Kaderli v Chief Public Prosecutor's Office of Gebeze, Turkey

For the appellant, Hugh Southey QC and Malcolm Hawkes (who did not appear below) relied on the observation of Thomas LJ in Dudko v Russia [2010] EWHC 1125 (Admin) at [42] that the fairness of the trial process depends on the “integrity, impartiality and good-faith of the prosecutor” and that “it may be difficult to see how the rights under article 6 of their trial can be met, if the prosecutor is corrupt or acts in bad faith”. Systemic judicial corruption means that no tribunal that operates within the system can be relied upon to be independent and impartial. This can amount to a flagrant breach of Article 6 ECHR: Kapri v Lord Advocate, [32]. Whilst there is a rebuttable presumption in the case of states parties to the ECHR that Convention rights will be upheld, “an international consensus” of the kind the judge held to exist here can rebut it: Krolik, [7]. Extradition is prohibited where there is a real risk of a flagrant breach of Article 5. That will be the case where there is a risk of a substantial period of detention following a flagrantly unfair trial: Othman v United Kingdom (2012) 55 EHRR 1, at [233].

In this case, having found that there was an international consensus that corruption was common in the Turkish judicial system, the judge erred in holding that it was for the appellant to prove on the balance of probabilities that the corruption alleged had occurred. The true test involved the application of a lower standard: whether there was a real risk that the appellant’s conviction was based on a trial tainted by corruption. This was consistent with the approach to the fact-finding in the immigration context (see Karanakaran v Secretary of State for the Home Department [2000] 3 All ER 449) and the approach to questions of apparent bias in domestic law (which requires a “real possibility” that the tribunal was biased): see Porter v Magill [2001] UKHL 67, [2002] 2 AC 357, [103].

Once the international consensus is accepted, and it could be shown that there were unexplained flaws in the approach of the Turkish court, strong inferences should have been drawn about corruption. It was for the Turkish authorities to rebut these inferences. The material provided does not do so.

This means that there is a real risk that the appellant’s trial was flagrantly unfair and, therefore, that his detention in Turkey on the basis of a conviction entered at that trial would constitute a flagrant breach of Article 5 ECHR. In addition, refusing to pay a state official a bribe is or could be perceived to be the expression of a political opinion. The test under s. 81 is whether it “appears” that the requested person “might” be prejudiced on account of his political opinion. The burden is on the requested person to show that there is “a reasonable chance” or “reasonable grounds for thinking” or “a serious possibility” that he might be prejudiced at trial or punished for a political reason: Adamescu v Romania [2020] EWHC 2709 (Admin), [67].

After the hearing, Mr Southey drew to my attention the judgment of the Grand

For the respondent, Mr dos Santos submitted that, to show that his extradition is barred under s. 87 of the 2003 Act, the appellant must establish that “he has suffered or risks suffering a flagrant denial of a fair trial in the requesting country”: Soering v United Kingdom (1989) 1 EHRR 439, [113]. This was described as an “exacting test”: R (Ullah) v Special Adjudicator [2004] UKHL 26, [2004] 2 AC 323, [24] (Lord Bingham).

In Othman, the Strasbourg Court said at [259] that the term “flagrant denial of justice” was synonymous with a trial which is “manifestly contrary to the provisions of article 6

In a case where the trial has already occurred, “[i]t is clear that the requested person must establish that his trial was flagrantly unfair, not merely that it contravened Article 6”: Elashmawy v Court of Brescia, Italy [2015] EWHC 28 (Admin), [38] (emphasis in original). Thus, the judge was correct to apply the ordinary civil standard of proof to the question whether the prosecutor had asked for a bribe.

But, even if the ordinary civil standard did not apply, it would still be necessary to adduce evidence showing that there were substantial grounds for believing that there was a real risk that the corruption alleged occurred. The appellant has not met this burden. His proof of evidence is vague. It contains no detail, for example, about when, where or how the request for a bribe was made. There is no evidence from the appellant’s lawyer. There is nothing to show that complaint was ever made about this, or about the respects in which the court’s reasons are said to be incoherent, whether in the first instance proceedings or on appeal before the Court of Cassation. General references in the evidence to the fact corruption did at the relevant time occur in Turkey does not absolve the appellant of the burden of adducing evidence showing that it occurred here (or at least that there is a real risk that it did).

Mr dos Santos submitted that this court should not seek to assume the role of an appellate court. He relied on R (Sobczak) v Judicial Authority in Poland [2011] EWHC 284, where Toulson LJ said this at [13]:

“It is no longer any part of the function of the requesting state to investigate if there is a sufficient case on which to prosecute the person concerned in cases of an accusation warrant. It would equally inconsistent with the framework of the Convention if it were for the requested state to investigate the fairness of the conviction in the case of a conviction warrant. Those are matters for the requesting state. In this case, the matter was considered at an appellate level. Ultimately, of course, a citizen aggrieved by the trial process in the courts of the country subscribing to the European Convention can bring a complaint to the Strasbourg Court.”

This was applied in R (Georgiev) v Sofia Prosecutor’s Office [2012] EWHC 3979 (Admin), [24] (Wilkie J).

Even if the court were to scrutinise the Turkish trial process in the manner deprecated by Toulson LJ in Sobczak, the appellant’s principal basis for the submission that something had gone awry was the fact that Messrs Mutlu and Basaran had been acquitted while he had been convicted. But once it is understood that the factual basis for the charges against them was quite different from the basis for the charge against the appellant, the different verdicts cannot be relied upon as showing inconsistency. Certainly, there was nothing that would meet the test applied by the Court of Appeal

In short submissions contained in an email after the end of the hearing, Mr dos Santos submitted that the decision of the ECJ in P and L did not, on a proper analysis, lend to support to the proposition that the “real risk” test applied to the question whether a trial that has already happened was flagrantly unfair.

The respondent’s submissions on ground 3 were made by Hannah Hinton. She accepted that, under s. 85 of the 2003 Act, the respondent bore the burden of proving that the appellant was convicted in his presence. In this regard, the Strasbourg Court had held that, although proceedings that take place in the accused’s absence are not themselves incompatible with Article 6 ECHR, a denial of justice undoubtedly occurs where a person convicted in absentia is unable subsequently to obtain from a court a fresh determination of the merits of the charge, unless it is established that he waived his right to appear and defend himself: Sejdovic v Italy (App. No. 56581/00), 1 March 2006 [GC], [82]. But an accused who has instructed a lawyer to represent him in the trial is not, however, absent from it: Cretu v Local Court of Suceava, Romania [2016] EWHC 353 (Admin), [34(iii)]. In this case, the information from the Turkish authorities establishes that the appellant was represented from 29 February 2008 onwards.

In this case, the victim made statements to the police and the prosecutor. The only thing she is recorded as having said at the hearing on 29 February 2008 in Uskudar was that she did not wish to be a complainant or intervener. This was favourable to the appellant. In any event, it is not unknown even in this jurisdiction for evidence in chief to be taken other than in the presence of the accused. Achieving Best Evidence (“ABE”) interviews are an example of this. There is no evidence to suggest that the appellant’s lawyer did not know what was happening or that he asked to question the girl or complained about not having been given the opportunity to do so, whether in the first instance proceedings or on appeal.

As to Article 6 ECHR, Ms Hinton relied on the judgment in Murtazaliyeva v Russia (App. No. 36658/05), 18 December 2018 [GC], where the Strasbourg Court rejected an allegation that a trial had been unfair because of a failure to allow questioning of a witness. The rejection was based in part on the fact that the applicant’s lawyer had not objected to the witness’s statement being read: see at [122]. In this case, the court can draw the inference that there was no objection. If so, the complaint about nonattendance on 29 February 2008 is not made out.

The focus of the appellant’s submissions was not on any trial that might happen in the future, but on the trial that has already taken place in 2007 and 2008. As the Strasbourg Court has made clear, however, the UK would be in breach of Article 5 ECHR if the appellant were extradited to a State where he was “at real risk of a flagrant breach of that article”. A flagrant breach might occur if an applicant would be “at risk of being imprisoned for a substantial period in the receiving state, having previously been convicted after a flagrantly unfair trial”: Othman, [233].

Normally, where a trial has already taken place, there will be no dispute about what happened, even if there is a dispute about whether that made it “flagrantly unfair”. If the court concludes that the trial was indeed flagrantly unfair, the question will be whether there is a real risk that the applicant will be detained on the basis of it (rather than, for example, having a retrial). Here, however, the dispute is about what happened at the trial (in the past), not about whether he will be detained on the basis of it (in the future). The formulation used by the Strasbourg Court at [233] of its judgment in Othman is unclear as to whether the “real risk” test applies to the question whether a trial that has already taken place was flagrantly unfair or only to the risk of what is still to happen (i.e. the risk of being imprisoned for a substantial period in the future).

In my judgment, the authorities show that a requested person sought pursuant to a conviction warrant who claims that extradition would be contrary to Article 5 because of a flagrant breach of Article 6 standards in the trial which led to his conviction must establish that the trial was flagrantly unfair, not merely a real risk that it was:

A requirement to show that the trial was flagrantly unfair is also consistent with this principle. As Toulson LJ said at [13] of his judgment in Sobczak, a person who claims that his trial was flagrantly unfair can be expected to bring his complaint before the appellate courts of the State in which he was convicted and, if that State is party to the ECHR, can apply to the Strasbourg Court. It would be surprising if an allegation of the sort advanced here could found a successful domestic appeal unless it were established that the bribe had in fact been requested. The appellant’s analogy with the test for apparent bias in domestic law rests, in my judgment, on a confusion between two stages of the enquiry in a case where bias is alleged. The first stage is to ascertain the facts; the second is to ask whether, on those facts, an objective observer would conclude that there was a real possibility that the judge was biased. See, for example the excerpt from

Nor do I consider apt the analogy with the approach in Karanakaran to proof in immigration cases. There, the court is concerned primarily with an evaluative question: “does the applicant have a well-founded fear of persecution for a convention reason?” This is a question about the likelihood of future events. Here, by contrast, the entire focus of the appellant’s challenge to extradition is on an alleged historic breach of ECHR rights. In this kind of case, the court must reach a conclusion about what happened. In general, this is done by “segregat[ing] past events and apply[ing] the civil standard of proof to them”: see at 477e (Sedley LJ). Only once the facts are ascertained can one reach a view about whether it amounted to a flagrant denial of justice.

The same approach would be applied by the Strasbourg Court if a complaint were made against the State where the trial took place. If an applicant alleges a breach, he must establish that it occurred, not merely a real risk that it did. It would be surprising if an appellant could avoid extradition to an ECHR Contracting State on the basis of a lower test than would apply if he had complained directly to the Strasbourg Court about the fairness of the proceedings in which he was convicted.

I would therefore hold that the judge applied the right test by asking whether the appellant had proved the corruption alleged on the balance of probabilities.

The decision of the ECJ in P & L does not, in my view, affect this conclusion. Mr Southey is correct to say that one of the two cases (L) involved a conviction warrant. The ECJ said at [66] that in an accusation warrant case, the executing judicial authority must examine “to what extent the systemic or generalised deficiencies… are liable to have an impact at the level of that Member State’s courts with jurisdiction over the proceedings to which that person will be subject”. At [67], it said that this would also be the case in a conviction warrant case “when, following his or her possible surrender, he or she will be subject to new court proceedings…” (emphasis added). At [68], the ECJ said that, in L’s case, the executing judicial authority must “also examine to what extent the systemic or generalised deficiencies… have, in the particular circumstances of the case, affected the independence of the court… which imposed the custodial sentence or detention order the execution of which is the subject of that European arrest warrant”. This language is consistent with the analysis at para. 51 above: the “real risk” test is appropriate for cases where new proceedings are envisaged after surrender; but when looking backwards at a conviction or sentence already imposed, the question is to what extent the deficiencies relied on have in fact affected the independence of the court.

I have gone on to consider whether, if the judge had applied the test advanced by Mr

First, even where the “real risk” applies, “it is for the applicant to adduce evidence capable of proving that there are substantial grounds for believing that, if he is removed from a Contracting State, he would be exposed to a real risk of being subjected to a flagrant denial of justice” (emphasis added): Othman, [261].

Second, the matter on which the judge found an international consensus was “that corruption has [been] and continues to be a problem in Turkey”. This was not, however, a finding of “systemic judicial corruption” of a kind that “affects everyone who is subjected to [the system]” such that “[n]o tribunal that operates within it can be relied on to be independent and impartial”: see Kapri v Lord Advocate [2013] UKSC 48, [2013] 1 WLR 2324, [32]. For example, Prof. Bowring’s evidence established that in 2007 Turkey ranked 64^th out of 179 on Transparency International’s Corruption Perceptions Index. This was very much better than Albania in 2012, the time relevant to the decision in Kapri (see [27]), and (incidentally) better than some European Union countries.

Third, the judge’s criticisms of the credibility of the appellant’s evidence about what happened in this case were, in my judgment, justified:

Fourth, a close analysis of the reasoning of the Turkish court does not establish any incoherence or irrationality sufficient to support the inference that the appellant’s conviction was the result of his failure to pay a bribe:

Fifth, it is not surprising that points (c), (d) and (e) above leave some questions to which the answers are not obvious to this Court. The courts of a sending State do not have access to all the trial papers – and they are not expected to be experts in the law of the requesting State. In any event, as Toulson LJ said in Sobczak, points of this kind are ordinarily for the appellate courts of the requesting State or for the Strasbourg Court. Although those remarks were made in a European arrest warrant case, they are equally applicable to cases involving other ECHR contracting States. Here, the points in (c), (d) and (e) could and should have been advanced on appeal in the Court of Cassation if there was any merit in them. In the absence of any evidence to show that they were, they supply no proper basis to conclude that the verdicts were irrational.

Mr dos Santos took an additional point that, even if it were shown that the prosecutor had requested a bribe (or there was a real risk that he had), this would not in and of itself show that the trial was flagrantly unfair, unless it could be shown that the judges who decided the case were also tainted by corruption. I accept that prosecutorial misconduct would not necessarily make a trial flagrantly unfair, particularly if the trial and/or appeal process provided an opportunity to complain about that misconduct. But it would be unsafe to place too much weight on this point in the light of Thomas LJ’s observations at [41]-[43] of his judgment in Dudko. In this case, the relationship between prosecutorial misconduct and the fairness of the trial does not matter, because the evidence does not establish reasonable grounds for believing that there is even a real risk that the prosecutorial corruption alleged actually occurred.

Fordham J granted permission to appeal on ground 2 making clear that it was “entirely parasitic on the Court first arriving at a conclusion that the real risk of corruption aspect of Articles 6 and 5 is made out in this case”: see [2020] EWHC 2893 (Admin), [13]. The appellant made reference to the Divisional Court’s judgment in Adamescu v Bucharest Appeal Court, Criminal Division, Romania [2020] EWHC 2709 (Admin), where Holroyde LJ and Garnham J held at [67] that s. 13 of the 2003 Act (the equivalent of s. 81 for Part 1 cases) required the appellant to show that “there is a reasonable chance (alternatively expressed as reasonable grounds for thinking, or a serious possibility) that he might be prejudiced at his trial or punished, detained or restricted in his personal liberty on account of his political opinions”. This applied both to what might happen in the future and to the reason for it happening.

I do not regard Adamescu as authority for the proposition that the argument under s. 81 in a case such as the present could succeed without a finding that the prosecutor did in fact request a bribe. As the court in Adamescu made clear, the test of “reasonable chance” or “reasonable grounds for thinking” or “serious possibility” is appropriate when considering what may happen in the future. But where the claim is based on a factual allegation about past corruption, the court would expect to see the allegation proved before considering the likelihood that some future action might be taken based on it.

In any event, it was not suggested that there was any material difference between the tests set out at [67] of Adamescu and the “real risk” test derived from Soering. If, as I have found, the evidence in this case did not establish a real risk that the corruption alleged occurred, it would also be insufficient to satisfy the Adamescu test.

Ground 3 challenges the judge’s conclusion that s. 85 of the 2003 Act did not bar extradition in this case. That section prescribes the questions that must be asked and answered in conviction cases. The judge was required to ask:

In each case, Ms Hinton accepts that, because of s. 206, it is for the respondent State to prove to the criminal standard that the answer is “Yes”. If it cannot prove that the answer to at least one of these questions is “Yes”, the judge must order the person’s discharge: s. 85(7).

In this case, the judge dealt with the issue briefly by holding at [223] that the “clear and unequivocal evidence” was that the appellant “was present at some hearings but when he was not it was [because] he chose to deliberately absent himself”. That being so, it was not necessary to consider whether he would be entitled to a retrial.

The difficulty with this conclusion was crisply identified by Fordham J at [15] of his permission judgment: it does not account for what happened on 29 February 2008. On the information supplied by the liaison judge on 22 January 2020, the appellant was on that day “brought before the court” at Gebze pursuant to an arrest warrant. The extradition request from the Gebze Chief Public Prosecutor says that the appellant had been arrested on 24 February 2008 and was released on 29 February 2008 during his trial at Gebze. It would therefore appear that he was produced from custody. If that is so, it is not obvious how he could have attended the hearing that took place on 29 February 2008 at another location, Uskudar, where the girl gave evidence. Nor is it obvious how his lawyer could have attended, given that the lawyer was appointed at court in Gebze, where he remained with the appellant for his “main trial session”.

Ms Hinton submitted that the only thing the girl said on that occasion was favourable to the appellant. But, although there is a discrepancy in the records about exactly what she said, the court, in its judgment, recorded her as saying: “I repeat the statement I gave at the prosecution”, before going on to say that she did not wish to be a complainant or to intervene as a party in the proceedings. If the appellant or his lawyer had been there, questions could have been asked of her with a view to showing that the statement she had given to the prosecutor, and avowed before the court, was unreliable. In evaluating the submissions now made on behalf of the respondent, it is important to recall that the Chief Judge did not say that the proceedings in Uskudar were purely formal, nor that there was no need for the appellant to attend because no evidence adverse to him was taken there. On the contrary, he said: “The accused person did not attend the hearing of the assisting court [at Uskudar]. Therefore the accused person did not use in fact his right to pose questions to the victim.”

Bearing in mind that the burden is on the respondent to prove to the criminal standard that the appellant deliberately absented himself from this hearing, and despite the very clear identification of the problem in Fordham J’s permission judgment, the respondent had no satisfactory answer to this point. On the material produced to the court, the respondent has failed to prove that the appellant deliberately absented himself from the hearing in Uskudar on 29 February 2008, the only occasion on which the victim and main witness for the prosecution gave evidence.

Ms Hinton’s reference to the Strasbourg Court’s decision in Murtazaliyeva does not assist her in relation to s. 85. In that case, the issue was whether the applicant had waived her Article 6 right to question a witness by failing to object to the reading of her evidence. But the question posed by s. 85 is not whether there was a breach (or flagrant breach) of Article 6 rights. It is whether the respondent has proved that the appellant deliberately absented himself. The answer is “No”. It follows that the appellant must be discharged under s. 85 unless the respondent can prove that he has a right to a retrial or (on appeal) a review amounting to a retrial. In these circumstances, it is not necessary to consider the arguments under Article 6 ECHR separately at this stage.

On the basis of the information currently before the court, it is not possible to say that the appellant will be entitled to a retrial. The judge at [220] recorded Prof. Bowring’s evidence to the contrary. There was an email from Prof. Basak Cali to the same effect. However, the further information dated 22 January 2020 from the liaison judge contains what is in effect an invitation to the UK authorities, if they consider that there has been a breach of the appellant’s procedural rights, to request a retrial pursuant to Article 3 ECE 2AP.

The respondent has been on notice since Fordham J’s permission judgment that ground 3 would be considered on a rolled-up basis. Mr Southey says that, if the respondent has not satisfied the court as to the matters arising under s. 85, the appellant is entitled to be discharged. However, the offence of which the appellant was accused was serious. There is a powerful public interest in favour of extradition. Moreover, the Turkish authorities have expressly referred to Article 3 ECE 2AP as creating “a binding rule… directly applicable in the Turkish jurisdiction on condition that the requested Party (the UK in this case) finds the proceedings leading to the judgment in the jurisdiction of the requesting Party (the Republic of Turkey in this case) did not satisfy the minimum rights of defence”.

In those circumstances, I consider it appropriate to adopt the procedure followed in Grecu v Cornetu Court (Romania) [2017] EWHC 1427 (Admin), [49]-[51] and Zelenko v Latvia [2019] EWHC 3840 (Admin), [25]-[26]. I shall grant permission to appeal on ground 3 and adjourn the appeal without making a final order, so as to allow a final opportunity to the Turkish authorities (i) to supply an undertaking that, in the light of this judgment, the appellant will be offered a retrial and (ii) to identify the domestic legal provisions under which this undertaking will be given effect. The respondent will have 28 days in which to provide the undertaking and explanation, together with any written submissions. There will then be an opportunity for the appellant to respond in writing as to the adequacy of the undertaking within a further 21 days. The appeal will be finally determined thereafter.