AustLII Home | Databases | WorldLII | Search | Feedback

Melbourne University Law Review

Melbourne Law School
You are here:  AustLII >> Databases >> Melbourne University Law Review >> 2020 >> [2020] MelbULawRw 34

Database Search | Name Search | Recent Articles | Noteup | LawCite | Author Info | Download | Help

Groves, Matthew --- "Clarity and Complexity in the Bias Rule" [2020] MelbULawRw 34; (2020) 44(2) Melbourne University Law Review 565


CLARITY AND COMPLEXITY IN THE BIAS RULE

Matthew Groves[1]*

An apprehension of bias will arise if courts find a fair-minded and reasonably informed observer might consider that a decision-maker might not approach their task with a sufficient level of impartiality. This twofold test to assess bias claims through the eyes of a hypothetical observer reflects the combined effect of Webb v The Queen [1994] HCA 30; (1994) 181 CLR 41 (‘Webb’) and Ebner v Official Trustee in Bankruptcy [2000] HCA 63; (2000) 205 CLR 337 (‘Ebner’). This article examines issues that were not resolved by the apparently clear approach created by Webb and Ebner. The article argues that notions of impartiality are central to the bias rule but not capable of easy definition. It also argues that judicial assessment of bias claims through the eyes of an objective observer is necessarily limited. The article also analyses the recent decision of CNY17 v Minister for Immigration and Border Protection [2019] HCA 50; (2019) 375 ALR 47, to illustrate the difficulties in assessing the impact of irrelevant, prejudicial information in claims of bias.

Contents

I Introduction

The rule against bias is one of the two constituent parts of natural justice. The other is the hearing rule.[1] The hearing and bias rules are complementary ones that require ‘fairness and detachment’ in government decision-making.[2] The rule against bias requires that judges and other public officials be impartial in their decision-making. The rule has been described to be of ‘almost universal application’,[3] which reflects the small number of officials or decisions that appear to be outside its scope.[4] The bias rule has a constitutional dimension in its application to courts and judges,[5] but it has long extended to other forms of government decision-making. The rule applies to jurors,[6] tribunal members,[7] and administrative officials.[8] It also applies to elected officials, such as government ministers and local councillors, in their capacity as public law decision-makers.[9] The operation of the bias rule beyond the courts should ‘recognise and accommodate’ the many distinctions between judicial and other forms of decision-making.[10] There is no simple binary distinction between judicial and non-judicial decision-making for these purposes, but instead a flexibility that enables the rule to adapt to the nature of the decision-maker and the circumstances of a case.[11]

Two bias cases have been particularly influential. The first was

Webb v The Queen (‘Webb’), which replaced subjective judicial assessment of bias claims with an objective test. Webb required that bias claims be determined by the conclusions that a fair-minded and reasonably informed observer might reach.[12] The second key bias case was Ebner v Official Trustee in Bankruptcy (‘Ebner’), in which the High Court cast aside the longstanding rule of automatic disqualification for pecuniary interest and adopted a single test for all claims of apprehended bias.[13] That test requires claimants to first identify the source of the claimed bias and then articulate how that source might have the effect claimed. Ebner’s twofold test is viewed through the lens of Webb’s objective and informed observer.[14] Bias claims succeed if a court finds that the observer might reasonably apprehend that the decision-maker might not be suitably impartial because of the claimed source of bias.

These key aspects of Webb and Ebner have continued to guide and divide the High Court. They have guided the Court because the central principles of both cases have never been seriously questioned. At the same time, however, judicial divisions of opinion continue within the High Court and between that Court and intermediate appellate courts about the principles arising from Webb and Ebner.[15] Why is that so? This article will argue that the clarity introduced by Webb and Ebner on some aspects of the bias rule diverted attention from those parts of the rule that remained uncertain. Those uncertainties mean that judges may agree about matters of principle yet easily disagree about their application. The article analyses the High Court’s recent decision in

CNY17 v Minister for Immigration and Border Protection (‘CNY17’) to explain how many of the guiding principles of the bias rule continue to divide courts. The article suggests that this potential level of difference in judicial assessments of bias claims is inevitable.

II Webb and Ebner: Objective Assessment of Bias Claims and Public Confidence

The bias rule has a long history in the common law.[16] A cornerstone of the modern law governing bias is the statement by Lord Hewart CJ that it ‘is of fundamental importance that justice should not only be done, but should manifestly and undoubtedly be seen to be done’.[17] The notion that justice be ‘seen’ to be done, or that the ‘appearance’ of bias be studiously avoided, draws attention to the importance of public perception to the bias rule. That perception has long been linked to confidence. The bias rule has long been explained as ensuring public confidence in the administration of justice.[18] The bias rule long measured this question, which is tethered to public confidence, by reference to the subjective views of presiding judges. The House of Lords explained that any assessment of public perceptions, in the guise of an objective observer, were redundant because the court itself could act as the ‘reasonable person’ but with the added advantage of possessing knowledge about the case at hand, and legal issues more generally, which members of the public would not hold.[19]

Chief Justice Mason and McHugh J noted in Webb that this approach assumed that public confidence in the ‘administration of justice will be maintained because the public will accept the conclusions of the judge’.[20] Their Honours thought it was ‘more likely’ that public confidence in the administration of justice would be maintained by a test ‘that reflects the reaction of the ordinary reasonable member of the public to the irregularity in question’.[21] Judicial use of a fictional member of the public is obviously an artificial process,[22] but it at least ensures that the judgments made in bias cases are no longer ‘based purely upon the assessment by some judges of the capacity or performance of their colleagues’.[23] Chief Justice Mason and McHugh J made clear that a test focused on the view of a reasonable member of the public, who was reasonably informed of the circumstances, ‘does not mean that the trial judge’s opinions and findings are irrelevant’.[24] They explained that ‘[t]he fair-minded and informed observer would place great weight on the judge’s view of the facts. Indeed, in many cases the fair-minded observer would be bound to evaluate the incident in terms of the judge’s findings.’[25]

This acceptance that judicial opinions would remain very influential, sometimes even binding, upon the informed observer invites questions about the extent of change caused by use of the informed observer. Chief Justice Mason and McHugh J also accepted that use of the informed observer might not necessarily narrow the scope of disagreement in bias cases. That possibility arose because any judgment about a reasonable apprehension, which is the standard by which the informed observer determines bias claims, ‘allows a margin for error in evaluating the facts as elicited’.[26] That margin of error, or difference of opinion, is inherent in the notion of reasonableness.

The role of the informed observer was strongly endorsed in Ebner, where the High Court set aside the longstanding rule of automatic disqualification for pecuniary interest in favour of a single test for all claims of apprehended bias.[27] This refashioned rule involved two steps applicable to all the different interests, influences and forms of conduct that could give rise to a claim of bias.[28] The first step required identification of the source of the claimed bias.[29] The second step required ‘articulation of the logical connection between the matter and the feared deviation from the course of deciding the case on its merits’.[30] This explanation of how bias might arise was important because ‘[o]nly then can the reasonableness of the asserted apprehension of bias be assessed’.[31] Whether that point provides a distinct third step is considered later in this article.

The High Court made clear that the relevant issue was not the nature of any interest or influence, but its possible effect. The assessment of this issue was best done through a single ‘general principle rather than a set of bright line rules which seek to distinguish between the indistinguishable’,[32] as was previously done by a separate rule of automatic disqualification for pecuniary interest. Bright conceptual lines may be blunt instruments but they are at least easy to see. Ebner’s rule was instead a single one, entirely dependent on the particular circumstances of each case. The bright lines of automatic disqualification were replaced with the haze of context. Ebner also made clear that its unified approach was of universal application. The two-step process would apply to all decision-makers but its application outside the courts ‘must sometimes recognise and accommodate differences between court proceedings and other kinds of decision-making’.[33] The High Court did not explain how, or to what extent, principles fashioned for courts and judges should vary in other decision-making environments, though it added the important rider that ‘few administrative decision-makers would enjoy the degree of independence and security of tenure which judges have’.[34] The important question left open by those remarks is how the lesser level of independence of an official might affect judgments about bias. It is argued later in this article that the High Court has not yet confronted this issue.

Like most bias cases, Ebner was about apprehended rather than actual bias. The distinction carries important consequences. Actual bias requires a finding that the decision-maker was ‘so committed to a conclusion already formed as to be incapable of alteration, whatever evidence or arguments may be presented’.[35] Any such conclusion is antithetical to fair and impartial decision-making, particularly in the exercise of judicial power. Accordingly, a finding of actual bias is a ‘grave matter’, which courts should ‘not lightly’ reach.[36] Claims of actual bias must also be ‘distinctly made and clearly proved’.[37] A further difficulty associated with actual bias arises from the very particular nature of its test, which is a subjective one about the actual state of mind about the actual decision-maker.[38] The English Court of Appeal explained that gathering ‘proof of actual bias is very difficult, because the law does not countenance the questioning of a judge about extraneous influences affecting [their] mind’.[39] These various difficulties associated with actual bias led Callinan J to suggest that some claims of apprehended bias ‘may be no more than a polite fiction for no doubt unintended, unconscious and ultimately unprovable, but nonetheless actual bias’.[40] There are occasional calls for courts to cast such fictions aside and be more ready to make findings of actual bias,[41] but for practical reasons judges remain reluctant to make such findings.[42]

Claims of apprehended bias involve a subtle but important shift in focus because existence of the apprehension is determined according to the assessment of the hypothetical fair-minded and reasonably informed observer. The subjective focus of actual bias, with its requirement of a fairly certain prediction about the reasoning or conduct of a particular decision-maker, is replaced with a less demanding assessment of the view that the hypothetical observer might reach. Justice Hayne explained that restricting focus to only a reasonable

apprehension

avoids any need for a court ... to attempt some analysis of the likely or actual thought processes of the decision-maker. It objectifies what otherwise would be a wholly subjective inquiry and it poses the relevant question in a way that avoids having to predict what probably will be done, or to identify what probably was done, by the decision-maker in reaching the decision in question.[43]

Later parts of this article will show that this avoidance of the actual thought processes of individual decision-makers is an understandable but imperfect solution because the judgment of the objective, fictional observer gives rise to

new difficulties.

III Just How Much Impartiality Does the Bias Rule Require?

The impartiality required in judicial and other forms of decision-making that are governed by the bias rule has been described as ‘the absence of a predisposition to favour the interests of either side in the dispute’.[44] Justice Scalia suggested three possible forms of judicial impartiality. The first was the ‘root meaning’ of impartiality, which was a ‘lack of bias for or against either party’ in a case.[45] The second was a ‘lack of preconception in favour of or against a particular legal view’.[46] The third was ‘open-mindedness’, which meant a judge was ‘willing to consider views that oppose [their] preconceptions, and remain open to persuasion, when the issues arise in a pending case’.[47] Justice Scalia thought the second form of impartiality was not possible or desirable and saw little value in courts pretending otherwise.[48] The real vice in this context is impartiality without good cause. Justice Scalia explained in another case that improper bias arises when a decision-maker held a ‘favourable or unfavourable disposition or opinion that is somehow wrongful or inappropriate, either because it is undeserved, or because it rests upon knowledge that the subject ought not to possess’.[49] Such reasoning anticipates the possibility that prejudice may sometimes be justified. A person convicted of many violent murders, for example, can fairly be regarded as a dangerous person of bad character. But even that conclusion could be inappropriate if the question at hand is whether an inquiry should be conducted into the convictions of that person. Such an inquiry could not be fair or impartial if conducted by a person who proceeded on the assumption that the outcome was necessarily governed by the fact that the convicted person was dangerous and evil because that belief would prejudge the very

issues to be investigated.

More mundane instances of predispositions that are appropriate, arguably even desirable, arise from our reliance on precedent. The doctrine of precedent, although operating at the level of law rather than fact, nonetheless requires a level of prejudgment that could theoretically offend the bias rule. However, Hayne J explained:

The principles about apprehension of bias must be understood in the context of a judicial system founded in precedent and directed to establishing, and maintaining, consistency of judicial decision so that like cases are treated alike and principles of law are applied uniformly. The bare fact that a judicial officer has earlier expressed an opinion on questions of law will therefore seldom, if ever, warrant a conclusion of appearance of bias, no matter how important that opinion may have been to the disposition of the past case or how important it may be to the outcome of the instant case. Fidelity to precedent and consistency may make it very likely that the same opinion about a question of law will be expressed in both cases. But that stops well short of saying that the judicial officer will not listen to and properly consider arguments against the earlier holding.[50]

A requirement of impartiality, however expressed, does not mean decision-makers should be devoid of preconceptions, attitudes or beliefs. A century ago, Benjamin Cardozo questioned whether this was ever possible when he conceded that ‘[w]e may try to see things as objectively as we please. None the less, we can never see them with any eyes except our own’.[51] While we can easily accept in principle that judges and other decision-makers are inevitably influenced by their own perspective, even when attempting to apply an objective standard of some sort, the legitimacy of personal judicial experience can easily become contentious. Justice Cory of the Supreme Court of Canada suggested that, even if judges could achieve the herculean task of casting aside their life experience, any ‘judge free of this heritage of past experience would probably lack the very qualities of humanity required of a judge’.[52] His Honour continued: ‘True impartiality does not require that the judge have no sympathies or opinions; it requires that the judge nevertheless be free to entertain and act upon different points of view with an open mind.’[53]

This acceptance that judges are not empty vessels, devoid of life experience, has a loose parallel with the decline of the declaratory theory of law. No judge or scholar now seriously claims that courts do no more than declare, or somehow uncover, law that is waiting to be found.[54] It is probably fair to suggest that most jurists similarly agree that traditional notions of judges as neutral ciphers of the law are another fiction now consigned to history.[55] This realism has given rise to new questions about whether and why judges should acknowledge they cannot be entirely neutral. Justices L’Heureux-Dubé and McLachlin appeared to believe that the solution, or perhaps the most intellectually honest step, was for judges to draw upon their own experiences and seek to identify and correct social injustice and inequality.[56] This ‘contextual judging’ conceives many attitudes and predispositions that could be deemed improper prejudices as useful influences upon decision-making.[57] One might question whether judges should always seek to identify and minimise their unconscious bias. One American study, for example, found that judges who had one or more daughters were more likely to be mindful and supportive of feminist issues in their decision-making than judges whose children were only male.[58] Such conduct could simply be described as empathy and thus an understandable human response, but the obvious difficulty with any empathy shown to one party or cause is that it is necessarily not shown to the other.[59]

Some have rejected contextual judging as an approach that would simply replace old fallacies about judicial impartiality with new ones.[60] Others suggest that an awareness by judges of their own predispositions and other forms of unconscious bias makes them much more able to control the effect of those possibly improper influences.[61] That suggestion is contrary to an enormous body of research which has found that the faith of judges in their own ability to recognise and control their personal biases is somewhat misplaced.[62] One detailed American study, for example, found that judges struggle to cast aside their emotional reactions to parties and witnesses.[63] Another American study found that judges are less able to cast aside subconscious racial prejudices than perhaps judges and lawyers might think.[64]

One American scholar concluded that such studies make clear that ‘[j]udges are subject to the same cognitive realities of human thought that sustain and plague all of us’.[65] If so, many theories about bias in decision-making that compel this conclusion also pose particular difficulty for judges. One is confirmation bias, which refers to the tendency of people to unwittingly seek out or interpret information that reinforces an existing belief they hold or conclusion they favour.[66] This theory would suggest that the ability of judges and scholars, including this author, who accept that the soundness of key elements of the bias rule may struggle to truly consider fundamental doctrinal change. Another relevant theory is the ‘anchoring effect’, which is the idea that decisions made about uncertain issues are greatly influenced by any initial information that is typically or easily reduced in our thinking to numerical terms (including estimates and assessments of probability).[67] This occurs because decision-makers are ‘anchored’ to, and influenced by, the initial value gleaned from information even if it is not relevant or admissible.[68] Numbers and estimates of different sorts abound in legal proceedings and many studies have shown that judges can easily fall prey to the anchoring effect, such as when faced with sentencing guidelines,[69] or decisions about awards of damages.[70] If individual judges are prone to anchoring in a variety of situations that seem relevant to their professional tasks, there is no reason to assume judges in their appellate role would be immune to the anchoring effect as it applies in groups of decision-makers.[71]

This vast body of research in behavioural psychology appears to be gaining greater attention among jurists, though not necessarily great support.[72] The solutions that have been suggested recently include greater research into scientific issues surrounding the impact of bias and cognition upon decision-making.[73] Another recent analysis suggested any such research should be accompanied by a partial codification of governing principles, which would introduce some level of structure.[74] Lord Sumption was more dismissive when he examined the large number of studies about the effect of diversity on judicial decision-making.[75] He thought that most of this research was ‘unenlightening’, often ‘rather crude’, and typically made ‘little allowance, for the possibility that the outcome, however classified, may actually be attributable to the facts of individual cases or the state of the law’.[76]

Whatever the value or limits of such research, its use in bias claims raises several difficulties. An important one is the fictional nature of the informed observer. Empirical research on cognition, predictive behaviour or aspects of decision-making typically focuses on real people rather than fictional ones.[77] It is not clear that scientific research about specific social groups can greatly illuminate bias claims, which require an abstract judicial assessment of the conclusions of a fictional person.[78] If the many calls for the greater use of empirical scientific evidence about bias in legal decision-making were somehow adopted, it is possible that they could only provide significant assistance in bias claims if the legal preference for a fictional observer was discarded in favour of a class of real rather than hypothetical decision-makers (over whom research had

been conducted).

IV Impartiality and the Nature and Structure of an Administrative Process

Application of the bias rule to decision-makers outside the courts must reflect ‘the nature of the decision and its statutory context, what is involved in making the decision and the identity of the decision-maker’.[79] This wider application of the bias rule requires unpicking of the requirement that decision-makers be independent and impartial.[80] Independence and impartiality are often blurred but they are not synonymous.[81] Baroness Hale explained their distinct nature in Gillies v Secretary of State for Work and Pensions, where questions arose about the impartiality of a disability support tribunal.[82] The tribunal contained medically qualified members who provided reports to the tribunal in some cases (while acting as private medical advisers) and adjudicated over cases involving the reports from other medical experts (while acting as tribunal members). The House of Lords rejected arguments that a reasonable apprehension of bias could arise because doctors who provided medical reports would not be sufficiently impartial when considering the reports of other medical experts.[83] Baroness Hale explained that any judgments of medically qualified tribunal members would be enhanced by their expertise, rather than hampered by an imagined professional solidarity with other medical experts, enabling them to make an ‘informed but dispassionate judgment’.[84] According to this view, the personal independence of tribunal members would overcome any looser professional ties they might have.

Baroness Hale explained the distinction between the independence of the tribunal and the impartiality of its members in the following terms:

Impartiality is the tribunal’s approach to deciding the cases before it. Independence is the structural or institutional framework which secures this impartiality, not only in the minds of the tribunal members but also in the perception of the public.[85]

This distinction between the position of the tribunal and the disposition of its members was considered by the High Court in CNY17. That case involved the review of a visa decision, conducted by the Immigration Assessment Authority (‘IAA’). Individual officials who acted for the IAA were appointed as independent statutory officers, but their statutory functions are closely intermingled with those of the departmental Secretary. Those intermingled functions included one requiring the Secretary to provide information they considered relevant to the IAA, which the IAA was then bound to consider. Justices Nettle and Gordon reasoned that these interconnected duties made it difficult to accept that the IAA ‘could or would’ put aside irrelevant material provided by the Secretary.[86] That reasoning recognises that the ability of officials to act freely may be constrained by statutory duties that do not directly compromise their independence or impartiality. The approach of Kiefel CJ and Gageler J drew a distinction between independence and impartiality similar to that of Baroness Hale. Their Honours explained:

The requisite independence is decisional independence, most importantly from influence by the Secretary or the Minister. The requisite impartiality is objectivity in the findings of facts, in the exercise of procedural discretions, and in the application of the ... legislated criteria for the grant or refusal of a protection visa.[87]

What might constitute ‘decisional independence’ was not explained, but Kiefel CJ and Gageler J gave some insight when they distinguished the proper and improper communications that the Secretary could make to the IAA. Their Honours thought the informed observer would be reluctant to accept the IAA was subconsciously influenced by irrelevant material if it had been provided in accordance with statutory requirements.[88] Just as the use of proper channels tended against the possibility of improper influence, any use of improper channels would have the opposite effect. Chief Justice Kiefel and Gageler J noted such conduct was normally prohibited in the courts but suggested it would be viewed even more seriously in the framework to review fast track visa

applications because of

the marked discrepancy in hierarchical position between the Secretary, on the one hand, and a Reviewer engaged under the Public Service Act, on the other. That discrepancy would make any communication from the Secretary to the Authority that might smack of instruction, advice or opinion concerning the conduct or outcome of a review a matter of grave concern.[89]

This reasoning draws useful attention to the different levels of power within administrative environments. Whatever independence may be accorded to those acting for the IAA, the singular and powerful position of a departmental secretary could have an impact. It is also arguable that the existence of improper communications from such a senior official to an independent but arguably low-level decision-maker would have an impact on the informed observer.

The reasoning of Kiefel CJ and Gageler J does little to address the wider institutional pressures that can be brought to bear on decision-makers. Institutional pressures are often diffuse and so powerful that they can operate without any form of direct communication. The power of reappointment is an obvious form of this power. Most tribunal members, including those in the IAA who are part of the Administrative Appeals Tribunal (‘AAT’), are appointed for a limited term. They are also eligible for reappointment. Concerns have long been expressed about the perceived political nature of appointments and reappointments to the AAT,[90] including the possibility that members whose decisions displease the government of the day are tacitly punished by not gaining reappointment.[91] The focus of the bias rule upon the specific context of individual cases means it cannot address these institutional forms of influence.[92] It must be conceded, however, that the difficulties presented by the institutional influences that may come to bear on decision-makers has long been problematic to other administrative law principles.[93]

V The Fair-Minded and Informed Observer

The adoption of the fair-minded and informed observer standard marked an important step in the bias rule because it signalled a move away from open reliance upon subjective judicial assessment of bias claims, in favour of judicial assessment of the likely assessment that the wider public might make of bias claims. The observer has slowly been given flesh of sorts over many cases but remains an idealised creature. The observer is attributed with general personality traits, though not key personal characteristics such as a particular gender, age, or religious or political affiliation.[94] Justice Kirby has suggested the observer is a ‘reasonable member of the public’ who is ‘neither complacent nor unduly sensitive or suspicious’.[95] But neither, Lord Mance later added, is the observer ‘unduly compliant or naïve’.[96] Any judgment made by the observer is, therefore, a considered one rather than a snap judgment,[97] which will not be made ‘until both sides of any argument are apparent’.[98] The observer will listen and think, though perhaps not for as long as the parties may wish. A judge can deal more quickly with claims or points that are clearly weak, but should nonetheless be prepared to listen to even apparently weak arguments with an

open mind.[99]

The device of the informed observer becomes more difficult to apply as the courts attribute relevant knowledge from the case or decision at hand. When stated in general terms, the attribution of contextual knowledge appears sensible. For example, the observer will place all relevant knowledge about a matter within ‘its overall social, political or geographical context’.[100] But this is done from the outside. The reason the observer cannot be an ‘insider’ of the relevant court, tribunal or public agency, Baroness Hale explained, is that an inside observer ‘would run the risk of having the insider’s blindness to the faults that outsiders can so easily see’.[101] The level of knowledge attributed from this position is limited, which means the observer is adequately rather than fully informed.[102] That limitation on the level of knowledge to be attributed lessens the possibility that the observer would also take on the actual knowledge, and presumably also the perspective, of the court or one of the parties.[103]

The greatest difficulty with this contextual knowledge is how it should take account of the observer’s knowledge of the law.[104] The observer is neither a judge nor a lawyer but will know enough to make judgments ‘in the context of ordinary judicial practice’.[105] That context includes the different nature of different processes and enables the observer to understand the procedural variations that occur outside the courts, such as the investigative and other practices often employed in coronial inquiries,[106] or the more interventionist conduct that tribunal members may use in hearings that do not operate according to more adversarial practices.[107] In addition, the observer ‘is taken to understand the dynamics of modern judicial practice’,[108] such as the nature and consequences of the docket system employed by many courts,[109] or the increased willingness of judges to make comments and sometimes actively deal with the parties and issues and not maintain a sphinx-like silence until judgment

is delivered.[110]

This contextual knowledge becomes more difficult to accept when it includes issues that might best be described as ones particular to legal culture or experience. Justice of Appeal Basten conceded, for example, that ‘[t]he idea that any fair-minded lay observer would have sat through six weeks of the trial may be an oxymoron’,[111] yet his Honour felt he had little choice but to attribute the observer with an understanding of all the main points from that lengthy hearing.[112] Such a wearying experience can be useful because it may, for example, allow the observer to sensibly accept that the terse or intemperate remarks of judges can occur during lengthy or difficult hearings without necessarily creating an apprehension of bias.[113] But it can also lead courts to attribute a sympathy on the part of the observer to supposed legal norms that most reasonable members of the public would surely reject. A striking example is

Taylor v Lawrence (‘Taylor’), where the English Court of Appeal held that the observer would not apprehend a possibility of bias after a judge had received free legal advice from a firm appearing before him while that trial was on foot.[114] The wholly unconvincing reason for the conclusion that the observer would be untroubled by such trifles as free legal advice was that the observer would accept the long history of close social and business relations between the bar and bench as enhancing, rather than damaging, judicial independence.[115] At a superficial level, the attribution of knowledge of procedural minutiae or professional habits is difficult to square with the caution that the observer’s views should ‘not be as informed as the perception of a lawyer, particularly a litigation lawyer’.[116] The deeper problem is that such reasoning provides a flimsy veil for judges to transpose their own habits and beliefs onto that of the observer.[117] As the Taylor case demonstrates, judges who draw that veil ascribe to the informed observer both knowledge and acceptance of practices that any sensible observer would be greatly troubled by and almost certainly reject.

These problems have led to many criticisms of the informed observer. The observer is so weighed down with these many virtues, so saintly in character, that some have suggested that that person is now quite removed from the general public they are intended to represent.[118] Professor Abimbola Olowofoyeku has argued the ‘complexities of the factual situations or the “grey areas known to the observer have grown ‘in order for the courts to be able to arrive at the right outcome’.[119] He concluded that courts should discard the fiction of the reasonable observer because ‘despite the pitch on objectivity and the view that the apprehensions of bias must have an objective basis, it is the opinion of the reviewing court on this issue that matters’.[120] Chief Justice French acknowledged Olowofoyeku’s criticisms in British American Tobacco Australia

Services Ltd v Laurie, but reasoned that the device of the observer should be retained because it provided judges with a timely reminder to view things ‘as best they can, though the eyes of non-judicial observers’ and that continued use of the observer could provide ‘utility as a guide to decision-making in this

difficult area’.[121]

Some of the artifice associated with the informed observer could be lessened by the suggestion of Gageler J that Ebner’s test has a distinct third step,[122] conducted after the party claiming bias has explained how the identified interest may have its claimed effect.[123] This possibility arises from the suggestion in Ebner that consideration of the reasonableness of the claimed apprehension could only occur after its claimed effect has been explained. Justice Gageler treated this consideration as a distinct step in Isbester.[124] Lower courts appear hesitant about this possible further step in the Ebner test,[125] which is perhaps understandable given that the High Court has not yet considered the issue.[126] Aronson, Groves and Weeks argue that addition of this separate step to Ebner’s test ‘may do little more than articulate openly a commonsense judgment that can and should occur as part of the second step of Ebner’.[127] Those authors argue that claims that would fail a distinct further requirement of reasonableness would in any case inevitably fail Ebner’s existing second step.[128] It has also been argued that Gageler J’s suggested new step raises the danger of confirmation bias because it only arises after a finding that a source of bias might have its claimed effect.[129] Judges who have reached that conclusion are hardly likely to then declare that the possibility was not reasonable.[130] This third step could, however, introduce a greater level of transparency into the bias rule. A third step, in which the reasonableness of a claim is assessed in a holistic or overall manner, could enable judges to provide their own assessment of bias claims. That approach might usefully enable judges to step back from a strained use of the informed observer and introduce their own assessment as part of this

possible third step in Ebner.

VI Irrelevant Material and Subconscious Bias

The discussion in earlier parts of this article about the extent to which decision-makers can or should discard their experience and wider knowledge draws attention to the difficulties of subconscious prejudices. The notion of contextual judging seeks to distinguish between those subconscious and undesirable prejudices that many, if not most, officials may carry with the life experiences that may usefully inform decision-making. The distinction here is not one of good or bad prejudices, but rather an acceptance that some attitudes or prejudices may usefully add perspective or insight to decision-making. The different views on contextual judging would almost certainly agree that this debate is concerned with the general attitudes and beliefs of decision-makers that transcend individual cases. Different considerations arise when the source of claimed prejudice arises from the material before the decision-maker rather than any experience or attitude that person may hold. The question in such cases is not simply whether a decision-maker can cast aside the influence of such material, but what subconscious effect such material might have and what the informed observer might make of that possibility.

Some guidance can be taken from the hearing rule, where courts have long grappled with the question of whether irrelevant or prejudicial material should be disclosed. This line of cases proceeds on the assumption that decision-makers may be influenced by information they have decided is irrelevant. The question that follows for the purposes of the hearing rule is whether that material should be disclosed to an affected person. Justice Brennan explained the rationale for disclosure in such cases in Kioa v West (‘Kioa’).[131] His Honour cautioned that decision-makers should not ‘be clogged by inquiries into allegations to which’ they ‘would not give credence, or which are not relevant ... or which are of little significance to the decision’.[132] Justice Brennan thought that a different approach was required for material that was ‘credible, relevant and significant’.[133] Where material of this kind was placed before an official, his

Honour reasoned:

It is not sufficient for the repository of the power to endeavour to shut information of that kind out of [their] mind and to reach a decision without reference to it. Information of that kind creates a real risk of prejudice, albeit subconscious, and it is unfair to deny a person whose interests are likely to be affected by the decision an opportunity to deal with the information.[134]

The solution suggested by Brennan J is the fairly easy one of disclosing the relevant material.[135] Disclosure enables the prejudice that material may create to be countered because an affected person can contradict or explain that material. In some cases, however, difficulty arises from the very receipt of material rather than its non-disclosure. This problem occurs when irrelevant, prejudicial material is received by decision-makers. Such information can create an unfavourable view that cannot always be overcome through disclosure.[136] The question in such cases is whether officials can be presumed to be able to cast aside the effect of such material. The more credible, relevant and significant the material, the harder it becomes for officials to dismiss it from their thoughts. The solution of the hearing rule is disclosure, depending on whether or not material meets the criterion of credible, relevant and significant.[137]

The High Court explained this criterion in Applicant VEAL of

2002 v Minister for Immigration and Multicultural and Indigenous Affairs, where a migration tribunal did not disclose an unfavourable letter about the applicant it had received but instead put the letter out of its consideration.[138] The Court held that that information would be credible, relevant and significant (and thus be required by fairness to be disclosed) if it could not be ‘dismissed from further consideration’ by the decision-maker.[139] The Court explained that

the decision-maker cannot dismiss information from further consideration unless the information is evidently not credible, not relevant, or of little or no significance to the decision that is to be made. References to information that is ‘credible, relevant and significant’ are not to be understood as depending upon whatever characterisation of the information the decision-maker may later have chosen to apply to the information when expressing reasons for the decision that has been reached.[140]

This reasoning confirms several aspects of disclosure under the hearing rule which are relevant to the bias rule. One is that any disclosure requirement does not depend on whether a decision-maker actually took material into account.[141] Another is that, if material cannot be dismissed immediately and must be considered, at least to determine if it has bearing on the decision to be made, fairness would normally require its disclosure.[142] Yet another is that such material can have an unconscious prejudicial effect through the lingering impression it can leave, even when discarded or disavowed.[143]

The bias cases have long proceeded on the assumption that decision-makers will place prejudicial material out of their thoughts. In one of the leading English bias cases, the Court of Appeal explained that ‘[j]ustice is portrayed as blind not because she ignores the facts and circumstances of individual cases but because she shuts her eyes to all considerations extraneous to the

particular case.’[144]

The ability of judges and other decision-makers to cast aside irrelevant information and its possible influence is difficult to gauge. The pragmatic approach of the bias cases has been to accept that the informed observer does not have absolute faith in the ability of justice and its decision-makers to remain blind. An influential statement of that pragmatism was made by Deane J in Webb, where his Honour acknowledged that an apprehension of bias could arise from ‘knowledge of some prejudicial but inadmissible fact or circumstance’.[145] The bias claim in Webb was against a juror and later cases have distinguished jurors and other laypeople with professional decision-makers, with the latter thought to be more able to cast aside irrelevant material and the unconscious prejudicial effect it may have.[146] In one such case, the New South Wales Court of Appeal held that an apprehension of bias was not created by the tender of a statutory declaration that might have been forged by a medical practitioner facing disciplinary proceedings.[147] The Court was influenced by the pointed dismissal of the document by the tribunal, but also by the requirement that the chair of the tribunal be a judicial officer.[148] The Court of Appeal reasoned that the informed observer would know and accept that judicial officers regularly ruled upon the admissibility of the material sought and could be ‘assumed to be able to put out of their minds irrelevant or prejudicial material which is excluded’.[149] The Court also reasoned that other members of the tribunal could be expected to ‘follow the direction’ of judicial members of the tribunal about the rejection of irrelevant evidence from their deliberations.[150] The Court of Appeal placed weight on the experience of the medically qualified member of the tribunal and the tertiary qualifications of the lay member.[151]

The Full Federal Court in Crowley v Holmes reached a similar conclusion about the investigation of a medical practitioner.[152] In that case, a doctor had been investigated for ‘inappropriate practice’, arising from him supposedly overbilling the Medicare scheme.[153] The investigating committee was given material explaining that the doctor had previously been counselled and disqualified. Although the material was prejudicial and irrelevant to the case at hand, the Court held that the informed observer would not hold an apprehension of bias simply because the committee knew those it was investigating had previous contact with regulatory officials.[154] The Court also thought that the observer would accept that courts and tribunals had to reach decisions after ‘weighing evidence, often rejecting or discounting some of it’.[155] Accordingly, there was no reason that the observer would accept that a ‘professional tribunal’ would be ‘unable or unwilling to set aside material which, for one reason or another, is not proper for its consideration’.[156] Justice Madgwick agreed with this reasoning and the wider proposition that ‘doctors are as capable as lawyers, including judges, of putting merely prejudicial matters from their minds’.[157] However, his Honour also noted that psychological research suggested that such professionals ‘are actually not very good at doing that at all’.[158]

A key difficulty in any assessment of subconscious bias is the need to look beyond the statements and actions of a decision-maker, to make judgments about unspoken issues. In Islam v Minister for Immigration and Citizenship (‘Islam’), Finn J suggested that, while the possibility of subconscious bias should be treated with ‘circumspection’, the ‘essential’ concern was to ‘maintain the integrity of ... processes and procedures and to provide public reassurance of that integrity’.[159] The question in that case was whether a tribunal member who had authorised the issue of a warrant to install a listening device, to monitor an applicant for a temporary business visa, was subconsciously influenced by the material provided to seek the warrant and thus precluded from determining a merits review claim against the refusal of the applicant’s visa.[160] Justice Finn

accepted that decision-making could be compromised

whenever a close relationship or association exists between the subject matter of an administrative or judicial decision to be taken by that official, and adverse information obtained (or an adverse opinion formed) relating to that subject matter in the course of, or as a result of, a separate earlier decision taken by the same official for other purposes.[161]

Justice Finn held that allowing a tribunal member, who had authorised the issue of a warrant, to review a subsequent decision that was based in large part on the fruits of that warrant, would enable the member to exercise ‘incompatible functions’ and thus justifiably breach the bias rule.[162]

Islam was an instance where the source of subconscious bias arose from an earlier proceeding,[163] but the same issue can arise during a single hearing or administrative process. The Full Federal Court faced such a problem in Minister for Immigration and Border Protection v AMA16 (‘AMA16’), where the IAA had made a ‘fast track’ assessment of an application for a protection visa.[164] The departmental Secretary was obliged to provide ‘review material’ to the IAA,[165] including any material in the possession or control of the Secretary which they considered relevant to the review.[166] The material forwarded by the Secretary included documents stating that the applicant was charged with several criminal offences. That prejudicial material was not mentioned in the decision of the IAA refusing the claim but the Full Court accepted that an apprehension of bias arose because the material might have had a subconscious influence on the IAA.[167] All members of the Full Court accepted that the informed observer would understand the key features of the legislative scheme, including the duty of the Secretary to provide information considered relevant to the review.[168] Justice Dowsett reasoned that the observer’s judgement on the possible effect of the prejudicial material would be influenced by knowledge that it was considered relevant by the Secretary when transmitted.[169] His Honour suggested that the effect of this statutory imprimatur was decisive and doubted whether knowledge of prejudicial allegations would itself create an apprehension of bias.[170] Justice Charlesworth reasoned that it did not matter that the Secretary was wrong about the relevance of the material and thus ‘had no statutory warrant’ to transmit it.[171] The informed observer would not enter finely grained judgment about possible procedural errors of the Secretary, and would instead take note of the Secretary’s subjective judgment about the material.[172]

Two important issues were left unsettled by the Full Court in AMA16. One was the relevance of the legal qualifications of the IAA official. That point was not examined in detail because the Full Court refused to admit further evidence of those qualifications, but Griffiths J cautioned that evidence of ‘a legal qualification alone’ might not attract the presumptions used by the New South Wales Court of Appeal explained above.[173] Reference to legal qualifications ‘alone’ draws attention to the distinction between judicial decision-makers and legally qualified ones. The informed observer would surely hold different views about each. The relevance of legal qualifications to judgments about bias is especially relevant to large tribunals, such as the AAT and the various state tribunals of general jurisdiction, which contain large numbers of members who are or are not legally qualified.[174] If the informed observer is deemed to hold assumptions about those with legal qualifications, different assumptions might apply to different types of members within a single tribunal. The other issue left unresolved by the Full Court was the effect, if any, of the clear disavowal of irrelevant or prejudicial material by decision-makers. Justice Dowsett held that the observer would be influenced by the failure of the IAA to mention the irrelevant material, or to expressly disclaim reliance upon it.[175] Justice Griffiths, with whom Charlesworth J agreed on this point, held that the silence of the IAA made it unnecessary to decide ‘what significance, if any’, should be given to statements by decision-makers that they have not considered or somehow rejected

irrelevant, prejudicial material.[176]

The High Court faced similar issues in CNY17, where the Secretary also provided review material to the IAA. The information included 48 pages of material that detailed the applicant’s involvement in disturbances while he was held in immigration detention. This material variously described or hinted the applicant was difficult and a troublemaker. A majority of the High Court found that the material created an apprehension of bias and that the informed observer would be mindful of the subconscious effect that the prejudicial material might have had. Justices Nettle and Gordon, and Edelman J, held that the accusations and innuendo contained in the material were clearly capable of creating an apprehension of bias even though the material was clearly irrelevant.[177] Their Honours accepted that an apprehension could clearly arise from the subconscious effect of the material.[178] Justices Nettle and Gordon reasoned that this effect ‘cannot be cured by putting the information aside’.[179] Their Honours thought that the possibility of casting aside information was ‘difficult to reconcile’ with legislative duties that required the Secretary to transmit information determined as relevant and required the IAA to consider whatever had been provided.[180] Their Honours also reasoned that placing the material aside ‘does not overcome the subconscious bias which might result from seeing

the material’.[181]

Justices Nettle and Gordon noted that the IAA had not disavowed the material in its reasons, but their Honours did not consider the effect that such a statement might have had.[182] Justice Edelman, by contrast, placed considerable weight on the ‘deafening silence’ of the reasons of the IAA about what it made of the material.[183] His Honour thought that the sheer quantity and prejudicial nature of the material, as well as the statutory imprimatur by which it was given, meant the informed observer could accept that an apprehension of bias had been created.[184] The emphasis of Edelman J on the related issues of the possible subconscious effect of the material and the silence in the reasons of the IAA on what it thought of the material, suggests that decision-makers can dispel an apprehension of bias if their reasons explain how they sought to guard against the influence of prejudicial material that was wrongly provided to them.[185]

Chief Justice Kiefel and Gageler J accepted that the material was lengthy, irrelevant and often prejudicial but thought it contained nothing ‘so shocking’ that it could ‘give rise to the realistic possibility that knowing of them would play on the subconscious of the Authority to the detriment of the appellant’.[186] Their Honours adopted a slightly different view of the legislative context in which the material was required to be provided and considered, holding that the informed observer might wonder why the material was provided and question the judgment of whoever thought it should be transmitted.[187] Chief Justice Kiefel and Gageler J concluded that the informed observer would accept the IAA could still perform an independent review of the case ‘untainted by the dashes of colour added’ by the irrelevant material.[188] Their Honours conceded that would not always be the case. They held that the informed observer would accept that the IAA ‘can ordinarily be expected’ to discard irrelevant and

prejudicial information, but cautioned that the observer would

recognise that even a professional decision-maker is not ‘a passionless thinking machine’ and that information consciously and conscientiously discarded might still sometimes have a subconscious effect on even the most professional of decision-making.[189]

The differing opinions expressed in CNY17 about the potential effect of irrelevant material highlight the difficulty of deciding when an issue might create an apprehension of bias. Justices Nettle and Gordon noted that difficulty when emphasising that their conclusions depended heavily on the facts of the case, and accepted there could be instances where material was ‘somewhat prejudicial ... but not such as might lead a fair-minded lay observer to apprehend a lack of impartiality’.[190] Their Honours explained that the difficulty presented by such cases was that the ‘particular point at which prejudicial information will lead to apprehended bias cannot be identified in the abstract’.[191] That reasoning makes clear that, where irrelevant information is not at one end of two possible extremes (either highly prejudicial or fairly obviously not), any assessment of its possible effect on the thinking of the informed and fair-minded observer will be an impressionistic one upon which judges can easily differ. CNY17 illustrates that the same point extends to judicial decisions as to whether information is, or is not, highly prejudicial. In simple terms, Kiefel CJ and Gageler J viewed the material and its effect one way. Justices Nettle and Gordon, and Edelman J, viewed it another. The distinctions between these two differing assessments was not one of principle, but rather a differing perception of its possible practical impact. Just as members of the High Court may disagree in their assessment of the nature of material and its possible effect on the subconscious, so may lower court judges and individual decision-makers.

VII Bias and the Materiality of an Error

One final aspect of the bias rule that remains unsettled is when and why an apprehension of bias may (or may not be) regarded as material for the purposes of jurisdictional error. The recent recognition by the High Court of materiality as an element of jurisdictional error is beyond the scope of this article, but the key features of materiality require brief explanation.[192] It was long thought that an error that is jurisdictional in nature would normally provide the basis for relief, subject to various discretionary issues that could provide a reason to refuse relief. The High Court recently recast this approach, by reference to a requirement of materiality.[193] An error will attract relief unless the court is satisfied that it is not material, which in this context means that there was no possibility that the decision would have been different had the error not

been made.[194]

Much about materiality remains unsettled,[195] including its application to claims of bias. Ebner provides strong reason to doubt that it has any such application. The plurality in Ebner stressed ‘the importance of the basic principle, that the tribunal be independent and impartial’.[196] It followed, the plurality reasoned, that ‘even the appearance of departure’ from these standards is prohibited.[197] The High Court also made clear that the key question was not what had or would happen, but what could or might happen.[198] Chief Justice Kiefel and Gageler J explained in CNY17 that this focus of the bias test ‘differs in concept and in nature from the question of whether’ the material wrongly provided in that case created a reasonable apprehension of bias.[199] Their Honours

continued:

The difference is that materiality is a question of counter-factual analysis to be determined by the court as a matter of objective possibility as an aspect of determining whether an identified failure to comply with a statutory condition has resulted in a decision that has in fact been made being a decision that is wanting in statutory authorisation. The question is not one of perception to be determined by the court by reference to the reasonable apprehension of a hypothetical fair-minded lay observer, which apprehension if established is of itself sufficient to result in a want of statutory authorisation.[200]

Justices Nettle and Gordon, and Edelman J, found it unnecessary to decide the possible role, if any, that materiality might play in claims of bias,[201] but the reasoning of Kiefel CJ and Gageler J suggests that materiality has very little role

to play.[202]

There are several further reasons why materiality could be counterproductive if applied to claims of bias. One arises from the connection drawn between the bias rule and public confidence in the courts and the administration of justice. Use of materiality would leave open the possibility that an apprehension of bias could arise and yet be found, in some circumstances, not to be material.[203] The very existence of that possibility can only undermine public confidence in the law.[204] A separate but closely related point arises from the role of the informed observer. What point would the observer serve if an apprehension of bias was identified through the eyes of that person, yet able to be effectively cast aside by a court that found it was not material? Chief Justice Kiefel and

Gageler J may have been mindful of this very concern. A possible role for materiality in bias claims would be especially difficult if the third step in Ebner’s test suggested by Gageler J was accepted. Recognition of a third step and a role for materiality would require four distinct steps to determine bias claims. It might also be difficult to distinguish between the third step suggested by

Gageler J (which requires consideration of the reasonableness of the claimed apprehension of bias) with determination of the materiality of an error (which requires a slightly different objective assessment by the court).[205]

VIII Concluding Observations

Do appearances really matter? The bias rule is anchored on the assumption that they do because of its emphasis on the need that justice be seen to be done, but even this assumption can be challenged. In a speech quoted earlier in this article, Lord Sumption did precisely that when he brushed aside a concern expressed by a fellow member of the United Kingdom Supreme Court.[206] Baroness Hale JSC had suggested during her judgment in a case about the validity of prenuptial agreements that ‘there is a gender dimension to the issue which some may think ill-suited to decision by a court consisting of eight men and one woman’.[207] Baroness Hale JSC was not directly invoking the informed observer but instead posing an obvious question — what might ordinary people think of this situation? The reply of Lord Sumption in his subsequent speech was revealing, though perhaps not intentionally so. He reasoned that the question was quite suited for a court of mostly male judges to decide, largely because this, and many other issues that particularly affected women, had been decided by benches comprised mostly or entirely by men.[208] Lord Sumption’s underlying point seemed to be that the Court in which Baroness Hale JSC raised her query had made the right decision, as superior courts in Britain generally appeared to do so in cases including those involving issues of gender equity, so what was

the problem?

At a superficial level, the problem was one of confirmation bias. Lord Sumption appeared to think that, as the decision in the relevant case was right, the concerns Baroness Hale JSC expressed about the composition of the bench deciding the case must be wrong. The deeper problem was revealed in Lord Sumption’s lengthy assessment of judicial appointments processes.[209] He explained the United Kingdom’s system in useful detail, made many interesting comparisons to other European countries, and identified the reasons he thought relatively fewer women and members of ethnic minority groups received judicial appointments.[210] It followed, according to this analysis, that diversity as an express criterion to encourage more judicial appointments of women and members of ethnic minority groups, should not be adopted. None of Lord Sumption’s lengthy analysis truly confronted, let alone effectively contradicted, Baroness Hale JSC’s simple and compelling point — what would observers think of a court comprised almost entirely of male judges deciding an issue that had particular significance to women? The same question may be posed about use of the informed observer, not simply in CNY17 and Isbester, but the many other cases that have attributed detailed knowledge of acceptance to that person about matters of statutory interpretation, quirky and questionable features of legal culture, or judicial conduct during a lengthy trial. So much of this knowledge seems attributed to the informed observer in order that judges may reach the decisions that they appear to feel are correct. What would the informed observer make of this judicial manipulation of the informed observer? Probably that it is a lot of nonsense that does not become more credible by adding further complications to the Ebner test.

One question that remains is whether the acceptance of Gageler J’s third step in Ebner would complicate the bias rule, or would instead simply acknowledge the practical limits to Webb and the informed and fair-minded observer that that case endorsed. The doubts expressed by judges and commentators about the knowledge and attitudes ascribed to the observer are ones that almost certainly cannot be resolved. Whatever knowledge or qualities the observer is deemed to have, or not have, the judgment of that fictional construct will inevitably be gauged through a judicial lens. That lens is necessarily a subjective one because judges are different from each other. Each judge may perceive things differently and thus reach different conclusions about what the informed observer might think. If we accept that all judges are the sum product of their particular life experiences, this level of difference is not simply inevitable, it

is desirable.


* Alfred Deakin Professor of Law, Deakin Law School, Deakin University. Thanks are due to Greg Weeks and the anonymous referees for helpful comments.

[1] The bias rule has been described in some recent cases as ‘an aspect’ of procedural fairness. See, eg, CNY17 v Minister for Immigration and Border Protection [2019] HCA 50; (2019) 375 ALR 47, 52 [16] (Kiefel CJ and Gageler J) (‘CNY17’); Isbester v Knox City Council [2015] HCA 20; (2015) 255 CLR 135, 146 [23] (Kiefel, Bell, Keane and Nettle JJ) (‘Isbester’).

[2] Plaintiff S157/2002 v Commonwealth [2003] HCA 2; (2003) 211 CLR 476, 490 [25] (Gleeson CJ). The two rules can intersect, such as when excessive judicial intervention is claimed to have caused both unfairness (by precluding a party from adequately presenting its case) and an apprehension of bias (because the interventions are made only to one party): see, eg, Jorgensen v Fair Work Ombudsman [2019] FCAFC 113; (2019) 271 FCR 461, 483–4 [93]–[96] (Greenwood, Reeves and Wigney JJ);

Gambaro v Mobycom Mobile Pty Ltd [2019] FCAFC 144; (2019) 271 FCR 530, 538 [22] (Greenwood and

Rangiah JJ); Dennis v Commonwealth Bank of Australia [2019] FCAFC 231; (2019) 272 FCR 343, 350 [25] (Greenwood, Besanko and Reeves JJ); Serafin v Malkiewicz [2020] UKSC 23; [2020] 1 WLR 2455, 2464 [38] (Lord Wilson).

[3] Matthew Groves, ‘The Rule against Bias’ (2009) 39(2) Hong Kong Law Journal 485, 485, cited with apparent approval in CNY17 (n 1) 61 [55] n 61 (Nettle and Gordon JJ).

[4] One such official may be the Governor-General in the exercise of their functions in extraordinary circumstances. Prior to his appointment to the High Court, Solicitor-General Stephen Gageler SC suggested that there was ‘no foothold in the structure or text of the Constitution’ to support arguments that the Governor-General was constrained by familial or other relationships that might otherwise trigger the bias rule: see Stephen Gageler, In the Matter of the

Governor-General (SG No 33 of 2010, 26 August 2010) 2 [3] <https://old.gg.gov.au/sites/default/files/files/media/2010/SG_Letter_26_8_2010.pdf>, archived at <https://perma.cc/DS43-HLP8>. The large assumptions made in that advice take it beyond the scope of this article.

[5] In Ebner v Official Trustee in Bankruptcy [2000] HCA 63; (2000) 205 CLR 337, 363 [81]–[82] (‘Ebner’),

Gaudron J explained that ch III of the Australian Constitution operates to guarantee both impartiality and the appearance of impartiality in Australia’s judicial system. That reasoning was cited with approval in North Australian Aboriginal Legal Aid Services Inc v Bradley [2004] HCA 31; (2004) 218 CLR 146, 162–3 [27]–[29] (McHugh, Gummow, Kirby, Hayne, Callinan and Heydon JJ).

[6] Webb v The Queen [1994] HCA 30; (1994) 181 CLR 41 (‘Webb’).

[7] See, eg, Re Refugee Review Tribunal; Ex parte H [2001] HCA 28; (2001) 179 ALR 425, 435 [32] (Gleeson CJ, Gaudron and Gummow JJ) (‘Ex parte H’); Re Minister for Immigration and Multicultural Affairs; Ex parte Epeabaka [2001] HCA 23; (2001) 206 CLR 128, 138 [27] (Gleeson CJ, McHugh, Gummow and Hayne JJ).

[8] See, eg, Isbester (n 1) where the rule was applied to local council officers.

[9] The elected status of decision-makers is relevant. The High Court has accepted that application of the bias rule to ministers can take account of their political functions and accountability: Minister for Immigration and Multicultural Affairs v Jia Legeng (2001) 205 CLR 507,

528–9 [61]–[63], 539 [102] (Gleeson CJ and Gummow J), 545 [122]–[124], 555 [155]–[156] (Kirby J), 563–5 [182]–[187] (Hayne J) (‘Jia Legeng’).

[10] Ebner (n 5) 344 [4] (Gleeson CJ, McHugh, Gummow and Hayne JJ).

[11] A contrary view is taken in Simon Young, ‘The Evolution of Bias: Spectrums, Species, and the Weary Lay Observer’ [2017] MelbULawRw 42; (2017) 41(2) Melbourne University Law Review 928. That author describes a flexible or context-sensitive approach to the bias rule as the ‘spectrum approach’, which he argues has begun to decline in Australian law.

[12] Webb (n 6) 51–2 (Mason CJ and McHugh J), 57 (Brennan J), 67–8 (Deane J), 87–8 (Toohey J).

[13] Shortly before that step was taken, the House of Lords affirmed and extended the principle of automatic disqualification: R v Bow Street Metropolitan Stipendiary Magistrate; Ex parte Pinochet Ugarte [No 2] [1999] UKHL 1; [2000] 1 AC 119 (‘Pinochet [No 2]’).

[14] The House of Lords adopted a similar position a few years later: Porter v Magill [2001] UKHL 67; [2002] 2 AC 357, 494 [103] (Lord Hope); Lawal v Northern Spirit Ltd [2003] UKHL 35; [2004] 1 All ER 187, 196 [20]–[21] (Lord Steyn for the Court).

[15] A useful illustration of both occurred in CNY17 (n 1). The High Court overturned the orders of the Full Federal Court and there was sharp division within both the High Court and the Full Court: at 54 [21] (Kiefel CJ and Gageler J), 61–2 [57] (Nettle and Gordon JJ), 76–7 [132]–[136] (Edelman J). See also CNY17 v Minister for Immigration and Border Protection [2018] FCAFC 159; (2018) 264 FCR 87, 101 [66] (Mortimer J), 116–7[134]–[135] (Moshinsky J), 125 [171], 125–6 [174], 126 [177] (Thawley J) (‘CNY17 Full Federal Court’). Divisions between superior courts occurred in Isbester (n 1) and Michael Wilson & Partners Ltd v Nicholls (2011) 244 CLR 427 (‘Michael Wilson & Partners’). In each case the High Court unanimously overturned a unanimous decision of a state appellate court.

[16] In Ebner (n 5) 371 [111], Kirby J said the rule was ‘long held sacred’ since the time of Dr Bonham’s Case [1572] EngR 107; (1610) 8 Co Rep 113b; 77 ER 646. The common law origins of the rule are traced to ‘at least’ the 14th century in Lord Woolf et al, De Smith’s Judicial Review (Sweet & Maxwell, 8th ed, 2018) 537–8 [10.008].

[17] R v Sussex Justices; Ex parte McCarthy [1923] EWHC KB 1; [1924] 1 KB 256, 259. In the same year, Atkin LJ stated that ‘[n]ext to the tribunal being in fact impartial is the importance of its appearing to be so’: Shrager v Basil Dighton Ltd [1924] 1 KB 274, 284.

[18] An early example was Serjeant v Dale [1877] UKLawRpKQB 82; (1877) 2 QBD 558, 567 (Lush J for the Court), which referred to the need to ‘promote the feeling of confidence in the administration of justice which is so essential to social order and security’.

[19] R v Gough [1993] 2 AC 646, 668–70 (Lord Goff).

[20] Webb (n 6) 51.

[21] Ibid.

[22] However, it is one regularly used, such as with the reasonable person on the Clapham omnibus or the Bondi tram, used sometimes in negligence, which were mentioned in the same category as the fair-minded observer of the bias rule in Steinmetz v Shannon [2019] NSWCA 114; (2019) 99 NSWLR 687, 697–8 [44] (White JA).

[23] Johnson v Johnson [2000] HCA 48; (2000) 201 CLR 488, 493 [12] (Gleeson CJ, Gaudron, McHugh, Gummow and Hayne JJ) (‘Johnson’).

[24] Webb (n 6) 52.

[25] Ibid. This statement has a loose parallel in a criminal case decided by the High Court in the same year. When considering whether, upon the whole of the evidence, it was open to a jury to find the defendant guilty, the High Court explained that ‘[i]n most cases a doubt experienced by an appellate court will be a doubt which a jury ought also to have experienced’:

M v The Queen [1994] HCA 63; (1994) 181 CLR 487, 494 (Mason CJ, Dawson, Deane and Toohey JJ).

[26] Webb (n 6) 52.

[27] Automatic disqualification for pecuniary interest was long traced to Dimes v Proprietors of the Grand Junction Canal [1852] EngR 789; (1852) 3 HLC 759; 10 ER 301 (‘Dimes’). The majority in Ebner (n 5) held that Dimes (n 27) had been wrongly applied but conceded that, while financial interests no longer formed a separate category in the bias rule, they continued to attract particular public concern: Ebner (n 5) 351 [34], 357–8 [55]–[56] (Gleeson CJ, McHugh, Gummow and

Hayne JJ).

[28] Ebner (n 5) 345 [8].

[29] Ibid.

[30] Ibid. Their Honours also later confirmed that ‘[t]he apprehension of bias principle requires articulation of the connection between the asserted interest and the disposition of the cause which is alleged’: at 357 [55].

[31] Ibid 345 [8].

[32] Ibid 350 [32].

[33] Ibid 344 [4]. That approach was affirmed in Jia Legeng (n 9) 563 [181] (Hayne J) and Isbester (n 1) 146 [22] (Kiefel, Bell, Keane and Nettle JJ).

[34] Ebner (n 5) 344 [4].

[35] Jia Legeng (n 9) 532 [72] (Gleeson CJ and Gummow J).

[36] Sun v Minister for Immigration and Ethnic Affairs [1997] FCA 1488; (1997) 81 FCR 71, 127, 133 (Burchett J).

[37] Jia Legeng (n 9) 531 [69].

[38] Michael Wilson & Partners (n 15) 437–8 [33] (Gummow ACJ, Hayne, Crennan and Bell JJ). Though it should be noted that decisions about the actual approach of the actual decision-maker are ultimately made with an element of subjective judgment by courts of review and appeal.

[39] Locabail (UK) Ltd v Bayfield Properties Ltd [1999] EWCA Civ 3004; [2000] QB 451, 472 [3] (Lord Bingham CJ, Lord Woolf MR and Scott V-C). This position is echoed in The Council of Chief Justices of Australia and New Zealand, Guide to Judicial Conduct (Australian Institute of Judicial Administration, 3rd ed, 2017) ch 3.5(e) 17–18, which cautions that it is ‘not appropriate for a judge to be questioned by parties or their advisers’.

[40] Johnson (n 23) 517 [79]. The same concerns were evident in the majority opinion in

Caperton v AT Massey Coal Co Inc, 556 US 868, 883 (Kennedy J for the Court) (2009).

[41] James Goudkamp, ‘Facing up to Actual Bias’ (2008) 27(1) Civil Justice Quarterly 32.

[42] A finding of actual bias is a very unfavourable one made about the conduct of the judge in question.

[43] Jia Legeng (n 9) 564 [184]. See also Ebner (n 5) 345 [7], where Gleeson CJ, McHugh, Gummow and Hayne JJ noted that ‘[n]o attempt need be made to inquire into the actual thought processes of the judge or juror’ in claims of apprehended bias, and Isbester (n 1) 156 [61] where Gageler J noted that there need not be any inquiry into state of mind of either the decision-maker or anyone else involved in the process.

[44] Gillies v Secretary of State for Work and Pensions [2006] UKHL 2; [2006] 1 WLR 781, 789 [23] (Lord Hope) (‘Gillies’). My adoption of this relatively simple definition for present purposes is not intended to deny the value of far more detailed explanations of the concepts, such as that offered in William Lucy, ‘The Possibility of Impartiality’ (2005) 25(1) Oxford Journal of Legal Studies 3.

[45] Republican Party of Minnesota v White, [2002] USSC 3399; 536 US 765, 775 (2002) (emphasis in original).

[46] Ibid 777 (emphasis in original).

[47] Ibid 778.

[48] Sir Anthony Mason concluded that Scalia J was ‘unquestionably right’ to reject his second form of impartiality and that his third variant ‘best captures the meaning of “impartiality” in a modern judicial context’: Sir Anthony Mason, ‘Judicial Independence in Australia: Contemporary Challenges, Future Directions’ in Rebecca Ananian-Welsh and Jonathan Crowe (eds), Judicial Independence in Australia: Contemporary Challenges, Future Directions (Federation Press, 2016) 7, 10.

[49] Liteky v United States, 510 US 540, 550 (Scalia J for Rehnquist CJ, O’Connor, Thomas and

Ginsburg JJ) (1994) (emphasis in original). See also Flaherty v National Greyhound Racing Club Ltd [2005] EWCA Civ 1117, [28] (Baker LJ), where the English Court of Appeal described bias as ‘predisposition of prejudice against one party’s case or evidence on an issue for reasons unconnected with the merits of the issue’.

[50] Helljay Investments Pty Ltd v Deputy Commissioner of Taxation [1999] HCA 56; (1999) 166 ALR 302, 307 [12].

[51] Benjamin N Cardozo, The Nature of the Judicial Process (Yale University Press, 1921) 13.

[52] RDS v The Queen [1997] 3 SCR 484, 534 [119] (‘RDS’).

[53] Ibid, quoting Canadian Judicial Council, Commentaries on Judicial Conduct (Éditions Yvon Blais, 1991) 12. Justices L’Heureux-Dubé and McLachlin quote the same passage from the Canadian Judicial Council: RDS (n 52) 504 [35]. A similar conclusion is reached in Lucy (n 44) 15.

[54] For present purposes, the more subtle recent defences of the declaratory theory are not relevant. See, eg, Allan Beever, ‘The Declaratory Theory of Law’ (2013) 33(3) Oxford Journal of Legal Studies 421. That author argues that the declaratory theory rejected by most modern jurists is effectively a caricature of the true theory. One could reply that recasting the theory is as much an admission of defeat as it is a defence, though this criticism does not do full justice to Beever’s thoughtful argument.

[55] But see Lee Epstein, ‘How Institutions Structure Judicial Behaviour: An Analysis of Alarie and Green’s Commitment and Cooperation on High Courts’ (2019) 69 (Spring) University of Toronto Law Journal 275, 276. That author contrasted models of judicial behaviour, such as the attitudinal model (in which judges’ decisions reflect their ideological views), the strategic model (in which judges decide in a goal-oriented manner, taking account of many competing factors) and the labour market model (in which judges are guided by pecuniary and non-pecuniary costs and benefits) with the ‘traditional legal model’ in which judges use various methodologies such as textualism and originalism to reach supposedly value-neutral outcomes.

[56] RDS (n 52) 505–7 [38]–[44]. Justice L’Heureux-Dubé made that point clear in a provocative paper delivered after the judgment: Claire L’Heureux-Dubé, ‘Reflections on Judicial Independence, Impartiality and the Foundations of Equality’ (1999) 7(2) Canadian Journal of International Law Yearbook 95. A far more restrained approach was evident in Yukon Francophone School Board, Education Area #23 v A-G (Yukon) [2015] 2 SCR 282. In delivering judgment for a unanimous court, Abella J explained that ‘impartiality and neutrality do not mean that a judge must have no prior conceptions, opinions or sensibilities. Rather, they require that the judge’s identity and experiences not close [their] mind to the evidence and issues.’: at 300 [33]. This approach has the obvious advantage of recognising the likely practical impact of greater diversity in judicial appointments.

[57] The phrase comes from Richard F Devlin, ‘We Can’t Go on Together with Suspicious Minds: Judicial Bias and Racialized Perspective in R v RDS’ (1995) 18(2) Dalhousie Law Journal 408, 413, who referred to the approach of the majority in RDS (n 52) as ‘contextual judicial method’.

[58] Adam N Glynn and Maya Sen, ‘Identifying Judicial Empathy: Does Having Daughters Cause Judges to Rule for Women’s Issues?’ (2015) 59(1) American Journal of Political Science 37. Interestingly, a subsequent study found no support for such conclusions in male members of Congress: Mia Costa et al, ‘Family Ties? The Limits of Fathering Daughters on Congressional Behaviour’ (2019) 47(3) American Politics 471, finding that, once party affiliation was taken into account, congressmen with daughters were not more likely to support feminist issues.

[59] It is perhaps for this reason that an alleged lack of empathy by a judge does not provide a factual foundation for a bias claim: Kennedy v Secretary, Department of Industry [No 2] [2016] FCA 746, [43] (Flick J).

[60] This suggestion that judges should become agents of social change was rebutted in DA Ipp, ‘Judicial Impartiality and Judicial Neutrality: Is There a Difference?’ (2000) 19(3) Australian Bar Review 212, 221–2. See also Andrew Higgins and Inbar Levy, ‘Judicial Policy, Public Perception, and the Science of Decision Making: A New Framework for the Law of Apprehended Bias’ (2019) 38(3) Civil Justice Quarterly 376, 385, who note the possibility of judges opening their eyes to the world, so to speak, ‘raises fascinating questions about when contextualised judging stops being good judging and becomes bias instead’.

[61] Justice Keith Mason, ‘Unconscious Judicial Prejudice’ (2001) 75 (November) Australian Law Journal 676, 686.

[62] See, eg, the meta-analysis of such studies in Jeffrey J Rachlinski and Andrew J Wistrich, ‘Judging the Judiciary by the Numbers: Empirical Research on Judges’ [2017] (13) Annual Review of Law and Social Science 203.

[63] Andrew J Wistrich, Jeffrey J Rachlinski and Chris Guthrie, ‘Heart versus Head: Do Judges Follow the Law or Follow Their Feelings?’ (2015) 93(4) Texas Law Review 855. The authors of this remarkable study interviewed 1,800 state and federal judges in the United States.

[64] Jeffrey J Rachlinski et al, ‘Does Unconscious Racial Bias Affect Trial Judges?’ (2009) 84(3) Notre Dame Law Review 1195. This novel study interviewed 133 American judges and found judges held implicit bias (the term used by those authors for unconscious bias) which influenced judicial decision-making and that judges could generally control these influences if aware of the need to do so. A subsequent study found similar racial bias in prosecutorial decisions:

Robert J Smith and Justin D Levinson, ‘The Impact of Racial Bias on the Exercise of Prosecutorial Discretion’ (2012) 35(3) Seattle University Law Review 795.

[65] Anna Spain Bradley, ‘The Disruptive Neuroscience of Judicial Choice’ (2018) 9(1) UC Irvine Law Review 1, 7. Bradley makes a strong case for greater research into cognitive psychology and judicial work.

[66] Raymond S Nickerson, ‘Confirmation Bias: A Ubiquitous Phenomenon in Many Guises’ (1998) 2(2) Review of General Psychology 175.

[67] The landmark study that coined the term ‘anchoring’ is Amos Tversky and Daniel Kahneman, ‘Judgment under Uncertainty: Heuristics and Biases’ (1974) 185(4157) Science 1124, 1128. A subsequent study by those authors suggested that anchoring is not limited to information that is typically reduced to numerical or probability terms: Amos Tversky and Daniel Kahneman, ‘Extensional versus Intuitive Reasoning: The Conjunction Fallacy in Probability Judgment’ (1983) 90(4) Psychological Review 293. That study created ‘Linda the bank teller’ who was, among other things, outspoken, very bright, marched in anti-nuclear demonstrations and was a fine arts graduate who had majored in psychology: at 297. They asked what was more

likely — that Linda was a bank teller, or a bank teller who is also an active feminist? Eighty-five per cent chose the latter option, even though all probability theories suggested that level of response should go to the former option: at 299.

[68] Two pioneering scholars of anchoring illustrated the concept by using a simple test about the possible length of the Mississippi River, asking one group whether the river was longer or shorter than 2,000 miles and using a figure of 70 miles for another group. Members of the first group guessed a longer median figure due to the anchoring effect of the higher figure given to them: Karen E Janowitz and Daniel Kahneman, ‘Measuring of Anchoring in Estimation Tasks’ (1995) 21(11) Personality and Social Psychology 1161, 1163.

[69] A persuasive study to this effect is Mark W Bennett, ‘Confronting Cognitive “Anchoring Effect” and “Blind Spot” Biases in Federal Sentencing: A Modest Solution for Reforming a Fundamental Flaw’ (2014) 104(3) Journal of Criminal Law and Criminology 489. An obvious qualification to this study is that legislative guidelines for sentencing, much like judicial precedents, are clearly legitimate anchors.

[70] Jeffrey J Rachlinski, Andrew J Wistrich and Chris Guthrie, ‘Can Judges Make Reliable Numeric Judgments?: Distorted Damages and Skewed Sentences’ (2015) 90(2) Indiana Law Journal 695.

[71] See Tim RW de Wilde, Femke S Ten Velden and Carsten KW de Dreu, ‘The Anchoring-Bias in Groups’ [2018] (76) Journal of Experimental Social Psychology 116. Justice Gageler has argued that such dangers depend in large part on whether concurring judges simply agree with the decision of a colleague, in which case they may simply repeat possible errors of a colleague, or instead form their own true view before joining any joint judgment: Stephen Gageler, ‘Why Write Judgements?’ [2014] SydLawRw 9; (2014) 36(2) Sydney Law Review 189, 195–6.

[72] One notable response was the enactment in 2019 of art 33 of the French Justice Reform Act. That provision introduced a criminal offence of using data that may identify judges or clerks ‘with the object or effect of evaluating, analysing, comparing or predicting their actual or supposed professional practices’: LOI n° 2019-222 du 23 mars 2019 [Law No 2019-222 of 23 March 2019] (France) JO, 24 March 2019, 33. The implications of using data are examined in Adrian Zuckerman, ‘Artificial Intelligence: the Implications for the Legal Profession, Adversarial Process and the Rule of Law’ (2020) 136 (July) Law Quarterly Review 427.

[73] An argument made in Gary Edmond and Kristy A Martire, ‘Just Cognition: Scientific Research on Bias and Some Implications for Legal Procedure and Decision-Making’ (2019) 82(4) Modern Law Review 633, 663–4.

[74] Higgins and Levy (n 60) 394–8. Those authors suggest that any code should identify those instances where the judge faced with a bias claim should not determine that claim. A similar approach was suggested by Callinan J in Ebner (n 5) 397–8 [185] but was pointedly rejected by Gleeson CJ, Gummow, McHugh and Hayne JJ: at 361 [74]. A variant of this approach was also suggested in the United Kingdom after Pinochet [No 2] (n 13): see Kate Malleson, ‘Judicial Bias and Disqualification after Pinochet [No 2]’ (2000) 63 (January) Modern Law Review 119, 126.

[75] Lord Sumption, ‘Home Truths about Judicial Diversity’ (Bar Council Law Reform Lecture, 15 November 2012).

[76] Ibid 17. Lord Sumption also noted that most of this research was American and did not appear to intend that description to be complimentary.

[77] It should be noted that research in this sense is more than the simple collation of statistics. The importance of analysis was made clear in ALA15 v Minister for Immigration and Border Protection [2016] FCAFC 30. In that case, evidence of the proportion of decisions made by a judge in migration cases was so overwhelming and consistently against applicants for protection and other visas that the judge’s conduct created an apprehension of bias. The material was tabulated to outline the number of cases and their disposition but contained no analysis or explanation of individual cases. Chief Justice Allsop, Kenny and Griffiths JJ held that ‘raw statistical material ... normally would need to be accompanied by a relevant analysis ... in order that the statistics were placed in a proper context’: at [38]. The absence of such analysis meant it was not capable of informing the judgment of the observer used in the bias rule. A differently constituted Full Federal Court subsequently noted that earlier rulings of a judge may not be admissible in a later case where it is sought to argue that those earlier cases created an apprehension of bias: CMU16 v Minister for Immigration and Border Protection [2020] FCAFC 104, [21]–[28] (Jagot, Yates and Stewart JJ).

[78] In theory, this research could be useful to judicial assessments made in bias claims. The many studies about judicial behaviour might provide insight into how judges apply objective tests, such as the informed observer of the bias test. Whether such studies would help or hinder the law of bias is a question beyond the scope of this article.

[79] Isbester (n 1) 146 [23] (Kiefel, Bell, Keane and Nettle JJ). See also Ebner (n 5) 344 [4] (Gleeson CJ, McHugh, Gummow and Hayne JJ); Jia Legeng (n 9) 563 [181]–[182] (Hayne J).

[80] Ebner (n 5) 343 [3], 345 [6]–[7].

[81] The majority in Ebner (n 5) noted that ‘[b]ias, whether actual or apprehended, connotes the absence of impartiality. It may not be an adequate term to cover all cases of the absence of independence.’: at 348 [23].

[82] Gillies (n 44) 793 [37]–[38].

[83] Ibid 790–1 [27] (Lord Hope), 792 [33] (Lord Rodger), 794 [43] (Baroness Hale). Lord Rodger also noted that to accept the possibility that medically qualified members of the tribunal would struggle to look beyond their own clinical judgment would inevitably require acceptance that disabled members of the tribunal were similarly unable to look beyond their own experience: at 791 [32].

[84] Ibid 794 [44].

[85] Ibid 793 [38].

[86] CNY17 (n 1) 69 [98].

[87] Ibid 52–3 [18].

[88] Ibid 53–4 [23]–[26].

[89] Ibid 54 [25].

[90] See, eg, Paul Karp, ‘“Full of Liberal Mates”: Labor Accuses Coalition of “Stacking” Tribunal’ The Guardian (online, 21 February 2019) <https://www.theguardian.com/australia-news/2019/feb/21/full-of-liberal-mates-labor-accuses-coalition-of-stacking-tribunal>, archived at <https://perma.cc/6H7M-GLWR>.

[91] See, eg, Ben Doherty, ‘Refugee Tribunal Members Lost Contracts after Peter Dutton Criticisms’ The Guardian (online, 1 June 2017) <https://www.theguardian.com/australia-news/2017/jun/02/contracts-wont-be-renewed-for-refugee-review-tribunal-members>, archived at <https://perma.cc/TD5T-PF4E>.

[92] Concerns about the apparent politicisation of many federal tribunal appointments were not addressed in the recent detailed review of federal tribunals: IDF Callinan, Review: Section 4 of the Tribunals Amalgamation Act 2015 (Cth) (Report, 23 July 2019) (‘Callinan Report’). This is despite many submissions to the Review raising this issue: see, eg, the submissions of the Refugee Council of Australia, Submission to Attorney-General’s Department, Statutory Review of the Tribunals Amalgamation Act 2015 (29 July 2019) [1.6], the concerns of which were acknowledged in the report but not addressed: at 95 [6.127]. Solutions to the often politicised nature of appointments to the Administrative Appeals Tribunal (‘AAT’) are suggested in James Morgan, ‘Securing the Administrative Appeals Tribunal’s Independence: Tenure and Mechanisms of Appointment’ (2018) 43(4) Alternative Law Journal 302, 305–8.

[93] An example is the rule that precludes officials from exercising discretionary powers at the behest of another, which is otherwise known as the ‘dictation’ ground of judicial review: Mark Aronson, Matthew Groves and Greg Weeks, Judicial Review of Administrative Action and Government Liability (Lawbook, 6th ed, 2017) 305 [5.340]. The extent to which a minister can lawfully influence or make ‘suggestions’ to an official vested with a discretionary power has long been unclear: at 305–10 [5.340]–[5.400].

[94] Earlier cases often used male pronouns, which have been replaced with ‘person’. Justice Kirby suggested that ‘in contemporary Australia, the fictitious bystander is not necessarily a man nor necessarily of European ethnicity or other majority traits’: Johnson (n 23) 508 [52]. While that assessment is clearly correct, it also invites the question of how a court might apply the reasoning of an observer who could be male or female, of European or other heritage and so on. The choice between one of each possibility is easier than some sort of blend of both.

[95] Ibid 509 [53].

[96] R v Abdroikov [2007] 1 WLR 2679, 2706 [81] (‘Abdroikov’).

[97] Johnson (n 23) 494 [14] (Gleeson CJ, Gaudron, McHugh, Gummow and Hayne JJ).

[98] Almazeedi v Penner [2018] UKPC 3, [20] (Lord Mance). See also Meerabux v A-G (Belize) [2005] EWCA Civ 634; [2005] 2 AC 513, 528 [25] (Lord Hope for the Court).

[99] See, eg, Antoun v The Queen (2006) 224 ALR 61, 77–8 [85]–[87] where Callinan J held that a judge erred when indicating that a clearly weak submission was bound to fail. The judge should have indicated a clear willingness to consider submissions, even if weak.

[100] Helow v Secretary of State for the Home Department [2008] UKHL 62; [2008] 1 WLR 2416, 2418 [3] (Lord Hope).

[101] Gillies (n 44) 793 [39].

[102] Saxmere Company Ltd v Wool Board Disestablishment Co Ltd [2009] NZSC 72; [2010] 1 NZLR 35, 52 [37] (Tipping J).

[103] The Full Federal Court has noted that ‘[t]he more informed a hypothetical bystander may be, the more difficult it may be to sustain an argument that an apprehension of bias is reasonable’: Reece v Webber [2011] FCAFC 33; (2011) 192 FCR 254, 273 [55] (Jacobson, Flick and Reeves JJ).

[104] In CNY17 (n 1) 62 [59], Nettle and Gordon JJ reasoned that where ‘the statutory context is complex, the fair-minded lay observer at least must have knowledge of the key elements of that scheme’.

[105] Johnson (n 23) 493 [13] (Gleeson CJ, Gaudron, McHugh, Gummow and Hayne JJ).

[106] See, eg, R v Doogan; Ex parte Lucas-Smith [2005] ACTSC 74; (2005) 193 FLR 239, 249–51 [44]–[50] (Higgins CJ, Crispin and Bennett JJ). The observer would know enough of coronial proceedings to understand that their early stages could include investigative activities undertaken without involvement of the parties.

[107] Ex parte H (n 7) 434–5 [27]–[29] (Gleeson CJ, Gaudron and Gummow JJ). The further relevant practice in that case was hearings in the tribunal were not open to the public.

[108] AJH Lawyers Pty Ltd v Careri [2011] VSCA 425; (2011) 34 VR 236, 243 [23] (Warren CJ, Hansen JA and

Almond AJA).

[109] The attribution of that knowledge to the observer enabled the High Court to find that the strong statements made by a judge at earlier stages of a case did not create an apprehension of bias: Concrete Pty Ltd v Parramatta Design & Developments Pty Ltd [2006] HCA 55; (2006) 229 CLR 577,

634–5 [174]–[175] (Callinan J) (‘Concrete’).

[110] Johnson (n 23) 493 [13] (Gleeson CJ, Gaudron, McHugh, Gummow and Hayne JJ). Justices Gordon and Edelman acknowledged that case management duties and other obligations of judges that have arisen in recent times ‘recognise that passive participation in litigation is no longer an option’ for judges: Rozenblit v Vainer (2018) 262 CLR 478, 501 [76].

[111] Lee v Cha [2008] NSWCA 13, [43].

[112] Ibid [45].

[113] IOOF Australia Trustees Ltd v Seas Sapfor Forests Pty Ltd [1999] SASC 249; (1999) 78 SASR 151, 183 [192],

184 [194] (Doyle CJ). See also Li v Minister for Immigration and Multicultural Affairs [2000] FCA 19; (2000) 96 FCR 125, 134 [42] (Drummond J); Michael v Western Australia [2007] WASCA 100, [90], [95] (Steytler P).

[114] [2002] EWCA Civ 90; [2003] QB 528, 554–5 [72]–[73] (Lord Woolf CJ for the Court).

[115] Ibid 548 [61].

[116] Concrete (n 109) 635 [177] (Callinan J).

[117] An excellent example is Scalia J’s reasons for his refusal to recuse himself in a case involving the Vice-President of the United States, whom his Honour acknowledged was a good friend, in Cheney v US District Court for DC, [2004] USSC 193; 541 US 913 (2004). The reasons given by Scalia J demonstrate an utter lack of insight about how an observer might perceive his close friendship with the Vice-President, whose conduct and honesty was challenged in a case before the Supreme Court. The many extraordinary assumptions made by Scalia J about the extent to which observers would accept without question his close contact with members of the executive government are analysed in Monroe H Freedman, ‘Duck-Blind Justice: Justice Scalia’s Memorandum in the Cheney Case’ (2004) 18(1) Georgetown Journal of Legal Ethics 229.

[118] See, eg, Abimbola A Olowofoyeku, ‘Bias and the Informed Observer: A Call for a Return to Gough’ (2009) 68(2) Cambridge Law Journal 388. Olowofoyeku suggested that the Archangel Michael would struggle to match the many good qualities of the fictional observer: at 395. See also Philip Havers and Alasdair Henderson, ‘Recent Developments (and Problems) in the Law on Bias’ (2011) 16(2) Judicial Review 80, 82 [9]. Justice Kirby described the informed observer as a ‘paragon’ in Smits v Roach (2006) 227 CLR 423, 457 [96].

[119] Olowofoyeku (n 118) 404.

[120] Ibid 396.

[121] (2011) 242 CLR 283, 306–7 [48] (‘British American Tobacco Australia Services’).

[122] Cf Ebner (n 5) 345 [8] (Gleeson CJ, McHugh, Gummow and Hayne JJ).

[123] Isbester (n 1) 155–6 [59].

[124] Justice Gageler also took the further step of considering whether the reasonableness of the apprehension of bias had been negatived by various circumstances: ibid 159 [69].

[125] See, eg, CNY17 Full Federal Court (n 15) 92 [11], where Mortimer J put this issue ‘to one side’.

[126] This point was not considered in CNY17 (n 1), but the uncertainty surrounding it was acknowledged when special leave was sought in that case: Transcript of Proceedings, CNY17 v Minister for Immigration and Border Protection [2019] HCATrans 101, 13–34 (Kiefel CJ and LG De

Ferrari SC), 38–40 (LG De Ferrari SC).

[127] Aronson, Groves and Weeks (n 93) 652 [9.20].

[128] Ibid. This criticism was echoed in Mackinnon v Partnership of Larter, Jones, Miraleste Pty Ltd [No 4] [2018] NSWSC 147, [198] (‘Mackinnon’), where Stevenson J reasoned that the new third step suggested by Gageler J was similar to the existing second step because both ‘are the articulation of how the identified factor might cause, in the ultimate determination of the case, a deviation from a neutral evaluation of its merits and the reasonableness of an apprehension that this will be the outcome of the case’.

[129] See Matthew Groves, ‘A Reasonably Reasonable Apprehension of Bias: CNY17 v Minister for Immigration and Border Protection[2019] SydLawRw 16; (2019) 41(3) Sydney Law Review 383, 393–4.

[130] Ibid 394.

[131] [1985] HCA 81; (1985) 159 CLR 550, 629 (‘Kioa’).

[132] Ibid 628.

[133] Ibid 629. Justice Brennan analysed the relevancy grounds of judicial review by reference to the ‘[c]redibility’ and ‘[s]ignificance’ of the information in Minister for Aboriginal

Affairs v Peko-Wallsend Ltd (1986) 162 CLR 24, 61–2.

[134] Kioa (n 131) 629.

[135] Ibid.

[136] In CNY17 (n 1) 69 [102], Nettle and Gordon JJ suggested that decision-makers ‘may need to invite an applicant to comment on adverse information to counteract the apprehension of bias’. This possibility assumes that a right of reply may cure that problem. It is one thing to require a right of reply to ensure fairness. It is quite another to assume it can counteract an apprehension of bias. A further difficulty is precisely what comment an affected person may make. Is it limited to the content and perceived effect of the material on the decision-maker? Or can it extend to the wider processes, so that the affected person can comment upon, or even strongly question, an administrative system that allows material that has created an apprehension of bias to be wrongly put before an official?

[137] Kioa (n 131) 629. Though Brennan J did not suggest a rigid approach. His Honour reasoned that material which was not credible, relevant and significant would not normally have to be disclosed: at 628.

[138] [2005] HCA 72; (2005) 225 CLR 88, 91–2 [4]–[5] (Gleeson CJ, Gummow, Kirby, Hayne and Heydon JJ).

[139] Ibid 96 [17].

[140] Ibid.

[141] Ibid 97 [19], quoting NIB Health Funds Ltd v Private Health Insurance Administration Council [2002] FCA 40; (2002) 115 FCR 561, 583 [84] (Allsop J).

[142] Disclosure of material may alter its relevance because an affected person may offer comment or further related material that casts the material in a different light: Minister for Immigration and Citizenship v Maman [2012] FCAFC 13; (2012) 200 FCR 30, 46 [50] (Flick and Foster JJ).

[143] The effect can include the impact of the material upon the willingness of decision-makers to accept or reject other evidence: see, eg, Jagroop v Minister for Immigration and Border Protection [2014] FCAFC 123; (2014) 225 FCR 482, 497 [70] (Dowsett, Murphy and White JJ).

[144] Locabail (n 39) 471 [2] (Lord Bingham CJ, Lord Woolf MR and Scott V-C).

[145] Webb (n 6) 74.

[146] See, eg, Abdroikov (n 96) 2702–3 [67], 2705 [78] (Lord Carswell). Lord Carswell suggested that the ‘number and diversity of people on a criminal jury’ was a safeguard against any single juror exercising undue influence on the outcome of a trial. That same diversity could operate to mitigate unconscious bias but only if there was open discussion by a jury of such issues:

at 2702–3 [67].

[147] O’Sullivan v Medical Tribunal (NSW) [2009] NSWCA 374, [43]–[45] (Basten JA) (‘O’Sullivan’).

[148] A requirement in the now repealed Medical Practice Act 1992 (NSW) s 148(9), as repealed by Health Practitioner Regulation Amendment Act 2010 (NSW) sch 3.

[149] O’Sullivan (n 147) [42].

[150] Ibid [43].

[151] Ibid.

[152] (2003) 132 FCR 114, 125 [36] (Dowsett J, Madgwick J agreeing at 115 [1], Finkelstein J agreeing at 116 [3]) (‘Crowley’).

[153] Ibid 119–20 [15]–[17]. The different forms of inappropriate practice are contained in the Health Insurance Act 1973 (Cth) s 82.

[154] Crowley (n 152) 125 [36].

[155] Ibid.

[156] Ibid.

[157] Ibid 115 [2].

[158] Ibid. His Honour felt precluded by the weight of Full Court authority to consider this issue in detail: at 115–16 [2].

[159] [2009] FCA 1526; (2009) 51 AAR 147, 158 [49], [51] (‘Islam’).

[160] Ibid 149–50 [9]–[10], 156–7 [41]–[43].

[161] Ibid 159 [52].

[162] Ibid. Reliance on the incompatible nature of functions was central to a finding of apprehension of bias in Isbester (n 1) 153 [49] (Kiefel, Bell, Keane and Nettle JJ), 157 [63] (Gageler J).

[163] Islam (n 159) 151 [20] (Finn J). See also British American Tobacco Australia Services (n 121) 333 [145] (Heydon, Kiefel and Bell JJ).

[164] [2017] FCAFC 136; (2017) 254 FCR 534, 538 [9]–[10] (Griffiths J) (‘AMA16’). A ‘fast track reviewable decision’ was provided for under the Migration Act 1958 (Cth) pt 7AA (‘Migration Act’). This process is explained in Emily McDonald and Maria O’Sullivan, ‘Protecting Vulnerable Refugees: Procedural Fairness in the Australian Fast Track Regime’ [2018] UNSWLawJl 34; (2018) 41(3) University of New South Wales Law Journal 1003.

[165] AMA16 (n 164) 539–40 [15] (Griffiths J), discussing s 473CB(1) of the Migration Act (n 164).

[166] AMA16 (n 164) 539–40 [15] (Griffiths J), quoting s 473CB(1)(c) of the Migration Act (n 164).

[167] AMA16 (n 164) 552 [75] (Griffiths J, Dowsett J agreeing at 536 [1], Charlesworth J agreeing

at 557 [97]).

[168] Ibid 536 [1] (Dowsett J), 552 [73] (Griffiths J), 557 [99] (Charlesworth J).

[169] Ibid 537 [4].

[170] Ibid 536 [3].

[171] Ibid 557 [99].

[172] Ibid 557–8 [99].

[173] Ibid 551 [72].

[174] Section 7(3)(b) of the Administrative Appeals Tribunal Act 1975 (Cth) allows for the appointment of non-legally qualified members where they have ‘special knowledge or skills’ that are relevant to the AAT.

[175] AMA16 (n 164) 537 [4].

[176] Ibid 552 [77] (Charlesworth J agreeing at 558 [100]).

[177] CNY17 (n 1) 60 [51] (Nettle and Gordon JJ, Edelman J agreeing at 71 [110]).

[178] Ibid 60 [51] (Nettle and Gordon JJ), 71–2 [111] (Edelman J).

[179] Ibid 69 [97]. Their Honours elsewhere noted that, in some instances, officials should invite affected people to comment on unfavourable or adverse material: at 69 [102]. This reasoning appears directed at less prejudicial material. Justices Nettle and Gordon were mindful of the difficulties in identifying when material moved from being somewhat or merely prejudicial (the effect of which could be counteracted by inviting comment from an affected person) as opposed to material of a more serious impact: at 69 [101].

[180] Ibid 69 [98].

[181] Ibid 69 [99].

[182] Ibid.

[183] Ibid 79 [141].

[184] Ibid 71 [110], 78 [138]–[140].

[185] Ibid 77 [135]. Justice Edelman acknowledged the cautionary remarks against the use of reasons to ‘confirm, enhance or diminish the existence’ of an apprehension of bias that were made in Michael Wilson & Partners (n 15) 466 [67] (Gummow ACJ, Hayne, Crennan and Bell JJ). Justice Edelman thought that caution should not be taken literally, noting that it ‘would be absurd if, on the one hand, remarks made by the decision-maker during the course of a hearing could be considered as part of an assessment of the presence of reasonable apprehension of bias but, on the other hand, remarks made at the conclusion of the proceedings could not’: CNY17

(n 1) 77 [135].

[186] CNY17 (n 1) 58 [43].

[187] Ibid 57–8 [40]–[42]. Justices Nettle and Gordon expressed similar concerns: at 69 [100].

[188] Ibid 58 [43].

[189] Ibid 55 [28] (citations omitted).

[190] Ibid 69 [101] (emphasis omitted).

[191] Ibid (emphasis in original).

[192] The concept of materiality in Australian administrative law is usefully examined in Paul Daly, ‘A Typology of Materiality’ (2019) 26(3) Australian Journal of Administrative Law 134.

[193] See, eg, Hossain v Minister for Immigration and Border Protection [2018] HCA 34; (2018) 264 CLR 123,

134–5 [28]–[31] (Kiefel CJ, Gageler and Keane JJ) (‘Hossain’); Minister for Immigration and Border Protection v SZMTA [2019] HCA 3; (2019) 264 CLR 421, 445 [45], 452–3 [72] (Bell, Gageler and

Keane JJ) (‘SZMTA’). Justices Nettle and Gordon provided a strong argument against the introduction of a requirement of materiality: at 458–60 [88]–[95].

[194] SZMTA (n 193) 445 [45], 451–3 [70]–[72] (Bell, Gageler and Keane JJ).

[195] The High Court revisited the issue of materiality in ABT17 v Minister for Immigration and Border Protection [2020] HCA 34; (2020) 383 ALR 407. That case did not directly consider how the principle might apply to the bias rule, though Gordon J drew attention to one aspect of materiality that was relevant to claims of bias. Her Honour noted that, as a general rule, judicial consideration of the validity of any exercise of statutory power had a ‘temporal element’ because courts were concerned with the manner in which a power was exercised. Justice Gordon held that this inquiry would normally concerned with requirements at the time the power was exercised but noted that, in claims of apprehended bias, ‘this may be before active steps are taken by a decision-maker’: at 437 [101] n 122. This suggestion of Gordon J did not appear to be directly connected to the requirements of materiality.

[196] Ebner (n 5) 345 [7] (Gleeson CJ, McHugh, Gummow and Hayne JJ).

[197] Ibid.

[198] Ibid 344–5 [6]–[7]. This approach is consistent with the focus of the test for apprehended bias on whether the informed observer might think that an official might not be suitably impartial.

[199] CNY17 (n 1) 59 [47].

[200] Ibid.

[201] Ibid 70 [104] (Nettle and Gordon JJ), 76 [129] (Edelman J). Their Honours each noted the Minister had conceded that, if the Court found an apprehension of bias arose, relief should be granted.

[202] This possibility may be qualified by the their Honours’ suggestion that, whether the irrelevant material ‘could realistically have made any difference ... falls to be answered by reference to the same considerations as inform the answer already given to the question of whether a hypothetical fair-minded lay observer, acting reasonably, would entertain as realistic the possibility’ that the decision-maker might have departed from required standards of impartiality: at ibid 59 [48] (emphasis added). This statement is curious for two reasons. First, the emphasised ‘would’ departs from the ‘double might’ standard that applies in bias claims. Second, this passage is somewhat at odds with the one quoted in the text because it appears to blur the difference between the informed observer acting reasonably and a decision-maker acting lawfully: see above n 200 and accompanying text.

[203] This is different to cases of waiver of bias, in which an apprehension of bias is found to have arisen but the affected party has waived the right to complain about it: see generally Aronson, Groves and Weeks (n 93) 715–21 [9.350]–[9.380]. Waiver can be found when a party is fully aware of the key issues that could support a claim of bias but fails to raise the matter in a timely manner: at 715 [9.360], citing Matthew Groves, ‘Waiver of the Rule against Bias’ [2009] MonashULawRw 19; (2009) 35(2) Monash University Law Review 315. See also Woolf et al (n 16) 572–3 [10.066]–[10.069].

[204] This possibility is different to the suggestion of Edelman J that an extreme instance of a denial of fairness will be material, irrespective of whether it deprived the person of the chance of a different outcome: Hossain (n 193) 147–8 [72]. Justice Edelman has elsewhere suggested that an extreme denial of fairness occurs when a ‘party receives no hearing at all’:

DWN042 v Republic of Nauru [2017] HCA 56; (2017) 350 ALR 582, 588 [21]. This reasoning appears to assume that public confidence in the law would be damaged if instances of grave unfairness were not quashed.

[205] That criticism proceeds on the assumption that materiality would not be determined within the third step suggested by Gageler J in Isbester (n 1) 155–6 [59]. Materiality might be considered as part of any assessment of the reasonableness of an apprehension of bias. If adopted as part of the bias test, however, materiality would likely require separate consideration. Materiality is ultimately about the effect of an error. The reasonableness of an apprehension of bias is not.

[206] Sumption (n 75) 19–20.

[207] Granatino v Radmacher (formerly Granatino) [2010] UKSC 42; [2011] 1 AC 534, 577 [137]. The novelty of this case was that, unlike most cases where prenuptial agreements are used to the advantage of men, it was the wife whose family wealth vastly outweighed her husband’s and she who wished to limit the impact of any future divorce on her assets by use of a prenuptial agreement:

at 544 [13] (Lord Phillips PSC).

[208] Sumption (n 75) 19–20. A contrary argument was made earlier in Mason (n 61) 687.

[209] Sumption (n 75) 3–8. Another equally important issue is the suggestion by Baroness Hale that ‘women may judge differently by virtue of being women’: Lizzie Barmes and Kate Malleson, ‘Lifting the Judicial Identity Blackout’ (2018) 38(2) Oxford Journal of Legal Studies 357, 361–2. According to Barmes and Malleson, this suggestion, ‘however tentatively framed, is profoundly disruptive because it highlights that men cannot escape the influence on their attitudes of their identity as a man in a world in which the male gender is privileged any more than a female judge can escape that of being a woman in a world in which women are the second sex’: at 362.

[210] Baroness Hale provided a more convincing assessment of this issue, as far as it affects the Supreme Court, in ‘Appointments to the Supreme Court’ in Graham Gee and Erika Rackley (eds), Debating Judicial Appointments in an Age of Diversity (Routledge, 2018). See also Margaret Thornton, ‘“Otherness” on the Bench: How Merit Is Gendered’ [2007] SydLawRw 16; (2007) 29(3) Sydney Law Review 391.


AustLII: Copyright Policy | Disclaimers | Privacy Policy | Feedback
URL: http://www.austlii.edu.au/au/journals/MelbULawRw/2020/34.html