Certification of Class Actions: A 'Solution' in Search of a Problem?

Justice Lee 20 October 2017

RTF version - 157 kb

A paper presented to the Commercial Law Association Seminar "Class Actions – Different Perspectives"


The programme identified my talk as being on the topic of "Multiplicity of class actions: a Judge's perspective on managing competing claims and assessing proposed settlements". At the risk of engaging in conduct that suggests a want of reasonable grounds in forecasting the nature of my talk, I am not going to address the settlement aspect of this topic at all (but rather refer you to my recent paper, "Varying Funding Agreements and Freedom of Contract: Some Observations" which can be found at http://www.fedcourt.gov.au/news-and-events/8-june-2017) and I will only address the topic of competing claims briefly in my final remarks.

What I propose to do is to deal with a more current question, which is tangentially connected to the rise of multiple class actions, but is a matter of some present importance. It is the issue of whether there should be a threshold criterion for commencing class actions, being a pre-commencement hearing to consider and then determine whether certain conditions have been met.

This proposal can be more simply described as whether or not a certification regime should be introduced. The reason why this is a matter of current discussion is that in December 2016, the Victorian Attorney-General asked the Victorian Law Reform Commission (VLRC) to report on a number of issues with respect to group proceedings conducted pursuant to the Supreme Court Act 1986 (Vic). The stated reason for the reference was to ascertain whether reform was necessary to ensure that litigants, who seek to enforce their rights using the services of litigation funders and/or through group proceedings, are not exposed to unfair risks or disproportionate cost burdens.

A discussion paper was distributed by the VLRC, which raised for comment a number of issues, including the possible merit of a certification regime.[1] For obvious reasons, even proposed or possible developments with regard to class actions in Victoria are of potential significance to those involved with other class action regimes throughout the Commonwealth – in particular, Part IVA of the Federal Court of Australia Act 1976 (Cth) (FCAA) which is the foundational Australian representative proceeding regime upon which the cognate New South Wales, Victorian and Queensland legislative schemes have been based. It is readily appreciated that the broad consistency of the federal and various state schemes is an enormous benefit – allowing each court charged with case managing and hearing class actions to draw upon the accumulated experience reflected in decisions made by other courts on largely identical provisions – no doubt it would only be for compelling reasons that the matchless blessing of near uniformity would be abandoned. But, in the end, this is for others to judge.

I stress that what follows are my personal views. Although they do not, in any way, reflect the views of the Federal Court, they are the views of someone who was briefed in most of the largest representative proceedings over the last decade and has been required to give detailed consideration to the structure of representative proceedings and has been involved, in one way or another, with the cases that have given rise to a number of the procedural developments in class actions in recent years.

The Present Position

It is appropriate to start by a short excursus on the present position.

Part IVA, unlike its North American equivalents, does not contain a certification requirement. The suggestion that such a certification criterion be adopted is neither novel nor new. The Australian Law Reform Commission in its report, Grouped Proceedings in the Federal Court (Report No. 46, 1988) (ALRC Report), recommended that the FCAA not include a certification requirement and noted:

In class actions in the United States and Quebec, the preliminary matter of the form of the proceedings has often been more complex and taken more time that the hearing of the substantive issues. Because the court's discretion is involved, appeals are frequent, leading to delays and further expense. These expenses are wasteful and would discourage use of the procedure. There is no need to go to the expense of a special hearing to determine that the requirements have been complied with as long as the respondent has a right to challenge the validity of the procedure at any time. [2]

In summary, the ALRC Report reflected the considered view that adequate protection of group members would be achieved by the Court giving effect to its supervisory and protective role as to group members, which is reflected by aspects of the class action mechanism, including opt-out provisions, mandatory notice requirements and the ability, in certain circumstances, to remove representative applicants.

The ALRC rejected a certification procedure in definitive terms and concluded that it saw "no value in imposing an additional costly procedure, with a strong risk of appeals involving further delay and expense, which will not achieve the aims of protecting parties or ensuring efficiency".[3]

Importantly, this rejection must been seen in the broader context of the work conducted by the ALRC and it progeny, Part IVA (as enacted). In broadly implementing the ALRC's recommendations, a deliberate legislative choice was made to adopt a new form of proceeding. As was recognised in Wong v Silkfeld Pty Ltd (1999) 199 CLR 255, the purpose of the enactment of Part IVA was not to narrow access to the new regime and that the "threshold requirements" of commencing a proceeding were required to be construed accordingly.[4] The narrow and technical requirements of the pre-existing rules developed in Chancery were swept away and the new scheme was to be construed to give effect to this statutory purpose.

Any proper examination of Part IVA practice and procedure (the detailed understanding of which is a starting point for any suggestion of reform) must take account of the central objects of Part IVA, which is to allow:

(a) access to justice for all persons whose claims are in respect of or arise out of the same, similar or related circumstances and which give rise to a substantial common issue of law or fact (s 33C(1));

(b) a reduction in litigation expense, per claim, by the aggregation of claims and the absence of resources thrown away in individually contacting and contracting with each group member;

(c) a means for the efficient resolution of a very large number of claims; and

(d) the provision of a private mechanism for regulatory enforcement of the rights of the persons in the position of group members that is efficient, effective and appropriately adapted to the practical environment in which the choses in action have accrued.

A textual analysis of Part IVA demonstrates how these policy ends are served: in essence, all that is required for a Part IVA proceeding to pass through the threshold, in effect, is that there be a substantial issue of law or fact arising in similar circumstances, common to a group membership which comprises seven or more persons: s 33C(1). Section 33H complements this provision by facilitating the assessment of whether a representative proceeding satisfies the requirements of s 33C(1). As can be seen (as has been said many times) the "gateway" requirements to Part IVA are undemanding; so undemanding that Part IVA requires only that a representative party surmount these s 33C criteria and have a sufficient interest to afford standing to bring the action; this is true even if the representative party subsequently ceases to have a claim against the respondent: s 33D.

Consistently with this structure, it is unnecessary for consent to be obtained from a group member for a person to be included as a member of the class: s 33E. Quite disparate claims may be involved as it is unnecessary that claims of group members be the same as the claim of the applicant – simply that such claims give rise to commonality in at least one not insubstantial respect.

Given no consent is required to be obtained from a group member and little might be known of the details of individual group member claims, it is unsurprising that specific protections were afforded to group members. These protections are threefold: a right to opt out, a right that must be provided by the Court (s 33J); the group member's right to make an application seeking substitution or related orders in the event of inadequate representation (s 33T); and the right to be notified in certain circumstances, for example, proposed settlement, want of prosecution or the proposed withdrawal of an applicant (s 33Y). Importantly, no provision requires group members to make any application or do anything with their claim against their will or oblige them to take any active step prior to an initial trial.

Moreover, and critically for present purposes, in addition to these specific protections for group members, another more general safeguard exists – this is s 33N.

 Section 33N(1) provides as follows:

(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 continue 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.

Importantly, the point of departure for this safeguard is that a properly constituted representative proceeding, passing through the s 33C and s 333H "gateway" provisions, is before the Court. A common misstep for those inexperienced in class actions, at least initially, was seeking to invoke s 33N when the real complaint was a want of compliance with s 33C (which meant there was no class action properly before the Court, allowing the representative aspects of the proceeding to be struck out or dismissed). Constitution and continuation are two distinct matters that must be kept quite separate – a distinction elided in 'informal consultations' referred to in part of the VLRC consultation paper.[5]

Together with ss 33L and 33M, s 33N is one of three sections of Part IVA that provide for "declassing" orders. They all provide discretions but they arise in different ways. Unlike s 33L[6] and s 33M[7], the question posed by s 33N is whether the interests of justice require that the proceeding not continue under Part IVA because of the existence of one or other of the circumstances listed in ss 33N(1)(a)-(c) or the Court is satisfied that it is otherwise inappropriate for the claims to be pursued as a representative proceeding under Part IVA: s 33N(1)(d).

Four matters can be immediately observed about s 33N(1):

(a) first, the section is part of an integrated scheme; its role is partly as a safeguard against inapt or maladroit use of the class action regime in circumstances where that scheme allows class actions to be commenced without the constraints of consent or permission of group members or the Court;

(b) secondly, the discretion to be exercised is a statutory discretion which requires formation of one or other particular opinions or judgments of the type referred to in Coal & Allied Operations Pty Ltd v Australian Industrial Relations Commission (2000) 203 CLR 194.[8] At first glance, the discretion created in the opening words of the section may be thought to be a discretion limited only by subject matter because the controlling discretionary opinion incorporates a concept of some breadth, viz. "the interests of justice". This would be an error, however, as consideration of this broad evaluative concept only arises if it is established that some or all of the circumstances described in ss 33N(1)(a)-(d) are present;

(c) thirdly, the applicant does not bear an onus of demonstrating that the interests of justice favour a representative proceeding either at the time of commencement of the proceeding or at the time of the s 33N application: Bright v Femcare Ltd (2002) 195 ALR 574[9];

(d) fourthly, the inquiry is whether the Court is satisfied that it is in the interests of justice to order the discontinuance of the proceeding under Pt IVA because of the existence of one or other of the circumstances listed or the Court is satisfied that it is otherwise inappropriate for the claims to be pursued as a representative proceeding.

The policy choice was to adopt a scheme providing ease of commencement, tempered by a discretionary 'control' mechanism. The colourfully expressed criticism of one North American commentator (quoted by the VLRC in its consultation paper)[10] seems to proceed from the fundamental misunderstanding that there is no measure of court control imbedded in Part IVA. A sophisticated textual and contextual analysis reveals the scheme has inbuilt protections and balances. Most importantly, the proof of the pudding is in the eating: it works.

Returning to the topic of certification, what then are the arguments for law reform of what appears to have been a remarkably successful example of law reform?

Outline of arguments in favour of a certification requirement in Part IVA

In March 2014, a report entitled Ripe for Reform: Improving the Australian Class Action Regime (Report) was released by King & Wood Mallesons, a firm of solicitors very experienced in acting in class actions. The Report was prepared for the US Chamber Institute for Legal Reform (a body, which according to its website, "advocates for civil justice reform to curb excessive litigation in both the US and abroad"),[11] and proposed reforms to Part IVA.

The Report argues that the ALRC's basis for recommending the exclusion of a certification regime in Part IVA has not been borne out by experience. The Report contends:

The decision not to include a class certification process has failed the test of time. The introduction of such a procedure would reduce costs, reduce the risk of inappropriate actions and safeguard the interests of group members and respondents.[12]

In making this argument, the Report further asserts that:

There is real value in having a structured and formal occasion on which the class action is subject to scrutiny, such as is provided by the U.S. certification system…. In assessing certification, U.S. federal courts exercise a high degree of oversight at an early initial stage of the proceeding to ensure that it progresses efficiently going forward.[13]

The Report also argues that "the proliferation of interlocutory applications" has, to a large extent, been the product of deficiencies in the drafting of pleadings by representative applicants and that such multitude of applications would be unnecessary if an early certification hearing was required.[14] Similarly, it suggests that the current regime permits a class representative to commence a speculative case which barely meets the requirements of Part IVA, "in the hope that their case will be improved or supplemented at the discovery stage".[15]

These concerns as to interlocutory disputation have followed others. In P Dawson Nominees Pty Ltd v Multiplex Ltd (2007) 242 ALR 111, Finkelstein J stated that:

Experience of class actions suggests that the absence of a certification process is itself the cause of numerous interlocutory applications with resultant expense and delay.[16]

Similarly, Professor Rachael Mulheron has written that:

Of all the significant class action regimes around the world, Australia's federal class action opted for the path of 'no certification'...As an experiment, it has been singularly unsuccessful. Litigation under Pt IVA has been mired in numerous interlocutory applications about issues that could better have been addressed at a certification hearing…The omission of a certification hearing has hardly achieved the cost-efficient and streamlined process that the ALRC hoped for when it recommended against certification. [17]

Leaving aside the perceived benefit in reducing interlocutory disputation (Reduction of Disputation Argument), the VLRC has summarised the other principal argument as to the benefit to be achieved from a certification process. This reflects one of the comments made noted above, being the "low threshold for commencing a class action" which has, according to some, resulted in class actions that are not suited to the class action mechanism proceeding. This is said to have caused little cohesion or commonality among claims of class members with an increase in costs and delays, and that any savings in time and cost by allowing such class actions to continue are illusory. For this reason it is said that an early certification hearing will assist in ensuring that claims which are not apt to be maintained are addressed with celerity at an early certification hearing (Suitability Argument).

I deal with both of these arguments below.

An Assessment of the Perceived Problem

Reduction of Disputation Argument

Let me commence my consideration of this contention by repeating something that I said in the first paragraph of a recent judgment (Dillon v RBS Group (Australia) Pty Ltd [2017] FCA 896):

Part IVA… has reached its silver anniversary. Born in controversy, in its turbulent infancy procedural disputation was its almost constant companion. As it matured, and the more the adjectival law and the lore of Part IVA developed, the disputation lessened. Despite this, procedural arguments still recur; indeed, they are often the same arguments in different guises….

The suggestion that the introduction of a certification regime would filter out litigation not suited to class actions, and would avoid unnecessary costs as a consequence of so-called 'satellite litigation' and numerous interlocutory applications has been advanced and repeated by academic commentators.[18]

Leaving aside the special topic of competing class actions (to which I will return below) these concerns do not accord with my personal experience, nor does it seem they reflect the empirical data, when properly analysed.

The concerns raised by Professor Mulheron (referred to above and by the VLRC), regarding the prevalence of interlocutory disputes in Australian class action litigation, were published in 2007.[19] This was once a justifiable criticism, but the courts, over the last decade and a half, have quelled a range of disputes not only in terms of the practice and procedure of Part IVA, but also how aspects of the substantive law intersect with Part IVA. This law and lore, to which I made reference in Dillon v RBS, means that areas of controversy such as: the true nature of a 'claim'; closed classes; the scope of the power in s 33ZF; the ability to 'convert' inter partes proceedings into class actions; group member discovery prior to an initial trial; the proper relationship between individual claims and common issues; the power of the Court to make common fund orders; whether group members must have claims against each respondent; the appropriate time and method for identifying common issues; and whether group members can be 'shut out' of participating in a judgment (in the event they do not register pursuant to a deadline imposed in a registration notice), have all been resolved. At the very least, the cases have attenuated the scope of disputation, with remaining issues to be worked out incrementally on a case by case basis. This reflects an approach once observed to be the genius of the common law.

In my experience, the bulk of so-called early 'satellite litigation' which arose at the beginning of cases (at the time certification would take place), and which was the subject of much criticism, revolved around the question of whether or not closed classes were permissible under the Federal Court's class actions regime in Part IVA. That question was resolved by the Full Court of the Federal Court (French, Lindgren and Jacobson JJ) in Multiplex Funds Management Ltd v P Dawson Nominees Pty Ltd (2007) 164 FCR 275, a decision handed down in December 2007. The issue simply went away to be replaced by another agitated in the same litigation – the managed investment scheme problem, which was then also simply resolved by regulatory intervention.

Although, occasionally, some arguments do recur in different guises (particularly relating to the nature of the claims made in class actions), the now established case law of Part IVA allows such arguments to be resolved by the application, adaptation or extension of the reasoning and principles explained in those cases.

In its consultation paper, the VLRC has identified[20] that certification will not do anything to address the many interlocutory disputes that are an incident of complex litigation which would not be addressed in a certification hearing. What is relevant for present purposes is the type of interlocutory disputation that is apt to be the subject of argument at the proposed certification stage. The best guide as to the prevalence and reality of a problem requiring resolution is assessing how often there is interlocutory disputation directed to the question of whether a class action ought to continue as a representative proceeding and how successful these attacks have been.

As explained above, a respondent (or the Court on its own motion) may seek an order under s 33N that a proceeding no longer continues as a group proceeding. The respondent or Court-initiated 'declassing' process has three important advantages: first, it is invoked only where there is a perceived problem (hence avoiding costs being spent in cases where there is no issue); secondly, it can occur at different stages, and hence can be adapted to deal with problems which may arise or become evident at differing stages of the proceeding; and thirdly, its scope transcends 'problems' with the proceeding and can be invoked after an initial trial as a case management tool to provide for the most effective mechanism of determining the individual claims of group members after all common issues (and issues of commonality) have been resolved.

The empirical data suggests that only 28.4% of the group proceedings issued in the Federal Court between 1992 and 2009 were subject to formal applications for declassing.[21] It appears that very few of these were related to 'declassing' after an initial trial (given the very rare examples where cases have proceeded to secondary and tertiary stages of decision-making).

The low rate of challenge, however, does not mean that respondents are deterred from making legitimate declassing applications. In the same period, only 19.1% of applications were successful,[22] which accounts for about 5% of all class actions. Given these rates of success, a conclusion is available to be drawn that many of the applications made may not have been soundly based.

Moreover, in addition to an overall low rate of success, importantly, the data demonstrates a diminishing success rate over time: from a high-water mark of a 40% success rate in the first five years, to a position where no applications were successful in the period 2004 to 2009. I have not done a detailed search and there may be some more, but I am personally aware of only four or so successful challenges in the Federal Court since[23] and one in the Supreme Court of Victoria[24]; I am also aware of number of challenges that were raised but, on mature reflection, not pursued. Irrespective of the precise numbers, it is less a flood than a trickle and (at least on my anecdotal observation), a diminishing trickle at that.

The conclusion seems obvious: in the 'turbulent infancy' respondents tested (with some evident success) the metes and bonds of the declassing power but, over time, those boundaries became clearer. Allied to this is the education of lawyers for the applicants whose experience has presumably assisted them in refining both case selection and class definition.

The notion there is some significant problem does not seem to be justified on the facts. There is much to be said for the view that the blunt instrument of legislative change in introducing a certification regime would unsettle a landscape that has become largely settled and potentially herald a new form of costly and time-consuming interlocutory disputation that will retard the progress of determining on the merits what are largely properly constituted claims that ought to continue as representative proceedings. This would be an outcome which is hard to reconcile with the policy informing case management imperatives in Part VB of the Act (and cognate provisions) which require primacy be given to the just resolution of disputes as quickly, inexpensively and efficiently as possible.

Suitability Argument

Given that in approximately 95% of cases[25] (with the percentage increasing) the class action procedure has been selected without controversy (or where controversy has apparently been unjustified), the suggestion that the undemanding "gateway" provisions have often resulted in class actions that are unsuitable is difficult to justify.

The further notion that little commonality among claims of class members in some cases has increased costs and that there are no savings in time and cost is similarly difficult to understand. Intuitively, even in small class actions, the determination of common issues (and issues of commonality) can be done cost effectively. It is hardly going to be cheaper for the issue to be resolved serially between parties to a number of individual proceedings. The fact that individual issues may then also need to be determined is neither here nor there. They would always have to be determined in any counterfactual proceeding. Moreover, the lack of secondary trials in even small class actions demonstrates the way the resolution, or pending resolution, of common issues, can spur non-curial quelling of disputes. In any event, the s 33N safeguard exists if (atypically) there is a bespoke problem.


The only circumstance where I consider a Court-initiated formal or informal certification may have merit is where two competing open class proceedings have been commenced and, even then, debate should include whether there is any need to add to case management powers which already exist. In any event, this is a large topic and any consideration should only follow detailed consideration of the decision of Beach J in McKay Super Solutions Pty Ltd (Trustee) v Bellamy's Australia Ltd [2017] FCA 947 (18 August 2017) and the comments of Foster J (noting that it was inappropriate to impose a 'one size fits all' approach to Part IVA proceedings, and that permitting two proceedings to run in parallel in that matter had not resulted in undue cost, confusion or delay) in Cantor v Audi Australia Pty Ltd (No 2) [2017] FCA 1042 (1 September 2017) at [74]-[75].

It follows from the above that to this interested bystander, the introduction of a one size fits all pre-commencement certification hearing seems likely to:

  • impose significant and unnecessary costs on those vast majority of group proceedings which are appropriately constituted;
  • delay the progress of the resolution of substantive disputes; and
  • encourage a new world of adjectival disputation and a potential return to the 'satellite litigation' that characterised the early years of the class actions regime as different boundaries are explored.

If these concerns were realised, the result might be described, with some understatement, as a suboptimal outcome.

To those eager to tinker with a system which might not be perfect, but which seems to be working, despite imperfection, it may be useful to recall the sage words of Hilaire Belloc who famously said: "always keep a-hold of nurse; for fear of finding something worse".

[1] Victorian Law Reform Commission, Access to Justice – Litigation Funding and Group Proceedings, Consultation Paper (July 2017).

[2] At 63 [146].

[3] Ibid at 63-4 [147].

[4] At 267 [28].

[5] See at [6.60].

[6] Which is enlivened when it appears likely that there are fewer than seven group members.

[7] Which is enlivened when it appears that if judgment were to be given, the cost to the respondent of identifying the group members and distributing to them the amounts ordered to be paid would be excessive.

[8] In Coal & Allied at 204-5 [19], Gleeson CJ, Gaudron and Hayne JJ observed that "'Discretion' is a notion that "signifies a number of different legal concepts". In general terms, it refers to a decision-making process in which "no one [consideration] and no combination of [considerations] is necessarily determinative of the result". Rather, the decision-maker is allowed some latitude as to the choice of the decision to be made. The latitude may be considerable as, for example, where the relevant considerations are confined only by the subject matter and object of the legislation which confers the discretion. On the other hand, it may be quite narrow where, for example, the decision-maker is required to make a particular decision if he or she forms a particular opinion or value judgment". (footnotes omitted)

[9] At 601 [128] per Kiefel J.

[10] At [6.31] – "At least one country – Australia – has apparently devised a class procedure that does not require court certification of the class. It is difficult to believe, however, that a group action can be maintained on any basis other than pure opt-in without some measure of court control. The risks of sloppy class definition are too great, including fundamental conflicts of interest and indeterminate res judicata consequences. The risks of indifferent or incompetent representation both by named class member parties and by class counsel are too great": Edward H Cooper, 'Class Action Advice in the Form of Questions' (2001) 11 Duke Journal of Comparative and International Law 215, 231-2.

[11] <https://au.linkedin.com/company/instituteforlegalreform>.

[12] Report at 1.

[13] Ibid at 5.

[14] Ibid at 6.

[15] Ibid at 7.

[16] At 116 [18].

[17] Rachael Mulheron, 'Justice Enhanced: Framing an Opt-Out Class Action for England' (2007) 70 Modern Law Review 550, 568.

[18] Vince Morabito and Jane Caruana, 'Can Class Action Regimes Operate Satisfactorily without a Certification Device? Empirical Insights from the Federal Court of Australia' (2013) 61 American Journal of Comparative Law 579; John Emmerig and Michael Legg, 'Twenty Five Years of Australian Class Actions – Time for Reform' (2017) 36 Civil Justice Quarterly 164, 169.

[19] Rachael Mulheron, 'Justice Enhanced: Framing an Opt-Out Class Action for England' (2007) 70 Modern Law Review 550.

[20] At [6.37].

[21] Vince Morabito and Jane Caruana, 'Can Class Action Regimes Operate Satisfactorily without a Certification Device? Empirical Insights from the Federal Court of Australia' (2013) 61 American Journal of Comparative Law 579, 594.

[22] Ibid, 597.

[23] Meaden v Bell Potter Securities Ltd (No 2) (2012) 291 ALR 482 (Edmonds J); Larsson v WealthSure Pty Ltd [2013] FCA 926 (Buchanan J) (although, properly analysed, this was a s 33C problem); Carr v Commins Hendriks Pty Ltd [2016] FCA 1282 (Rares J); Pampered Paws Connection Pty Ltd v Pets Paradise Franchising (Qld) Pty Ltd (No 11) [2013] FCA 241 (Mansfield J) (at a later time in the proceeding).

[24] A S v Minister for Immigration & Ors (Ruling No 7) [2017] VSC 137 (J Forrest J).

[25] Vince Morabito and Jane Caruana, 'Can Class Action Regimes Operate Satisfactorily without a Certification Device? Empirical Insights from the Federal Court of Australia' (2013) 61 American Journal of Comparative Law 579, 597.