ABDIHAKIM ELMI v REX

Article 1A of the Refugee Convention defines “refugee”:

“For the purposes of the present Convention, the term ‘refugee’ shall apply to any person who:

…

(2)

…owing to well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group or political opinion, is outside the country of his nationality and is unable or, owing to such fear, is unwilling to avail himself of the protection of that country; …”

Article 31 of the Refugee Convention provides:

“The contracting states shall not impose penalties, on account of their illegal entry or presence, on refugees who, coming directly from a territory where their life or freedom was threatened in the sense of article 1, enter or are present in their territory without authorisation, provided they present themselves without delay to the authorities and show good cause for their illegal entry or presence.”

In R v Uxbridge Magistrates’ Court, ex p Adimi [2001] QB 667, Simon Brown LJ confirmed the purpose of article 31 in the following passage (p 677 G-H) (with emphasis added):

“What, then, was the broad purpose sought to be achieved by article 31? Self-evidently it was to provide immunity for genuine refugees whose quest for asylum reasonably involved them in breaching the law. In the course of argument, Newman J suggested the following formulation: where the illegal entry or use of false documents or delay can be attributed to a bona fide desire to seek asylum whether here or elsewhere, that conduct should be covered by article 31. That seems to me helpful. That article 31 extends not merely to those ultimately accorded refugee status but also to those claiming asylum in good faith (presumptive refugees) is not in doubt. Nor is it disputed that article 31’s protection can apply equally to those using false documents as to those (characteristically the refugees of earlier times) who enter a country clandestinely.”

Newman J’s formulation has been adopted in subsequent cases, see R v AM and others [2010] EWCA Crim 2400; [2011] 1 Crim App R 35 at [6] and R v Mateta and Others [2014] 1 WLR 1516 at [7] (see below).

A person not entitled to refugee status may nevertheless be entitled to “subsidiary” protection pursuant to the Qualification Directive (2004/83/EC), as that has now been implemented into domestic law under the head of “humanitarian protection” by paragraphs 339C and 339CA of the Immigration Rules (HC 395):

“Grant of humanitarian protection

339C. An asylum applicant will be granted humanitarian protection in the United Kingdom if the Secretary of State is satisfied that:

(i)

they are in the United Kingdom or have arrived at a port of entry in the United Kingdom;

(ii)

they are not a refugee within the meaning of Article 1 of the 1951 Refugee Convention;

(iii)

substantial grounds have been shown for believing that the asylum applicant concerned, if returned to the country of origin, would face a real risk of suffering serious harm and is unable, or, owing to such risk, unwilling to avail themselves of the protection of that country; and

(iv)

they are not excluded from a grant of humanitarian protection.

339CA. For the purposes of paragraph 339C, serious harm consists of:

(i)

the death penalty or execution;

(ii)

unlawful killing;

(iii)

torture or inhuman or degrading treatment or punishment of a person in the country of origin; or

(iv)

serious and individual threat to a civilian’s life or person by reason of indiscriminate violence in situations of international or internal armed conflict.”

These provisions were explained and held to be equivalent in some respects to the protection afforded to refugees by the Court of Appeal in FA (Iraq) v Secretary of State for the Home Department [2010] 1 WLR 2545, see [2]-[4], [23] and [25] per Longmore LJ.

Adimi exposed a “serious lacuna” in domestic law; section 31 of the 1999 Act was subsequently implemented in order to meet the UK’s obligations under article 31 (cf R v Asfaw [2008] UKHL 31; [2008] 1 AC 1061 per Lord Bingham at [23] and [26]).

Section 31 provides as follows (taking the version in force on 7 September 2010):

“31.— Defences based on Article 31(1) of the Refugee Convention.

(1)

It is a defence for a refugee charged with an offence to which this section applies to show that, having come to the United Kingdom directly from a country where his life or freedom was threatened (within the meaning of the Refugee Convention), he—

(a)

presented himself to the authorities in the United Kingdom without delay;

(b)

showed good cause for his illegal entry or presence; and

(c)

made a claim for asylum as soon as was reasonably practicable after his arrival in the United Kingdom.

(2)

If, in coming from the country where his life or freedom was threatened, the refugee stopped in another country outside the United Kingdom, subsection (1) applies only if he shows that he could not reasonably have expected to be given protection under the Refugee Convention in that other country.

(3)

In England and Wales and Northern Ireland the offences to which this section applies are any offence, and any attempt to commit an offence, under—

(aa)  section 25(1) or (5) of the Identity Cards Act 2006;

…

(5)

A refugee who has made a claim for asylum is not entitled to the defence provided by subsection (1) in relation to any offence committed by him after making that claim.

(6)

“Refugee” has the same meaning as it has for the purposes of the Refugee Convention.

(7)

If the Secretary of State has refused to grant a claim for asylum made by a person who claims that he has a defence under subsection (1), that person is to be taken not to be a refugee unless he shows that he is.

…”

This Court has considered a number of cases where appellants have complained that they were not informed about the availability of a section 31 defence and have sought to vacate their convictions as a result. The general principle is that erroneous legal advice does not automatically result in a successful appeal and the Court can only intervene only where the conviction is unsafe as a result. In Boal (1992) 95 Cr App R 272, Simon Brown LJ stated at p 278 that:

“Only most exceptionally will this Court be prepared to intervene … Only, in short, where it believes the defence would quite probably have succeeded and concludes, therefore, that a clear injustice has been done”.

That approach has been confirmed recently by this Court in R v Tredget [2022] EWCA Crim 108, at [158]-[159].

The approach to be taken by a Court when a defendant, following incorrect legal advice, has failed to raise the defence and instead has pleaded guilty to an offence under s 25 of the Identity Cards Act 2006 (at issue in this appeal) or its successor provision, s 4 of the Identity Documents Act 2010, was summarised by this Court in Mateta (Leveson LJ, Fulford LJ and Spencer J):

“21.

To summarise, the main elements of the operation of this defence are as follows:

i)

The defendant must provide sufficient evidence in support of his claim to refugee status to raise the issue and thereafter the burden falls on the prosecution to prove to the criminal standard that he is not a refugee (section 31 Immigration and Asylum Act 1999 and Makuwa [26]) unless an application by the defendant for asylum has been refused by the Secretary of State, when the legal burden rests on him to establish on a balance of probabilities that he is a refugee (s.31(7) of the Asylum and Immigration Act 1999 and Sadighpour [38]-[40]).

ii)

If the Crown fails to disprove that the defendant was a refugee (or if the defendant proves on a balance of probabilities he is a refugee following the Secretary of State’s refusal of his application for asylum), it then falls to a defendant to prove on the balance of probabilities that

a)

that he did not stop in any country in transit to the United Kingdom for more than a short stopover (which, on the facts, was explicable, see (iv) below) or, alternatively, that he could not reasonably have expected to be given protection under the Refugee Convention in countries outside the United Kingdom in which he stopped; and if so:

b)

he presented himself to the authorities in the UK “without delay”, unless (again, depending on the facts) it was explicable that he did not present himself to the authorities in the United Kingdom during a short stopover in this country when travelling through to the nation where he intended to claim asylum;

c)

he had good cause for his illegal entry or presence in the UK; and

d)

he made a claim for asylum as soon as was reasonably practicable after his arrival in the United Kingdom, unless (once again, depending on the facts) it was explicable that he did not present himself to the authorities in the United Kingdom during a short stopover in this country when travelling through to the nation where he intended to claim asylum. (s.31(1); Sadighpour [18] and [38]-[40]; Jaddi [16] and [30].)

iii)

The requirement that the claim for asylum must be made as soon as was reasonably practicable does not necessarily mean at the earliest possible moment (Asfaw [16]; R v MA (Footnote: ¹) [9]).

iv)

It follows that the fact a refugee stopped in a third country in transit is not necessarily fatal and may be explicable: the refugee has some choice as to where he might properly claim asylum. The main touchstones by which exclusion from protection should be judged are the length of the stay in the intermediate country, the reasons for delaying there and whether or not the refugee sought or found protection de jure or de facto from the persecution from which he or she was seeking to escape (Asfaw [26]; R v MA [9]).

v)

The requirement that the refugee demonstrates “good cause” for his illegal entry or presence in the United Kingdom will be satisfied by him showing he was reasonably travelling on false papers (ex p. Adimi at 679 H).”

The statute provides that if, by the time of trial in the Crown Court, a defendant’s claim for asylum has been refused by the Secretary of State, then he bears the burden of proving at trial that he is a refugee, see section 31(7).

If, by the time an appeal based on ignorance of the section 31 defence reaches this Court, the appellant’s asylum claim has been decided by the FTT, then it is appropriate for the Court of Appeal to assess the prospects of an asylum defence succeeding by reference to the tribunal’s findings. This was explained in Ali Reza Sadighpour v R [2012] EWCA Crim 2669; [2013] 1 Cr App R 20 (Treacy LJ, Mackay J and HHJ McCreath, the Recorder of Westminster) (cited in Mateta at [23]):

“35.

We are therefore satisfied that it is appropriate to have regard to the tribunal's decision in assessing the appellant's prospects under s.31 on any retrial. After all, the tribunal is a properly constituted judicial body. Its members have particular specialist experience in dealing with matters pertaining to immigration and asylum. The appellant was able to deploy his full arguments and call relevant witnesses. The evidence was fully tested. Both parties made their respective submissions, and a fully reasoned judgment was reached.

36.

As already stated, para.31(7) provides if the Secretary of State has refused to grant a claim for asylum made by a person who claims that he has a defence under subs.(1), that person is taken not to be a refugee unless he shows that he is.”

It is common ground in this appeal that, at the time he was prosecuted, the Appellant was a presumptive refugee (applying Newman J’s formulation in Adimi, see paragraph 24 above). There are cases where a defendant’s application for asylum cannot be said to be made in good faith, for example R v Evans (Fabian) [2013] EWCA Crim 125; [2013] 1 Cr App R 34, where the Court of Appeal upheld the judge’s ruling that the section 31 defence was not available to a Jamaican national who could not produce any evidence to support his case that he was a member of a particular social group susceptible to persecution. But here, the Crown accept that the appellant’s asylum claim was made in good faith to meet the required standard (although the Crown’s case is that the appellant was not a refugee, so that his defence would have failed – see below).

Further, it is common ground that the appellant was denied the opportunity of relying on the section 31 defence because of erroneous legal advice, and that if he had advanced that defence, he would have stood a good prospect of establishing those elements he was required to prove, listed at [21 ii)] of Mateta, namely (a) that he did not stop in any transit country for more than a short stopover, (b) that he presented himself to the UK authorities without delay, (c) that he had a good cause for his illegal entry or presence in the UK because he was fleeing from a real risk of serious harm, and (d) that he made an asylum claim as soon as was reasonably practicable following his arrival in the UK. That means that the sole issue for the jury at trial, if the appellant had run this defence, would have been whether the appellant was a refugee.

It was common ground that the burden would have been on the prosecution to disprove the appellant’s refugee status (see again, Mateta at [21 i) and ii)]. The parties agreed that the jury, in this hypothetical trial, would have been directed along these lines: “Has the prosecution made you sure that the defendant is not a refugee? If you conclude that he is or might be a refugee, you must return a verdict of not guilty.”

The parties give different answers to that hypothetical question asked of the hypothetical jury. The appellant’s case is that if he had raised the section 31 defence in 2010, he would quite probably have succeeded in that defence and so there has been a clear injustice, justifying this conviction being quashed, applying Boal.

The Crown argue that this defence could not have succeeded. The mainstay of their case is that the FTT has since adjudicated his asylum application and found him not to be a refugee. If he was retried now, the section 31 defence would not succeed.

To rebut the Crown’s case, the defence relied on three arguments, set out in a skeleton argument filed in advance of the appeal:

At the hearing, Mr Thomas KC (who, with Mr Newton, acted for the appellant) pursued an additional line of argument, or, possibly, a variant on the second argument outlined above, based on the content of the FTT’s decision, to the effect that the FTT had not addressed the appellant’s situation in 2010, when these criminal proceedings took place, but had found only that the appellant was not a refugee in 2013, although they had accepted that he had been subject to persecution for a Convention reason at some time unspecified before that. The appellant’s argument was that the Crown could not therefore rely on the FTT’s decision to assert that he was not a refugee in 2010, he may very well have been. Further, the FTT found that the appellant’s account was credible and correct, and that he had entered the UK while fleeing, with a fear of suffering serious harm if returned to Somalia, and it was reasonable to suppose that the jury would have reached the same broad conclusions if they had been asked to determine whether the appellant was a refugee for the purposes of his notional section 31 defence. In all the circumstances, the Crown would not have been able to rebut the defence and this Court could be satisfied that the defence would quite probably have succeeded.

Mr Douglas-Jones KC (who, with Mr Johnson, represented the Crown) maintained that this Court could not go behind the FTT decision which held that the appellant was not a refugee. Section 31 only applied to refugees. No analogy can be drawn with the non-prosecution regime for victims of human trafficking or by any other analogy to extend the plain words of section 31(6). The jury were bound to conclude he was not a refugee, given the FTT’s conclusions to that effect. This prosecution had been properly undertaken in the public interest and there was no wider public interest justification for concluding that this conviction was unsafe.

We are grateful to all counsel involved in this appeal for the assistance they provided to the Court. Between them, they seem to have appeared in nearly all the cases cited to us. Their expertise in this area was conspicuous and valuable.

There are, in our judgment, three issues falling to be determined:

Mr Thomas submitted that s 31 should be read expansively. There were two strands to the submission. The first and more ambitious argument was that s 31 should be construed to include those with humanitarian protection within the definition of refugees, on the basis that the two types of protection are equivalent. He relied on FA (Iraq), where the Court found that section 83 of the Nationality, Immigration and Asylum Act 2002, which provided a right of appeal from the refusal of asylum but was silent on appeals from refusal of humanitarian protection, should be read to encompass appeals from both types of decision:

“47.

The rights of a refugee, as now provided in national law, and the rights of a person with a subsidiary protection status, as provided by the Directive, are in many respects similar. They are sufficiently similar, in my judgment, to require national law to provide the person seeking international protection of that kind to have the same remedy of recourse to an independent tribunal against an adverse decision of the Secretary of State as has a person seeking international protection as a refugee. That requires section 83 to be read as applying to a person who has sought [humanitarian protection] as it applies to a person who has sought asylum.” (per Pill LJ).

He also took us to paragraph 327EC of the Immigration Rules, which deems a person claiming humanitarian protection to be an asylum applicant so that any claim for humanitarian protection will be recorded as an application for asylum and provides that: “If the application for refugee status is refused, then the Secretary of State will go on to consider the claim as a claim for humanitarian protection”. This, he said, showed that the two types of protection were treated as related, even equivalent, by the Secretary of State for administrative purposes. He invited us to have regard to the UNHCR mandate which, he submitted, extends to the whole protective regime for displaced persons, whether they are refugees or those in need of humanitarian protection. In short, he submitted that there is or should be equivalence of treatment between the two types of protection.

His second and narrower argument was that section 31 must extend substantively not only to refugees, but also to “presumptive refugees”. He took as his starting point the Crown’s acceptance that it was open to a presumptive refugee to advance this defence at trial. From that starting point, he submitted that it was or should be sufficient for the defence to succeed that the defendant was indeed a presumptive refugee. The anchor for this submission was the passage in Adimi which confirmed the broad purpose of article 31 to protect not only refugees but also presumptive refugees (see paragraph 24 above).

In answer, Mr Douglas-Jones submits that section 31(6) presents an insuperable hurdle to this argument. It defines refugee by reference to the Convention and leaves no space for a more expansive definition to include those entitled to humanitarian protection or to those who are merely presumptive refugees. He emphasised the distinction between the two concepts drawn by paragraph 339C(ii) of the Immigration Rules.

We are with Mr Douglas-Jones. It is not possible to construe section 31 as if it applied to those with either sort of protection. The defence only applies to refugees, but, consistent with its statutory purpose, may be advanced at trial by those who are at that time presumptive refugees. It is for the jury to determine whether the defence is made out and the issue for the jury, posed in terms recorded at paragraph 37 above, is whether the defendant is a refugee.

The issue in this case is whether this appellant, who was in 2010 a presumptive refugee, but who did not persuade the FTT in 2013 that he was at that date a refugee, would quite probably have succeeded in his section 31 defence at trial in 2010.

We begin by considering some of the previous cases where this Court has considered whether to quash a conviction because a section 31 defence was overloooked. In those cases, the FTT has typically decided whether the appellant is, or is not, a refugee for immigration purposes by the time of the appeal.

In Sadighpour [2012] EWCA Crim 2669, the FTT had rejected the appellant’s asylum claim and had found that he had fabricated much of his story. The Court of Appeal dismissed his appeal, holding the conviction to be safe, because there was “no reason to think that prosecuting counsel at trial would not make the same or similar points [as had been made in the FTT and had led to the adverse credibility finding] to good effect” (see [44] and [45]).

In Mateta, the Court allowed the appeals of five appellants. Each of those appellants had succeeded in their FTT appeals and, by the time their appeal to this court was heard, had been accorded refugee status. The Court (per Leveson LJ) noted that each appellant had been granted refugee status (see [29], [35], [42], [52]) and the Court was prepared to accept, in all the circumstances, that the section 31 defences would quite probably have succeeded and that therefore a clear injustice had been done, applying Boal. In none of those cases did the Crown resist that course.

This case is very different. The FTT refused this appellant’s asylum claim, but not because the judge did not believe him. To the contrary, the FTT judge (Mr O’Brian) did accept his story, but held that in light of the political and security situation as it stood in 2013 in his home country, he would not be persecuted on return for a Convention reason, although he would face a serious risk of harm for non-Convention reasons. Neither counsel, despite their considerable experience in this area of work, was able to show us any previous case which had involved facts like these.

The question raised by the hypothetical section 31 defence would have been whether, in 2010 when the appellant used the false Norwegian passport, he was a refugee. We have considered the FTT decision carefully to see what assistance it provides to this Court in assessing whether the appellant probably would have succeeded in that defence. We conclude that the FTT does not offer us much help. The FTT found that the appellant was not a refugee in July 2013. It can be inferred that he had not been a refugee since 2011, when the country guidance case of AMM was released. The FTT found that he had been a refugee (or at least had been entitled to claim protection under the Refugee Convention) at some time before that, but it is unclear when. Accordingly, the FTT does not offer assistance on the issue which would have faced the jury: was the appellant a refugee when he came into the UK on a false passport in July 2010?

Still, the Crown submits that the FTT decision is determinative of the outcome of this appeal. To quote their skeleton (see [21]), that defence “could not have succeeded”. As we understand it, that is because, on their case, the FTT decision would apply in retrospect to matters as they stood in 2010, and the same result would and should have been reached in 2010 as was in fact reached in 2013, namely that the appellant was not a refugee.

We are unable to accept that argument, given the content of the FTT’s decision in this case. This is not a case where the FTT’s decision will help to determine the position as it was at the time of the criminal proceedings: so, in Sadighpour, the FTT found that the appellant’s account lacked credibility, and it was reasonable for this Court to conclude that the jury would have reached the same conclusion and similarly rejected that account; in Mateta, the appellants were established to be refugees at their subsequent FTT hearings and in consequence the Crown did not resist the quashing of their conviction, so this Court could reasonably conclude that, if the FTT had found that they were refugees since entry into the UK, that logically would have been the outcome if they had advanced a section 31 defence at trial. This case is different. The FTT in this case does not help us to determine whether the appellant was a refugee in 2010. In this case, the answer to the Boal question does not necessarily align with the outcome of the subsequent FTT.

We have to answer the Boal question without having an FTT decision to assist on the appellant’s immigration status at the time of the criminal proceedings. This is surely a highly artificial exercise. The FTT addressed the appellant’s refugee status in 2013 at a full day’s hearing, with the benefit of 12 bundles of documents, after hearing oral evidence from the appellant and two witnesses. The FTT took time for its decision, which is long and detailed, and involved analysis of complex legal materials. In R v Makuwa [2006] EWCA Crim 175; [2006] 1 WLR 2755, the Court (Moore-Bick LJ, Lloyd Jones J and Judge Findlay Baker QC) said (with emphasis added):

“26.

…We have come to the conclusion that,…provided that the defendant can adduce sufficient evidence in support of his claim to refugee status to raise the issue, the prosecution bears the burden of proving to the usual standard that he is not in fact a refugee.

…

37.

… It may well be that, in many cases where the defendant claims to be a refugee the Crown, while not accepting the claim, will not seek to establish that he is not.”

We were shown extracts from the CPS guidance on when prosecutors might not seek to contest refugee status in the Crown Court. If ever there was a case where the Crown might have opted not to contest the point, this surely is it.

However, Mr Douglas-Jones says that we should assume that the jury would have been asked to consider the refugee issue. He says that this is how the Boal test works. We will therefore do as asked and assess the prospects of that defence succeeding.

Although the FTT decision does not assist on whether the appellant was a refugee in 2010, it does offer some insight into the sort of evidence and material that would have been available to the hypothetical jury. The FTT found the appellant to be credible, based in part on the fact that he had given a consistent story throughout. We think it likely that the appellant would have chosen to give evidence to support his defence, and in doing so he would have told the jury the same story as he had given to the Secretary of State in his screening interview and two asylum interviews, and as he ultimately provided to the FTT. This story contained harrowing details of the treatment meted out to him and his family in Somalia, including the murder of his father, the rape of his sister, his own near death by shooting, bombs being thrown into his house and further attempts to shoot him. He would have explained that the reason for these attacks was that he was a member of the Mehari sub-clan.

We think it is probable that the jury would have believed this evidence, or would, to put it at its lowest, not have been sure that it was untrue. We note that the FTT, applying the lower standard of proof, accepted the appellant’s account as true. The jury would have been asked whether the Crown had made them sure that it was untrue, which is, if anything, a lower hurdle for the appellant than he faced in the FTT. This is not to suggest that the credibility findings of the FTT are in some way determinative of what the jury would have decided, because the jury would have reached their own conclusions. But the FTT’s acceptance of the appellant’s account does offer some support for our view that the appellant’s story would not have been rejected as incredible.

The argument would then have shifted to a discussion of whether those facts, and the fear consequent upon repeat of them if the appellant was returned to Somalia, were sufficient to make him a refugee. We note the way in which the Court in Makuwa suggested that the jury should be directed:

“37.

… Where the Crown disputes the defendant’s claim it will be necessary to explain what a refugee is for the purpose of section 31 of the 1999 Act. We would suggest that is best done by drawing on the language of the Refugee Convention itself, using words of the following kind:

“a refugee is a person who has left his own country owing to a well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group or political opinion”.

“38.

It will also be necessary to give the jury some assistance on the meaning of “well-founded fear”. We would suggest that the concept can best be conveyed by directing that a fear of persecution is well-founded if there is a serious possibility that the defendant will suffer persecution if returned to his own country. Finally, it will be necessary to direct their attention to the fact that in order to be a refugee the defendant must fear persecution for one of the reasons mentioned in the Refugee Convention, that is, race, religion, nationality, membership of a particular social group or political opinion. …”

We conclude that, faced with directions of that sort, adapted as necessary to the particular facts of this case, the jury would probably have concluded that they could not be sure that the appellant was not a refugee, and they would therefore have returned a verdict of not guilty. We know from the FTT that the reason the appellant ultimately failed in his claim for asylum was because of improvements in the security situation in Modagidishu by 2013. There was no evidence before us as to how matters stood in that city in 2010, and Mr Douglas-Jones did not suggest that issue would have been raised by the Crown.

In our judgment, the appellant would quite probably have succeeded in his defence.

Mr Douglas-Jones submitted that the two elements of the Boal test were distinct. His primary case was that the defence would probably not have succeeded. But even if it he failed on that primary case, still he suggested there had not been any clear injustice, and the conviction remained safe. As we understood it, that was because of the FTT decision which subsequently determined that the appellant was not a refugee.

Mr Thomas took issue with that proposition and submitted that the formulation in Boal, using the connector “therefore”, means that once it is established that the defence would quite probably have succeeded, the clear injustice is made out and it follows that the conviction must be quashed.

In our judgment, the answer lies somewhere between those two extremes. There may be cases where no clear injustice has occurred even though an appeal court concludes that the overlooked defence would quite probably have succeeded at trial. But if there are such cases (as to which we reach no firm conclusion), this is not one of them. In this case, it cannot be suggested that the FTT’s subsequent conclusion that he was not a refugee somehow moderates or cures the injustice to the appellant of not being made aware that he could raise the defence in 2010. We come back to the gap in the timing to which the FTT’s conclusion related: the FTT only dealt with matters in 2013, and not before. It recognised that the appellant had refugee status at some, albeit unspecified, earlier stage. The position as it might have stood in 2010 cannot be extrapolated from the FTT’s decision three years later.

We conclude that the failure to raise the section 31 defence in this case, which defence would quite probably have succeeded, has led to a clear injustice. This appellant would probably have been acquitted if he had raised the section 31 defence. Nothing that has occurred subsequently casts doubt on the fairness of that outcome, for example, by showing that he was not, in fact and law, a refugee at that time.

This conviction is not safe.

This appeal is allowed. We quash the conviction.

We do not need to deal with the wider public policy points raised by Mr Thomas and answered by the Crown; we leave those for another day when they may be material to the outcome.