16. There is no straitjacket formula in a case of personal bias, for deciding, if, there is personal bias or not. But acceptable tests are (i) the real likelihood of bias; (ii) whether a reasonable intelligent man, fully apprised of all the circumstances, would feel a serious apprehension of bias and (iii) not to the extent that there should be clear evidence of bias/existence of bias. In Halsbury’s Laws of England, 4th Edn., Vol.2, para 551, it has been indicated that the test of bias is whether a reasonable intelligent man, fully apprised of all the circumstances, would feel a serious apprehension of bias. This principle had been accepted by the Apex Court in Manak Lal v. Dr. Prem Chand: AIR 1957 SC 425. De Smith in his Judicial review of Administrative Action, (1980) P.262 had considered what are “real likelihood of bias or reasonable suspicion of bias” and observed that:

“Real Likelihood of Bias or Reasonable Suspicion of Bias A “real likelihood” of bias means at least substantial possibility of bias. The court, it has been said, will judge of the matter “as a reasonable man would judge of any matter in the conduct of his own business”. The test of real likelihood of bias, which has been applied in a number of leading cases in magisterial and liquor licensing, is based on the reasonable apprehensions of a reasonable man fully apprised of the facts. It is no doubt desirable that all judges, like Caesar’s wife, should be above suspicion; but it would be hopeless for the courts to insist that only “people who cannot be suspected of improper motives” were qualified at common law to discharge judicial functions, or to quash decisions on the strength of the suspicions of fools or other capricious and unreasonable people.

What is the position if the court is satisfied, on the evidence before it, that there was no real likelihood of bias but is nevertheless of the opinion that a reasonable man, at the time when the decision under review was made, could well have suspected that the tribunal would be biased? Does the public interest nevertheless demand that the original decision be set aside? The cases do not speak with one voice on this matter. The courts have often quashed decisions strength of the reasonable suspicions of the party aggrieved, without having made any finding that a real likelihood of bias in fact existed. In 1954 the Divisional Court of the Queen’s Bench Division, after having reviewed the authorities, held that “real likelihood” was the proper test, and that a real likelihood of bias had to be “made to appear not only from the materials in fact ascertained by the party complaining, but from such further facts as he might readily have ascertained and easily verified in the course of his inquiries. “This might be regarded as but another formulation of the “reasonable suspicion” test, for it is the duty of a reasonable man to make reasonable inquiries; if, after having made his inquiries, he is still left with the impression that there was a real likelihood that the tribunal would be biased against him, it is the duty of the Court to quash the decision although, on the full information available to it, it is satisfied that his suspicions were unfounded? The answer given by Devlin L.J. in a subsequent case was in the negative. Whether a real likelihood of bias existed was to be “determined on the probabilities to be interfered from the circumstances, not upon the basis of the impressions that might reasonably be left on the minds of the party aggrieved or the public at large.

However, the pendulum has now swung towards a test of reasonable suspicion, founded on the apprehensions of a reasonable man who had taken reasonable steps to inform himself of the material facts. “Reasonable suspicion” tests look mainly to outward appearances, “real likelihood” tests focus on the court’s own evaluation of the probabilities; but in practice the tests have much in common with one another, and in the vast majority of cases they will lead to the same result. For the courts to retain both tests as alternative methods of approach is unlikely to cause serious uncertainty, and there may be advantages in preserving a measure of flexibility. It would be surprisingly, surely, if a court were to refuse to set aside a decision on the ground that a reasonable observer could not have discovered facts that subsequently came to light and which indicated to the court that there was a real likelihood of bias in the adjudicator.”

17. In the case of personal bias, two tests are there: (i) likelihood of bias and (ii) that there is reasonable case of bias. But nevertheless, the opinion of likelihood of bias is that of a reasonable man, at the time the decision under review was made, could well have suspected that the Tribunal would be bias? There is another formula “reasonable suspicion” tests for it is the duty of a reasonable man to make reasonable inquiries, if, after having made his inquiries, he is still left with the impression that there was a real likelihood that the tribunal would be biased against him, is it is the duty of the Court to quash the decision. De Smith observed that the apprehension of a reasonable man who had taken reasonable steps to inform himself of the material facts and “reasonable suspicion” tests look mainly to outward appearances. The inferences of malafide can be drawn only on the basis of factual matrix and not merely on the basis of insinuations, conjectures and surmises. (Ref:- M Sankarannarayan, IAS v. State of Karnataka and others: (1993) 1 SCC 54).