addressed how the issues in question arose on their cases, to meet the test case point, and to refute the lack of utility contention. What the applicants’ conduct of the appeals did not do was to contend that the issues in question justified the rejection of the declassing application on the basis that the issues could be determined at the initial trial even if they did not arise on the representative applicants’ cases.

19    It is also necessary to clarify the analytical significance of the adjective “substantial” (and its synonyms), when used in describing issues of commonality (or common issues). In addressing the declassing application, the primary judge considered whether issues that transcended individual claims, and so had some degree of commonality, could be characterised as “substantial common issues of law and fact which necessitate a contested hearing”: PJ [102] and [115] (emphasis in original).

20    Once the gateway provisions in s 33C(1) have been satisfied, there is no statutory requirement that the continuation of a proceeding as a class action depends on issues of commonality being “substantial”. There is nothing objectionable in examining whether identified issues of commonality are “substantial” (or “significant”), provided it is understood that consideration of issues of commonality in terms of their significance is a tool that may be deployed in exposing whether it is in the interests of justice to declass the proceeding because of one or more of the sets of circumstances set out in ss 33N(1)(a)–(d); the significance of issues of commonality should not be misunderstood as effecting a de facto statutory prerequisite to defeating a declassing application. Indeed, considering the nature and significance of issues of commonality will often be critical to determining whether or not to declass a class action pursuant to s 33N. As Lindgren J observed in Bright v Femcare at [18]:

[O]rdinarily one would expect that, in an attempt to give effect to the legislative intention, a means will be sought, by case management techniques, to enable a representative proceeding to continue to the stage of resolution of the substantial common issues on the basis that after that stage is completed, an order under s 33N or directions under s 33Q will be made …

21    Conversely, the identification of issues of commonality, even if “substantial”, does not mean that it cannot be in the interests of justice to declass the proceeding. Subsections 33N(1)(a)–(d) set out a number of disparate circumstances which may lead the Court to conclude that it is in the interests of justice to declass the proceeding. Of particular relevance in the present appeals is s 33N(1)(c), which applies where the representative proceeding will not provide an “efficient and effective means” of dealing with group members’ claims. Section 33N(1)(d) refers to it being “otherwise inappropriate” that the claims be pursued by means of a representative proceeding. This provision recognises that the Court may consider a declassing application — or indeed consider declassing on its own motion — in a wide variety of circumstances. There are no hard and fast rules that will determine when the interests of justice will warrant declassing a representative proceeding.

22    By its terms, s 33N(1)(a) invites a comparison between the costs if the representative proceeding continues as such, and the costs if “each group member conducted a separate proceeding”. Section 33N(1)(b) directs attention to whether all the relief sought can be obtained by means of a “proceeding” other than a representative proceeding under Pt IVA of the Act. By contrast, ss 33N(1)(c)–(d) are broader; they do not direct attention only to other “proceedings”. This is particularly relevant in the present appeals, where the insurers’ internal claims handling processes, and the capacity to bring complaints to the Australian Financial Complaints Authority (AFCA), forms part of the consideration of whether the representative proceedings provide an “efficient and effective means” of dealing with the claims of group members.

ISSUES IDENTIFIED BY THE APPLICANTS AS WARRANTING CONTINUATION OF THE PROCEEDINGS AS CLASS ACTIONS

The Issues Table

23    What was specified in the Issues Table needs to be understood in its procedural context.

24    At the initial hearing of the declassing applications, the primary judge heard argument on those applications but did not determine them as his Honour considered that further information was required to examine: the extent of interest of the individual insureds in the class actions, being a relevant consideration to the utility of the class actions; and the specific common issues and issues of commonality that remained to be determined: PJ [59]ff.

25    The declassing applications were adjourned part-heard, and a process was put in place to obtain that information.

26    One aspect of that process was the issuance of a notice to group members inviting them to register their interest in the class action, which occurred between February and March 2024. The primary judge explained at PJ [64] that: “My underlying reason for requiring notification was not only to inform group members as to their rights and as to the extant applications, but I was cognisant that if there was minimal interest in group members having their claims determined in a class action then this would be a potentially determinative factor in the exercise of any discretion under s 33N.”

27    A separate aspect of the process concerned the identification of the issues said to remain for determination following the test cases. In his Honour’s words (at PJ [66]): “The second aspect of the applications that required further clarification relates to the Efficiency Ground and arose because of the need of the class action applicants to be far more specific in identifying the common issues and issues of commonality that they say remain to be determined in the class actions.” That exercise ultimately gave rise to an initial draft of the Issues Table (referred to as MFI-1), which underwent a further round of input from the parties under the so-called “order 2(e) process” (as explained below), and eventually culminated in the final Issues Table.

28    The hearing of the declassing applications resumed on 12 June 2024. Paragraph 2(d) of the orders made by the primary judge after the hearing required the applicants to identify whether each issue in MFI-1 was said to be a common issue or an issue of commonality. Paragraph 2(e) of the orders then required the parties to identify, for each issue in MFI-1, whether it had been addressed in the test cases, whether the respondents contended that the issue was inappropriately framed, the parties’ proposed answers to the issue, and the utility of determining it in the class actions. These matters were addressed by the parties in the columns headed “Applicants’ Responses to Order 2(e) Matters”, and “Respondents’ Responses to Order 2(e) Matters” in what eventually became the final Issues Table.

29    The primary judge described the Issues Table and its genesis as follows (PJ [70]–[72]) (emphasis in original):

70    Following the applicants belatedly attending to the task of comprehensively setting out the common issues and issues of commonality that they say remain to be determined, I required the respondents to provide a detailed response indicating whether the proposed question was appropriately framed; whether it was dealt with by the test cases; and whether there was utility in answering it.

71    This has proved a useful exercise and set out as Annexure B to this judgment is a schedule of the questions identified by the applicants marked MFI-3 and the principal submission made on behalf of the relevant respondent as to why that issue is insufficiently common or otherwise lacks utility.

72    Regrettably, obtaining clarity as to these two matters has significantly delayed the resolution of the declassing applications. But having identified the level of group member interest and having obtained clarity as to the remaining issues said to have a degree of commonality, a firmer basis has been established by which the Court can make an assessment as to whether, in particular, the Efficiency Ground is established, and whether the Court should proceed to declass the proceedings. I will return to both these matters below.

30    The Issues Table was divided into parts, as follows:

(a)    “Issues relevant to all wordings” (questions 1–14);

(b)    “Issues that affect disease outbreak clauses” (questions 15–24);

(c)    “Issues that affect hybrid clauses generally” (questions 25–32);

(d)    “Issues that affect hybrid outbreak clauses” (questions 33–35);

(e)    “Issues that affect disease hybrid clauses” (questions 36–37); and

(f)    “Issues that affect discovery hybrid clauses” (questions 38–43).

The test cases

31    The broad outline of the test cases was described by the primary judge (at PJ [85]–[91]) as follows (emphasis in original):

85    There have been two test cases conducted in Australia. Unimaginatively, they have been called the First Test Case and the Second Test Case. As noted above, both were organised by the insurers, or more specifically, the Insurance Council of Australia (ICA) with the consent of AFCA, the external dispute resolution scheme with power to make determinations of up to $542,500 (prior to 1 January 2024) for direct financial loss, which determinations are not binding unless the insured accepts the determination. That limit has since increased for complaints lodged on or after 1 January 2024 to $631,500, according to the AFCA webpage entitled “Outcomes AFCA Provides”.

86    The First Test Case was announced in July 2020. It related to whether exclusions which referenced the repealed Quarantine Act rather than the current Biosecurity Act 2015 (Cth) excluded claims arising from the COVID-‍‍19 pandemic. In November 2020, the decision in the First Test Case was delivered by the Court of Appeal of New South Wales, being HDI Global Specialty SE v Wonkana [2020] NSWCA 296; (2020) 104 NSWLR 634. The Court of Appeal found that the exclusions did not apply to COVID-19. Special leave to appeal was refused.

87    After judgment in the First Test Case, the ICA announced that it was considering a further test case “that explores outstanding policy matters, including proximity and prevention of access, relating to the pandemic and business interruption insurance”.

88    The Second Test Case, then comprising nine individual proceedings, was commenced in this Court in February 2021, with a further proceeding commenced by QBE in April 2021.

89    The Second Test Case was heard by Justice Jagot at first instance, with her Honour being asked to consider certain questions relating to policy response (no case involved a loss or quantum assessment). Judgment was delivered in October 2021 in Swiss Re International SE v LCA Marrickville Pty Ltd [2021] FCA 1206; (2021) 394 ALR 461. In nine of the ten cases, it was found that the insured’s policy did not respond. In the tenth case (Meridian v IAL), which concerned an infectious disease clause, Jagot J found that the insuring clause was triggered. I am informed that this case has recently settled.

90    There were then appeals and cross-appeals in respect of five of the cases, with the Full Court (Moshinsky, Derrington and Colvin JJ) delivering judgment in LCA Marrickville Pty Limited v Swiss Re International SE [2022] FCAFC 17; (2022) 290 FCR 435. The Full Court largely upheld the conclusions of Jagot J. Special leave to appeal was subsequently refused.

91    The Second Test Case, obviously enough, was determined in relation to the policies, and on the facts (both agreed and proven), before the Court. Although the applicants “embrace the reasoning in the Second Test Case in relation to policy construction in so far as it can be applied by analogy” they contend “substantial common issues of law and fact” remain.

The applicants accepted that his Honour’s description was accurate in relation to the Second Test Case and did not contend otherwise in relation to the First Test Case.

32    As the primary judge explained, by the time it was tried, the Second Test Case comprised 10 individual proceedings and was heard and determined by Jagot J (then a judge of this Court): see Swiss Re International SE v LCA Marrickville Pty Ltd (Second COVID-19 Insurance Test Cases) [2021] FCA 1206; (2021) 394 ALR 461. Her Honour determined that, in nine of the ten cases, the insured’s policy did not respond. Five of the ten cases were the subject of appeals and cross-appeals, which were largely rejected by the Full Court. Special leave to appeal was refused.

The policies and group memberships

33    For QBE, there were 28 forms of policy wording with four different types of insuring clauses: infectious disease clauses, hybrid outbreak clauses, disease hybrid clauses, and discovery hybrid clauses: see PJ [18], which also describes these types of clauses. Based on available figures, there were approximately 36,000 insureds with relevant policies. There were 4,133 registrations (plus 241 “late” registrations) pursuant to the process referred to at paragraph 26 above: PJ [19]. By 12 July 2024, QBE had received 580 claims, lodged by insureds, of which over 500 related to hybrid clauses with a radius requirement.

34    For Lloyd’s, there were only four policy wordings, although it seems no policies were issued in respect of one of those wordings. There were only 77 policies, applying to 110 insured locations. 46 businesses registered pursuant to the process referred to at paragraph 26 above: PJ [22]. Only 10 claims under the relevant policies have been submitted; each was declined.

35    Counsel for the applicants acknowledged in oral submissions that, because the group member definitions require group members to have experienced business interruption as a result of the pandemic, not all policy-holders would be group members.

36    We turn now to the questions identified by the applicants as demonstrating the utility of the continuation of the proceedings as representative proceedings and, thereby, error by the primary judge. We address the questions in the same order as the parties addressed them in submissions.

Questions 17 and 25

37    Question 17 was as follows:

Do clauses (Disease Clauses) providing cover for business interruption or interference:

(a)     “due to”; or

(b)     “as a result of”

the outbreak of disease within a 20-kilometre or other specified radius of the insured premises or business (Required Area), require that the outbreak be the proximate cause of the interruption or interference? If not, what causal nexus is required?

38    Only the “as a result of” language is relevant to the Lloyd’s and QBE policies (cf the Hollard and IAL policies, which used the “due to” wording, as well as the “as a result of” wording in IAL’s case).

39    Question 25 was as follows:

Do hybrid clauses that respond to loss or interruption:

(a)     “due to”;

(b)     “in consequence of”;

(c)     “caused by or results from”;

(d)     “caused by”;

(e)     “as a result of”; or

(f)     “directly or indirectly arising from”,

closure or evacuation of the premises or business, require that the closure or evacuation be the proximate cause of the loss or interruption? If not, what causal nexus is required?

40    This question was only relevant to QBE’s policy wording.

QBE proceeding

41    The second applicant in the QBE class action ran a wine bar in Hampton, Victoria and had a policy with a hybrid outbreak clause. The wording of that clause referred to the “[l]oss of business income or gross rentals incurred as a result of: (a) the closure or evacuation of the whole or part of the premises by an order of a competent government, public or statutory authority as a result of … (iv) the outbreak of a notifiable human infectious or contagious disease occurring within a twenty (20) kilometre radius of the premises…” (underlined emphasis added; other emphasis omitted). The applicants noted that the parties had already agreed that the second “as a result of” causal link means something more than a remote causal link, but contended that there was no agreement in respect of the first “as a result of” causal link. The applicants submitted that that kind of causal link also arose, and was the only relevant causal link, in the infectious disease clauses (as distinct from the hybrid outbreak clauses where it was one of two links).

42    The third applicant in the QBE class action ran a cafe in Dandenong in Victoria and held a policy which provided indemnity in respect of “interruption or interference with your business in consequence of … (1)(c) closure or evacuation of all or part of the business premises by order of a competent government, public or statutory authority as a result of: … (ii) infectious or contagious human diseases; within a fifty (50) kilometre radius of your business premises” (emphasis added).

43    The fourth applicant in the QBE action ran gyms in Victoria and Western Australia and held two forms of insuring clause. The insuring clause provided that: “Interruption to or interference with the business as a result of the following shall be deemed to be Damage to property used by you at the situation”, including “[t]he outbreak of a human infectious or contagious disease occurring within a 20 kilometre radius of the situation”, and “[c]losure or evacuation of the business by order of a government, authorised public or statutory authority consequent upon: (a) the discovery of an organism likely to result in a human infectious or contagious disease…” (emphasis added).

44    It should be noted that the causal link required in order for closure or evacuation of premises by a relevant authority to be “as a result of” or “consequent upon” a disease outbreak was the subject of a different question (question 26). By contrast, question 25 was addressed to the causal connection between “loss or interruption” and the closure or evacuation of the premises or business, and question 17 concerned the causal connection between “business interruption or interference” and the outbreak of disease within the stipulated radius of the insured premises or business.

45    In relation to question 17, the applicants identified in the Issues Table that their proposed answer to the question was “No. The causal connection required is something more than a remote causal link.”

46    In relation to question 25, the applicants accepted that the wording “due to” or “caused by” was “proximate cause language”, but described the other expressions in submissions as expressions that constitute “broader or more generous causal links”. The applicants’ response in the Issues Table to question 25 was, accordingly, “Yes for ‘due to’ and ‘caused by’” and, for the remaining expressions, “No … something more than a remote causal link is all that is required”.

47    Although not part of the group of questions that the primary judge addressed on the basis that they constituted the “high watermark” of the applicants’ case, question 17 was part of a group of questions in relation to which his Honour made some observations. The primary judge considered that there was a “high degree of artificiality about the alleged commonality of some of the issues identified” (including question 17) (PJ [98]) and observed that question 17 (which was linked to question 25) was also part of a series of questions “which have a similar theme and have elicited a common response from the insurers” (PJ [101]). However, the primary judge reasoned that “[t]he reality is that following the test cases, what remains to be resolved are some potential issues which transcend individual claims, but upon close analysis I accept the insurers’ submission that they cannot be properly characterised as substantial common issues of law and fact which necessitate a contested hearing” (PJ [102]) (emphasis in original).

48    The applicants submitted, noting those observations, that there was “substantial commonality” and the “court would practically and conveniently deal with all these questions — that is, questions 17 and 25 across all the policies at once, noting that it will really need to deal with it in relation to four policy wordings” (the three applicants in the QBE proceeding, and the applicant in the Lloyd’s proceeding). The applicants did not suggest, however, that all of the contested wordings referred to in question 25 were to be found in the wordings of the second, third or fourth applicants’ policies.

49    QBE’s position in relation to question 17, at first instance, and on the appeal, was that the causal relationship required was not a matter that could be stated universally and, in the abstract, divorced from the facts of a particular case. QBE’s position in relation to question 25 was to the same effect. QBE also submitted that, as detailed guidance was given in the Second Test Case (both at first instance and by the Full Court) as to the principles for construing these phrases, there was unlikely to be any genuine dispute about those matters.

50    In oral submissions, counsel for QBE characterised questions 17 and 25 as being “in the nature of advisory opinions, which don’t arise on any of the applicants’ claims”. QBE submitted that care must be taken to distinguish between the causal link required between loss and the insured peril — which counsel described as “the classic causation in an insurance context” — and causal links within the insured peril. Using the second applicant’s policy as an example, the distinction was between the causal link required between the loss of business income and the insured peril — the loss needing to be incurred “as a result of” the peril — and the causal link required between the closure or evacuation of the premises by order of a competent authority and what caused that order.

51    QBE’s counsel submitted that the Full Court in the Second Test Case addressed the principles concerning causation and proximate cause in the insurance context. The Full Court confirmed, by reference to the decision of McColl JA (with whom Ipp and Tobias JJA agreed) in Lasermax Engineering Pty Ltd v QBE Insurance (Aust) Ltd [2005] NSWCA 66; (2005) 13 ANZ Ins Cas 61-643 (Lasermax), that absent a contrary intention, it is presumed that insurance contracts adopt a proximate cause requirement in relation to the standard between the incurring of the insured peril and the suffering of loss: LCA Marrickville Pty Ltd v Swiss Re International SE (2022) 290 FCR 435; [2022] FCAFC 17 (Second Test Case FC) at [106]–[111] (Derrington and Colvin JJ, with whom Moshinsky J agreed). The Full Court then contrasted concepts of causation where they exist within the definition of an insured peril: Second Test Case FC at [112]. As QBE explained the Full Court’s judgment, where causal words such as “as a result of” are used as an element in defining the insured peril then the presumptive proximate cause standard does not apply and a lower or less demanding test is ordinarily to be expected, and was found in the case before the Full Court: Second Test Case FC at [531]–[535] (Derrington and Colvin JJ, with whom Moshinsky J agreed).

52    QBE’s submission was that, to the extent that matters of construction relevant to questions 17 and 25 can be addressed at a level of generality, they have been addressed through the Second Test Case and that whether or not the causal link is evident inevitably turns on the facts of each case. Furthermore, QBE submitted that the questions posed are hypothetical because there has been no suggestion, on any of the applicants’ claims, that cover has been denied due to QBE applying a proximate cause analysis, as opposed to some looser standard and that, in respect of the fourth applicant, it has been accepted by QBE that the business was closed as a result of the discovery of an organism (COVID-19). Drawing those matters together, QBE’s submission was that:

the point is, whether at the first stage it’s proximate cause, or something less, is not an issue that arises on any of the applicants’ claims, because it’s not in any way determinative of their outcome. And for that reason, we do say it is abstract and it is hypothetical.

53    Questions 17 and 25 are expressed in an abstract manner, and embed two questions (which we will, for want of a better term, refer to as the first and second sub-questions). The first sub-question asks whether the nexus required is a “proximate cause” nexus (ie, between the business interruption or interference and the outbreak of disease in a certain area, according to question 17, and between the loss or interruption and the closure or evacuation of the premises or business, in the case of question 25). The second sub-question is premised on the answer to the first sub-question being “no” and then asks “what causal nexus is required?”.

54    As QBE submitted, the Full Court in the Second Test Case rejected the suggestion that the “proximate cause” causal relationship — as it is described and presumptively applied in relation to the causal relationship between insured peril and loss — is also to be presumptively applied to causal relationships embedded in clauses describing the insured peril: Second Test Case FC at [109]–[111], [530]–‍[535] (Derrington and Colvin JJ, with whom Moshinsky J agreed). As to the second sub-question, we observe that the Full Court’s reasons (at [112]) make it apparent that close attention must be paid to the wording of clauses that will often have more than one causal link in the description of the insured peril:

In particular, the insured peril in the hybrid clauses is the occurrence of a series of causally-related, sequential events. Often they are (a) the occurrence of a disease causing (b) the actions of a relevant government authority causing (c) the closing or restricting of access to business premises. Since the causal nexus between (a) and (b) and between (b) and (c) are expressed in varying terms, it is necessary in each case to identify the precise nature of the cause referred to, the precise result to which it is to be associated, the degree of causal relationship that is necessary, all according to the language used for the particular provision, and whether that is present in the circumstance of the claim.

55    As we have noted above, the reasons of the Full Court in the Second Test Case go on to address, to the extent capable of generality, the approach to be taken to construing causal language in insuring clauses and, going further, the nature of the causal link required where the expression “as a result of” is used in describing an insured peril: Second Test Case FC at [532]–[535] (Derrington and Colvin JJ, with whom Moshinsky J agreed).

56    The second sub-questions posed by questions 17 and 25 invite answers that could, at most, only describe the causal nexus required in general language. To the extent that anything useful can be said at that level of abstraction, it appears to have been said by the Full Court in the Second Test Case.

57    While it may still be possible to view questions 17 and 25 as identifying issues of some commonality, in view of the way in which the matters covered by both sub-questions posed by questions 17 and 25 were addressed in the Second Test Case, we consider that there is likely to be only limited utility in the continued pursuit of the QBE proceeding as a class action in order to address questions 17 and 25. Moreover, for the reasons given by QBE, it is not apparent that those questions would clearly be answered through the determination of the applicants’ claims, less so that any answers given specific to the circumstances of those applicants would be of material benefit in resolving the claims of group members.

Lloyd’s proceeding

58    The Lloyd’s class action may be dealt with in short order. First, it should be recalled that only question 17 arises in the Lloyd’s action.

59    The applicant in the Lloyd’s proceeding operated gemstone retail businesses, one in each of Sydney’s CBD and Melbourne’s CBD. Its policy referred to the business being interrupted or interfered with “as a result of Loss or Damage”, and provided that there would be deemed to be Loss or Damage where there was an “outbreak of a notifiable human infectious or contagious disease occurring within a 20 kilometre radius of the Situation”.

60    However, Lloyd’s has accepted that the “as a result of” wording in the cover clause does not require the establishment of a proximate cause, and relies on the observations of Jagot J in the Second Test Case as to the application of a proximate cause standard not making any difference. Lloyd’s submitted that something less than a proximate cause will be sufficient, but establishing whether or not that is made out requires a commonsense evaluation of the causal chain. Given that matter, and the fact that what can otherwise usefully be said on the question has already been said by Jagot J in the Second Test Case, and by Garde J in Sayers Property Holdings Pty Ltd v AIG Australia Ltd [2024] VSC 139 at [32]–[41] (subsequently overturned on appeal but not on this point: see Sayers Property Holdings Pty Ltd v AIG Australia Ltd [2025] VSCA 294), Lloyd’s maintained that there would be no utility in the pursuit of the proceeding as a class action.

61    In view of the small number of group members in the Lloyd’s action, the even smaller number who have even lodged claims, and Lloyd’s acceptance that a proximate cause nexus is not required, question 17 does not present as a common issue, or an issue of commonality, warranting the continuation of the proceeding as a class action. Lloyd’s does not contend that the outbreak of disease must be the proximate cause of the business interruption or interference. To that extent, the action will not resolve any live controversy. To the extent that accepting that proximate cause is not required leaves unanswered what causal nexus is required, we accept Lloyd’s submissions that, beyond general statements of a kind already made in the cases to which reference has been made (and the Full Court’s judgment in the Second Test Case), the enquiry would be specific to the facts of the interruption or interference to the business of each insured.

Question 11

62    Question 11 asked, for each of over 20 sets of lockdown orders, “[w]hat caused the following orders to be made, or otherwise informed the making of them, including whether they were caused or informed by one or more occurrences or outbreaks of COVID-19, or were consequent upon one or more discoveries of SARS-CoV-2, and if so what occurrences, outbreaks or discoveries…”.

63    The pleaded claims concerned business closure orders, rather than lockdown orders. Perhaps this point was recognised by the applicants, as their submissions on the appeal described question 11 as “seek[ing] to resolve common questions regarding the cause of government lockdown/business closure orders”. While QBE, in its submissions, described question 11 as concerned with the causes of “lockdown orders”, it accepted that the cause of some of the lockdown orders raised by the applicants were relevant to their individual claims, did not otherwise take issue with the distinction between business closure orders and lockdown orders, and did not contend that the cause of “lockdown” (as distinct from business closure) orders robbed question 11 of utility. Rather, QBE contended that only a small number of the lockdown orders were relevant to the applicants’ claims, that the question, as framed, was divorced from the different relational language appearing in the 28 different QBE policy wordings, and that the cause of the March 2020 lockdown orders had been determined by the test cases, as had the approach to be taken to determining the cause of lockdown orders.

64    As summarised by the applicants, their case is that the orders can be traced to certain outbreaks, whereas the insurers say that the question has been determined in the Second Test Case by Jagot J and that none of the March 2020 orders were made as a result of particular outbreaks.

65    The cause of the March 2020 orders was the subject of factual findings made by Jagot J in the Second Test Case. As the primary judge recognised (PJ [10]), those factual findings were not strictly binding on the applicants in the QBE and Lloyd’s proceedings, as they were not parties to the Second Test Case. The applicants embraced the reasoning in that case as to the cause of orders being determined objectively but took the position that they would seek to secure a different outcome on the facts, by relying on evidence that was not before the Court in the Second Test Case. The applicants submitted that they would seek to rely on material, which they would invite the Court to infer was before the decision-makers and motivated the making of the decisions that were taken at a national level, specifically advice that identified particular outbreaks. It was suggested that it may be necessary to obtain further information, and it was expected it could be obtained on subpoena or pursuant to freedom of information requests.

66    In an exchange with the Court, counsel for the applicants accepted that determining the cause of each set of orders at large would only constitute the first step of the analysis, as it would then be necessary to establish, in respect of each insured’s business or premises, whether it was located within the stipulated radius of the location of any outbreaks referred to in the advice that was before the decision-maker. In submitting that the first stage of the analysis would still be useful, notwithstanding that the second stage referred to would be required, it was submitted that the second stage analysis would still need to be determined on the applicants’ claims, where the applicants were located (variously) in Sydney’s CBD, Melbourne’s CBD, Hampton and Dandenong in Victoria, and one applicant also had premises in Western Australia.

67    In respect of orders that were in place for extended periods of time, the applicants would invite an inference that the reason an order was implemented in the first instance remained the reason for its maintenance, but that individual insureds not within the geographic area in question on that basis would still be able to advance a case that the maintenance of an order occurred on a basis other than that on which it was initially instituted, and that reason brought them within the insuring clause. Even more issues were also recognised as arising where premises were originally closed, for example as a result of the March 2020 orders, but then permitted to reopen months later, but only on a limited basis, and whether the limited reopening would be regarded as a different event.

68    Although the policy wordings contain different relational wording, as QBE submitted, in our view, the question of whether particular orders were made because of outbreaks having occurred at particular locations is capable of determination on a common basis. To that extent, we accept that question 11 raises some issues of commonality.

69    However, the cause of the March 2020 lockdown orders was determined by the Second Test Case. While the applicants are not strictly bound by that finding, and have pointed to a theoretical pathway by which they would ultimately present different evidence and so hope to secure a different factual conclusion, that pathway is speculative. The likelihood is that any attempt to achieve a different outcome on the fact of the cause of the March 2020 lockdown from that achieved in the Second Test Case would be time consuming, yet forlorn.

70    In relation to lockdown orders other than the March 2020 orders, to the extent that a small subset of the large collection of lockdown orders listed in question 11 would be determined on the applicants’ claims, it is not apparent that the answers to those questions would have much utility in respect of the wider group, even assuming that some group members in the QBE class action are likely to be located in the same general geographic areas as the applicants.

71    In the case of the Lloyd’s class action, we accept the submission of Lloyd’s that the lack of apparent utility is amplified by the relatively small size of the class, and the geographical spread of the “Situation” of group members’ businesses.

Questions 36 and 40

72    Question 36 was as follows:

For disease hybrid clauses which respond to disease without any expressed locational requirement:

(a)     does there need to be (or be perceived by a relevant authority to be) an occurrence of disease at all?

(b)     if not:

(i)     is a potential occurrence (or perceived potential occurrence) of disease sufficient? And can that potential occurrence be (or be perceived to be) anywhere in the world?

(ii)     is a risk of an occurrence (or perceived risk) of disease sufficient? And can that risk of occurrence be (or be perceived to be) anywhere in the world?

(c)     if so (to (a)):

(i)     does the occurrence need to occur (or be perceived by the relevant authority to occur) at the premises?

(ii)     can the occurrence (or perceived occurrence) occur anywhere in the world?

73    Question 40 was as follows:

For discovery hybrid clauses with no locational requirement:

(a)    does the discovery need to occur at the location?

(b)    if not, can the discovery occur anywhere in the world?

74    These questions were only relevant to the QBE proceeding. On the figures given by the applicants’ counsel:

(a)    591 policies had been issued relevant to question 40, and a further 2,156 polices were relevant to question 36;

(b)    only 14 claims had been made under question 40 policies, and 50 claims under question 36 policies (although these were “at least” figures, derived from QBE’s evidence); and

(c)    it appears that no claim has been approved under the question 40 policies, but four claims have been approved under question 36 policies (although one was for $0).

75    Questions 36 and 40 concerned only a sub-set of policies (six of QBE’s 28 different policy wordings, two in the case of question 36 and four in the case of question 40). The only applicant’s claim in respect of which either question was said to arise was that of the fourth applicant, who held a question 40 policy.

76    Questions 36 and 40 received some attention in the primary judge’s reasons. His Honour observed (at PJ [101]) that “[c]learly there remain some real or potential disputes, which might be thought to have some degree of commonality. For example, as to Issues 35, 36 and 40, there is clearly an issue and the parties have been unable to agree an answer.” (emphasis in original). The applicants took issue with the primary judge’s conclusion (at PJ [102]) that the questions his Honour referred to in PJ [101] could not properly be characterised as “substantial common issues of law and fact which necessitate a contested hearing” (emphasis in original), on the basis that why that was the case was not explained and was in error.

77    QBE emphasised that questions 36 and 40 were said to affect only a small sub-set of its policy wordings, that question 36 was not said to arise on any applicant’s claim, and question 40 only arises in some way on the fourth applicant’s claim. QBE further submitted that the questions are framed in an abstracted and hypothetical fashion, “unmoored from actual policy wording or actual events alleged to engage that policy wording”. QBE contended that, even if the questions could be reframed so as to be capable of a binding answer, “neither question is a substantial question of law or fact common to the claims of all group members”.

78    Insofar as QBE’s submissions tended to suggest that a question must be “substantial” in order for the answer to it to have utility, we reiterate our observations (see paragraphs 8–22 above) to the effect that the substantiality (or otherwise) of the question is relevant to the assessment of whether or not a proceeding should be declassed under s 33N of the Act. We also do not accept QBE’s contention that, because questions 36 and 40 are not common to “all” group members, that establishes the lack of utility of the class action proceeding insofar as it concerns those two questions (see our observations above at paragraph 11 regarding the distinction between common questions and questions of commonality).

79    However, rejection of these elements of QBE’s argument does not much assist the applicants. As we have outlined above, the parties argued the appeal on the basis that an initial trial would determine the claims of the applicants. An initial trial of the applicants’ claims would, on any view, not address question 36.

80    The wording referred to in question 40 only concerns four of QBE’s 28 policy wordings, and it appears that only 14 claims have been made (noting that this figure was presented as an “at least” figure, so may be an understatement to some extent). While QBE accepted that question 40 could conceivably arise on the fourth applicant’s claims, because QBE had accepted that the insured peril had occurred, there was in fact no controversy on this point that would be determined as part of the fourth applicant’s claims.

81    While the possibility of sample group members being appointed had been left open, the applicants’ submissions on questions 36 and 40 did not suggest that their utility was to be supported on the basis of sample group members. Added to which, and as QBE submitted, question 36 in particular is convoluted and divorced from specific policy wording.

82    For these reasons, while it is possible to identify an element of commonality in questions 36 and 40, it is unlikely that an initial trial would address either question in a way that lends utility to the continuation of the QBE action as a class action.

Question 6(a)

83    Question 6(a) was as follows (emphasis in original):

For the purpose of the basis of settlement provisions, where an insured received:

(a)     a rent waiver or deferral under the following:

(i)     COVID-19 Legislation Amendment (Emergency Measures) Act 2020 (NSW), Retail and Other Commercial Leases (COVID-19) Regulation 2020 (NSW); Retail and Other Commercial Leases (COVID-19) Regulation (No 2) 2020 (NSW); Retail and Other Commercial Leases (COVID-19) Regulation (No 3) 2020 (NSW); Retail and Other Commercial Leases (COVID-19) Regulation 2021 (NSW);

…

is this “savings made”, a “sum saved”, or “expenses saved” of a kind referred to in question 5 above?

84    The various legislative instruments referred to in question 6(a) required, among other things, the renegotiation of leases upon request where certain eligibility criteria were met (usually based on the tenant’s annual turnover and whether they qualified for JobKeeper).

85    The applicants submitted that question 6(a) was not answered by the Second Test Case. Their submission was that although Jagot J held that certain insureds would (if they had cover) need to give credit for rent relief, that conclusion was reached on the basis that the savings arose from the same underlying cause as the insured peril and by reference to the indemnity principle, and her Honour’s findings on rent relief were not appealed, even though her Honour’s findings on the need to give credit for JobKeeper payments were appealed and overturned by the Full Court.

86    In written submissions in the QBE action, the applicants contended that whether or not rent relief under the statutory instruments was a “saving” to be credited arose on the third applicant’s case. However, in oral submissions, it was accepted that this claim was in error and the matter did not arise for the third applicant in the QBE action. Nevertheless, the applicants submitted that the issue did arise on the claim of the applicant in the Lloyd’s action, and could arise generally in relation to the QBE policies if the cases were heard together. The applicant’s submissions in the Lloyd’s action stated that the applicant in that proceeding had received a “rent deferral” (we infer, then, not a rent waiver).

87    QBE’s position was that question 6(a) did not arise on any of the QBE applicants’ individual claims and, even if it did, it would not be capable of common resolution because, when they applied, the regulations in question required landlords to renegotiate the terms of commercial leases, which may involve deferral (and not just the waiver of rent), and would depend on the reasons why the rent waiver or deferral was agreed between the parties to the lease, being an individual issue. QBE also submitted, by reference to its answers in the Issues Table, that question 6(a) was divorced from the policy wordings as references to sums saved were, in many cases, to be found in clauses that qualified or limited what would constitute a “sum saved”, eg, requiring that the sum saved be “in respect of such charges and expenses of the business as may cease or be reduced in consequence of the interruption or interference”.

88    In its submissions, Lloyd’s pointed to the decision of the Full Court in Second Test Case FC (at [461]), in which their Honours considered that whether or not JobKeeper payments constituted a sum saved would, for one of the insureds (Meridian), depend on whether or not the payments were “in consequence of” the interruption or interference. Applying that to the class action, Lloyd’s submitted that, for any particular insured, it would be necessary to address whether the rent reduction or deferral would have been received irrespective of the insured peril, which would require close attention to the circumstances of each case and was not something that could be answered on a common basis.

89    Contrary to the applicants’ submissions, the fact that qualifying for JobKeeper was common to insureds in respect of whom the obligation to renegotiate arose does not mean that “the question can therefore be answered on a common basis”. The legislative material to which question 6(a) refers did not provide for rent relief. Rather, those provisions required parties to commercial leases to renegotiate, in certain circumstances. The product of such renegotiations was, necessarily, the product of negotiations between two parties to a leasing arrangement.

90    Even if the ultimate outcome of the negotiation was a reduction in rent, for the reasons given by the insurers, identifying rent reductions as the product of a negotiation would not answer the question of whether rent saved was to be accounted for pursuant to the terms of the policies. Moreover, the issue of rent savings did not arise on the claims of any of the applicants in the QBE action, and there was no proposal for issues outside their claims to be addressed as issues of commonality in this respect. In the case of the Lloyd’s action, the applicant was said only to have had rent deferred. The circumstances of a rent deferral, and the terms on which such deferral occurred, present even more individualised issues than is the case where rent is waived.

91    While not accepting that question 6(a) could not be answered on a common basis, as the insurers submitted, the applicants submitted that “whether or not it can be answered on a common basis is perhaps a common question to be addressed within the class action rather than at this stage, but we’ve addressed in our submissions why there is in fact a common source”, the common source being eligibility for JobKeeper. That submission rather exposes the limited utility in the determination of question 6(a) through a contested initial trial of the class actions.

DISPOSITION

92    Section 33N(1) confers a discretion on the Court to declass a class action proceeding where the Court is satisfied that to do so is “in the interests of justice” because of one or more of the matters listed in ss 33N(1)(a)–(d), of which (c) and (d) are of present relevance. Section 33N(1)(c) will be established where “the representative proceeding will not provide an efficient and effective means of dealing with the claims of group members”, and s 33N(1)(d) will be satisfied where “it is otherwise inappropriate that the claims be pursued by means of a representative proceeding”.

93    While recognising that the discretion to declass proceedings will only arise where the Court is persuaded that one or more of ss 33N(1)(a)–(d) is made out, the circumstances relevant to the satisfaction of ss 33N(1)(a)–(d) overlap with the matters going to the assessment of whether declassing a proceeding because of those matters is “in the interests of justice”. On the facts of this case, we do not consider there to be a separate set of matters to be considered in respect of the final exercise of discretion as compared with the matters arising in determining whether the statutory power to declass has been enlivened.

94    We have set out above our conclusions concerning why it is that the questions on which the applicants staked the appeals are variously unlikely to arise for determination on an initial trial of the applicants’ claims, tend to pose hypothetical questions or questions whose answers necessarily involve individual analysis, go over ground largely covered by the Second Test Case, or are otherwise questions whose determination lacks sufficient utility in relation to the determination of the claims of group members.

95    While the Court has wide procedural powers to fashion appropriate trial structures and, at times, some creativity is required, it was not suggested that the actions with which the appeals are concerned would proceed other than through the standard class action procedural approach. Pursuant to those ordinary procedures, there is an initial trial, following which orders are made pursuant to s 33ZB setting out the findings binding on group members, by reference to identified questions. Thereafter, the claims of individual group members are addressed.

96    Whether or not the representative proceeding provides an “efficient and effective means of dealing with the claims of group members”, and consideration of whether it is in the interests of justice to declass the proceedings invites consideration of matters that include the means by which group members’ claims may be addressed if the proceedings are declassed.

97    Compared with the number of insureds holding relevant policies issued by QBE — approximately 36,000 — only a small number of insureds have made claims — 580, as at July 2024. Although we are mindful that there may be reasons behind such relatively low claim numbers, and although the number of potential group members expressing interest by the registration process was higher, we consider it is still significant that such a relatively small proportion of potential group members have actually claimed on their policies.

98    The figures are just as, if not more, stark in respect of the Lloyd’s action, with only 10 potential group members having lodged claims. The number of claims lodged by potential group members in the Lloyd’s action is to be considered in the context of Lloyd’s having issued 77 relevant policies, applying to 110 insured locations.

99    Within the cohort of insured, there was substantial variation in the QBE action, arising from factors including the existence of 28 different relevant forms of policy wording.

100    Whatever views might be held regarding whether the test case approach was sensible — a matter on which the primary judge expressed views (see PJ [6]ff) — the fact remains that the test case procedure was adopted. Following that test case process, the insurers’ internal claims handling mechanisms stand ready to address such claims as arise for consideration.

101    The group proceedings processes of this Court are well adapted to the determination of common issues and issues of commonality. Use of those processes is typically efficient where determination of such issues directly resolves large parts of the claims of group members, or provides a sufficient binding framework of findings to allow the parties to settle the balance of group member claims. But this Court is not, nor would it be efficient to render it, in effect, an insurance claims handling clearing house, still less so where it is likely an initial trial would have limited utility in providing a framework of binding findings in respect of the identified questions.

102    By contrast, handling individual claims is what insurers do, day in, day out. The insurers’ internal claims handling procedures present a much more efficient mechanism for the determination of the claims of individual group member insureds (to the extent they actually made, or may yet make, claims) than any process this Court could adopt, whether that be by judicial assessment of individual claims by one or more judges, or by procedures involving the reference of claims to one or more referees for report.

103    The insurers’ internal processes were explained in greater detail by the primary judge (at PJ [73]ff). In summary, QBE’s internal claims handling processes provide for an insured who is dissatisfied with QBE’s response to a claim to submit a complaint, which will be handled by an internally independent “Dispute Resolution Specialist” according to an established process. The outcome of that process will also notify the insured of the right to contact AFCA if the insured is still dissatisfied. Lloyd’s’ internal processes provide for two stages following receipt of a complaint. If Lloyd’s does not require further information, the complaint will be determined in 15 business days and, if the insured remains dissatisfied, the complaint will be “reviewed” by Lloyd’s Australia if it falls within certain terms of reference, or otherwise it will be referred to an overseas internal complaints team. Subject to claims limits (AFCA may award up to $631,500 compensation for complaints submitted on or after 1 January 2024: PJ [85]), insureds who are dissatisfied with the outcome of the insurers’ internal procedures can submit a complaint to AFCA for determination. If AFCA’s determination is accepted by the complaining insured, it is binding. Any group member who is still dissatisfied following AFCA’s determination could pursue its claims by litigation.

104    We did not find the applicants’ submissions, by which they disputed the comparative utility of the insurers’ internal processes, to be persuasive. While it may be accepted that the proceedings may settle before an initial trial, the facts of this case leave no cause for optimism, particularly given the insurers’ positions following the test cases, and the significance they would be likely to attribute to the very low number of claims actually made by insureds who are potential group members.

105    Further, as referred to above, the enhanced potential for settlement following an initial trial is a benefit that is usually associated with the class action structure on the basis that the initial trial is expected to deliver a meaningful framework for determination of other group members’ claims. As such, the possibility for settlement at any stage does not, in this case, undermine the conclusions underpinning a determination to declass the proceedings. Further, even assuming, in the applicants’ favour, that most group members will not receive anything through the insurers’ internal processes, and that many group members will not advance claims outside a group proceeding context, those matters beg more questions than they answer. In some circumstances, it may be said that the likelihood that many group members with ostensibly meritorious claims will not recover and/or will not pursue their claims supports continuation of a group proceeding. Those conclusions cannot be drawn here, where there may be many reasons the vast majority of potential group members have not made claims on their policies, and it cannot be assumed that the insurers’ internal claims handling processes will deny recovery of claims that would, ultimately, succeed through the pursuit of the proceedings as group proceedings.

106    For these reasons, we consider that the QBE and Lloyd’s representative proceedings do not provide an efficient and effective means of dealing with the claims of group members, and that it is in the interests of justice that the proceedings no longer continue under Pt IVA of the Act. While we consider that the above matters likewise support the same outcome on the basis that it is “otherwise inappropriate that the claims be pursued by means of a representative proceeding”, it is not necessary formally to so conclude in order to support the disposition of the appeals.

CONCLUSION

107    Leave to appeal should be granted, but the appeals dismissed. While the applicants have succeeded in establishing error on the part of the primary judge, the error is confined to the primary judge having proceeded by reference to a different set of questions from those which the applicants said exposed the continued utility of the group proceedings (see paragraphs 3–5 above). We have addressed the questions upon which the applicants staked the appeals.

108    Having regard to the foregoing, it is not necessary to determine the notices of contention.

Associate:

Dated:    5 August 2026

SCHEDULE OF PARTIES