Yesterday the Guardian published a top secret policy in place for the UK intelligence service since 2002 in several versions, dealing with their obtaining intelligence from detainees in the custody of foreign services who may be subject to mistreatment. The document is entitled ‘Agency Policy on Liason with Overseas Security and Intelligence Services in Relation to Detainees Who May Be Subject to Mistreatment,’ and is available with a few redactions here. The policy is sure to be prove controversial; as readers are aware, an inquiry is underway in the UK regarding complicity of UK services in overseas torture, while a judicial review case has recently been brought before the High Court (see Dapo’s recent post). The policy has been replacedy a more anodyne and public version in 2010 by the new coalition government.
Note that there is at least one more prior policy that is yet to be disclosed, which covered circumstances in which UK agents were ‘directly involved’ in the questioning of a detainee in the custody of a foreign intelligence service (see para. 2 of this policy). That other document may prove to be at least as interesting.
One may find a number of things morally objectionable upon reading the policy – and it’s precisely these parts of the document that caught the eyes of the press, e.g. the explicit reference to negative publicity for the UK as a factor in a balancing exercise in deciding whether or not to give permission for UK agents to provide information to the overseas agency even when there is a risk of mistreatment. But what I found striking about the old policy was how legalistic it was, i.e. how legal advice was used to put limits (or not) on UK participation in overseas interrogation. The new 2010 policy is very different – it seemingly quite deliberately omits any substantial legal discussion.
What is even more striking is how the policy concludes that human rights law as such does not apply to the issues at hand, although it then proceeds to import some human rights standards in the rules it sets out. Crucially, as Dapo pointed out in his earlier post and as I mentioned in a post from a couple of years ago, the key question here is the extraterritorial application of human rights treaties. Do, say, detainees in Pakistani custody questioned by a Pakistani intelligence officer who are being fed questions or information by UK agents have rights vis-a-vis the UK under the human rights treaties to which it is a party, or is it only Pakistan which has obligations in this situation?
In para. 21, the policy answers that question in the negative:
Under section 6 of the Human Rights Act 1998 it is unlawful for a public authority to commit torture , or to inflict inhuman or degrading treatment, as this would be incompatible with a Convention right [Article 3 ECHR]. However, in order for the Act to apply to a detainee held overseas, the UK would need to have “effective control” of the area in which the detainee is located, as the primary jurisdiction of the Act is territorial. The Act is therefore unlikely to apply in situations covered by this policy.
Note how the policy refers to the spatial conception of Art. 1 ECHR state jurisdiction, as control of an area. We have seen e.g. in Al-Saadoon how that spatial conception can apply to ever smaller ‘areas’, such as a detention facility – but in this scenario, of course, the UK is not in control of the detention facility. The policy then extends the same reasoning to the ICCPR, in para. 25. It is a bit more circumspect with the CAT, in para. 24, but note how the CAT itself obliges states to prevent torture and inhuman treatment only in territories under their jurisdiction. Indeed, it was precisely the argument of the US Office of Legal Counsel in its infamous torture memos that the CAT did not apply to interrogation in CIA black sites (e.g. a former riding school in Vilnius, Lithuania) because they did not take place in areas under US control – see more here on the water cure with saline solution.
Unpalatable as it is morally, this is not a bad legal argument. Under the spatial model of jurisdiction as in Loizidou or Bankovic, even if a UK agent were to personally stab a detainee in Pakistani custody with a hot poker the ECHR would not apply, let alone if the UK agent ‘merely’ fed the Pakistani agent with questions or information. But what of the personal model of jurisdiction, as authority and control over individuals, as e.g. in Issa, which was recently used to such effect by the Grand Chamber of the European Court in Al-Skeini?
The policy does not even entertain the possibility that jurisdiction attaches on such grounds – but of course, Bankovic would have seemed to preclude such a result. If killing is not ‘jurisdiction’, why would torture or mere feeding of questions to the torture constitute ‘jurisdiction’? Now, post Al-Skeini, the situation might be different, but we are still in uncharted waters – everything would hinge on the interpretation of the nebulous ‘public powers’ criterion that the Court has introduced there.
To my mind, there is only one sensible solution to this problem – to say that negative obligations under the treaties (i.e. the obligation to respect human rights, and not to be complicit in mistreatment or expose an individual to real risk thereof) are territorially unlimited, and that only positive obligations under the treaties (i.e. the obligation to secure or ensure human rights even against violations by third parties) would apply to ares under the state’s effective control. To the extent that such negative obligations are not explicit in the treaties, they should be read into them implicitly, as in the ICJ’s 2007 Genocide judgment. This isin my view the only model of extraterritorial application which is stable in the long term and does not produce arbitrary results – see more here, at 209-221. Whether courts and other decision-makers will be prepared to adopt this model anytime soon is a different matter. We’ll of course see how the UK inquiry and complicity case progress.