THE TEST OF ACTUAL BIAS
24 In Abebe v Commonwealth (1999) 197 CLR 510, a majority of the High Court upheld the constitutionality of Part 8 of the Migration Act, insofar as it restricts the jurisdiction of the Federal Court to review decisions made under the Act. The consequence of this decision is that in certain important respects the jurisdiction conferred on the Federal Court under Part 8 of the Migration Act is narrower than the jurisdiction conferred on the High Court under s 75(v) of the Constitution, narrower than the scope for review of administrative action at common law and narrower than the grounds for review of administrative action set out in the Administrative Decisions (Judicial Review) Act 1977 (Cth): Abebe v Commonwealth, at 522, per Gleeson CJ and McHugh J. See M Crock, "Of Fortress Australia and Castles in the Air: The High Court and the Judicial Review of Migration Decisions" (2000) 24 Melb Uni LR 190, at 202-203. More specifically, the statutory scheme precludes the Federal Court from reviewing a Tribunal's decision on the ground of apprehended bias. In a case where bias is alleged, the Court can grant relief only if the decision is affected by actual bias.
25 In Sun v Minister, a majority of the Full Court (Burchett and North JJ; Wilcox J not deciding) held that a Tribunal's decision had been affected by actual bias. Burchett J pointed out (at 126) that actual bias, like any other conclusion of fact, may be established as an inference from the circumstances. His Honour continued (at 126-127):
"I think [the Court] should interpret the words of s 476(1)(f) in their natural sense. The use of the word 'actual' strongly suggests that the legislature was endeavouring to get away from the somewhat special concept of bias which is immanent in the case law. The cases use expressions such as 'real likelihood of bias' in a sense designed to ensure 'that it is not necessary that actual bias should be proved": R v Barnsley Licensing Justices; Ex parte Barnsley and District Licensed Victuallers' Association [1960] 2 QB 167 at 187, per Devlin LJ. In Australia, 'real likelihood' has itself been abandoned in favour of an even more special test, based on the principle that justice should be undoubtedly seen to be done, of 'whether fair-minded people might reasonably apprehend or suspect that the judge has prejudged or might prejudge the case': Webb v The Queen (1994) 181 CLR 41 at 47; Gaisford v Hunt (No 2) (1996) 71 FCR 187. The doctrine that has been developed, however it is stated, is commonly called the rule of apprehended bias. In my opinion, the statute, when it used Devlin LJ's expression 'actual bias', submitted a test that looks to whether the Tribunal has, at least in some respect, prejudged the case, so as to be unable or unwilling to decide it impartially. I say 'at least in some respect' because the statute extends to the situation where 'the decision was…affected…by actual bias'. The observation that it is sufficient that a decision be affected by bias adds weight to the conclusion of Finkelstein J in Bilgin [v Minister for Immigration and Multicultural Affairs (1997) 149 ALR 281], with which I agree, that actual bias need not be confined to an intentional state of mind. Bias may be subconscious, provided it is real. Nevertheless, a finding of bias is a grave matter, different in kind from a finding of mere error, or even wrong-headedness, whether in law, logic, or approach. It would be a sad reflection on administrative tribunals, and certainly on courts which exist in the name of justice, if it were to be seen as other than exceptional."
26 North J stated (at 134) that
"[a]ctual bias exists where the decision-maker has prejudged the case against the applicant, or acted with such partisanship or hostility as to show that the decision-maker had a mind made up against the applicant and was not open to persuasion in favour of the applicant."
His Honour relied on a Canadian authority (Re Gooliah and Minister of Citizenship and Immigration (1967) 63 DLR (2d) 224 (Man CA)) to support two propositions (at 135):
"proof of actual bias by inference from the facts and circumstances of the case will usually involve an assessment of a series of actions by the decision-maker which, when taken together, form a whole picture leading to the conclusion of pre-judgment. It is unlikely that one single action, as distinct from a pattern of conduct, will demonstrate actual bias.
Gooliah also demonstrates that actual bias does not necessarily involve deliberate, knowing, or wilful prejudice against an applicant."
27 North J considered that since a finding of actual bias may be made without a judge being aware of his or her prejudices or prejudgments, any "special reticence" in pursuing a case of actual bias should be diminished. His Honour characterised (at 136) an unintended prejudgment as an error similar to other appellable errors, involving no impropriety or misconduct on the part of the decision-maker. He stated that applicants "should not feel restricted by the nature of the complaint from pursuing" the ground of actual bias.
28 There seems to be a significant difference of emphasis between the judgments of Burchett J and North J, although both stated the primary test of actual bias in similar terms. Both judgments also seem to accept that errors in reasoning and fact-finding may be so egregious as to warrant an inference that the decision-maker has prejudged the case to the point of being unable to decide it impartially. Burchett J nonetheless regarded a finding of actual bias as a "grave matter", different from "mere error, or even wrong-headedness". North J, extrapolating from the notion that decision-makers may not recognise their own prejudices, appears to suggest that the boundaries are less clear. On one view, his Honour's judgment extends an invitation to practitioners to challenge poor quality decision-making by invoking "unintended actual bias" as a ground of review. If taken up, the invitation may require this Court to examine closely the factual findings made by the decision-maker in order to determine whether there are errors and, if so, whether their cumulative effect is to suggest unconscious prejudice or irreversible pre-judgment.
29 Later decisions have not needed to resolve the apparent differences between the majority judgments in Sun. They have, however, tended to emphasise that actual bias involves a closed mind or an unalterable pre-judgment. In Jia v Minister for Immigration and Multicultural Affairs (1999) 93 FCR 556, a majority of the Full Court held that the Minister's decision to cancel a visa was affected by actual bias in consequence of statements he had made about the particular case he had to decide. Spender J considered (at 566) the test of actual bias to be "plain":
"A decision-maker is biased if the decision-maker has made up his or her mind on the matter to be decided. If a decision-maker has prejudged the matter, a decision giving effect to that prejudgment is a biased decision."
R D Nicholson J, the other member of the majority, after considering the judgments in Sun v Minister, stated the principle this way (at 598):
"for actual bias to be established there must be evidence of 'a closed mind to the issues raised', 'preliminary views incapable of alteration', prejudgment of the case at least in some respect, real although not necessarily intentional."
30 In Li v Minister for Immigration and Multicultural Affairs (2000) 96 FCR 125, Drummond J summarised the effect of the authorities as follows (at 133-134):
(a) Actual bias exists where the decision-maker has prejudged the case against the applicant, or has acted with such partisanship or hostility as to show that the decision-maker had a mind made up against the applicant and was not open to persuasion in favour of the applicant.
(b) The emphasis is upon the state of mind which affects the decision-making rather than elements of the process of decision-making taken in isolation.
(c) Proof of an intentional state of mind adverse to the party is not the only way of proving actual bias. Such bias may be subconscious, provided it is real.
(d) It is not sufficient proof of actual bias to show that the decision-maker has expressed views adverse to the party's position at an early stage of the proceedings unless there is also proof that those views were incapable of being changed in the course of the proceedings. Though relevant to proof of actual bias, displays of irritation or impatience and the use of sarcasm by the decision-maker during the hearing are not, without more, generally sufficient to establish such bias, proof of which requires a finding on a question of fact, having regard to all the circumstances of the case." (Citations omitted.)
31 It is not necessary in the present case to explore the significance of the differences in emphasis between the majority judgments in Sun v Minister. As will be apparent, I think that on the most generous view of "actual bias" the present applicant has failed to establish it. But I wish to associate myself with the distinction drawn by Burchett J between actual bias and "mere error, or even wrongheadedness, whether in law, logic or approach".
32 One of the foundations of judicial review of administrative action in Australia is, in the words of Brennan J in Attorney-General v Quin (1990) 170 CLR 1, at 36, that
"the court has no jurisdiction simply to cure administrative injustice or error. The merits of administrative action, to the extent they can be distinguished from legality, are for the repository of the relevant power and, subject to political control, for the repository alone".
In my view, it would threaten that principle, as well as do violence to the restrictive language of s 476(1)(f) of the Migration Act, to make the leap too readily from factual error or faulty reasoning (even serious factual error or misconceived reasoning) to a finding of actual bias. Such a finding implies a closed mind on the part of the decision-maker, or an irreversible prejudgment of some aspect of the case such as to make the decision-maker unable or unwilling to decide the case impartially. The fact that the decision-maker does not realise that he or she has acted in this way does not detract from the seriousness of the finding.
33 It is also important to appreciate, as Drummond J pointed out in Li v Minister, at 134-135, that a claim of actual bias must take into account the legislative framework within which tribunals operate. In a court the impartiality of the judge is protected and preserved by rules limiting his or her involvement in the forensic process, including the gathering of evidence. By contrast, tribunals have a variety of fact finding and information gathering functions that are characteristic of an administrative decision-making process. They are often required to play an active role not only in eliciting information but in testing the reliability of evidence presented or relied upon by a particular applicant. Their procedures necessarily differ from those of courts.
34 The point can be illustrated by the present case. Division 5 of Part 5 of the Migration Act, confers a variety of powers and imposes a variety of duties and limitations on the Tribunal: see s 359 (power to seek relevant information); s 359A (duty to give to the applicant particulars of information that the Tribunal considers would be the reason, or part of the reason, for affirming the decision under review); s 360 (the applicant must be invited to appear and give evidence); s 361 (the applicant may request witnesses to be called, but the Tribunal is not bound to comply); s 366D (a person is not entitled to examine or cross-examine any person appearing before the Tribunal to give evidence).
35 In accordance with the requirements of s 359A of the Migration Act, the Tribunal in the present case sent a letter to the sponsor on 19 August 1999. The letter specified allegations that had been made against him and outlined evidence adverse to his claim to have maintained a genuine marital relationship with the visa applicant. To an observer unfamiliar with s 359A of the Migration Act, or with the functions of the Tribunal, this procedure might perhaps suggest that the Tribunal had already formed an adverse view of the sponsor's claims. Certainly, the role performed by the Tribunal is not one that a court exercising judicial functions would be expected to perform. But it cannot be indicative of actual bias to follow the very procedure mandated by the statute.
36 There may, of course, be cases where the decision-maker exhibits such overt partisanship or hostility towards a person as to warrant an inference that the decision-maker has a closed mind or has prejudged the case. In the absence of such partisanship or hostility, however, I think that considerable care should be taken before inferring from errors in reasoning or fact finding that a claim of actual bias has been made out. I agree with Burchett J that such cases are likely to be truly exceptional.