REASONS FOR JUDGMENT
1 The question is whether I should recuse myself from hearing three shareholder class actions brought against the Centro group of companies. The central allegations are that the respondents made false statements to the market and failed to comply with their continuous disclosure obligations with the result that their securities traded at an inflated price. Motions for my recusal have been brought on the basis of the ownership by Motown Investments Pty Ltd, the trustee of a self-managed superannuation fund of which I am a member, of shares in one of the respondents, Centro Properties Limited. An additional problem is that the trustee was a member of the group on whose behalf one action is brought. For the reasons that follow, I am satisfied I should recuse myself.
2 First, some background. I am one of two members of the self-managed superannuation fund, the trustee of which is Motown Investments Pty Ltd. Motown has two directors; I am one. On 8 August 2007, Motown purchased 2,400 shares in Centro Properties Group (ASX ticker symbol: CNP). The price per share was $8.17, and the total cost was $19,608. At the time, this constituted a very small percentage of the fund's overall value.
3 Since their purchase the value of the shares steadily declined, as has the value of all the shares in the fund. As at 30 June 2008, the value of the CNP shares had fallen to something in the neighbourhood of $600. According to the ASX website, the value of a CNP share as at 10 October 2008 (the day I informed the parties of the Motown shareholding) was $0.068, making their then total market value something in the order of $200. I should also add that, according to CNP's annual report issued on 25 September 2008, the number of CNP securities issued and outstanding is over 800,000,000, meaning that the superannuation fund's holding amounted to a relatively insignificant proportion of the company's outstanding securities. In short, both in relative and absolute terms, Motown's financial interest in CNP has been minimal.
4 In company with many other judges, in various capacities (personally, as trustee and as a beneficiary) I have an interest in shares in a range of listed companies. As a general rule (but, as with all rules, there are exceptions), prior to dealing with any case I or my staff check to see if I have an interest in relation to either the parties or the subject matter of the proceeding. In this case, due to an administrative oversight it did not immediately come to my attention that Motown held CNP shares. What happened was that shares held by Motown are listed in a "Portfolio Valuation" prepared by Motown's investment adviser. (Almost all purchases and sales of shares, including the purchase of shares in CNP, are transacted on their advice). The shares are listed in several categories (eg "Aust Shares", "Property Listed", "Int'l Shares Managed") and appear in alphabetical order in each category. The category in which CNP shares were listed ("Property Listed") was inadvertently overlooked at first instance.
5 When a subsequent check, after the writing and just prior to the delivery on 10 October 2008 of my reasons for decision in relation to the respondents' motion for a stay (Kirby v Centro Properties Limited [2008] FCA 1505), revealed Motown's holding, I promptly disclosed it to the parties in open court. In response to further inquiries from Freehills, who act for CNP and CPT Manager, the respondents in two actions, by letter dated 17 October 2008 (a Friday), my staff provided most of the historical information set out above in an email sent on Monday, 20 October 2008.
6 On 24 October 2008, Freehills wrote to my staff stating that, notwithstanding this disclosure, CNP remained concerned that "there may be a reasonable apprehension by fair minded lay observers, including other CNP security holders, that [I] might not bring an impartial mind" to the proceedings. The concern was expressed as follows:
We consider that there is a logical connection between [the loss in value of my fund's CNP shares] and the feared affect [sic] on impartiality in that a lay observer may reasonably apprehend that a loss of approximately $19,000 in one's personal superannuation fund might create a disposition against the respondents, who are alleged to have caused the loss, or an inclination toward other security holders who may have suffered similar loss.
Freehills also expressed the view that even if Motown was removed from the class, the apprehension of bias would remain because of the realistic possibility that the outcome of the litigation would affect the value of the shares and the fact that Motown could still bring its own independent action. Freehills' recusal request was joined in by Middletons, solicitors for the other two Centro respondents.
7 Given the relatively small nature of my interest in CNP through the superannuation fund, I initially thought that disclosure of the interest and an appropriate amendment to the statement of claim to exclude Motown as a group member would be sufficient to address any appearance of bias. This was what I proposed both in court on 10 October and in a letter from my executive assistant dated 21 October. My reasons for thinking this would solve the problem were twofold. First of all, the High Court has held that a judge's holding of a relatively small stake in a corporate party to litigation does not give rise to an appearance of impropriety such as to found a reasonable apprehension of bias requiring the judge's recusal: Ebner v Official Trustee in Bankruptcy (2000) 205 CLR 337. Although disclosure of even a de minimis financial interest is no doubt the better practice, Ebner indicates that no disclosure is legally required and that certainly belated disclosure of such an interest that was the result of inadvertence (as here) or excusable neglect will suffice: at [70] (per Gleeson CJ, McHugh, Gummow and Hayne JJ) and [167] (per Kirby J). Second, having regard to the nature of a class action it is inevitable that, from time to time, a judge (and not only a judge who holds shares in listed companies) will find that he or she, or a related party, is a member of a group on whose behalf proceedings have been brought. In many cases the appropriate remedy is to amend the definition of the group to exclude the judge or the related entity. That is what occurred, for example, in Johnson Tiles Pty Ltd v Esso Australia Ltd (2000) 104 FCR 564. Nonetheless, I informed the parties that the only proper way for the matter to be dealt with was in open court. Indeed, that matters had reached the stage they had in itself required the matter to be heard publicly.
8 In the meantime, I began to examine the authorities for guidance. I could find no Anglo-Australian authorities that considered my precise position. But in the United States, a jurisdiction which takes a much stricter view on judicial shareholding than our own, the point is not novel. Many US jurisdictions have statutory, rule-based and case law authority for the proposition that, subject to certain exceptions including the common law rule of necessity (see US v Will 449 US 200, 217 (1980); for the Australian position on necessity, see Ebner at [64]-[65], [101]-[103] and [172]), any financial interest of the judge in the litigation requires recusal: see eg 28 USC s 455(b)(4) read with subsections (d)(4) and (e)(4) (which requires a judge to recuse himself if he or she holds a financial interest in the litigation, no matter how small, and regardless of whether the parties offer to waive any objection); ABA Model Code of Judicial Conduct (2004) Canon 3E(1)(c) (same); see also Tramonte v Chrysler Corp 136 F3d 1025, 1029-30 (5th Cir 1998) (holding that s 455(b)(4) imposes a per se recusal rule where the judge or someone in his or her family is a member of a class seeking financial relief). There is also a practice in the US Supreme Court for justices to recuse themselves without regard to the size of their financial interest and even if it results in the court being unable to hear a case for lack of a quorum or unable to decide a case for lack of a majority: see eg American Isuzu Motors Inc v Ntsebeza 128 SCt 2424 (2008) (affirming the judgment below where four judges were recused and the Court lacked a quorum); Warner-Lambert v Kent 128 SCt 1168 (2008) (affirming the judgment below where the chief justice was recused and the remaining eight justices were equally divided).
9 There is an exception to this recusal rule both in statute, in the case law and according to recent US Supreme Court practice. On occasion, a justice has divested himself of the stock prior to oral argument in the case, and returned to participate in the hearing and decision of the case: see Tony Mauro, Roberts' Recusal Is Poison Pill for Drug Case Before Supreme Court (Legal Times, 4 March 2008), available at http://www.law.com/jsp/article.jsp?id= 1204544937246. This practice is authorised by statute: see 28 USC s 455(f) (which provides that where substantial judicial time has been expended on the matter, a judge need not recuse based on the discovery of a financial interest if he or she divests himself or herself of the interest); see also Kidder Peabody & Co v Maxus Energy Corp 925 F2d 556 (2d Cir 1991) (where it was held that s 455(f) obviated the need for recusal even where the judge held a large share of the stock of one of the parties where he promptly divested his interest upon learning of it and three years of the parties' and court's time and resources had already gone into the litigation). While there is some criticism of this approach in the cases, the majority view is that divestiture cures any appearance of impropriety: compare In re Initial Public Offering Securities Litigation 174 FSupp2d 70, 80-81 (SDNY 2001) (declining to recuse based on s 455(f) and criticizing Tramonte); Key Pharmaceutical Inc v Mylan Labs Inc 24 FSupp2d 480 (WDPa 1998) (using s 455(f) curative divestiture as a basis for declining to recuse) with Gordon v Reliant Energy Inc 141 FSupp2d 1041 (SDCal 2001) (relying on Tramonte and holding that the conflict was incurable).
10 I came across two decisions that were particularly important. In re Certain Underwriter 294 F3d 297 (2d Cir 2002), the appellate decision in the IPO case just referred to, was a securities class action in which investors in an initial public offering during the stock market boom of the late 1990s sued to recover losses against the issuers, underwriters and brokers, alleging widespread manipulation in the IPO market. The class on whose behalf the action was brought included investors who had bought stock from a particular date. The judge was a member of the class who had purchased stock in one of the defendants as well as in other companies that may have been connected to the defendants. The judge sold her stock in the companies and waived any interest in pursuing claims against them. Still, she was asked to recuse herself. She declined. The Court of Appeals for the Second Circuit refused to issue mandamus against the judge. The court referred to 28 USC s 455(f) and (at 305) said a judge with a minor interest in a class action is not considered a party once she has divested herself of that interest. Nor according to the judges did the shareholding suggest partiality to an objective disinterested observer: 292 F3rd at 306.
11 In re Literary Works in Electronic Databases Copyright Litigation 509 F3d 136 (2d Cir 2007) concerned the settlement of a class action brought on behalf of freelance authors whose works had been reproduced without consent. The settlement was approval by the District Court. An appeal from that approval to the Second Circuit Court of Appeals came before a panel of three judges, two of whom were members of the plaintiff class and entitled to share in the settlement sum. The judges did not recuse themselves. Instead they said in open court that they would forgo any interest in the settlement and this was sufficient to get rid of any appearance of bias.
12 At the time, it seemed to me that this was the approach I should adopt. Accordingly, I arranged for Motown to sell its CNP shares. I assumed, in line with the US cases, that this, plus a waiver of all claims arising out of or related to the facts alleged in the pleadings, would cure any vestigial appearance of bias. With those steps attended to, I had the strong conviction that the respondents' position was wrong.
13 When the matter came on for hearing, I advised the parties that the shares had been sold. This information did not, as I anticipated it would, appease anyone. To the contrary, it seemed to have an inflammatory affect, although, as one would expect, counsels' submissions were restrained. The principal point made by counsel (Dr Pannam QC for Centro Retail Limited and Centro MCS Manager Limited, and Mr Uren QC for CNP and CPT Manager Limited) was that as Motown had now crystallized a loss of some $19,000 and I was, to that extent, a "victim" of what the applicants say is the respondents' wrongful conduct, a fair-minded lay observer might still think I would favour the applicants' interests over those of the Centro group. That is to say, they submitted that even though I had divested myself of my pecuniary interest in both the party CNP and in the cause itself (ie in the outcome of the litigation), the reasonable apprehension of bias (said to stem from the grudge I might be perceived to bear due to the now forever unrecoverable $19,000 loss) was real and could not be cured.
14 Frankly, I had not given these contentions much weight when they were first raised in Freehills' letter. One reason was that the "victim" point had not occurred to the members of the Full Court in Johnson Tiles, two of whom were, prior to amendments to the pleadings, in positions similar to the members of the represented group. Second, having read the United States cases, I was of the view that the contentions had no merit where curative divestiture is made. Third, the cases seemed to say that the lay observer (Mr Uren QC referred to him as "Joe Sixpack") must regard the chance of the judge bearing a grudge (which is the way Dr Pannam QC put it) as being fairly high - that is, there must be a high probability of favouritism inconsistent with the fair performance of the judge's duties: R v Australian Stevedoring (1953) 88 CLR 100, 116. As the High Court said in that case (at 116), the apprehended bias must be "real". See also Re Polites (1991) 173 CLR 78, 86. At the time, I did not see how anyone might reasonably ascribe a high probability that I would show favouritism based on the facts I have set out.
15 However, by the time counsel had concluded their submissions, I confess to having doubt about the matter. What had earlier seemed quite clear was becoming murky. My view was that the lay observer would appreciate that, having regard to the experience and qualification of judges, as well as the high standards of independence to which they must conform, judges would be able to decide cases fairly even if they were "victims" of the actions the subject of the complaint. No doubt the judges in Johnson Tiles were of that view. But it is, I suppose, possible that the lay observer would see things differently. The problem, I think, is that it is sometimes difficult for a judge to know what a lay observer might think about a particular issue. Moreover, a closer reading of the High Court cases suggested that the bias test is much stricter - that is, it is not necessary to establish a high probability of bias to demonstrate that a reasonable apprehension is real, but instead some lower probability will suffice - in cases involving allegations of interest as opposed to, say, allegations of prejudgment: see eg Australian Stevedoring at 116 (noting that cases involving allegations of "bias through interest" are excepted from the "high probability" formulation of the test).
16 In the face of these considerations, I decided in the end that I should recuse myself. I consider that in a case such as the present, where there is real doubt about the matter, the better course is to be cautious: see Ebner at [20] (counselling that "[i]n a case of real doubt, it will often be prudent for a judge to decide not to sit in order to avoid the inconvenience that could result if an appellate court were to take a different view on the matter of disqualification"). As suggested in Ebner, I should as a first instance judge consider (and have considered) the possibility that the respondents would have appealed a refusal to recuse. I should make it clear that I am not at all troubled by the possibility that an appellate court might have taken a different view had I refused to recuse myself. What does concern me, however, is the delay an appeal would have caused, as well as the additional cost the parties would have incurred in what is already an expensive piece of litigation.
17 What this highlights is several problems. First, while the test for recusal laid down by the High Court is stated in quite clear terms, its application to particular facts is far from clear. Some refinement is necessary, perhaps even by statute. If the United States model were adopted at least everyone will know where they stand: but see Ebner at [156] (per Kirby J) (counselling against adoption of a statutory, US-style approach to recusal). Second, the problem with murky cases is that they give parties the opportunity to go judge-shopping. It is because of this possibility, among others, that on several occasions the High Court has said that a judge should not lightly recuse himself: see eg Ebner at [20] (stating that "if the mere making of an insubstantial objection were sufficient to lead a judge to decline to hear or decide a case, the system would soon reach a stage where, for practical purposes, individual parties could influence the composition of the bench" and "[t]hat would be intolerable").
18 Handing the cases to another judge, as I will now do, takes away the opportunity for me to deal with some quite interesting procedural issues that arise in these class actions. Still, there is one interesting procedural matter that I can address before concluding. I said at the outset that I have before me several motions for my recusal. There is authority for the view that a judge ordinarily "must not" entertain formally a motion for disqualification on the grounds of bias: see eg Idoport Pty Limited v National Australia Bank Ltd [2004] NSWSC 270 at [8]-[9] (citing, among other cases, Barton v Walker (1979) 2 NSWLR 740, 748; Australian National Industries Ltd v Spedley Securities Ltd (in liq) (1992) 26 NSWLR 411, 436; and Rajski v Wood (1989) 18 NSWLR 512, 518). This line of authority (the most recent of which is Lee v Cha [2008] NSWCA 13) has been taken to support the view that although a judge may publish reasons for refusing to recuse, no independent appeal will lie nor is any prerogative writ available; rather, a party must wait until a subsequent appealable order is made, at which time the failure to recuse can be raised as a point of error underlying that subsequent order.
19 Whether the Barton line of cases is binding on me is a question I shall come to in a moment. I should start, however, by saying that the reasoning in those cases is not without its problems, notwithstanding that I was a member of a full bench of the Federal Court which approved Barton: see Parramatta Design & Developments Pty Ltd v Concrete Pty Ltd (2005) 144 FCR 244 at [36]-[37], reversed on other grounds (2006) 229 CLR 577. For one thing, the reasoning is circular - the primary reason given for there being no order to appeal from is because the party seeking disqualification is not allowed to bring a motion. To say, as was said in Idoport (at [8] quoting Barton), that formal application for disqualification must not be entertained because it "is beyond doubt that a judge of any court who might reasonably be suspected of bias should not hear the cause" is not a justification for the rule but rather a restatement of the problem. The question, as it is here, is always whether and when the suspicion of bias is reasonable, not what should be done if it is. In my view, the public interest in seeing justice done openly is best served by allowing (but not always mandating) recusal applications to be brought in public by motion and whatever process is required to see that the allegations are fully ventilated.
20 The solution offered in Idoport to allow an indirect appeal against a refusal to recuse by having the judge simply make whatever order might be convenient so that the parties may take an appeal from that order is not attractive. Once the concession is made (as numerous courts have, see Idoport at [9]; Gas & Fuel Corporation Superannuation Fund v Saunders (1994) 52 FCR 48; Parramatta 144 FCR 244 at [37]) that there will be circumstances where an interlocutory appeal based on refusal to recuse is desirable, why should the court be forced to make a pretextual order in order to facilitate appellate review? It is much cleaner simply to allow an appeal from what everyone knows is the actual decision being challenged. This would also make it unnecessary for a party to apply to the High Court for prohibition.
21 It seems, in any event, that Barton is of doubtful authority. In Brooks v The Upjohn Company (1998) 85 FCR 469, 476-77, the Full Federal Court gave various grounds for abandoning the rule in Barton. See also Bahonko v Nurses Board of Victoria (No 2) [2007] FCA 351 at [41] (citing authorities in support of the view that it is permissible to seek disqualification of a judge by notice of motion); Witness v Marsden (2000) 49 NSWLR 429 at [96] (acknowledging the disagreement and suggesting that Barton might require reconsideration); but see Lee v Cha [2008] NSWCA 13 at [23] (discussing the contradictory authorities, reaffirming Barton and rejecting Brooks,albeit in dictum). I should also point out that in Parramatta the Full Court cited Barton but neither acknowledged nor addressed the tension between that case and Brooks. Most likely this was because, to the extent that the Court in Parramatta appeared to approve Barton, such approval was peripheral and unnecessary to the relevant holding, which was that a trial judge's earlier decision in relation to a recusal request does not foreclose him or her from revisiting that decision in light of fresh material: Parramatta at [38]. This proposition is both correct and unremarkable - as a general matter, any interlocutory order may be the subject of reconsideration by the trial court prior to the making of final orders.
22 It is difficult to know what a trial judge should do when the Full Court of his own jurisdiction (Brooks) has cast doubt in dictum on the authority of another Full Court's judgment outside his jurisdiction (Barton) and yet a third Full Court of his jurisdiction in dictum (Parramatta) preferred Barton.
23 Fortunately, while I have entertained the motions I do not think there is a need for any order to be made. It would be odd (if not legally unsound), as Mr Uren pointed out, to direct myself not to hear the cases. On the other hand, I see nothing that would prevent me from circumventing that problem by simply framing my recusal in terms of an order granting the respondents' motions. In the end, however, all I will do is make a direction for the cases to be allocated to another judge.
I certify that the preceding twenty-three (23) numbered paragraphs are a true copy of the Reasons for Judgment herein of the Honourable Justice Finkelstein.