COURT decisions identifying the applicable law for torts are surprisingly challenging. The rules are not apparently unclear, but their application leads to incomprehensibly different results. Judges disagree dramatically even in the simplest fact pattern of a car accident (e.g. Harding v Wealands [2007] 2 A.C. 1) as they did in Zubaydah v FCDO [2023] UKSC 50, [2024] 1 W.L.R. 290. The case raised the applicable law to determine liability of defendants based in the UK for damages resulting from torture by agents of the US. Mr. Zubaydah is a Palestinian captured in Pakistan and detained without trial by the US authorities since 2002. Between then and at least 2006 he was held in a number of countries, tortured by the CIA and rendered to Guantanamo Bay. The Secret Intelligence Service and the Security Service (the UK services) had asked the CIA to put questions to the claimant during questioning. He claimed damages in tort for misfeasance in public office, conspiracy to injure, trespass to the person, false imprisonment and negligence. The claimant alleged that the defendants, the Foreign, Commonwealth and Development Office (“FCDO”) and the Home Office, were vicariously liable for the actions and omissions of the UK services.

The issue of the applicable law to determine the liability of the defendants was raised as a preliminary matter. At this stage of the dispute process no facts are known or proved, so the applicable law is decided on the basis that the facts were as alleged by the claimant. The facts will have to be proved later at trial unless both the claimant and defendants agree them.

The applicable law for these torts fell to be determined by the Private International Law (Miscellaneous Provisions) Act 1995 not the current regime of the Rome II Regulation on the law applicable to non-contractual obligations (Regulation (EC) No 864/2007) (“Rome II”) which is part of retained EU law. Nonetheless, the case is still important. The two regimes have a common structure of a general rule and exception for a more closely connected law. The general rule for personal injury is similar, broadly being the law of the country where the injury occurred. In the PILA it is “the law of the country where the individual was when he sustained the injury”. Under Rome II it is “the law of the country in which the damage occurs”. Both regimes indicate that the exception is to be rarely applied after an evaluation of factors connecting the tort to a different country than that of the general rule (see s. 12(2) PILA: “If it appears, in all the circumstances … substantially more appropriate”; and Art. 4(3) Rome II: “Where it is clear from all the circumstances of the case that the tort/delict is manifestly more closely connected” (both my emphases)).

The identification of a more closely connected law than that identified by the general rule is complex. First the factors connecting the tort to the possible laws and countries must be identified: the connections with various countries of the events making up the tort and connections of each of the parties to any country. Then the significance of the factors must be evaluated. Significance involves weighing of factors, not merely counting them up. Some are more significant than others. The relevance of a factor may vary depending upon context.

Lastly, a conclusion must be reached on the relative weight of the general rule against the exception – that is, whether the threshold of “substantially” or “manifestly” has been crossed. Along the way, justifications for the application of the law of the place of the injury or for another relevant law are set against each other. The general rule is often justified as being what both the claimant and defendant probably expected. This is self-evident in the case of A negligently injuring B in a car accident, despite the fact that A and B are from different countries. The self-evidence of that expectation is challenged say when A and B come from the same country and only happen to be in the place of the injury because they are holidaying there. The lack of predictable connection to the place of the injury is outweighed either by the strength of their reasonable expectation due to their pre-existing connection or that the place of the injury is merely adventitious (as in Boys v Chaplin [1971] A.C. 356).

Orthodox conflict of laws analysis in England assumes the choice of law rule to be neutral between the claimant and defendant, in effect value-free. Therefore, the substantive outcome, once the domestic rules of the applicable law have been applied to determine the actual liability of the defendants, is not relevant. At the stage of identifying the applicable law by a choice of law rule, the domestic rules are not yet in play. This “jurisdiction-selecting” analysis may not subject defendants to proper regulation of their conduct by the law where they acted in cases where the claimant is not protected under the law of the place of injury. This is particularly so in situations, such as this, where the claimant has no control or foresight of where the injury will occur. US choice of law has taken a different approach, allowing a wider range of factors. The unpredictable consequences of that approach have led to criticism. Both the PILA and Rome II attempt to redress the balance towards predictability by placing great emphasis on the general rule.

The claimant had been tortured for information by the CIA in six countries. He had no connection to England at all. The factors connected to England were that UK services had asked the CIA to put questions to the claimant. These requests had, it was reasonably presumed, been made in England. The activities of the defendants are those of Departments of State of the UK Government. The factors connecting the torts to the six countries included the events of torture and imprisonment, the suffering of the claimant, and the actions of the CIA. The law of the US was agreed not to be an option.

The first instance judge, Lane J., had decided that the law of each of the countries where the claimant was when he was injured or detained applied using the section 11 PILA general rule. He performed a careful evaluation to decide that the section 12 PILA exception did not apply. The Court of Appeal permitted an appeal. Although appeals are not usual from the evaluative judgment of first instance judges, here Lane J. had erred, since he ought to have concluded that the significance of the factors connecting the torts to England was “very substantial”. Evaluating the factors it held that English law applied as a matter of the exception in section 12. The majority of the Supreme Court agreed that Lane J. had erred, held that the Court of Appeal had done so too, and so could perform its own evaluative judgment. They also concluded that the law of England and Wales applied as a matter of exception. Lord Sales disagreed. Lane J. had not erred and, having evaluated the factors with similar thoroughness, Lord Sales decided that the law of each of the six countries applied under the general rule. He emphasised the need for neutrality. Had the claimant achieved better results under the law of one of the six countries, no doubt he would have argued for that law and not English law. These differing conclusions were all reached with careful analysis.

The differences came down to the individual perspective on and significance of the factors. For example, Lord Lloyd-Jones for the majority placed weighty emphasis on the claimant not having control of his whereabouts; on the difficulty that the claimant was injured in “de facto exclaves” where laws of those countries did not run; and on the responsibility of the UK services (i.e. focussing on the request for information rather than the acts of the torturers). It was a justifiable conclusion. Whereas Lord Sales in dissent preferred to leave the facts of exactly where the claimant was injured and the rules of law there to be proven at trial; and agreed with Lane J. that the primary liability was of the CIA agents who were acting in each of those countries. Broadly, there was insufficient connection with England to displace the general rule. And it is difficult to argue with that.

Two primary conclusions come to mind. First, the use of “evaluative judgment” as a concept to control appeals is fanciful, as the cases on jurisdiction have (most recently FS Cairo (Nile Plaza) LLC v Brownlie [2021] UKSC 45 at [201]) already shown. Second, the adage that “hard cases make bad law” is true. This is a particularly hard case in which the UK state is allegedly involved in requests for evidence from someone subject to torture. The US has made itself impossible to be sued, leaving the claimant fighting for redress elsewhere. Clearly, the liability of these English defendants for those actions is a matter of concern for English regulatory rules. On the other hand, is it sufficient for English law to apply merely because the defendants acted in England despite no additional connection between the claimant and England? How could the claimant have foreseen English law? Torture is abhorrent which may justifiably skew this result. Outside the specific circumstances of this case, too much emphasis on the defendant’s action rather than the claimant’s harm could backfire in cases where the claimant is worse off under the law where the defendant acted. Alternatively, the reasoning may encourage a claimant to seek out a forum whose choice of law rules will give them the best result. There is a real danger of unpredictability and uncertainty. Unless judges are careful to note the special circumstances of the case, Zubaydah is in danger of taking the English rules towards what Dean Prosser described as “the dismal swamp” (“Interstate Publication” (1953) 51 Michigan Law Review 959, 971).