As the leading practitioner text explains ever since the Divisional Court’s decision in the case of Bryant and Bench v The Law Society [2007] EWHC 3043 (Admin) it has been settled law that the correct test to be applied in disciplinary proceedings is the two-stage test for dishonesty used in both the criminal and the civil jurisdictions, namely the R v Ghosh [1982] 1 QB 1053 and Twinsectra Limited v Yardley [2002] 2 AC 164 tests. This test comprises:

  • the objective test – whether according to the ordinary standards of reasonable and honest people, what was done was dishonest;

followed by

  • the subjective test – if it was, then the fact finder must consider whether the defendant himself realised that what he was doing was by those standards dishonest.

As practitioners will know, this has typically led to the use of the Ghosh test across a wide range of disciplinary jurisdictions seemingly without problem. However, in the recent case of R(Professional Standards Authority for Health and Social Care) v Health and Care Professions Council and others [2014] EWHC 4657 (Admin) Mr. Justice Popplewell concluded at section 42 that the advice of the Legal Assessor to the Conduct and Competence Committee of the Health and Care Professions Council (“the CCC”) that in considering the question of dishonesty “they should apply the civil standard of proof and that the test of dishonesty that they should apply was that set out in Ghosh” gave rise to a difficulty sufficient to result in the CCC applying the wrong standard of proof because the use of the word “must” in the subjective limb of the Ghosh test meant that the “the subjective element formulated in Ghosh  is expressed in the language of the criminal standard of proof; “The person must have realised” that what he was doing was by ordinary standards dishonest.”  The CCC of the HCPC are meant to apply the civil standard when considering the factual particulars in a fitness to practise case. Mr. Justice Popplewell concluded that as a result of the advice of the legal assessor the CCC had applied the criminal standard, see section 47.

This article considers the ramifications of the decision in R(PSA) and examines the appropriateness, or otherwise of the two alternatives to the Ghosh test suggested by Popplewell J in that case.

The Ghosh Test

The Ghosh test derives from the judgment of Lord Lane CJ (at p.1064) in the criminal case of R v Ghosh and has two limbs,

  1. the jury must first decide whether according to the ordinary standards of reasonable and honesty people what was done was dishonest and
  2. if it was dishonest by those standards then the jury must consider whether the defendant himself must have realised that what he was doing was by those standards dishonest.

The Twinsectra Test

By contrast the test derived from the House of Lords in Twinsectra whilst it also has two limbs, an objective limb and a subjective limb, does not require that in order to satisfy the subjective limb the fact finder has to be satisfied that the defendant “must” have realised that his actions were by the standards of ordinary people dishonest. In the test set out by Lord Hutton at section 27 the standard simply requires that “before there can be a finding of dishonesty it must be established that the defendant’s conduct was dishonest by the ordinary standards of reasonable and honest people and that he himself realised that by those standards his conduct was dishonest.” At section 36 Lord Hutton phrases the subjective limb of the test in the following way, “that dishonesty requires knowledge by the defendant that what he was doing would be regarded as dishonest by honest people.

It should be said that around the time that the Divisional Court made its decision in the case of Bryant there had been some dispute as to the correct interpretation of the Twinsectra test following the Privy Council’s decision in the case of Barlow Clowes International Ltd v Eurotrust International Limited [2005] UKPC 37, [2006] 1 WLR 1476 and the Court of Appeal decision in Abou-Rahmah v Abacha [2006] EWCA Civ 1492.

These latter cases stated that the interpretation of Twinsectra which required both an objective element and a subjective element in the test for dishonesty was wrong and that therefore it was not a requirement of the standard of dishonesty that a defendant should be conscious of his wrongdoing. At section 65-6 of Abou-Rahmah Arden LJ said this: 

The decision of the House of Lords in Twinsectra …. was widely interpreted as requiring both an objective and subjective test to be applied to the question of standard. … The Privy Council in Barlow Clowes had now clarified that this is a wrong interpretation of the Twinsectra decision. It is not a requirement of the standard of dishonesty that the defendant should be conscious of his wrongdoing. The test of dishonesty is predominantly objective: did the conduct of the defendant fall below the normally acceptable standard?

In Bryant, however, the Divisional Court was asked to consider how the decisions in Barlow Clowes and Abou-Rahmah affected the correct interpretation of the Twinsectra test in the context of disciplinary proceedings. In coming to their decision the Divisional Court analysed the competing interpretations of the Twinsectra test and the decisions of a number of High Court decisions in professional discipline appeals involving dishonesty, specifically the case of Court of Appeal’s decision in the case of Bultitude v The Law Society [2004] EWCA Civ 1853. Having done so the Court approved the interpretation of the Twinsectra test that had been applied in Bultitude namely that any finding of dishonesty required first a decision as to whether the solicitor had acted dishonestly by the ordinary standards of reasonable and honest people, and if so whether he aware that by those standards he was acting dishonestly. 

The Divisional Court in Bryant then effectively endorsed a specific interpretation of the Twinsectra test for the purposes of disciplinary proceedings saying at section 153 that:

“the test to be applied in the context of solicitors disciplinary proceedings is the Twinsectra test as it was widely understood before Barlow Clowes, that is a test that includes a separate subjective element. The fact that the Privy Council in Barlow Clowes has subsequently placed a different interpretation on Twinsectra for the purposes of the accessory liability principle does not alter the substance of the test accepted in Bultitude and does not call for any departure from that test.”

And expressing the appropriate test at section155 in the following terms:

Accordingly, the tribunal in the present case, should, in our judgment have asked itself two questions when deciding the issue of dishonesty: first, whether Mr. Bryant acted dishonestly by the ordinary standards of reasonable and honest people; and, secondly, whether he was aware that by those standards he was acting dishonestly.”

“Must have”

Mr. Justice Popplewell’s difficulty with the Ghosh test arose simply from the use of the word “must” in the subjective limb of the test, i.e that the Registrant “must have realised that what he was doing was …. dishonest.” The Concise Oxford Dictionary defines “must have” in conjunction with the verb to do as

  1. “surely did or has done” and
  2. “necessarily would have done.”

If one applies that definition to the context of disciplinary proceedings it is clear why the learned judge felt that to ask a panel to find, on the balance of probabilities that a Registrant surely did, or necessarily would have known something, a requirement which appears to necessitate the exclusion of all other possibilities, might prove if not impossible certainly contradictory and confusing.

It is certainly arguable that the use of the word “must” in the subjective limb of the Ghosh test, isn’t necessary and can, as it did in the R(PSA) simply confuse a committee about precisely what it is being asked to find. Certainly the use of the word “must” in the Ghosh test doesn’t change the elements that it is necessary to have found proved in order to support a finding of dishonesty. On those, the Ghosh test and the Twinsectra test (as applied in Bryant) are the same, namely

  1. that the defendant’s conduct was dishonest by the ordinary standards of reasonable and honest people and
  2. that the defendant knew / realised that what he was doing was, by those standards dishonest.

Rather the purpose of the word “must” may simply be to provide a route map to the jury to enable them to make a finding of dishonesty in a jurisdiction where it may be rare that a defendant will admit that he was dishonest e.g. he must have known because, even though he has not admitted it, given the circumstances he cannot but have known.

Popplewell J is not the first judge to speculate on whether or not the Ghoshtest, given its origin in criminal proceedings, is the appropriate test for use in tribunals which apply the civil standard of proof. In Uddin v General Medical Council [2012] EWHC 2669 Mr. Justice Singh, stated obiter at section 30 that: “care needs to be taken about applying a test which was devised in the context of criminal. As is clear from the standard of proof which is relevant in the present context, the standard of proof is the ordinary standard of the balance of probabilities and not the criminal standard. The question of dishonesty can arise in civil contexts as well as criminal ones.”

The alternative tests

If the above analysis of the relevance of the word “must” in the subjective limb of the Ghosh test is right, and it doesn’t change the elements necessary to prove dishonesty then, given the apparent scope for confusion that was identified by Popplewell J in R(PSA) the first alternative that the learned judge sets out at section 44 (“the modified Ghosh test” which in effect simply reiterates the test set out in the case of Bryant) may well be a more appropriate and clearer test to apply in disciplinary proceedings with a civil standard of proof.

Whilst the first alternative suggested by his Lordship is without apparent difficulty the second alternative is extremely problematic. His Lordship explains this second alternative, which is referred to in the judgment as the Twinsectra test, at section 45 where he states that:

“The alternative is the trust which applies in civil proceedings for dishonest assistance for the purposes of constructive trust. That test would be that the disciplinary tribunal must be persuaded on the balance of probabilities that (i) what was done was dishonest by the standards of reasonable and honest people and (ii) that the person knew the facts which made those acts dishonesty by the standards of reasonable and honest people. The latter is a lower standard because the subjective element does not require conscious dishonesty but only an awareness of the facts which render the conduct dishonest. That may be a distinction which is of some important in practice. People sometimes act in a way which they would recognise in hindsight would be characterized as dishonest but without really applying their mind at the time that they are acting to the morality of their conduct when that conduct is being committed. It is common to hear someone say by way of explanation “I did it without thinking in the circumstances of particular cases.”

Such a test would remove the need for the regulator to establish conscious dishonesty and in so doing would depart significantly from the well-established jurisprudence in this area, including the decision of the Divisional Court inBryant and the Court of Appeal’s decision in Bultitude. As their Lordships remark at section 154 in Bryant it is overwhelmingly in the interests of fairness that a finding of dishonesty in disciplinary proceedings requires committees to consider the two limb test:

“… in our view it is more appropriate that the test for dishonesty in the context of solicitors’ disciplinary proceedings should be aligned with the criminal test than with the test for determining civil liability for assisting in a breach of a trust. It is true… that disciplinary proceedings are not themselves criminal in character and that they may involve issues of dishonesty that could not give rise to any criminal liability […]. But the tribunal’s finding of dishonesty against a solicitor is likely to have extremely serious consequences for him both professionally (it will normally lead to an order striking him off) and personally. It is just as appropriate to require a finding that the defendant had a subjectively dishonesty state of mind in this context as the court in R v Ghosh considered it to be in the criminal context.”

The decision in R(PSA) provides no compelling reason why the sentiments expressed so clearly in Bryant are ripe for change.