Dishonesty divides a civil tax dispute from a criminal one. Ivey removed the limb of the old test that let a defendant escape by pointing to their own moral standards, and it did so in a case where the gambler genuinely believed he had done nothing wrong, and the court believed him.
On this page
Full name: Ivey v Genting Casinos (UK) Ltd (trading as Crockfords)
Citation: [2017] UKSC 67; [2018] AC 391
Court: Supreme Court (Lord Hughes giving the judgment)
Judgment: 25 October 2017 (unanimous)
Subject: Cheating at gambling; the test for dishonesty
Result: Mr Ivey’s appeal dismissed. The technique amounted to cheating, and the two-stage Ghosh test for dishonesty was disapproved.
Why a Gambling Case Governs Tax Fraud
Dishonesty is the dividing line between a civil tax dispute and a criminal one. It determines whether a matter proceeds under Code of Practice 9 or towards prosecution, whether a director faces a personal liability notice founded on dishonesty, whether an agent falls within the dishonest conduct regime, and whether allegations of fraud in civil proceedings can be made out.
Before 2017 the criminal test came from R v Ghosh and had two limbs, the second of which asked whether the defendant realised that ordinary honest people would regard their conduct as dishonest. Ivey removed that second limb. The consequence in a tax context is significant: a taxpayer can no longer escape a finding of dishonesty by saying that in their own moral universe what they did was acceptable.
Step 1 (subjective). Ascertain the defendant’s actual state of knowledge or belief as to the facts. The belief need not be reasonable, but the more unreasonable it is, the less likely it is to be genuinely held.
Step 2 (objective). Determine whether, given that state of mind, the conduct was dishonest by the standards of ordinary decent people.
There is no third requirement that the defendant appreciated their conduct was dishonest by those standards.
The Facts
Phil Ivey is a professional gambler. Over two days at Crockfords casino he played punto banco, a game of pure chance in which the player has no skill-based advantage.
He used a technique known as edge-sorting. The backs of the playing cards had a repeating diamond pattern which, because of a manufacturing imperfection, was not perfectly symmetrical. Mr Ivey and his companion persuaded the croupier, by requests presented as superstition, to rotate certain cards through 180 degrees, and to use the same shoe of cards for subsequent play. Once the high-value cards were oriented differently from the rest, he could identify them from the pattern on the back before they were dealt.
The technique turned a game of chance into one in which he had a substantial edge. He won approximately £7.7 million. The casino declined to pay, saying he had cheated.
Mr Ivey was entirely open about what he had done. He gave evidence that he did not regard edge-sorting as cheating; he considered it legitimate advantage play, and the trial judge accepted that he was truthful in saying so.
The Ratio Decidendi
The dishonesty holding
Because the Court decided the appeal on cheating, its treatment of dishonesty was strictly unnecessary. It nonetheless addressed the question fully, held that the second limb of Ghosh was wrong in principle, and set out the two-step test above.
The reasoning against the Ghosh second limb was that it produced the perverse result that the less a defendant’s standards conformed to society’s, the less likely they were to be convicted. It also created an unjustified divergence between the criminal and civil tests, since the civil law had never required the defendant to appreciate that their conduct was dishonest.
Obiter, and Why It Binds Anyway
- Strictly obiter. The dishonesty analysis was not required to dispose of the appeal, which turned on cheating.
- Treated as authoritative from the outset. The passage was fully reasoned, unanimous, and expressly directed at correcting the law. Civil courts applied it immediately.
- Confirmed for the criminal law. The Court of Appeal in R v Barton and Booth [2020] EWCA Crim 575 held that the Ivey test is to be applied in criminal proceedings in place of Ghosh. The point is therefore settled in both jurisdictions.
- The parallel with Tooth. This is the second major instance in recent years of a Supreme Court passage that was technically obiter being adopted as settled law, the other being the rejection of staleness in HMRC v Tooth. In both, the answer to “it is only obiter” is that the courts below have applied it regardless.
How It Operates in Tax Cases
Dishonesty and the tax concepts of “deliberate” behaviour are related but distinct, and conflating them is a common error in HMRC correspondence.
| Deliberate inaccuracy | Dishonesty | |
|---|---|---|
| Test | Knowledge of the error plus intention that HMRC rely on it (Tooth, Auxilium | Actual state of knowledge or belief, then objective standards of ordinary decent people) Ivey |
| Where it bites | Schedule 24 penalties; extended time limits; PLNs | Criminal prosecution; COP9 and the Contractual Disclosure Facility; agent conduct; civil fraud allegations |
| Does the taxpayer’s moral view matter? | No, the question is knowledge and intention | No, step 2 is objective |
A finding of deliberate behaviour does not automatically establish dishonesty, and HMRC should not be permitted to slide from one to the other. Conversely, where dishonesty is alleged (in a COP9 case, in a prosecution, or against an adviser) the Ivey framework governs and should be applied in two express steps.
Practitioner Application
Where dishonesty is alleged
- Fight step one, not step two. The objective standard is not realistically arguable. What is arguable is the client’s actual state of knowledge and belief about the facts: what they understood the position to be, what they were told, what records they had seen. Build the case there.
- Do not run “I did not think it was wrong”. After Ivey that submission is not merely weak; it concedes step one and invites the objective assessment. Ivey himself ran it and lost.
- Distinguish honest belief about facts from honest belief about morality. A genuine belief that income had already been taxed, or that an expense was allowable, goes to step one. A belief that under-declaring is acceptable does not.
- Watch for industry practice arguments. That everyone in the trade does the same thing is not an answer at step two, and it can be actively damaging.
- Keep the concepts separate in correspondence. Where HMRC alleges deliberate behaviour, answer Tooth and Auxilium. Where it alleges dishonesty, answer Ivey. Where it does both without distinguishing them, say so.
Where the client is considering disclosure
The Contractual Disclosure Facility under Code of Practice 9 requires the taxpayer to admit deliberate conduct bringing about a loss of tax. The Ivey test is the standard against which any allegation of dishonesty would be assessed if matters proceeded criminally, and it is the right framework for advising a client honestly about the risk before they decide whether to accept the CDF offer.
Frequently Asked Questions
What is the Ivey test for dishonesty?
Two steps. First, subjectively, ascertain the defendant’s actual state of knowledge or belief as to the facts, the belief need not be reasonable, though the more unreasonable it is the less likely it is to be genuinely held. Second, objectively, determine whether the conduct was dishonest by the standards of ordinary decent people. There is no third requirement that the defendant appreciated their conduct was dishonest by those standards.
Can I still argue that my client did not think what they did was wrong?
No, and after Ivey it is a damaging submission. It concedes the subjective step and invites the objective assessment. Mr Ivey himself genuinely believed his technique was legitimate advantage play, the court accepted he was truthful about that, and he still lost. What is arguable is the client’s actual belief about the facts, not about the morality.
Is the dishonesty holding binding, given it was obiter?
In practice yes. The passage was fully reasoned, unanimous and expressly directed at correcting the law, and the civil courts applied it immediately. The Court of Appeal in R v Barton and Booth [2020] EWCA Crim 575 confirmed that Ivey replaces Ghosh in criminal proceedings. The position mirrors the obiter rejection of staleness in HMRC v Tooth.
Is dishonesty the same as deliberate behaviour for a tax penalty?
No, and the concepts should not be conflated. Deliberate inaccuracy requires knowledge of the error and an intention that HMRC rely on it, under Tooth and Auxilium. Dishonesty applies the two-step Ivey framework and matters for prosecution, COP9, agent conduct and civil fraud allegations. A deliberate finding does not automatically establish dishonesty.
Does 'everyone in my industry does this' help?
No. Step two asks whether the conduct was dishonest by the standards of ordinary decent people, not by the standards of the trade. Evidence that a practice is widespread in a sector does not make it honest, and in a reasonable excuse context it is affirmatively unhelpful because the comparator is a trader who intends to comply.