Sydney Law Review
Before the High Court
A Reasonably Reasonable Apprehension of Bias: CNY17 v Minister for Immigration and Border Protection
In CNY17 v Minister for Immigration and Border Protection, the High Court of Australia will consider the test for apprehended bias. The current test was adopted in Ebner v Official Trustee in Bankruptcy  HCA 63; (2000) 205 CLR 337 and has long been taken to involve two steps. The first requires those claiming bias to identify the claimed source of bias. The second step is to explain how that influence will affect the impartiality of a decision-maker. Despite the apparent simplicity of this test, claims of apprehended bias remain impressionistic and difficult to judge. Whether an apprehension of bias is reasonable is a contextual question that can easily yield contestable judgments. Justice Gageler has suggested that Ebner requires a third step, which asks whether an apprehension of bias is reasonable in all the circumstances. This Before the High Court column argues that third possible step would add only confusion to the test for apprehended bias. It also explains how similar problems can arise if courts apply the Ebner test by assuming an unrealistic knowledge of relevant legislation when determining the reasonableness of any apprehension of bias.
Ebner v Official Trustee in Bankruptcy discarded the rule of automatic disqualification in favour of a two-fold test for bias that applied to all cases. A test that appeared simple and sensible has required repeated explanation by the High Court of Australia and will face the High Court yet again in CNY17 v Minister for Immigration and Border Protection. This column explains why the bias rule remains difficult to apply and why future developments in the law could complicate the rule even more. That possibility arises from two uncertainties in the bias rule, each of which is in play in CNY17. One is the extent to which the legislative context in which a decision is made should influence judgments about a reasonable apprehension of bias. The legislation governing the decision in CNY17 was extremely complex and difficult to understand, yet a majority of the Full Federal Court of Australia reasoned that the hypothetical observer, upon whose thinking claims of apprehended bias are tested, would have an expert’s knowledge of this law. This approach has the potential to enable judges to subsume entirely the role of the informed observer. The second uncertainty of the bias rule that is raised by CNY17 is the additional step to determine an apprehension of bias, proposed by Gageler J, which superimposes a question — whether an apprehension of bias established under the existing two-step test is reasonable in all the circumstances. This proposed third step would complicate the test and also enable judges to subsume the role of the informed observer. To consider the best approach to apprehended bias, it is first useful to examine the rationale of the bias rule and its recent evolution.
Bias is a frequently used term that eludes easy or precise definition, though some governing principles are fairly clear. The High Court has reasoned that bias ‘connotes the absence of impartiality’. It is well understood that this encompasses the reality and appearance of impartiality. When a claim concerns apprehended rather than actual bias, as virtually all bias claims do, the significance of the appearance of impartiality arguably assumes greater importance because any apprehension reflects how things may appear, rather than how they actually may be. The standard is breached if a fair-minded and informed observer might reasonably apprehend that decision-makers might not bring a sufficiently impartial mind to the task before them. Partiality can take many forms, such as prejudgment of the issues or parties of a case, possessing a connection to a party or issue that somehow affects the impartiality of decision-makers, or giving unjustified preferable treatment to a party. The rule against bias applies to all those who exercise public power of some form, such as judges, jurors, tribunal members, coroners, local councillors, government ministers, and bureaucrats. The bias rule fosters fairness and impartiality. Judges often suggest that impartiality has particular relevance to their role, but the bias rule is equally important to non-judicial officers and decision-makers. It fosters public confidence in the officials and institutions to which it applies, by enhancing the appearance and actuality of impartial decision-making.
The central requirement of the bias rule, which is that decision-makers be sufficiently impartial, can be traced back though several centuries of common law. In recent times, key elements of the test governing this requirement have been recast. One was the perspective from which claims of bias should be assessed. It was uncertain whether claims of bias should be decided by reference to the view of the judge or judges faced with the issue, or from some objective standpoint. Australian courts slowly moved towards an objective assessment of claims of bias, though slightly different explanations of that objective observer abounded. These included ‘fair-minded people’, a ‘fair-minded observer’, a ‘reasonable person’ a ‘reasonable or fair-minded observer’, a ‘lay observer’, and a ‘fair-minded, informed lay observer’. These different expressions all conveyed the same single concept, of a fictional observer constructed by the courts.
An important element of that fictional construct was settled in Webb, when the High Court decisively accepted that judgments about bias claims should be made by reference to the reasoning of fair-minded members of the public, rather than the subjective views of the judge faced with a claim of bias. Chief Justice Mason and McHugh J reasoned that, if the bias rule operates to foster public confidence in the administration of justice, questions of bias should be resolved by reference to the perceived views of that same public, rather than by judges’ own views. The High Court expressly rejected the contrary view reached just a year earlier by the House of Lords. It was not long before English law also adopted the objective device of the fair-minded and informed observer to determine bias claims.
However, an important new difference in the Australian approach to bias arose in Ebner. In that decision, the High Court disavowed the longstanding rule of automatic disqualification for pecuniary interest that had prevailed for 150 years. The House of Lords had affirmed and extended automatic disqualification only a year earlier in Pinochet (No 2), when it held that the professional connections of one of the Law Lords (his patronage and work with a charity that intervened in a case before the Lords) was not simply capable of supporting an apprehension of bias, but could also trigger automatic disqualification. Ebner rejected the bifurcated approach of automatic disqualification, under which some claims are deemed conclusive, but others subject to different reasoning, and instead applied a single approach to all claims of apprehended bias. The new approach, the majority explained ‘requires two steps’. These were:
First, it requires the identification of what it is said might lead a judge (or juror) to decide a case other than on its legal and factual merits. The second step is no less important. There must be an articulation of the logical connection between the matter and the feared deviation from the course of deciding the case on its merits.
The majority continued:
The bare assertion that a judge (or juror) has an ‘interest’ in litigation, or an interest in a party to it, will be of no assistance until the nature of the interest, and the asserted connection with the possibility of departure from impartial decision making, is articulated. Only then can the reasonableness of the asserted apprehension of bias be assessed.
This two-fold test requires identification then explanation. Those claiming bias must first identify what will affect a decision-maker’s impartiality. The next step is to articulate how the claimed interest or influence will have the suggested effect. Bias claims rarely fall at the first hurdle, largely as a consequence of Ebner’s rejection of automatic disqualification. Just as courts could no longer accept a narrow class of bias claims without further reasoning, they appeared reluctant to reject sometimes flimsy claims without furthering reasoning.
The two-step process of Ebner may appear clear in principle, but its application regularly gives rise to judicial disagreement. Different judges can easily take different views of how a claimed source of impartiality might be perceived by the informed observer. In more than one case, a unanimous High Court reached a different conclusion to an equally unanimous intermediate court. The most recent instance was Isbester, in which the majority applied Ebner’s two-fold test and found a reasonable apprehension of bias arose from the involvement of a council official in two different proceedings about an allegedly dangerous dog. An apprehension of bias was found to have arisen because the official who was the ‘moving force’ of the first case (to prosecute the dog owner) possessed enough of an interest in the outcome of the subsequent administrative hearing to require she not participate. Justices Kiefel, Bell, Keane and Nettle reached that conclusion by reference to Ebner’s two-step process, suggesting that the only novel element of the case was the proper characterisation of the official’s interest. Their Honours accepted that the official was akin, though not precisely equivalent, to a prosecutor in the first case. That role made her participation in subsequent administrative proceedings undesirable because the likely desire for vindication of anyone in such a position was an interest sufficient to create an apprehension of bias. Justices Kiefel, Bell, Keane and Nettle also noted that the rejection of automatic disqualification in Ebner meant the two-step test of that case applied to cases involving a clear or strong personal interest, such as the one at hand. The notable feature of such cases, their Honours reasoned, was that where the interest identified in Ebner’s first step is an ‘incompatibility of roles ... which points to a conflict of interest’, the decision required by Ebner’s second step ‘is obvious’.
Justice Gageler reached a similar conclusion, but by use of a third step his Honour identified from Ebner. The third step Gageler J identified was ‘consideration of the reasonableness of the apprehension of’ the deviation from impartiality as suggested by the party claiming bias. The key passage in Ebner provides some textual support for this further step because the majority concluded its explanation of its new two-step test for bias by noting that, after both steps were taken, ‘[o]nly then can the reasonableness of the asserted apprehension of bias be assessed.’ However, the nature of this third step is unsettled, as discussed further below. This is one of the difficulties in application of the Ebner test that face the Court in CNY17.
The applicant in CNY17 arrived in Australia, from Iraq by boat, in August 2013 and was detained on Christmas Island. On 20 March 2015, he was involved in a disturbance while in detention and charged with a relatively minor criminal offence. He was involved in another disturbance, which occurred in the tumultuous period after the death of a fellow detainee, and was charged with more offences.
A few days before this second disturbance, the applicant was notified that his claim for refugee status would be considered under the ‘Fast Track Assessment’ process. This meant the so-called ‘bar’ in s 46A of the Migration Act 1958 (Cth) (‘Migration Act’) was lifted and the applicant could seek either a temporary protection or safe haven visa. A few days after that second disturbance, the applicant was transferred to a prison in Western Australia. He later pleaded guilty to charges arising from the disturbances (for breaking a window) and received the minor penalties of a six-month good behaviour bond, with a requirement to pay $820.60 restitution and provide a security of $500.
The applicant sought a safe haven visa, but this was denied by a ministerial delegate, who essentially rejected key parts of the applicant’s claim and found he was not a refugee within the meaning of the Migration Act. The Minister then referred the case to the Immigration Assessment Authority (‘the Authority), as part of the ‘Fast Track Review Scheme’ under pt 7AA of the Act. The Authority conducts an odd and rather limited form of de novo review, in which it cannot seek or receive new material, or interview applicants. In observance of the detailed procedural requirements of that Fast Track regime, the Secretary of the Department provided specified material to the Authority that included all material before the delegate and ‘any other material that is in the Secretary’s possession or control and is considered by the Secretary ... to be relevant to the review’. That material included irrelevant and prejudicial information about the applicant’s involvement in disturbances while in detention, his transfer to prison, many comments from immigration officials about his supposedly difficult or aggressive behaviour and that he had been interviewed by the National Security Monitoring Section of the Department. The Authority is expressly obliged to conduct its review function by considering the material supplied by the Secretary. The Authority duly noted that it had done so, though it did not make any specific reference in its reasons to the prejudicial material just noted, when affirming the delegate’s decision.
The Authority’s decision was upheld by the Federal Circuit Court of Australia, and by a majority of the Full Court of the Federal Court. In the Federal Court, Moshinsky and Thawley JJ each held that the majority of the irrelevant and prejudicial material was already before the Authority. The only significant exception was material that mentioned officials of the National Security Monitoring Section had interviewed the applicant. Their Honours held that reference to that interview was not itself enough to create a reasonable apprehension of bias. Justice Mortimer dissented, in large part because her Honour found that the content of the prejudicial material and the context in which it was sent could create a reasonable apprehension of bias. These different judgments highlight difficulties in recourse to the statutory regime in question and in the role of the additional step that Gageler J proposed in Isbester for the Ebner test.
CNY17 sheds light on a particular aspect of the bias test and informed observer construct that has been largely neglected: that is, the importance of the statutory context in which decisions are made and the extent to which the observer should be attributed knowledge of that scheme. The particular statutory context in which any decision is made is typically mentioned as one of several relevant issues, along with the factual context and the wider legal environment of the decision. Chief Justice Spigelman explained that the relevant statute ‘must be part of the assessment from the outset and not treated as some kind of qualification’ to common law principles. That suggestion draws useful attention to the fact that the great majority of bias cases place little attention on the legislative context in which decisions are made, or the many differences in different statutory regimes, or the singular importance that a statute can play in the work of administrative officials. This issue is vital in CNY17 because the claim of apprehended bias arose in the context of an extremely detailed and complex regime of procedures governing the process of administrative review of an unsuccessful claim for a protection visa. Whether an apprehension of bias could reasonably be found to have arisen was not determined, in the words of Spigelman CJ, with simply an ‘assessment’ of the applicable legislative context, but instead an expert and microscopic analysis of that scheme. The attribution of such detailed knowledge of legislative complexity and detail enabled judicial interpretation of statute to overwhelm, arguably even replace, any real use of an assessment by the informed observer.
The Migration Act is complex and the provisions that establish the Fast Track Review Process in pt 7AA of the Act are no exception. They span twenty pages and contain a large number of precise procedural requirements and many particular restrictions. All members of the Full Federal Court accepted that the informed observer should be aware of this statutory context, but the difficulty lay in the detail that the observer should be held to know. Justice Moshinsky skirted over the detail of the legislative regime and instead drew attention to just two key features of the scheme: namely, the interrelated ones of the duty of the Secretary to provide information he deemed relevant to the Authority and the obligation of the Authority to consider the material provided by that process. Justice Moshinsky essentially found no reasonable apprehension arose because the informed observer would accept it was likely the Authority could and would put the prejudicial information aside and, in any case, those parts of the information the applicant complained most about were not sufficiently prejudicial to create an apprehension of bias.
The simplified view of the process and confidence in the ability of the Authority stands in strong contrast to the reasoning of Mortimer J. While her Honour limited her analysis to several features of the scheme she identified as ‘critical’, that lengthy assessment included very fine details about the role of the Secretary of the Department in giving review material to the Authority and his role as the ‘fundamental source’ of information for the Authority. Justice Mortimer also emphasised the Secretary’s role in holding and controlling material, and also determining what material he considers relevant to review by the Authority. Her Honour also explained the nature of the review process in great detail, including the many constraints on the process. Justice Mortimer concluded that the Court ‘can and should assume’ that the Authority knows how pt 7AA operates, especially the process by which the Secretary provided material to the Authority. Her Honour made clear that the informed observer ‘has the same understanding’. Her Honour also reasoned that the observer would understand the ‘overwhelming focus’ of the Authority fixed upon what had been regarded as relevant and provided by the Secretary. This knowledge almost certainly included Mortimer J’s later finding that the information provided by the Secretary ‘carried with it the premise that it was relevant’ to the Authority’s task and would presumably therefore be considered.
Justice Mortimer thought the informed observer would accept the relatively limited capacity of administrative members of the Authority to put prejudicial material out of their mind. That approach drew support from other cases, including ones about the Fast Track Review Process, which have accepted that the content of the bias rule and the expectations of the informed observer vary according to the character of decision-makers and the nature of the process in which they work. Some such cases have accepted that the particular qualifications or experience of decision-makers enables the informed observer to have considerable faith in the capacities of their skills. Such cases have clear logic. After all, the environments in which decisions are made and the particular processes by which they are made will vary enormously. An observer could and should be expected to be aware, and accepting, of many of those differences. But would the observer be as sceptical as Mortimer J appeared to be about the effect of a lack of legal qualifications of decision-makers? The observer could just as easily conclude that legal qualifications and experience might add little to the highly prescriptive task that the Migration Act creates for officials in the Authority. The observer might also think that the utility of legal qualifications is simply overrated. Such assessments about the qualities of the informed observer are ones judges are not well placed to make.
Justice Thawley also recounted the administrative process in some detail, listing 12 key features of the review process that his Honour thought relevant to the case. His Honour found that the irrelevant material was not sufficiently prejudicial and confirmed this by further reference to several key parts of the review process that the informed observer would know, including that many key tasks were performed on behalf of, rather than by, the Secretary himself, such as the actual referral of the applicant’s claim to the Authority. Justice Thawley also thought that the informed observer would understand the many details of the review process, including: the different tasks of the original decision-maker and the Authority in its limited review function; the reasons why the Secretary was required to provide information to the Authority; the statutory imperatives that the Secretary refer decisions and send on material as soon as practicable; and the overall mission of the Authority to act efficiently, quickly and free from bias.
The judgments of Mortimer and Thawley JJ both attributed the informed observer with detailed understanding of the key points of a complex administrative regime. On any measure, to attribute that level of knowledge to an observer who is meant to be reasonably, but not entirely, informed of relevant details is nonsense. The intricate procedures governing the Fast Track Review Process include details that elude most lawyers. An ordinary or reasonable person would surely ‘tune out’ many such details. Why should informed observers be any different? There are good reasons why they should not be. One is the caution of Kirby J that the observer had been stretched ‘virtually to snapping point’ by being attributed with far too much knowledge. That problem continues unabated if the observer is attributed with detailed knowledge about the finer details of complex administrative regimes such as the Fast Track process. That is the sort of knowledge that only judges, migration officials and specialist migration lawyers hold. Transposing such detailed knowledge to the informed observer simply enables that person, and his or her judgments about bias, to more closely align with that of the judge responsible for the transposing.
At this point, it is useful to consider the wider issue of detailed legislation and the problem it poses to the informed observer. While opinions might differ on the correct interpretation of the Fast Track Review Process, or the extent to which the informed observer should be attributed with knowledge of that procedure, there can be little doubt that this regime is symptomatic of the increasing volume and complexity of statutes. The Full Federal Court did not consider the more general questions of the extent to which the device of the informed observer could or should accommodate this growing complexity of statues, but the analysis of the Full Court invites some further questions. One is whether the informed observer could be expected to take particular care in administrative processes that have grave potential impact upon people. That approach would place judges in the difficult position of deciding which decisions or interests attract particular care, and which do not. Another question that flows from judgments about the growing detail and complexity of modern legislation is the equivalent growth in the complexity of administrative processes. The Fast Track Review Process is arguably just one of many novel forms of administrative decision-making that have recently arisen in our migration processes. What would the informed observer know of the bureaucratic processes that underpin this and other such regimes? Would the observer carry the scepticism of bureaucrats that Australians pride themselves upon? The focus of the Full Court on legislative detail meant that these difficult questions were overlooked.
In Isbester, Gageler J identified a third step in Ebner — ‘consideration of the reasonableness of the apprehension of’ the deviation from impartiality as suggested by the party claiming bias. The unsettled nature of this third step in the test for apprehended bias was acknowledged when special leave was sought for CNY17, and was treated with caution by the Full Federal Court in that case. Other courts have acknowledged the third step suggested by Gageler J, often in cautious terms and generally without any acknowledgement of the silence of the majority in Isbester on this matter. One case purported to adopt the approach of Gageler J, but in fact treated the new third step as no different to Ebner’s second step by compressing both together. Another applied the third step as a statement of conclusion rather than with detailed reasoning.
Aronson, Groves and Weeks criticised this ‘possible’ third step as one that ‘may do little more than articulate openly a commonsense judgment that can and should occur as part of the second step of Ebner’. Those authors also thought this proposed third step redundant because an apprehension that was not reasonable would also ‘surely be difficult, if not impossible, to articulate under Ebner’s second limb’. Aronson, Groves and Weeks further criticise this third possible step as a holistic assessment that is sometimes already used by courts faced with bias claims that are based on a wide range of matters.
The third step suggested by Gageler J may also be likely to raise the problem of confirmation bias. That concept refers to tendency of people to simultaneously seek out information that is consistent with their established views, while also ignoring or diminishing information that might contradict those established views. Studies suggest that judges are as prone to confirmation bias as members of other professions. The application by Gageler J of his proposed third step in Isbester itself arguably provides an example of confirmation bias. His Honour found that the apparently contradictory roles of the official in question would enable the fictional observer to hold a reasonable apprehension of bias. His Honour concluded by noting that the reasonableness of that apprehension ‘is not negatived by the [key] circumstances’ of the case. That conclusion is arguably akin to a suggestion that the conclusion reached was confirmed again because it met the requirement of the third step: namely, to be reasonable in the circumstances. When characterised in this way, his Honour’s conclusion edges towards confirmation bias. Even if not couched in those terms, the loaded nature of the third step may be identified another way. If an apprehension of bias is established after the second step, which means that the effect of the claimed source of impartiality has been explained and accepted as giving rise to a reasonable apprehension of bias, how then can the judge who reached that conclusion then disclaim it on the ground that it is somehow ‘not reasonable’ as required by the third step? The difficulty is obvious.
The same can be said about the use by both Mortimer J and Thawley J of the third step in the Ebner test suggested by Gageler J. Both used that step. Both did so to effectively confirm the judgement they had reached about the existence of a reasonable apprehension of bias. Both reached a different conclusion. That result indicates that the test suggested by Gageler J would add little but pointless complication to Ebner’s test. This complication would have consequences. The bias test would become more obscure. A further requirement of reasonableness would also enable courts to find that impartiality is thought by the observer to be imperilled (under Ebner’s second step), but then find that apprehension is somehow not reasonable (under Gageler J’s third step). How could that possibility enhance public confidence in the law?
The value of the informed observer to determine bias claims has long been doubted. A longstanding criticism is that courts imbue the observer with so much knowledge that the fictional person is a flimsy veil for the views of judges. One suggested solution is for courts to discard the informed observer and decide bias claims by open reference to the views and conclusions of the judge who decides the issue. Chief Justice French accepted the force of these criticisms in British American Tobacco Australia Services Ltd v Laurie, but concluded the observer retained ‘utility as a guide to decision-making in this difficult area’ by providing a means for judges to view issues ‘as best they can, through the eyes of non-judicial observers’. The adoption of a third possible step in Ebner’s test hampers that function because it provides an artificial means for judges to further transpose their views onto those of the observer, at the expense of public confidence in the courts and the administration of justice.
 Alfred Deakin Professor of Law, Deakin Law School, Deakin University, Victoria, Australia. Thanks are due to Emily Hammond, Before the High Court Editor of the Sydney Law Review, for helpful comments.
  HCA 63; (2000) 205 CLR 337 (‘Ebner’).
 High Court of Australia, Case No M72/2019, appealed from CNY17 v Minister for Immigration and Border Protection  FCAFC 159; (2018) 264 FCR 87 (‘CNY17 (FCA)’).
 Ebner (n 1) 348  (Gleeson CJ, McHugh, Gummow and Hayne JJ; Callinan J agreeing at 396 ).
 Claims of actual bias require investigation and assessment of the state of mind of the decision-maker. The key issue is not how things appear, but how they actually are: Michael Wilson & Partners Ltd v Nicholls (2011) 244 CLR 427, 437–8  (Gummow A-CJ, Hayne Crennan and Bell JJ) (‘Wilson & Partners’).
 Ebner (n 1) 344–5  (Gleeson CJ, McHugh, Gummow and Hayne JJ; Callinan J agreeing at 396 ). This test is remarkably similar to the American one that the decision-maker’s ‘attitude or state of mind leads a detached observer to conclude that a fair and impartial hearing is unlikely’: Liteky v United States, 510 US 540 (1994) 564 (Scalia J).
 R v Bow Street Metropolitan Stipendiary Magistrate; Ex parte Pinochet Ugarte (No 2) 
1 AC 119 (‘Pinochet (No 2)’); Ebner (n 1); Helow v Secretary of State for the Home Department  2 All ER 1031; Wilson & Partners (n 4).
 See, eg, Webb v The Queen  HCA 30; (1994) 181 CLR 41 (‘Webb’); R v Goodall  VSCA 63; (2007) 15 VR 673;
R v Abdroikov  1 WLR 2679.
 See, eg, Laws v Australian Broadcasting Tribunal  HCA 31; (1990) 170 CLR 70 (‘Laws v ABC’);
Re Minister for Immigration and Multicultural Affairs; Ex parte Epeabaka  HCA 23; (2001) 206 CLR 128; Re Refugee Review Tribunal; Ex parte H  HCA 28; (2001) 179 ALR 425; Gillies v Secretary of State for Work and Pensions  UKHL 2;  1 All ER 731.
 The inquisitorial nature of coronial proceedings and the novel powers granted to coroners for that jurisdiction greatly affect the bias rule in this context. See, eg, R v Doogan; Ex parte Lucas-Smith  ACTSC 74; (2005) 158 ACTR 1; Leahy v Barnes  QSC 226.
 See, eg, Porter v Magill  2 AC 357 (‘Porter’); Man O’War Station Ltd v Auckland City Council (No 1)  3 NZLR 577; McGovern v Ku-Ring-Gai Council  NSWCA 209; (2008) 72 NSWLR 504 (‘McGovern’).
 See, eg, Minister for Immigration and Multicultural Affairs v Jia (2001) 205 CLR 507 (‘Jia’);
Hot Holdings Pty Ltd v Creasy  HCA 51; (2002) 210 CLR 438 (‘Hot Holdings’).
 See, eg, Isbester v Knox City Council  HCA 20; (2015) 255 CLR 135 (‘Isbester’) (administrative official assisting council body); Kwan v Victoria Legal Aid  VSC 122 (legal aid official).
 Chief Justice Gleeson, for example, suggested that ‘to be judicial is to be impartial’: Chief Justice Murray Gleeson, The Rule of Law and the Constitution (ABC Books, 2000) 129. This statement echoes Bentham, who declared ‘[i]t is the duty of the judge to be impartial; — therefore it is his duty to be partial’: Jeremy Bentham, Rationale of Judicial Evidence, Works (1843) vol VI, 350.
 The vast research on this issue largely followed the work begun by Tyler in the 1970s and has demonstrated a close connection between perceptions of the legitimacy of officials and institutions and the level of fairness and impartiality they offer. Tyler sees judges and courts as specific examples of this phenomena: Tom Tyler, ‘Procedural Justice and Policing: A Rush to Judgment’ (2017) 13 Annual Review of Law and Social Science 29.
 Lord Woolf et al, De Smith’s Judicial Review (Sweet & Maxwell, 8th ed, 2018) 537–8 (fn 17) cite (1371) 45 Lib Ass 3, which was an assizes case involving a conflict of interest resulting in the disqualification of one judge.
 R v Watson; Ex parte Armstrong  HCA 39; (1976) 136 CLR 248, 263 (Barwick CJ, Gibbs, Stephen and Mason JJ).
 Livesey v New South Wales Bar Association  HCA 17; (1983) 151 CLR 288, 300 (Mason, Murphy, Brennan, Deane and Dawson JJ); Laws v ABC (n 8) 87 (Mason CJ and Brennan J).
 Vakauta v Kelly  HCA 44; (1989) 167 CLR 568, 576 (Dawson J) (‘Vakauta’).
 Ibid 585 (Toohey J).
 Ibid 573–4 (Brennan, Deane and Gaurdon JJ).
 Laws v ABC (n 8) 92 (Deane J).
 Webb (n 7).
 Ibid 52.
 R v Gough  UKHL 1;  AC 646.
 That change in UK law occurred in Porter (n 10).
 Ebner (n 1).
 That principle can be traced to Dimes v Grand Junction Canal Proprietors  EngR 789; (1852) 3 HL Cas 759 (‘Dimes’). Strictly speaking, the High Court did not reject automatic disqualification, but instead held that Dimes had been wrongly interpreted as equating the Lord Chancellor’s shareholding in a litigant company with a pecuniary interest that required automatic disqualification: Ebner (n 1) 355–8 – (Gleeson, McHugh, Gummow and Hayne JJ; Callinan J agreeing at 396 ). Justice Kirby flatly rejected this ‘ahistorical interpretation’ of Dimes: 378 .
 The basis of this finding was the long and close association of Lord Hoffmann with the intervening party meant that his Lordship was too closely identified with the causes of the intervener.
 Ebner (n 1) 345  (Gleeson CJ, McHugh, Gummow and Hayne JJ; Callinan J agreeing at 396 ) (emphasis added).
 This occurred in both Isbester (n 12) and Wilson & Partners (n 4).
 Isbester (n 12). The first proceeding was a criminal prosecution against the dog owner, who pleaded guilty in relation to attacks by the dog. The second proceeding, in which the apprehension of bias was found, was an administrative one to consider whether the dog should be destroyed, as allowed under the Domestic Animals Act 1994 (Vic) s 84P(e).
 Isbester (n 12) 151 .
 Ibid 146 –.
 Ibid 152–3 –.
 Ibid 153 .
 Ibid 155–6 . The three steps identified by Gageler J are distinct from the three contentions that Hayne J has suggested are raised in claims of prejudgment, which are that a decision-maker: ‘has an opinion on a relevant aspect of the matter in issue’; ‘will apply that opinion to that matter in issue’; and ‘will do so without giving fresh consideration’ in light of the circumstances at hand: Jia (n 11) 564 .
 Ebner (n 1) 345  (Gleeson CJ, McHugh, Gummow and Hayne JJ; Callinan J agreeing at 396 ).
 The ‘bar’ in s 46A is a general obstacle for those who arrive by sea (an ‘unauthorised maritime arrival’) and wish to apply for any form of visa. Applications are essentially prohibited by a provision that deems applications invalid: Migration Act s 46A(1). The Minister is granted a non-compellable discretionary power to vary that requirement, or ‘lift the bar’: Migration Act ss 46A(2), 46A(3)–(7).
 CNY17 (FCA) (n 2) 104  (Moshinsky J).
 Justice Moshinsky explained the reasons of the delegate: ibid 105–6 –.
 This regime is analysed in Emily McDonald and Maria O’Sullivan, ‘Protecting Vulnerable Refugees: Procedural Fairness in the Australian Fast Track Regime’  UNSWLawJl 34; (2018) 41(3) University of New South Wales Law Journal 1003.
 The fast track regime was described as providing ‘de novo consideration of the merits’ in Plaintiff M174/2016 v Minister for Immigration and Border Protection (2018) 92 ALJR 481, 487  (Gageler, Keane and Nettle JJ). The process was also described as a de novo one by Gordon J at 497  and Edelman J at 498 . The many procedural restrictions imposed upon the Authority prevent it from undertaking anything approaching a full consideration of the merits of a claim, so the process may be better described as a limited form of reconsideration.
 Migration Act (n 41) s 473DB(1)(a)–(b).
 Ibid s 473CB(1)(c).
 Ibid s 473DB(1).
 CNY17 v Minister for Immigration and Border Protection  FCCA 2731.
 CNY17 (FCA) (n 2) 116  (Moshinsky J), 124  (Thawley J).
 Ibid 116–7  (Moshinsky J), 124–6 – (Thawley J).
 Ibid 96–101 –.
 See, eg, Isbester (n 12) where Kiefel, Bell, Keane and Nettle JJ described the question of the judgement of the informed observer as ‘largely a factual one, albeit one which it is necessary to consider in the legal, statutory and factual contexts in which the decision is made’: at 146 .
 McGovern (n 10) 507 .
 CNY17 (FCA) (n 2) 115 .
 Ibid 116–17 .
 Ibid 94 .
 Ibid 94 .
 Ibid 95 .
 Ibid 95 .
 Ibid 101 . Justice Mortimer did not expressly state that the informed observer would be attributed with this assumption, but the tenor of her Honour’s analysis appears to assume it.
 See, eg, Minister for Immigration and Border Protection v AMA16, where Griffith J rejected arguments that decision-makers in the Fast Track Review Process had the qualifications or expertise of legally qualified or experienced members of specialist tribunals:  FCAFC 136; (2017) 254 FCR 534, 551 . This had led courts to find in other administrative regimes that the informed observer would trust in the ability of suitably qualified and experienced officials to cast aside irrelevant or prejudicial material.
 See, eg, Jia (n 11); Hot Holdings (n 11). Each case stressed the need to take account of the particular nature of ministerial decision-making in reaching assessments about apprehensions of bias.
 See, eg, O’Sullivan Medical Tribunal of NSW  NSWCA 374.
 CNY17 (FCA) (n 2) 120–1 –.
 Ibid 125–6 .
 Ibid 126 .
 The observer does not have detailed knowledge of the law: Johnson v Johnson (2000) 201 CLR 488, 493  (Gleeson, CJ, Gaudron, McHugh, Gummow and Hayne JJ), or legal knowledge that ‘ordinary experience suggests not to be the case’: Vakauta (n 18) 585 (Toohey J; Brennan, Deane and Gaudron JJ agreeing at 570).
 Smits v Roach (2006) 227 CLR 423, 457 .
 The Full Federal Court has also cautioned 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  FCAFC 33; (2011) 192 FCR 254, 273 .
 Isbester (n 12) 156 .
 Transcript of Proceedings, CNY17 v Minister for Immigration and Border Protection  HCA Trans 101, 10–15 (LG De Ferrari) (where counsel for the applicant acknowledged the ‘possible’ third step suggested by Gageler J in Isbester).
 CNY17 (FCA) (n 2) 92  (Mortimer J put ‘to one side’ the question of whether Gageler J’s third step ‘represents a new articulation of the established test’). Justice Thawley appeared less hesitant about the third step: at 120 .
 Mackinnon v Partnership of Larter, Jones, Miraleste Pty Ltd and Johnson (No 4)  NSWSC 147, –. Justice Stevenson noted that Ebner’s second step and the third one suggested by Gageler J each ‘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’: at .
 Marlinspike Debt Acquisitions Pty Ltd v Undone Pty Ltd (No 2)  NSWSC 72, .
 Mark Aronson, Matthew Groves and Greg Weeks, Judicial Review of Administrative Action and Government Liability (Thomson Reuters, 6th ed, 2017) 651–2.
 Ibid 652.
 The foundational study used a simple but ingenuous card game to demonstrate key issues:
PC Wason, ‘On the Failure to Eliminate Hypotheses in a Conceptual Task’ (1960) 12(3) Quarterly Journal of Experimental Psychology 129. Wason only used the term confirmation bias years later: PC Wason, ‘Reasoning About a Rule’ (1968) 20(3) Journal of Experimental Psychology 273.
 Confirmation bias is not dissimilar to the ‘lock in effect’, which is the unwillingness of officials to significantly revise earlier decisions during later proceedings that involve the earlier decision. Cognitive bias means early decisions greatly influence later ones taken by the same official. See Kevin J Lynch, ‘The Lock-In Effect of Preliminary Injunctions’ (2014) 66(2) Florida Law Review 779.
 Andrew J Wistrich and Jeffrey J Rachlinski, ‘How Lawyers’ Intuitions Prolong Litigation’ (2013) 86(3) Southern California Law Review 571. That study showed judges were slightly less prone to confirmation bias using a scenario about discovery in civil proceedings rather than Wason’s classic card game.
 Isbester (n 12) 158–9 –.
 Ibid 159 .
 Abimbola A Olowofoyeku, ‘Bias and the Informed Observer: A Call for a Return to Gough’ (2009) 68(2) Cambridge Law Journal 388, 395.
 (2011) 242 CLR 283.
 Ibid 306 .