dlvr.it
The Kingdom of Bahrain v Shehabi and another -
Find Case Law
- The National Archives
The Kingdom of Bahrain (Appellant) v Shehabi and another (Respondents)
LORD LLOYD-JONES, LORD HAMBLEN AND LADY SIMLER:
This appeal concerns the correct interpretation of section 5 of the State Immunity Act 1978 (“the SIA”). Section 5 is one of the listed exceptions to the general rule of state immunity from jurisdiction conferred by section 1 of the SIA. It provides that a state is not immune as respects proceedings brought in respect of death or personal injury or damage to or loss of tangible property “caused by an act or omission in the United Kingdom”.
The courts below held that the exception applies if it can be shown that an act or omission takes place in the United Kingdom which causes personal injury or property damage.
The appellant contends that they were wrong so to conclude. It submits that the exception only applies: (i) where the act or omission causing personal injury or property damage was the responsible or precipitating act or omission; or (ii) where every act or omission causing personal injury or property damage took place in the United Kingdom; or (iii) where the author of the personal injury or property damage was present in the United Kingdom at the time when the facts which occasioned the injury or damage occurred. If the appellant is correct about any of these submissions, state immunity would apply to these proceedings.
The alleged factual circumstances in which these issues arise concern the hacking by the servants or agents of the appellant of the respondents’ computers with spyware. On the assumed facts, this was initiated outside the United Kingdom, but the computers and the respondents were in the United Kingdom. The hacking allowed access to and exfiltration of information on the computers, interception of communications conducted using the computers and use of the computers’ microphones and cameras to surveil the respondents. It is alleged that this involved a course of conduct amounting to harassment, contrary to the Protection from Harassment Act 1997. It is further alleged that both respondents suffered personal injury, in the form of psychiatric injury, as a result of learning about the acts committed by the appellant.
The factual case
The facts are taken from the respondents’ pleaded case and are assumed to be correct for the purpose of the preliminary issue which is the subject of the appeal.
The first respondent, Dr Saeed Shehabi, is a pro-democracy journalist and activist who is a leading figure in the Bahraini opposition movement. He is a leader of the opposition group Bahrain Freedom Movement and the founder of a Bahraini pro-democracy organisation called Al Wefaq. He has lived in the United Kingdom since 1973. He was granted asylum in 1985 and British citizenship in 2002.
The second respondent, Mr Moosa Mohammed, is a photographer, videographer and an activist for human rights and democracy in Bahrain. He has lived in the United Kingdom since 2006, was granted refugee status in 2007 and has indefinite leave to remain. In 2012, his Bahraini citizenship was revoked.
From around September 2011, persons acting on behalf of the appellant hacked or infected the respondents’ computers with a spyware program known as “FinSpy”.
While the spyware was being installed and used: (i) the appellant’s servants or agents were likely operating remotely from outside the United Kingdom; (ii) the spyware was operated via a spyware command and control server located in Bahrain; and (iii) the respondents and their computers were in the United Kingdom.
The spyware allowed the appellant’s servants or agents, without the respondents’ knowledge: (i) to access, copy, and/or exfiltrate information stored on, available on and/or transmitted by the computers; (ii) to intercept textual, audio, and/or video communications conducted using the computers; and (iii) to use the computers’ microphones and/or cameras to surveil the respondents.
This course of conduct, which involved monitoring the respondents’ activities on a wide-ranging basis and by highly intrusive means, amounted to harassment contrary to the Protection from Harassment Act 1997.
The respondents learned of the use of spyware in or around August 2014, following the publication of information about the spyware program on WikiLeaks and/or by an organisation called Bahrain Watch.
Both respondents suffered personal injury, in the form of psychiatric injury, as a result of learning about the acts committed by the appellant. The first respondent developed adjustment disorder. The second respondent underwent a significant exacerbation of the adjustment disorder from which he suffered.
The legal background
The State Immunity Act 1978
The background to the SIA is described in paras 17–26 of this court’s recent decision in Argentum Exploration Ltd v The Silver and all Persons Claiming to be Interested in, and/or Have Rights in Respect of, the Silver [2024] UKSC 16; [2025] AC 555.
As there stated, “while it had become clear that international law in a general way gave support to a restrictive theory of state immunity, the precise limits of the doctrine were still in the course of development and were in many respects uncertain” (para 24). What was needed was “a new statutory scheme providing detailed and comprehensive rules governing both adjudicative and enforcement jurisdiction in cases involving foreign and Commonwealth states. There was also a commercial need to bring domestic law in the United Kingdom into line with the new international reality of restrictive immunity” (para 25). In addition, the United Kingdom wished to ratify the International Convention for the Unification of Certain Rules concerning the Immunity of State-owned Ships, Brussels, 10 April 1926 (“the Brussels Convention”) and the European Convention on State Immunity, Basle, 16 May 1972 (“the ECSI”).
The long title of the SIA reads as follows:
“An Act to make new provision with respect to proceedings in the United Kingdom by or against other States; to provide for the effect of judgments given against the United Kingdom in the courts of States parties to the European Convention on State Immunity; to make new provision with respect to the immunities and privileges of heads of State; and for connected purposes.”.
Part I concerns proceedings in the United Kingdom by or against other states. Section 1 establishes a general immunity from jurisdiction:
A state is immune from the jurisdiction of the courts of the United Kingdom except as provided in the following provisions of this Part of this Act.
A court shall give effect to the immunity conferred by this section even though the state does not appear in the proceedings in question.”
The following sections then set out exceptions to immunity in cases of submission to the jurisdiction (section 2), commercial transactions and contracts to be performed in the United Kingdom (section 3), contracts of employment (section 4), personal injuries and damage to property (section 5), ownership, possession and use of property (section 6), patents, trade-marks etc (section 7), membership of bodies corporate etc (section 8), arbitrations (section 9), ships used for commercial purposes (section 10) and value added tax, customs duties etc (section 11).
19.
Because they were referred to in argument, it is convenient to set out sections 3 and 4 dealing with exceptions from immunity in relation to different kinds of contracts. They provide so far as material as follows:
Commercial transactions and contracts to be performed in United Kingdom.
A State is not immune as respects proceedings relating to –
an obligation of the State which by virtue of a contract (whether a commercial transaction or not) falls to be performed wholly or partly in the United Kingdom.
any loan or other transaction for the provision of finance and any guarantee or indemnity in respect of any such transaction or of any other financial obligation; and
any other transaction or activity (whether of a commercial, industrial, financial, professional or other similar character) into which a State enters or in which it engages otherwise than in the exercise of sovereign authority; …”
A State is not immune as respects proceedings relating to a contract of employment between the State and an individual where the contract was made in the United Kingdom or the work is to be wholly or partly performed there.
Subject to subsections (3) and (4) below, this section does not apply if-
at the time when the proceedings are brought the individual is a national of the State concerned; or
at the time when the contract was made the individual was neither a national of the United Kingdom nor habitually resident there; or
Where the work is for an office, agency or establishment maintained by the State in the United Kingdom for commercial purposes, subsection (2)(a) and (b) above do not exclude the application of this section unless the individual was, at the time when the contract was made, habitually resident in that State. …”
20.
Of particular relevance to the present case is section 5 which provides:
A State is not immune as respects proceedings in respect of–
This Convention is currently in force between eight member states of the Council of Europe, including the United Kingdom.
22.
It provides (article 15) that a contracting state shall be entitled to immunity from the jurisdiction of the courts of another contracting state if the proceedings do not fall within articles 1 to 14. These articles set out various circumstances in which a contracting state cannot claim such immunity. Of relevance to the present case is article 11, which provides:
Article 24(1) allows a contracting state to make a declaration that its courts shall be entitled to entertain proceedings against another contracting state in cases not falling within articles 1 to 13 to the extent that its courts are entitled to do so against non-contracting states. This allows contracting states to retain and develop their own immunity regimes. Article 24(1) provides:
The United Kingdom made a declaration under article 24 at the time of the deposit of its instrument of ratification on 3 July 1979. It provides:
“1(a) In pursuance of the provisions of paragraph 1 of Article 24 thereof, the United Kingdom hereby declare that, in cases not falling within Articles 1 to 13, their courts and the courts of any territory in respect of which they are a Party to the Convention shall be entitled to entertain proceedings against another Contracting State to the extent that these courts are entitled to entertain proceedings against States not Party to the present Convention. This declaration is without prejudice to the immunity from jurisdiction which foreign States enjoy in respect of acts performed in the exercise of sovereign authority (acta jure imperii).”
The United Nations Convention on Jurisdictional Immunities of States and their Property, 2004 (“the UNCSI”)
This was adopted by the General Assembly of the United Nations on 2 December 2004. It is based on the work of the International Law Commission (“the ILC”). It has not yet received sufficient ratifications to enter into force. The United Kingdom has signed but has not yet ratified the Convention.
Article 12 is an exception from immunity for personal injury and damage to property. It provides:
“Unless otherwise agreed between the States concerned, a State cannot invoke immunity from jurisdiction before a court of another State which is otherwise competent in a proceeding which relates to pecuniary compensation for death or injury to the person, or damage to or loss of tangible property, caused by an act or omission which is alleged to be attributable to the State, if the act or omission occurred in whole or in part in the territory of that other State and if the author of the act or omission was present in that territory at the time of the act or omission.”
The judgments below
At first instance, Julian Knowles J dismissed the appellant’s application for a declaration of immunity and other relief—[2023] EWHC 89 (KB).
27.
He held that there was no requirement that the author of the personal injury or property damage be present in the United Kingdom (“a presence requirement”), although he noted that this was not contended for. He stated:
I am clear that it is the location of the act or acts causing the injury which is the issue under s 5, not the presence of the author of the act(s). ... There is no ‘presence requirement’ in s 5. When an instrument on immunity incorporates a presence requirement, this is done so expressly as in Article 11 of [the ECSI] and Article 12 of [the UNCSI].
Parliament’s omission of the presence requirement in s 5 could only have been intentional, given that it was legislating against the backdrop of [the ECSI] (which the UK signed in 1972). Parliament is sovereign and is free to legislate in a way which differs from an international treaty if it wishes to do so, as I shall explain in a moment. It plainly chose not to adopt the Article 11 model.” (Emphasis in original.)
28.
He further held that there was no requirement that every act or omission which is causative of the requisite injury or damage has to take place in the United Kingdom and upheld the respondents’ case that it is sufficient for there to be a substantial and effective causative act or omission. He followed the view he had earlier expressed in the case of Al-Masarir v Kingdom of Saudi Arabia [2022] EWHC 2199 (QB); [2023] QB 475 at para 120:
“In my judgment, the grammatical meaning of section 5, and in particular the use of the indefinite article (death or personal injury caused by ‘an act or omission’) (emphasis added) means what it says. There has to be an act or omission in the UK which is causative of the requisite damage on a more than de minimis basis. Parliament did not say ‘the act or omission’, still less, ‘acts or omissions occurring entirely within the UK’, both of which would have been more supportive of the defendant’s interpretation of section 5.” (Original emphasis)
He also observed in Shehabi at para 131:
“Taking a step back, it seems to me that to uphold the Defendant’s case would empty section 5 of much of its content. It would mean it would not apply except in the most straightforward of cases (eg, a road traffic accident involving a vehicle driven by an employee of a foreign embassy). But many, if not most, of the cases where a foreign state ought not to be immune will involve some tortious activity outside the UK…”
The Court of Appeal (Lady Carr CJ, Males and Warby LJJ) dismissed the appeal, the lead judgment being given by Males LJ—[2024] EWCA Civ 1158; [2025] KB 490.
31.
Ground 1 of the appeal was that when a person located abroad uses a computer to infect a computer in the United Kingdom with spyware, the act in question is to be regarded as having taken place abroad and not in the United Kingdom for the purposes of section 5 of the SIA. In dismissing this ground, Males LJ stated:
In my judgment, as a straightforward use of language, the remote manipulation from abroad of a computer located in the United Kingdom is an act within the United Kingdom. The true position in such a case is that the agents of the foreign state commit acts both in this country and abroad. To distinguish between what happens abroad and what happens here, characterising the former as an act and the latter as merely the effect of the act, is artificial and unprincipled. The reality is that a foreign state which acts in this way is interfering here with the territorial sovereignty of the United Kingdom.
In my judgment this conclusion is in accordance, not only with the language of section 5 of the 1978 Act, but also with the principles underpinning state immunity in international law. That is because the hacking by a foreign state of a computer located in this jurisdiction is an interference with the territorial sovereignty of the United Kingdom …
In modern terms, the hacking of a person’s computer is equivalent to burglars breaking in and stealing the contents of their safe. Just as the latter is an act within the United Kingdom, so too is the former.”
32.
Ground 2 was that all the acts of the foreign state must take place in the United Kingdom. In dismissing this ground Males LJ stated:
I agree with the judge that the language of section 5 is clear and unambiguous in this respect. A foreign state does not have immunity for personal injury caused by an act in the United Kingdom, even if other causative acts take place abroad. Since the language of the section is clear and unambiguous, there is no scope to arrive at a different interpretation based on the external aids on which Professor Sarooshi relied, although I also agree with Mr Silverstone that those aids do not support Bahrain’s interpretation.
55.
I consider that the claimants’ interpretation of section 5 is in accordance with the fundamental principles of international law on which state immunity is based. Once again, a foreign state which hacks a computer located in the United Kingdom interferes with the territorial sovereignty of the United Kingdom even if some of the acts in question take place abroad. Legislation which is broadly similar to the State Immunity Act 1978 has been enacted in numerous jurisdictions and there are international conventions to similar effect, even if such legislation does not (or does not yet) represent customary international law. Accordingly, if State A interferes with the territorial sovereignty of State B by doing an act in State B which is liable to cause death or personal injury to persons in State B, it takes the risk that it will be subject to civil proceedings in State B. Such proceedings are in accordance with principles of international comity.
As already explained, it is apparent from a comparison of section 5 of the 1978 Act with article 11 of the ECSI that Parliament deliberately departed from the terms of article 11. Section 5 is not concerned with ‘the facts’ which occasion the injury or damage, but with ‘an act’ of the foreign state, while the requirement for ‘the author of the injury or damage’ to be present in the forum state has been deliberately omitted. In these circumstances the terms of article 11 are of no real help in interpreting section 5. …”
33.
Against this background, the issues to be determined on the appeal to this court may be summarised as follows:
Does section 5 of the SIA apply only where the act causing personal injury or damage to or loss of tangible property was the responsible or precipitating act?
Does section 5 of the SIA apply only where every act causing personal injury or damage to or loss of tangible property took place in the United Kingdom?
Does section 5 of the SIA apply only where the author of the personal injury or damage to or loss of tangible property was present in the United Kingdom at the time when the facts which occasioned the injury, damage or loss occurred?
34.
It is convenient to address the issues under the following headings:
By a respondents’ notice, the respondents contend, if necessary, that the SIA should be construed, under section 2 of the Human Rights Act 1998, as not conferring immunity on the appellant in this case on the basis that the grant of immunity would breach the respondents’ right of access to a court under article 6 of the European Convention on Human Rights.
The applicable principles of statutory interpretation
The approach to statutory interpretation is well established. The court must seek to give effect to Parliament’s legislative purpose. It does so by ascertaining the meaning of the words used in the statutory provision in question in the light of their context and the purpose of the statute and the provision itself: see, for example, R (Quintavalle) v Secretary of State for Health [2003] UKHL 13; [2003] 2 AC 687, para 8 (per Lord Bingham of Cornhill); R (O) v Secretary of State for the Home Department [2022] UKSC 3; [2023] AC 255, paras 29-31 (per Lord Hodge). The controversial words of the statute are the primary focus.
37.
“They are the words which Parliament has chosen to enact as an expression of the purpose of the legislation and are therefore the primary source by which meaning is ascertained. There is an important constitutional reason for having regard primarily to the statutory context as Lord Nicholls explained in Spath Holme, p 397: ‘Citizens, with the assistance of their advisers, are intended to be able to understand parliamentary enactments, so that they can regulate their conduct accordingly. They should be able to rely upon what they read in an Act of Parliament.’”
(The reference to Spath Holme is to the passage in R v Secretary of State for the Environment, Transport and the Regions, Ex p Spath Holme Ltd [2001] 2 AC 349 at p 397.)
The controversial words of the provision should be read in the context of the section as a whole and in the wider context of the group of sections of which it forms part and of the statute read as a whole. This is the primary means by which Parliament’s meaning is to be ascertained: see R (O) para 30; R (PACCAR Inc) v Competition Appeal Tribunal [2023] UKSC 28; [2023] 1 WLR 2594 (“PACCAR”) at para 42 (per Lord Sales).
So far as the SIA itself is concerned, and as the Court of Appeal noted at para 22, in General Dynamics United Kingdom Ltd v State of Libya [2021] UKSC 22; [2022] AC 318 at para 59 Lord Lloyd-Jones explained that its provisions must be understood in the context of the twin principles of international law: sovereign equality between states and a state’s sovereignty over its own territory.
40.
The International Court of Justice (“the ICJ”) summarised the relationship between these principles in Jurisdictional Immunities of the State (Germany v Italy) [2012] ICJ Rep 99 at para 57, in the following terms. The rule of state immunity is one of the fundamental principles of the international legal order. It derives from the principle of sovereign equality of states and occupies an important place in international law and international relations. It should be viewed together with the principle that each state possesses sovereignty over its own territory and that there flows from that sovereignty the jurisdiction of the state over events and persons within that territory. The ICJ continued:
“Exceptions to the immunity of the State represent a departure from the principle of sovereign equality. Immunity may represent a departure from the principle of territorial sovereignty and the jurisdiction which flows from it.”
The natural meaning of section 5 of the SIA in its domestic statutory context
As the long title of the SIA makes clear, one of its purposes was “to make new provision with respect to proceedings in the United Kingdom by or against other States; …”. It did this in Part I of the SIA (the part in which section 5 appears). Part I is a complete code. Section 1 affords general immunity to foreign states from the jurisdiction of courts in the UK, save as provided for in the exceptions given by other provisions in Part I. If the case does not fall within one of the exceptions to section 1, the foreign state is immune.
As Lord Sumption explained in Benkharbouche v Embassy of the Republic of Sudan [2017] UKSC 62, [2019] AC 777 in a passage concerned with the UNCSI, it is necessary to read the grant of immunity (in article 5 of that Convention) with the exceptions which follow, “as an organic whole” since “[t]he exceptions are so fundamental in their character, so consistent in their objective and so broad in their effect as to amount in reality to a qualification of the principle of immunity itself rather than a mere collection of special exceptions. …” (para 39). The same approach applies to the interpretation of the SIA.
The exceptions to immunity in sections 3 to 8 do not preclude the exercise of adjudicative jurisdiction by UK courts in respect of the matters they identify. In some instances these relate to non-sovereign activities. It is, however, a striking feature of section 5 that it does not distinguish between sovereign and non-sovereign acts. The statutory exceptions also employ various jurisdictional linking factors. Overall, the scheme is intended to reflect both the sovereign equality of states and the United Kingdom’s sovereignty over its own territory.
The requisite jurisdictional linking factor is expressed in different terms in these exceptions. For example, section 3(1)(b) of the SIA provides that a foreign state is not immune as respects proceedings relating to an obligation of the state which, by virtue of a contract, falls to be performed “wholly or partly in the United Kingdom” and section 4(1) provides for an exception from immunity as respects proceedings relating to a contract of employment between the foreign state and an individual where “the work is to be wholly or partly performed” in the United Kingdom.
Section 5 of the SIA is expressed in straightforward language. It provides that a foreign state is not immune as respects proceedings in respect of personal injury caused by an act or omission in the United Kingdom. The section requires an act in the United Kingdom which is causative of personal injury or damage to property. As a matter of ordinary language, there is no requirement in section 5 for “the” act or “all” acts to be in the United Kingdom, still less is there any express requirement of presence in the United Kingdom for the author of the relevant act.
This was the view of the Court of Appeal and Julian Knowles J. Both courts held that the language of section 5 is clear and unambiguous and means that a foreign state does not have immunity for personal injury caused by an act or omission in the United Kingdom, even if other causative acts take place abroad (Court of Appeal, para 54).
This interpretation of section 5 does not involve reading in words. Although it is true that Julian Knowles J described the requisite act as having a causative effect which was more than minimal (para 97), it is not necessary to read in such a qualification, and nor do we read the Court of Appeal as having done so. The words “caused by an act or omission” naturally import the legal requirements of causation in the area of law concerned. In other words, a legally causative act is required. This will exclude acts which are not legally causative because too remote, insignificant or non-operative. But no further qualification is necessary.
48.
Academic commentary supports the conclusion that section 5 simply requires a causative event to occur within the forum state’s jurisdiction. For example, in “State Responsibility and Tort Proceedings against a Foreign State in Municipal Courts” (1989) 20 NYIL 3, pp 25–26, Lady Fox states:
“In cases such as the Lockerbie disaster where the whole damage and a substantive causative event occurs within the local court’s jurisdiction a strong argument can be made that immunity under section 5 of the SIA should be removed so as to enable the victims to sue in the English Court….”
Having referred to a line of cases dealing with multi-state events as supporting that argument (in particular, Distillers Co (Biochemicals) Ltd v Thompson [1971] AC 458 which concerned the thalidomide litigation commenced in New South Wales, where the critical point, as identified by the Privy Council at pp 468–469, was to ascertain whether the defendant’s act which gave the plaintiff his cause of complaint had occurred within the jurisdiction) she continues:
“Applying this line of cases to section 5 it is open to an English court to hold that an act or omission which causes the explosion of an aircraft and loss of life in British airspace constitutes proceedings in respect of death or personal injuries caused by an act or omission within the UK, and hence is sufficient to remove State immunity under section 5.”
The words of section 5 must, of course, be understood in their wider statutory context and in light of the purpose of the SIA. In this regard, it is common ground that one reason for the enactment of the SIA was to enable the UK to ratify the ECSI. It is also common ground that the claim brought by the respondents would not fall within the scope of the exception to immunity contained in article 11 of the ECSI, since under that provision “the facts which occasioned the injury or damage” must have “occurred in the territory of the State of the forum” and the author of the alleged tort must also be present in the forum state. Section 5 by contrast is not concerned, on the face of its wording, with “the facts” which occasion the injury or damage, but with “an act” of the foreign state which is causative of relevant harm, and no requirement for “the author of the injury or damage” to be present in the forum state is expressed in section 5.
We shall set out our detailed considerations of the context in which the SIA was enacted and its legislative purpose when addressing the appellant’s arguments based on article 11 of the ECSI and the relevant principles of international law below. At this stage, however, it is convenient to address the three alternative ways in which the appellant advances its challenge to the interpretation of section 5 of the SIA reflected above.
The appellant’s challenge to the interpretation of section 5
The appellant submits that there are three different ways to achieve the result for which it contends but, ultimately, they come down to a submission that section 5 requires the exclusive act that causes the personal injury or damage to property, or if more than one, all such acts, to take place in the United Kingdom.
52.
We can deal quite shortly with the first two alternative arguments presented, both of which focussed on the meaning of the word “act” in section 5. Taking them in reverse order, the second argument (originally the appellant’s primary argument below, but only faintly relied on at the hearing in this court) is that if “act” means every event occasioning the harm, each must be performed in the United Kingdom. The appellant relies on section 6(c) of the Interpretation Act 1978 (which provides that the singular includes the plural) and contends that its effect here is that the words “caused by an act …” also mean “caused by acts …” Section 5 therefore means both that a state is not immune in respect of proceedings in respect of specified harm “caused by an act … in the United Kingdom” and that a state is not immune in respect of harm “caused by acts … in the United Kingdom”. It submits that the result is that, if more than one act has caused the harm, these must all be acts within the United Kingdom for the exception to immunity to apply.
There is no justification for re-writing section 5 to introduce this limitation. We accept that the effect of section 6(c) of the Interpretation Act 1978 is that where there is more than one act in issue, section 5 means that a state is not immune in respect of harm “caused by acts … in the United Kingdom”, and in fact, the respondents do rely on multiple acts in this case. However, there is no basis in the Interpretation Act 1978 or otherwise for reading in a concept of exclusivity so that the harm must be caused only by an act or acts in the United Kingdom. Provided that the causative acts relied on by a claimant are in the United Kingdom, there is no reason why the claimant should also have to show that there were no other causative acts elsewhere.
54.
The appellant’s alternative argument is that the reference to “act” in section 5 means the responsible, precipitating or initiating act. The appellant relies on the reference in section 5 to “proceedings in respect of …” types of harm commonly protected against by the law of tort and submits that it is therefore concerned with legal actions in tort and the words “caused by an act or omission …” are a reference to the causal inquiry engaged in a tort claim. The appellant submits that events which are subsequent and secondary to the precipitating act are not properly to be regarded as “the cause” of the harm since one must trace the causation inquiry through such acts to identify the cause which is responsible. Once the responsible cause is identified, the causation inquiry ceases. Thus, in this case, the responsible cause, as pleaded, is the inputting of instructions into a computer system and/or computer programme which took place outside the United Kingdom. That resulted in technological processes occurring on the respondents’ personal computers (for example, the corruption of a computer file, or copying of data from a computer’s memory), but these are merely secondary events and not the cause of the alleged harm.
We do not accept this argument and can see no good reason why the word act (or omission) should be read as meaning only the initiating, precipitating or responsible act or omission. In the context of a tort claim, there is no requirement that there will only be one causative act capable of giving rise to a cause of action and it would be odd if section 5 of the SIA applied a different test of causation from the law of tort that is reflected in this provision. Moreover, this interpretation also requires words to be read into section 5 that are not there. Again, we see no justification for reading in the limitation contended for by the appellant in this argument.
56.
The fact that other exceptions from immunity (for example, section 3(1)(b) and 4(1) of the SIA referred to above) expressly provide for an exception from immunity in relation to events that occur “wholly or partly” in the United Kingdom does not support the conclusion that “all of a responsible act must occur in the United Kingdom for immunity to be lost by virtue of section 5”, as the appellant contends. As already indicated, section 3(1)(b) of the SIA is concerned with the place of performance of a contractual obligation and section 4(1) is concerned with the place of performance of an employment contract. In each case, the place of performance could readily occur in more than one (or even multiple) different jurisdictions, and it is therefore unsurprising that the SIA makes clear that only part of the relevant performance need be in the United Kingdom to establish a sufficient link with the United Kingdom. Section 5 is different. The necessary linking factor is “an act” in the United Kingdom which causes personal injury or damage to property, and it is irrelevant whether other acts or omissions occur elsewhere. There is therefore no need for the words “wholly or partly” in section 5.
57.
The appellant submits that the Court of Appeal’s interpretation of section 5 introduces uncertainty (by reference to a de minimis test) and produces arbitrary or absurd consequences and cannot therefore be an interpretation that is clear from the words used in the section. The appellant suggests, for example, that the Court of Appeal’s interpretation would include, in the present context, every automatic technological operation triggered by the instructions delivered from a computer located abroad or a computer programme originating from abroad, every electrical signal generated by those instructions or programme, the acts of the respondents themselves in opening malware or downloading files onto their computers, the failure of the respondents or others to install or update virus protections onto their computers, the acts of persons discovering malware and informing the respondents. Further, the appellant suggests that it would mean that if state A accesses a mobile device or laptop computer belonging to X within its own territory, but X enters the United Kingdom temporarily where more than minimal acts occur, state A would have no immunity in the United Kingdom. Similarly, if X was resident in the US, state A would have immunity from suit in the US but, if technologically significant acts happened to occur on UK servers, state A would have no immunity under English law.
58.
We do not accept these arguments. First, as we have indicated, the Court of Appeal’s interpretation does not involve reading in the words de minimis or any other words. It depends only on the plain meaning of the ordinary words used. Secondly, it is obviously the case that the relevant act or omission in the United Kingdom must not only be a causative act or omission, but also one that is attributable to the foreign state or its agent. There is nothing in the Court of Appeal’s interpretation of section 5 to suggest that remote or non-causative acts or acts done by others, including the respondents, are caught. Thirdly, it follows that the examples relied on by the appellant simply do not flow from the Court of Appeal’s interpretation. If state A interferes with the territorial sovereignty of state B by doing an act in state B which causes personal injury or damage to property, it takes the risk that it will be subject to civil proceedings and not entitled to adjudicative immunity. There is nothing arbitrary or absurd about that. Rather, this is a principled outcome that is consistent with the fundamental principles on which state immunity is based. Moreover, the fact that US law might, by reference to differently worded legislation, lead to a different finding on immunity, cannot affect or determine the proper interpretation of the SIA.
The central issue in these proceedings has therefore become the third way in which the appellant puts its case on the interpretation of section 5, namely that it requires the presence of the responsible actor in the United Kingdom at the time of the act that is causative of personal injury.
As the appellant accepts, unlike article 11 of the ECSI, there is no express presence requirement in section 5 of the SIA. This is to be contrasted with section 16 of the SIA, which deals with several so-called excluded matters in Part I and does have an express presence requirement in subsection (2). Thus section 16(2) provides: “This Part of this Act does not apply to proceedings relating to anything done by or in relation to the armed forces of a State while present in the United Kingdom and, in particular, has effect subject to the Visiting Forces Act 1952” (emphasis added).
The appellant submits that an interpretation of section 5 that is wider in scope than article 11 of the ECSI (because there is no presence requirement) will place the United Kingdom in breach of its obligation under article 15 to recognise immunity in cases not falling within articles 1 to 13 of the ECSI in cases concerning state parties to the ECSI and that this is inconsistent with the purpose and intent of Parliament in enacting the SIA. The appellant therefore contends that, since section 5 is reasonably capable of being read in a manner that complies with articles 11 and 15, section 5 must be read in that way. This is the argument to which we now turn.
One purpose of the SIA was to permit the United Kingdom to become a party to the ECSI. However, it is clear that while the SIA was intended to give broad effect to the ECSI it was not intended to implement it precisely. The long title of the SIA (set out at para 16 above) is significant in this regard. It refers to the ECSI but only in relation to the enforcement of judgments against the United Kingdom. The immunity established by the SIA and the exceptions thereto are simply described as the making of new provision.
That the SIA was not intended to implement precisely the ECSI rules on immunity has long been recognised. In La Générale des Carrières & des Mines v FG Hemisphere Associates LLC [2012] UKPC 27; [2013] 1 All ER 409 Lord Mance explained (at para 10) that the SIA was aimed at giving broad effect to, though not following precisely the wording of, ECSI. In Benkharbouche v Embassy of the Republic of Sudan [2017] UKSC 62; [2019] AC 777 Lord Sumption observed (at para 10) that while one purpose of the SIA was to give effect to the ECSI and thereby enable the United Kingdom to ratify it, by the time it did so in 1979 the Convention had been largely superseded by the adoption of the restrictive doctrine of state immunity at common law. The SIA therefore “dealt more broadly with state immunity”.
The exception to the statutory immunity established by section 5 SIA in the case of personal injury and damage to property corresponds broadly to article 11 of the ECSI but is not, nor is it intended to be, a precise implementation. This is permitted by a power of derogation conferred by article 24(1) of the ECSI and the UK declaration on ratification. (See paras 23 and 24 above.)
65.
The ability of contracting states to derogate from the ECSI is described in the Explanatory Report on the Convention (Misc No 31 (1972); Cmnd 5081 at paras 96–97):
Certain States which at present apply rules of qualified State immunity considered that Article 15, which provides that immunity must be accorded to States in all cases other than those falling within Articles 1 to 13, was too rigid either because some acts iure gestionis fall outside the cases covered by these articles, or because the connecting links prescribed in these articles do not correspond with rules of jurisdictional competence applied in those States. Article 24 permits States to derogate from the provisions of Article 15.
Pursuant to paragraph 1, Contracting States have the option of declaring, by notification to the Secretary General of the Council of Europe, that their courts are to be entitled to entertain proceedings against other Contracting States to the extent that they may entertain such proceedings against third States; for this purpose, treaties concluded with third States which relate to problems of immunity, should not be taken into account. In other words, the regime applied by the courts of a State which has made the declaration will not be affected by the Convention, and can even continue to develop along its own lines.
The declaration addressed to the Secretary General of the Council of Europe will not affect the immunity from jurisdiction enjoyed by foreign States in respect of acts done in the exercise of sovereign authority (acta iure imperii). …
However, the courts may not entertain proceedings within the “grey zone” (ie the matters not covered by Articles 1 to 13 which are subjected to jurisdiction in relations with non-Contracting States) if their jurisdiction can be based solely on an “exorbitant” ground of jurisdiction (paragraph 2: for further details, see the Annex).”
None of the exorbitant grounds of jurisdiction set out in the Annex to the Convention is relevant to the present case.
66.
The effect of this derogation is that the United Kingdom may, consistently with its obligations under the ECSI, exclude immunity in a wider range of cases than is permitted by articles 1 to 13 of the ECSI, to the extent that the exclusion is also applied to non-party states. This is, however, subject to a further requirement found in the last sentence of article 24(1), reflected in the UK declaration, namely that the declaration is “without prejudice to the immunity from jurisdiction which foreign States enjoy in respect of acts performed in the exercise of sovereign authority (acta jure imperii)”. On behalf of the appellant Mr Hickman KC initially submitted that these words exclude from the derogation all acts performed in the exercise of sovereign authority. He submitted, therefore, that to the extent that section 5 of the SIA goes beyond article 11 of the ECSI in denying immunity in respect of sovereign acts, it constitutes a breach of the ECSI. This submission was not maintained in reply, however. In our view, the appellant’s initial submission is a misreading of the final sentence of article 24(1). The words “the immunity from jurisdiction which foreign States enjoy” in respect of sovereign acts make clear that the exclusion from the derogation is limited to those sovereign acts which attract an entitlement to immunity in customary international law. Article 24(1) is not premised on there being immunity for all sovereign acts under customary international law. That not all sovereign acts attract immunity under the ECSI is readily apparent from the fact that the exception to immunity created by article 11 of the ECSI itself draws no distinction between sovereign and non-sovereign acts, as the appellant accepts. It would make no sense to read article 24(1) as requiring immunity for all sovereign acts when article 11 does not adopt that approach. In the context of section 5 of the SIA, therefore, article 24(1) of the ECSI permits the United Kingdom to deny immunity in wider circumstances than those identified in article 11 of the ECSI, provided that in doing so it does not deny immunity in respect of such sovereign acts for which customary international law requires immunity. Such sovereign acts for which customary international law does not require immunity are within the scope of the permitted derogation.
67.
The provisions of the SIA depart from those of the ECSI in a number of important respects. The following instances, on which we were addressed by Professor Philippa Webb, include departures employing different jurisdictional linking factors from those in the ECSI.
(1)
Section 3(1)(a) SIA creates an exception to immunity “as respects proceedings relating to … a commercial transaction entered into by the State”. The closest equivalent to this provision in the ECSI is article 7 which creates an exception in more limited terms. It only applies if the state “has on the territory of the State of the forum an office, agency or other establishment through which it engages, in the same manner as a private person, in an industrial, commercial or financial activity, and the proceedings relate to that activity of the office, agency or establishment”. By contrast, the exception to immunity under section 3(1)(a) extends to a commercial transaction entered into by a state with no territorial connection to the United Kingdom. In NML Capital Ltd v Republic of Argentina [2011] UKSC 31; [2011] 2 AC 495 at para 37, Lord Phillips PSC stated:
“When Parliament enacted the 1978 Act the exemption from immunity under section 3(1)(a) in respect of proceedings relating to a commercial transaction entered into by the state was not qualified by any requirement for a link between the transaction and the United Kingdom. This was not accidental.”
Lord Phillips explained that this was an exercise by the United Kingdom of its powers pursuant to its declaration under article 24(1) of the ECSI.
Section 4(1) provides that a state is not immune “as respects proceedings relating to a contract of employment between the State and an individual where the contract was made in the United Kingdom or the work is to be wholly or partly performed there”. Article 5(1) of the ECSI on the other hand provides for an exception to immunity only “if the proceedings relate to a contract of employment between the State and an individual where the work has to be performed on the territory of the State of the forum”. Under the ECSI immunity is not removed where the contract was made in the forum state. Furthermore, the words “wholly or partly” do not appear in article 5(1) with the result that the exclusion from immunity appears to apply only where the work has to be performed solely on the territory of the forum state.
Section 5 of the SIA provides that a state is not immune “as respects proceedings in respect of – (a) death or personal injury; or (b) damage to or loss of tangible property, caused by an act or omission in the United Kingdom”. By contrast article 11 of the ECSI excludes immunity “in proceedings which relate to redress for injury to the person or damage to tangible property, if the facts which occasioned the injury or damage occurred in the territory of the State of the forum, and if the author of the injury or damage was present in that territory at the time when those facts occurred.” The substitution in section 5 of “an act or omission” for “the facts” and the omission in section 5 of the presence requirement appear to expand the scope of the exception in section 5 beyond that of the exception in article 11.
“Where a State has agreed in writing to submit a dispute which has arisen, or may arise, to arbitration, the State is not immune as respects proceedings in the courts of the United Kingdom which relate to the arbitration.”
This exception to immunity is wider than that under the equivalent provision in article 12 of the ECSI which only applies to “a dispute which has arisen or may arise out of a civil or commercial matter” and only applies to “proceedings relating to: (a) the validity or interpretation of the arbitration agreement; (b) the arbitration procedure; (c) the setting aside of the awards, unless the arbitration agreement otherwise provides”. (See Svenska Petroleum Exploration AB v Government of the Republic of Lithuania [2006] EWCA Civ 1529; [2007] QB 886 at paras 122, 137.) In addition, section 9 of the SIA, unlike article 12 of the ECSI, does not require the arbitration to take place in the United Kingdom or according to the law of the United Kingdom and it also applies more generally to proceedings which relate to the arbitration.
Section 13(4) of the SIA creates an exception to the prohibition in section 13(2)(b) on enforcement against the property of a state. Section 13(4) permits enforcement of a judgment or arbitration award against property which is for the time being in use or intended for use for commercial purposes or, in an action in rem, for proceedings for the arrest, detention or sale of such property, save, in certain circumstances, in respect of the property of a contracting state to the ECSI. Article 23 of the ECSI, on the other hand, prohibits any measures of execution against the property of a contracting state except where that state has consented in a particular case. In General Dynamics United Kingdom Ltd v State of Libya [2021] UKSC 22; [2022] AC 318; Lord Lloyd-Jones drew attention to the deliberate divergence of the SIA from the ECSI in this regard (at paras 47–48, 76(4)).
It appears therefore that sections 3(1)(a), 4, 5, 9 and 13(4) SIA have a different scope (including the use of different jurisdictional connecting factors) from the corresponding provisions in the ECSI.
69.
The exception to immunity under article 11 of the ECSI in the case of proceedings which relate to redress for injury to the person or damage to tangible property does not distinguish between sovereign and non-sovereign acts and applies to both. Similarly, section 5 of the SIA does not draw any distinction between sovereign and non-sovereign acts and applies to both. (See generally, Argentum Exploration at paras 29, 30. See also Al-Masarir.) This view is held generally by commentators. For example:
Fox and Webb, The Law of State Immunity (3rd rev ed, 2015) states that “the UK legislation followed ECSI in allowing the exception for tortious acts, whether performed de jure imperii or de jure gestionis”.
Dicey, Morris and Collins on the Conflict of Laws (16th ed, 2022) states (at para 9–050) with regard to the exception to immunity created by section 5:
“It is suggested that the question is one of statutory construction and that there is no basis for excluding acts which are jure imperii. Consequently if agents of a foreign State on an intelligence mission in England commit acts within England that cause personal injury or death to someone in England, then neither they or the foreign State will be entitled to immunity.”
Similarly, Dickinson and Thompson, The State Immunity Act 1978 (2025) state (at para 8.6):
“On their face, those provisions [article 11 of the ECSI, article 12 of the UNCSI and section 5 of the SIA] are notable in representing, respectively, the two Conventions’ and the 1978 Act’s clearest departures from the traditional distinction between sovereign and private acts within the law on state immunity. They contain no express language that would restrict this exception to commercial or other non-sovereign activities.”
(4)
“These two limitations, the exclusion of pure economic loss and the territorial jurisdictional link, are the only express limitations to be found in the English express tort exception. As already noted, there is no requirement that the acts must be committed in the course of private law or commercial activities. On the face of it the section removes immunity for claims arising out of personal injuries caused by overtly political or administrative functions of the State. Thus the US decisions of Letelier v Chile 488 F Supp 665 (DDC 1980) where there was held to be no immunity for the assassination in the United States of the former Chilean ambassador, and in Olsen by Sheldon v Government of Mexico 729 F 2d 641 (9th Cir 1984), cert denied 105 Sup Ct 295 (1984) where a claim was allowed for the death of prisoners caused when they were being transferred from Mexico to the US authorities might provide relevant precedents for the English court.” (Fox, State Responsibility and Tort Proceedings against a Foreign State in Municipal Courts, (1989) 20 NYIL 3, 26–27.)
The exception created by article 11 is expressly limited to cases where the author of the injury or damage was present in the territory of the forum state at the time when the facts which occasioned the injury or damage occurred. No corresponding words appear in section 5 of the SIA.
71.
On behalf of the appellant Mr Hickman relies on the principle of consistent interpretation as expressed by Lord Diplock in The Eschersheim [1976] 1 WLR 430. That case concerned the interpretation of provisions in the Administration of Justice Act 1956 which had been enacted for the purpose, among others, of enabling the United Kingdom to ratify the International Convention Relating to the Arrest of Seagoing Ships (1952):
“As the Act was passed to enable Her Majesty’s Government to give effect to the obligations in international law which it would assume on ratifying the Convention to which it was a signatory, the rule of statutory construction laid down in Salomon v. Customs and Excise Commissioners [1967] 2 QB 116 and Post Office v. Estuary Radio Ltd. [1968] 2 QB 740 is applicable. If there be any difference between the language of the statutory provision and that of the corresponding provision of the Convention, the statutory language should be construed in the same sense as that of the Convention if the words of the statute are reasonably capable of bearing that meaning.” (at pp 436B–D)
The appellant also relies on the following statement by Lord Diplock in Garland v British Rail Engineering Ltd [1983] 2 AC 751 to similar effect:
“… it is a principle of construction of United Kingdom statutes, now too well established to call for citation of authority, that the words of a statute passed after the Treaty has been signed and dealing with the subject matter of the international obligation of the United Kingdom, are to be construed, if they are reasonably capable of bearing such a meaning, as intended to carry out the obligation, and not to be inconsistent with it.” (at p 771B)
Mr Hickman points to the difference in wording between article 11 of the ECSI and section 5 of the SIA. Despite the absence of any express requirement in the statute that the person performing the relevant act in the United Kingdom must be present in the United Kingdom at that time, he submits that section 5 is reasonably capable of bearing that meaning and it should therefore be interpreted accordingly. Otherwise, he submits, the exercise of jurisdiction in circumstances where the actor was not present in the United Kingdom at the time of the relevant act would constitute a breach of the ECSI.
We do not accept this submission. The words of section 5 are not reasonably capable of bearing the same meaning as the words of article 11 which impose the presence requirement. Death, personal injury or damage can be caused by an act in the United Kingdom without the author being present in the United Kingdom at the time of the act. The natural meaning of the word “act” in this context is not limited to the act of a person or the “physical movement of a human being”. This is demonstrated by the fact that, as further discussed below, both article 11 of the ECSI and article 12 of the UNCSI expressly employ two distinct linking factors making clear the possibility of an act being performed in the territory of a state without the author of that act also being present in the territory of that state.
74.
As article 24(1) permits the extension of an exception to immunity beyond that recognised by the Convention, the appellant’s submission also makes assumptions about the position in customary international law. The United Kingdom, having made an appropriate declaration under article 24(1), was entitled to apply wider exceptions to immunity than those in the ECSI, provided that this did not result in an exercise of jurisdiction in respect of sovereign acts which is forbidden by customary international law. We will return to this question later in this judgment. However, even if there is in customary international law no territorial tort exception which extends to sovereign acts or even if such an exception exists but is limited by a requirement that the actor must be present in the jurisdiction at the time of the relevant act, it is clear that the principle of consistent interpretation has no application here because the departure from the treaty scheme is deliberate. In our view, this is determinative of the appeal.
75.
Article 11 of the ECSI requires two distinct linking factors in the exception it creates: (a) the facts which occasioned the injury or damage must have occurred in the territory of the forum state and (b) the author of the injury or damage must have been present in that territory at the time when those facts occurred. The first condition is reflected in the express terms of section 5 of the SIA. The second is not. Contrary to the submission on behalf of the appellant, the words in section 5 “caused by an act or omission in the United Kingdom” do not in their natural meaning import the second condition. This is even more readily apparent when the words of section 5 are compared to those of article 11. If it had been the intention to require both linking factors in section 5, the obvious course would have been to recite both using the words of article 11. If that had been the intention it is inconceivable that the second linking factor should have been omitted. Nor can its omission be explained by a difference in drafting styles between the ECSI and the SIA; it was open to Parliament to include a presence requirement without adopting the exact language of article 11. In the circumstances we are driven to the conclusion that the omission was deliberate. Mr Hickman accepted in the course of argument that there had been a deliberate departure from the wording of article 11 of the ECSI but maintained that there can have been no intention to depart from its effect. We disagree. Whether or not the possibility of a transborder situation involving a remotely activated device causing personal injury or property damage was in contemplation, the simple step of adopting the second linking factor (the presence requirement) would have made the position clear if a presence condition was intended. It cannot have been regarded as unnecessary. Nor is it possible that nobody could have envisaged an act taking place other than where the actor was present. There was here a clear and deliberate departure from the presence requirement. The adoption of a single linking condition in section 5 was entirely consistent with the other instances, referred to above, in which the statute employs different jurisdictional linking factors from those employed by the ECSI.
76.
As a result, the principle of consistent interpretation has no application here. The statute is not intended to be a precise implementation of the ECSI and there has been a deliberate decision to depart from the ECSI scheme in this particular respect. The position was explained by Lord Hoffmann in R v Lyons [2003] 1 AC 976 (at paras 27, 28) in the following terms:
“Parliament may pass a law which mirrors the terms of the treaty and in that sense incorporates the treaty into English law. But even then, the metaphor of incorporation may be misleading. It is not the treaty but the statute which forms part of English law. And English courts will not (unless the statute expressly so provides) be bound to give effect to interpretations of the treaty by an international court, even though the United Kingdom is bound by international law to do so. Of course there is a strong presumption in favour of interpreting English law (whether common law or statute) in a way which does not place the United Kingdom in breach of an international obligation. As Lord Goff of Chieveley said in Attorney General v Guardian Newspapers Ltd (No 2) [1990] 1 AC 109, 283: ‘I conceive it to be my duty, when I am free to do so, to interpret the law in accordance with the obligations of the Crown under [the Convention].’
But for present purposes the important words are ‘when I am free to do so’. The sovereign legislator in the United Kingdom is Parliament. If Parliament has plainly laid down the law, it is the duty of the courts to apply it, whether that would involve the Crown in breach of an international treaty or not.”
For these reasons we conclude that the principle of consistent interpretation has no application and that article 11 of the ECSI does not mean that section 5 of the SIA has to be interpreted so as to require the author of the injury to have been present in the territory of the United Kingdom at the time when the acts which caused the injury occurred.
We now turn to the argument that section 5 of the SIA should be so interpreted as otherwise it would result in an exercise of jurisdiction in respect of sovereign acts forbidden by customary international law and that Parliament cannot have so intended.
For the reasons explained below, we have concluded that customary international law does not indicate still less compel a conclusion that section 5 must be interpreted, as the appellant contends, as containing a presence requirement. In our view there was a reasonable basis in customary international law for the United Kingdom to conclude that it could lawfully give effect to a territorial exception to immunity in section 5 applicable erga omnes (and not just by mutual consent with contracting states to the ECSI) which did not distinguish between sovereign and non-sovereign acts and which employed a different jurisdictional linking factor from that contained in article 11 of the ECSI. Moreover, even if no territorial tort exception exists in customary international law, that does not support the appellant’s interpretation of section 5.
The final sentence of article 24(1) of the ECSI prohibits the extension of non-immunity only in situations where the grant of immunity in respect of sovereign acts is required by customary international law. As a result, the focus shifts to customary international law. It is common ground before us that the conduct alleged in these proceedings would be a sovereign activity. The question then becomes whether there exists in customary international law an exception to immunity which applies to sovereign acts in the forum state causing injury or damage. For convenience, we will adopt the term “territorial tort exception” although such an exception would not necessarily be limited to causes of action in tort.
81.
To demonstrate a rule of customary international law requires both widespread, representative and consistent state practice and an acceptance by states that the practice is followed as a matter of legal obligation (opinio juris) (North Sea Continental Shelf Cases (Federal Republic of Germany v Denmark; Federal Republic of Germany v The Netherlands) [1969] ICJ Rep 3, para 77; Military and Paramilitary Activities in and against Nicaragua (Nicaragua v United States of America) [1986] ICJ Rep 14, para 186; Jennings and Watts (eds), Oppenheim’s International Law, vol 1, 9th ed, pp 25–36; Draft conclusions on identification of customary international law, with commentaries, ILC Yearbook [2018] vol II(2), p 100 (conclusion 8)). Establishing a rule of customary international law can be a difficult issue for municipal courts to address. Procedures in this jurisdiction do not readily lend themselves to the extensive survey of state practice necessary to establish practice and legal obligation. We also have firmly in mind the warnings in Jones v Ministry of the Interior of the Kingdom of Saudi Arabia [2007] 1 AC 270 (per Lord Bingham at para 22, per Lord Hoffmann at para 63) that national courts should exercise restraint and not force the pace by recognising new rules of customary international law where there is no consensus among states. As Lord Bingham observed “one swallow does not make a rule of international law”.
In this regard it is highly significant that in Jurisdictional Immunities of the State (Germany v Italy: Greece intervening) [2012] ICJ Rep 99 the International Court of Justice (at paras 64, 65) left unresolved the question whether there exists in customary international law a “territorial tort exception” to state immunity which would apply to sovereign acts in the forum state causing death, personal injury or damage to property. The case concerned the denial of immunity by Italian courts in proceedings against Germany in respect of acts of German armed forces in Italy during the Second World War. Italy submitted that customary international law had developed to the point where a state was no longer entitled to immunity in respect of acts occasioning death, personal injury or damage to property on the territory of the forum state, even if the act in question was performed jure imperii.
83.
With regard to the existence of a territorial tort exception to immunity the ICJ stated (at paras 64–5):
“The Court begins by observing that the notion that State immunity does not extend to civil proceedings in respect of acts committed on the territory of the forum State causing death, personal injury or damage to property originated in cases concerning road traffic accidents and other ‘insurable risks’. The limitation of immunity recognized by some national courts in such cases was treated as confined to acta jure gestionis (see, e.g., the judgment of the Supreme Court of Austria in Holubek v. Government of the United States of America (Juristische Blätter (Vienna), Vol. 84, 1962, p. 43; ILR, Vol. 40, p. 73)). The Court notes, however, that none of the national legislation which provides for a ‘territorial tort exception’ to immunity expressly distinguishes between acta jure gestionis and acta jure imperii. The Supreme Court of Canada expressly rejected the suggestion that the exception in the Canadian legislation was subject to such a distinction (Schreiber v. Federal Republic of Germany and the Attorney General of Canada, [2002] Supreme Court Reports (SCR), Vol. 3, p. 269, paras. 33–36). Nor is such a distinction featured in either Article 11 of the European Convention or Article 12 of the United Nations Convention. The International Law Commission’s commentary on the text of what became Article 12 of the United Nations Convention makes clear that this was a deliberate choice and that the provision was not intended to be restricted to acta jure gestionis (Yearbook of the International Law Commission, 1991, Vol. II (2), p. 45, para. 8). Germany has not, however, been alone in suggesting that, in so far as it was intended to apply to acta jure imperii, Article 12 was not representative of customary international law. In criticizing the International Law Commission’s draft of what became Article 12, China commented in 1990 that ‘the article had gone even further than the restrictive doctrine, for it made no distinction between sovereign acts and private law acts’ (United Nations doc. A/C.6/45/SR.25, p. 2) and the United States, commenting in 2004 on the draft United Nations Convention, stated that Article 12 ‘must be interpreted and applied consistently with the time-honoured distinction between acts jure imperii and acts jure gestionis’ since to extend jurisdiction without regard to that distinction “would be contrary to the existing principles of international law” (United Nations doc. A/C.6/59/SR.13, p. 10, para. 63).
The Court considers that it is not called upon in the present proceedings to resolve the question whether there is in customary international law a ‘tort exception’ to State immunity applicable to acta jure imperii in general. The issue before the Court is confined to acts committed on the territory of the forum State by the armed forces of a foreign State, and other organs of State working in co-operation with those armed forces, in the course of conducting an armed conflict.”
So far as that narrower issue was concerned, the ICJ went on to conclude (at paras 77–79) that customary international law continues to require that a state be accorded immunity in proceedings for torts allegedly committed on the territory of another state by its armed forces and other organs of state in the course of conducting an armed conflict.
There is, as the ICJ acknowledged in Jurisdictional Immunities, a substantial body of state practice supporting the existence of a territorial tort exception applicable to sovereign acts. States frequently claim the right to exercise jurisdiction over acts of foreign states performed in the forum state which cause injury or damage, without drawing any distinction between sovereign and non-sovereign acts. This view also finds support in other materials.
85.
Article 11 of the ECSI does not distinguish between sovereign and non-sovereign acts. (See para 22 above.) The same is true of article 12 of the UNCSI. The International Law Commission commentary on the draft article which became article 12 of the UNCSI makes clear that this was a deliberate choice:
“The basis for the assumption and exercise of jurisdiction in cases covered by this exception is territoriality. The locus delicti commissi offers a substantial territorial connection regardless of the motivation of the act or omission, whether intentional or even malicious, or whether accidental, negligent, inadvertent, reckless or careless, and indeed irrespective of the nature of the activities involved, whether jure imperii or jure gestionis. This distinction has been maintained in the case law of some States involving motor accidents in the course of official or military duties. While immunity has been maintained for acts jure imperii, it has been rejected for acts jure gestionis. The exception proposed in article 12 makes no such distinction, …” (ILC Yearbook [1991] vol II(2), p 45, para 8).
86.
As we have seen, in the Jurisdictional Immunities case the ICJ noted that none of the national legislation which provides for a “territorial tort exception” to immunity expressly distinguishes between sovereign and non-sovereign acts. (The ICJ was here referring to the legislation of the following states: United Kingdom (section 5, SIA); United States (para 1605(a)(5), Foreign Sovereign Immunities Act (“FSIA”) (which should be contrasted with its 2004 statement cited by the ICJ at para 64, cited at para 84, above)); Canada (section 6, State Immunity Act 1985); Argentina (Law No 24.488 (1995); Australia (section 12, Foreign States Immunities Act 1985); Singapore (section 7, State Immunity Act 1985); South Africa (section 6, Foreign States Immunities Act 1981); Israel (section 5, Foreign State Immunity Law 2008); Japan (article 10, Act on the Civil Jurisdiction of Japan with respect to a Foreign State 2009)). The Pakistan State Immunity Ordinance 1981 contained no tort exception. At the hearing of the present appeal we were referred to further national legislation none of which distinguishes between sovereign and non-sovereign acts in respect of its territorial tort exception: Malawi (section 7, Immunities and Privileges Act (1984)); Russia (article 11, Federal Law No 297–FZ (2015)); Spain (article 11, Organic Law on Privileges and Immunities of Foreign States (2015)); Hungary (section 84, Act on Private International Law (2017)); China (article 9, Foreign State Immunity Law (2024)). The last is particularly significant because the ICJ judgment in the Jurisdictional Immunities case stated (at para 64) that in 1990 China had been critical of what became article 12 UNCSI on the ground that it made no distinction “between sovereign acts and private law acts”. (UN doc. A/C.6/45/SR.25)
In Schreiber v Federal Republic of Germany and the Attorney General of Canada 2002 SCC 62, a decision on section 6(a) of the Canadian State Immunity Act 1985, the Canadian Supreme Court observed that “most of the international law authorities cited by the parties appear to accept that the personal injury exception does not distinguish between jure imperii and jure gestionis acts” (para 34). It referred to article 11 of the ECSI. It also referred to the draft article which became article 12 of the UNCSI and the ILC commentary on it. The court observed (at para 37) that an alternative interpretation would deprive the victims of the worst breaches of basic rights of any possibility of redress in national courts which would jeopardise the progress in the protection of human rights.
In Letelier v Republic of Chile 488 F Supp 665 the US District Court, District of Columbia, held that neither the FSIA nor the act of state doctrine would protect a foreign government from a civil suit if it ordered an assassination that took place in the United States. District Judge Green observed (at para 3) that nowhere in the FSIA was there an indication that the tortious acts to which it makes reference were only to be those formerly classified as private, thereby engrafting onto the statute a requirement to determine whether the given sovereign act was jure gestionis or jure imperii. (See also Olsen by Sheldon v Government of Mexico 729 F 2d 641 (9th Cir 1984), cert denied 105 Sup Ct 295 (1984).)
We have, nevertheless, not been referred to any decided case in which jurisdiction was exercised where the agent of the foreign state performing the act which caused injury or damage was not physically present in the territory of the forum state at the relevant time.
90.
Reference has been made above to writings of scholars describing how the distinction between sovereign and non-sovereign acts has been disregarded in this context. (See para 70 above.) Yang, State Immunity in International Law (2012) states:
“In current State practice, the jure imperii / jure gestionis dichotomy seems rather out of place with regard to jurisdiction over tortious conduct in the territory of the forum State. A number of states, chief among them the US and the UK, have abandoned the test altogether.” (at pp 207–208)
“The jure imperii / jure gestionis distinction has been discarded altogether in some States and is followed by others only to the extent that, when injury has been caused, causing injury will be regarded as jure gestionis, regardless of whether the act in the course of which injury has been caused is jure imperii or jure gestionis. Except where foreign armed forces are involved, it is rather doubtful if, in the event of actual injury being caused in the State of the forum, a court will deny jurisdiction on the sole ground that the act in question is jure imperii.” (at p 229) (Emphasis in original.)
91.
It may be objected that in interpreting section 5 of the SIA it is necessary to have regard to the state of customary international law as at 1978, the year of its enactment. Ascertaining precisely when a rule of customary international law became established can be problematic. Professor James Crawford has written of the ICJ’s reference to “nascent opinio juris” in the Nuclear Weapons Opinion (Legality of the Threat or Use of Nuclear Weapons, Advisory Opinion [1996] ICJ Rep 226, at para 73) as suggesting “that opinio juris does not spring into existence fully formed at a particular moment but must undergo a period of gestation” (James Crawford, Chance, Order, Change: The Course of International Law (2013) The Hague Academy Collected Courses, vol 365, p 61, para 70). In Military and Paramilitary Activities in and against Nicaragua (Nicaragua v United States of America) [1986] ICJ Rep 14 the ICJ stated:
“In order to deduce the existence of customary rules, the Court deems it sufficient that the conduct of States should, in general, be consistent with such rules, and that instances of State conduct inconsistent with a given rule should generally have been treated as breaches of that rule, not as indications of the recognition of a new rule” (para 186).
It also made the following particularly pertinent observation:
“[r]eliance by a State on a novel right or an unprecedented exception to the principle might, if shared in principle by other States, tend towards a modification of customary international law” (para 207).
While it is necessary to be cautious in attaching weight to the inaction of states in failing to react to practice by other states as support for the customary status of that practice, “[t]olerance of a certain practice may indeed serve as evidence of acceptance as law (opinio juris) when it represents concurrence in that practice”, in particular when states were in a position to react and the circumstances called for some reaction (Draft conclusions on identification of customary international law, with commentaries, ILC Yearbook [2018] vol II(2), pp 103–104 (Conclusion 10(3) and comment 8)). In this regard we note that “[w]hen the State Immunity Bill was before … Parliament copies of it were sent to all diplomatic missions in London on two occasions. … No State which was sent the legislation in draft offered substantive criticism of its terms.” The accompanying note explained that “the United Kingdom intended to apply the provisions of the Bill to all sovereign States in the belief that the provisions of the European Convention reflected with sufficient accuracy general State practice in the field of sovereign immunity”. (Geoffrey Marston (ed), United Kingdom Materials on International Law (1980) 51 British Yearbook of International Law 355, 424; ILC, Jurisdictional Immunities of States and Their Property: Information and Materials Submitted by Governments, 14 April 1981, UN Doc A/CN.4/343, pp. 27–28.) A reaction by the recipient states would have been possible and called for if they did not accept that the clause which became section 5 did not reflect the position in customary international law.
92.
It is clear that in 1978 the United Kingdom followed the example of the United States in enacting a territorial tort exception to state immunity which extended to sovereign acts of a foreign state. Some indication of where these changes sit in the context of state practice is provided by the retrospective view of the International Law Commission’s Special Rapporteur on jurisdictional immunities of states and their property and in the responses of states to the ILC.
In his Fifth Report published in 1983 (Fifth Report on jurisdictional immunities of States and their property; UN Doc A/CN.4/363) Mr S Sucharitkul drew attention to the limited evidence of state practice prior to the 1970s allowing or disallowing immunity in respect of proceedings for personal injuries and damage to property (at para 76). National legislation had been adopted “in a number of countries in the last decade or so” (at para 83). He referred to Letelier as “a clear precedent for the award of pecuniary damages against a foreign State in connection with proven acts of political violence in the United States” (at para 84). Although not all ILC members agreed with the Special Rapporteur’s assessment of practice, his proposed wording was substantially reflected in what subsequently became article 12 of the UNCSI.
At the 1762nd meeting on 17 May 1983 (ILC Yearbook [1983] vol 1, p 46, para 15) Mr Sucharitkul observed:
“In recent times, the law had changed somewhat. In the past, the essential difference had been between acta jure gestionis and acta jure imperii and State immunity had fully applied in cases where the person responsible for the accident had been acting in an official capacity. That distinction had been largely abandoned, although it was still applied in Belgium and Egypt. In that connection, it was interesting to note the developments in United States judicial practice …”
He then drew attention (at para 16) to section 5 of the SIA (“an exception that was indeed very broad”) and noted that similar legislation had been enacted or was being drafted in several other common law countries. Article 11 of the ECSI contained an analogous provision but he noted that the provisions in the SIA applied to all states and not merely those that were parties to the ECSI. With regard to international opinion, he observed that the International Law Association Draft Convention on State Immunity (adopted in 1982) included a provision akin to that in the SIA ((1983) 22 ILM 287, 290 (article III(F)).
His successor as Special Rapporteur, Mr Motoo Ogiso, ultimately maintained the version of the text that had been provisionally adopted by his predecessor (UN Doc A/CN.4/431, p 25).
(4)
Of the states that submitted comments and observations to the ILC on the lead up to the adoption of the ILC draft articles in 1991 about the provision in the ILC draft articles which became article 12 of UNCSI, very few expressed the view that it was not reflective of customary international law. Only Brazil clearly stated that “[t]he precedents of practice that can be invoked to justify the exceptions to State immunity contained in [the relevant provisions and elsewhere] do not indicate the existence of a general acceptance of those exceptions” (ILC Yearbook [1988] vol II (1), p 58, para 5). Chile also considered the relevant provision “very broad and imprecise” (p 63, para 11). Most states that expressed concerns about this provision had other concerns, including domestic courts becoming involved in determining questions of state responsibility and the law of state immunity diverging from that governing diplomatic agents. (See, in particular, Bulgaria (p 59, para 10), the Byelorussian SSR (p 61, para 12), Czechoslovakia (p 64, para 8), the German Democratic Republic (p 69, para 20), and the USSR (p 83, para 13.) France considered the relevant provision “too absolute” (p 66, para 28). Similar concerns seem to have informed the doubt expressed by certain ILC members about the relevant provision (for example, ILC Yearbook [1990] vol II(2), pp 35–36, para 189).
(5)
The position of delegates at the Sixth Committee of the United Nations General Assembly between 1989 and 1991 also provides some support for the existence of a territorial tort exception to state immunity which extends to sovereign acts. Chile indicated that the relevant draft article “was particularly important for [it]” (UN doc A/C.6/45/SR.30, p 9, para 38). While a small number of states raised concerns as regards the application of this exception in respect of sovereign acts (Bahrain: UN Doc A/C.6/44/SR.34, p 21, para 85; China: UN doc A/C.6/45/SR.25, p 2, para 3; and Czechoslovakia: UN Doc A/C.6/46/SR.23, p 5, para 21), and although Venezuela took the view that the draft article “should be deleted, basically because exceptions to State immunity must be reduced to a minimum” (UN doc A/C.6/45/SR.24, p 6, para 18), other states supported the draft article, with some even considering that it ought to go further: Federal Republic of Germany (UN doc A/C.6/44/SR.34, p 17, para 68), Italy (UN doc A/C.6/44/SR.35, p 4, para 10), and Uruguay (UN doc A/C.6/46/SR.24, p 7, para 27). According to Australia, the draft article in question “accorded with the practice of those States which had taken a public position on the issue of personal injuries and damage to property” (UN doc A/C.6/45/SR.24, p 11, para 51).
Following the adoption by the ILC of its draft articles on the jurisdictional immunities of states and their property in 1991, a second round of comments from states with a view to concluding a multilateral convention contained little opposition to the article in question (eg the particular concerns of the USA (UN doc A/47/326, pp 31–32, para 19)). Rather, further support emerged in favour of the article in question, with Austria considering its retention to be “essential” (UN doc A/47/326, p 12, para 12) and Belgium expressly “endors[ing] the principle set forth in article 12” (UN doc A/48/313, p 5).
Whether the practice referred to above is sufficient to establish as a rule of customary international law a territorial tort exception applicable to sovereign acts, in 1978 or even today, is, in any event, an issue on which we do not need to express a concluded view. It certainly appears that state practice has moved in the direction of accepting such an exception. It is not necessary, however, to express a view as to whether the point has yet been reached at which divergence from an established norm has gained sufficient support from state practice to become established as a new rule of customary international law. For present purposes it is sufficient that, as explained above, there was a basis on which the United Kingdom could reasonably conclude that it could, in conformity with customary international law, enact a territorial tort exception of general application which applies where a relevant act took place in the United Kingdom. Such a conclusion is reflected in section 5 of the SIA.
In any event, there are several reasons why we consider that, even if no territorial tort exception exists in customary international law, that does not support the appellant’s interpretation of section 5 of the SIA.
95.
First, as we have seen (para 70 above), both section 5 of the SIA and article 11 of the ECSI clearly apply, on any reading, to both sovereign and non-sovereign acts. It is undeniably the effect of section 5 that it does not distinguish between sovereign and non-sovereign acts performed in the United Kingdom. That is the natural meaning of the provision and the unanimous view of commentators. It would have been open to the United Kingdom to restrict immunity in this way in respect of the other states party to the ECSI as a matter of mutual agreement. However, section 5 is not so limited and applies to all states (erga omnes) and not just contracting states. As a result, denial of immunity for sovereign acts on a consensual basis cannot explain section 5. In enacting section 5 the United Kingdom must have intended to give effect to a territorial tort exception of general application, which applies to sovereign acts where a relevant act took place in the United Kingdom. For the reasons set out above, we consider that there was at least a basis on which the United Kingdom could reasonably have considered that such a course was consistent with customary international law.
Secondly, if, as the appellant maintains, there exists in customary international law no territorial tort exception to immunity applicable to sovereign acts, section 5 would permit the exercise of jurisdiction in breach of customary international law. That, however, would be the case even if section 5 were interpreted, as the appellant proposes, to conform with article 11 of the ECSI by reading in a requirement that the author of the injury or damage be present in the forum state. On this hypothesis, the appellant’s proposed interpretation of section 5 would not cure the resulting breach of customary international law.
97.
Thirdly, although Mr Hickman on behalf of the appellant made clear that it does not accept that there is in customary international law a territorial tort exception applicable to sovereign acts, he was bound to acknowledge that section 5 of the SIA (and article 11 of the ECSI) creates an exception to state immunity applicable to sovereign as well as non-sovereign acts which applies erga omnes. As a result, he was compelled to modify his submissions and to contend that the court should prefer a reading of section 5 which minimises the material scope for states to be subject to the jurisdiction of foreign domestic courts in relation to sovereign acts. In other words, the court should adopt an interpretation which would impinge less extensively upon sovereign acts. However, section 5 already requires satisfaction of the pre-eminent linking factor justifying a denial of immunity, namely an act by the foreign state within the territory of the forum state (see excerpt from the ILC commentary describing it as “a substantial territorial connection” regardless of other considerations, para 86 above).
98.
Fourthly, it is a particular feature of the present case that, on the assumed facts, the agents of the foreign state performing the act which caused personal injury or damage were not physically present in the territory of the forum state at the relevant time. The appellant assumes, erroneously, that, if in customary international law a territorial tort exception applies to sovereign acts, article 11 of the ECSI and article 12 of the UNCSI (and in particular their requirement of the presence in the territory of the forum state of the author of the injury or damage) necessarily define the scope of the exception. There is, however, no particular reason why presence should define the limits of the area of non-immunity in respect of sovereign acts in the forum state causing personal injury or damage. Article 11 of the ECSI and article 12 of the UNCSI are relevant in illustrating that the sovereign quality of acts does not give rise to immunity in respect of torts involving injury or damage caused by acts in the forum state, but those provisions do not necessarily define the limits of such an exception as a matter of customary international law. On the contrary, it appears that the requirement of the author’s presence in article 11 of the ECSI and article 12 of the UNCSI may serve the very different purpose of defining jurisdiction as opposed to immunity. This is considered in detail at paras 101–106, below.
Fifthly, the appellant’s argument proves too much. The appellant maintains that it is necessary to read down section 5 in order to comply with customary international law because in 1978 there was no established territorial tort exception. On this argument customary international law would also require that sovereign acts performed by agents of a foreign state present in the United Kingdom would be entitled to immunity. On this basis, a foreign state whose agents carried out an assassination in the United Kingdom would be entitled to immunity. Even if otherwise permissible, to read down section 5 in this way would clearly conflict with the intention of Parliament.
100.
The requirement of the author’s presence in the forum state arises from concerns as to jurisdiction as opposed to immunity. The Explanatory Report on the ECSI states (at para 47) that article 11 was drafted on the lines of article 10(4) of the Hague Convention of 1 February 1971 on the Recognition and Enforcement of Foreign Judgments in Civil and Commercial Matters (“the Hague Convention”). Article 10(4) employs the same jurisdictional linking factors as does article 11 of the ECSI but for the purpose of recognition and enforcement of judgments. Article 10(4) provides that the court of the state of origin shall be considered to have jurisdiction for the purposes of the Hague Convention in the case of injuries to the person or damage to tangible property, if the facts which occasioned the damage occurred in the territory of the state of origin, and if the author of the injury or damage was present in that territory at the time when those facts occurred. The presence requirement, therefore, has its origin in a Convention not concerned with state immunity but with the distinct subject of jurisdiction and enforcement of judgments. Furthermore, the Explanatory Report to the ECSI explains (at para 96) that article 15 was considered too rigid, either because some acts jure gestionis fall outside the cases covered by articles 1 to 13, “or because the connecting links prescribed in these articles do not correspond with rules of jurisdictional competence applied in those States”. (See para 66 above.) There is, therefore, no reason to view the presence requirement as an essential pre-condition to this exception to state immunity. (We also note that, by contrast, the Brussels Convention on Jurisdiction and Enforcement of Judgments in Civil and Commercial Matters, 1968 in article 5(3) recognises the jurisdiction of the courts of the place where the tortious event occurred, without requiring the author’s presence.)
101.
Turning to the work of the International Law Commission, it seems that, once again, a concern to avoid an overreaching of jurisdiction underpins the suggested need for a presence requirement.
“The existence of two cumulative conditions is needed for the application of this exception. The act or omission causing the death, injury or damage must occur in whole or in part in the territory of the State of the forum so as to locate the locus delicti commissi within the territory of the State of the forum. In addition, the author of such act or omission must also be present in that State at the time of the act or omission so as to render even closer the territorial connection between the State of the forum and the author or individual whose act or omission was the cause of the damage in the State of the forum.” (ILC Draft Articles on Jurisdictional Immunities of States and Their Property 1991; Commentary on article 12, comment (6))
The commentary on draft article 12 of UNCSI makes clear that the presence requirement was included to ensure the exclusion from the application of the article of cases of trans-boundary injuries or trans-frontier torts or damage (ILC Draft Articles on Jurisdictional Immunities of States and Their Property 1991; Commentary on article 12, comment (7)). This may have been unnecessary as a matter of international law where a test of territoriality is otherwise satisfied. In his Fifth Report (1983) the Special Rapporteur had observed (at para 95):
“In some more or less precise way, the locus delicti commissi appears to afford an internationally accepted criterion for the assumption of jurisdiction and a sound basis for its exercise, if ever a general exception to State immunity is to become universally recognized in future State practice.”
It should also be noted in relation to the ILC’s work that a number of states expressed doubt as to the need for a presence requirement. Australia commented:
The pre-condition that the author of the act or omission must be present in the territory at the time of the act or omission seems to be unnecessary; it adds nothing in terms of logic, may well be unduly restrictive and creates difficulties if there is more than one author, not all of whom are so present at the time.
In the context of the exclusion of immunity, the place where the injury occurred should establish the necessary jurisdictional basis, unless a closer relationship can be shown to exist with the foreign State pleading immunity. Such an approach would have the advantage of being in harmony with the principle, expressed in a number of Anglo-Australian cases, as well as by the Court of the European Communities, that a court, in deciding whether to exercise jurisdiction over an action in tort, should identify where the substance of the cause of action arose.” (ILC Yearbook [1988] vol II (1), p 54).
Of particular significance is the following observation by the Federal Republic of Germany:
“The exclusion of immunity in the case of injuries or damage the author of which was present in the State where the injuries or damage occurred is in line with article 11 of the 1972 European Convention. There is, however, a danger of the reverse conclusion being drawn that immunity can always be invoked for transborder injuries or damage. It must be made clear … that this is not the case.” (p 71, para 15)
See also the observations of Italy (p 73, para 9) and Thailand (p 81, para 11).
102.
We also note the following observation of Sir Ian Sinclair at the 1767th meeting of the ILC on 25 May 1983. He proposed a reformulation of the draft article and explained:
“Finally, he had deliberately omitted the additional phrase contained in article 11 of the European Convention about the author of the injury or damage being present in the territory at the time of the occurrence. The commentary to the European Convention did not shed much light on the genesis of that extra condition, which, in his view, added an unnecessary and undesirable complication by appearing to confer immunity on the foreign State where it might otherwise be liable vicariously for the injurious act or omission of its servant initiated outside the territory of the forum State but completed within it.” (ILC Yearbook [1983] Vol I, p 79, para 33)
103.
State practice supportive of a territorial tort exception employs a variety of different jurisdictional linking factors. Thus, for example, article 11 of the ECSI and article 12 of the UNCSI, neither of which distinguishes between sovereign and non-sovereign acts, both have a presence requirement but whereas article 11 of the ECSI requires that “the facts which occasioned the injury or damage occurred in the territory” of the forum state, it is sufficient for article 12 of the UNCSI that “the act or omission occurred in whole or in part in the territory” of the forum state. The corresponding exception to immunity in the US FSIA para 1605(a)(5), which does not distinguish between sovereign and non-sovereign acts, does not require an act or omission in the United States but requires “personal injury or death, or damage to or loss of property, occurring in the United States and caused by the tortious act or omission of that foreign state or of any official or employee of that foreign state while acting within the scope of his office or employment”. (The US courts have in interpreting this provision developed an “entire tort doctrine” as a result of which immunity will be lost only if the entire tort, including not only the injury but also the act precipitating the injury, occurs in the United States. However, as further discussed below, the entire tort doctrine is rooted in the wording and legislative history of the FSIA and, as the appellant accepts, forms no part of UK law.) In Canada section 6 of the Canadian State Immunity Act bases jurisdiction on the place of occurrence of loss of life or property, or damage to person and property, being in Canada (see paras 119–120 below).
On our reading of section 5 of the SIA there is no requirement that the state agent be present within the territory of the forum state. There is, however, a requirement that injury or damage has been caused by an act or omission within the jurisdiction. This, in itself, is a powerful linking factor underpinned by the important principle that states have jurisdiction over acts that take place in their own territory. This is particularly so in the present case where, on the assumed facts, there was covert surveillance and extraction and transmission of information within the United Kingdom. As Males LJ observed in the Court of Appeal (at para 55), a foreign state which hacks a computer located in the United Kingdom interferes with the territorial sovereignty of the United Kingdom even if some of the acts in question take place abroad. If established, such conduct by an agent of the appellant State would be an exercise of sovereign power in breach of international law.
105.
We note that, as Lord Sumption observed in Benkharbouche (at para 59), “the core principle of international law is that sovereignty is territorial and state immunity is an exception to that principle”. As we noted above (at para 41), in the Jurisdictional Immunities case the ICJ observed (at para 57) that the principle of state immunity has to be viewed with the principle that each state possesses sovereignty over its own territory and that there flows from that sovereignty the jurisdiction of the state over events and persons within that territory. The relevant consideration, for the purposes of the principles of territorial sovereignty and sovereign equality, is whether the jurisdictional linking factor establishes a territorial connection with the forum state. As Yang observes (at p 229):
“The most important requirement in the context of tort is a territorial connection between either the tortious act/omission or the resultant injury/damage, or both, and the forum State. In short, when torts committed in the forum State are concerned, the overriding consideration for a court is neither a jure imperii / jure gestionis test nor whether ‘insurable’ personal injury is at issue, but whether the tortious act or omission has a connection with the territory of the forum State.”
Similarly, in Kazemi Estate v Islamic Republic of Iran 2014 SCC 62, [2014] 3 SCR 176, a case concerning section 6 of the Canadian State Immunity Act, LeBel J, delivering the majority judgment in the Supreme Court of Canada, stated (at para 72):
“If a foreign state is committing torts within Canadian controlled boundaries, Canada has the competence (derived from its independence) to bring the foreign state within Canada’s adjudicative jurisdiction. There would thus be sufficient connection with the forum state to justify bringing the foreign state’s actions under Canadian scrutiny.”
Foreign act of state
It is necessary to say something about the UK foreign act of state doctrine which was invoked in the courts below. The exercise of jurisdiction in the present proceedings would not engage the UK foreign act of state doctrine. This principle is distinct from state immunity and is not required by international law. The term is used to describe a number of different rules identified by Lord Neuberger in his judgment in Belhaj v Straw [2017] UKSC 3; [2017] AC 964, of which two are of relevance here. While neither is any longer directly in play in the present appeal, their examination casts some light on considerations of comity.
107.
In the courts below it was submitted that to deprive the appellant of immunity would infringe principles of international comity because, contrary to the foreign act of state doctrine, it would involve the English court sitting in judgment on acts performed by a foreign state within its own territory. It was submitted that this reinforced the appellant’s case on state immunity. That submission was rejected by the Court of Appeal and is no longer maintained by the appellant. In Maduro Board of the Central Bank of Venezuela v Guaidó Board of the Central Bank of Venezuela (reported as Deutsche Bank AG London Branch v Receivers Appointed by the Court) [2021] UKSC 57; [2023] AC 156 the Supreme Court held that there exists in English law an exclusionary rule which limits the power of courts to decide certain issues as to the legality or validity of the conduct of the executive of a foreign state when acting within its proper jurisdiction. It operates not by reference to law but by reference to the sovereign character of the conduct which forms the subject matter of the proceedings. We agree with the Court of Appeal that this version of the act of state doctrine does not apply here. As Males LJ explained (at para 69), if these proceedings are allowed to continue they will be concerned with the act of infecting the respondents’ computers within the United Kingdom:
“The fact that the agents who caused this to happen were located in Bahrain at the time, if that proves to be the case, is immaterial. It is equally immaterial whether whatever was done in Bahrain was lawful under the law of Bahrain. The act of state principle does not protect a state which chooses, by an act committed within the United Kingdom, to infringe the territorial sovereignty of this country.”
To allow the present proceedings to continue would not involve UK courts in ruling on the legality or validity of executive acts of a foreign state performed within its own territory or proper jurisdiction. To employ the language used by LeBel J in the Supreme Court of Canada in Kazemi (at para 70), it would not put the foreign state’s decisions and actions in its own territory directly under the scrutiny of courts in this jurisdiction.
108.
Another variety of foreign act of state concerns the non-justiciability of transactions of sovereign states on the international plane (Buttes Gas and Oil Co v Hammer (No 3) [1982] AC 888 per Lord Wilberforce at pp 931G–932A; Belhaj v Straw per Lord Neuberger at paras 123, 129–130). It has not been invoked in the present proceedings, correctly in our view. It was however invoked in the Al Maktoum litigation. In family proceedings relating to the welfare of two children, the mother alleged that her mobile phone and those of some of her legal advisers, security staff and personal assistants had been hacked by agents of the Emirate of Dubai or the United Arab Emirates (“UAE”) acting on behalf of the father, HH Sheikh Mohammed bin Rashid Al Maktoum, ruler of Dubai and Vice-President and Prime Minister of the UAE. On behalf of the father, no point was taken on state immunity (at least after an early stage in the proceedings) but it was contended that this variety of the act of state doctrine applied with the result that the court lacked jurisdiction to adjudicate on these allegations.
109.
In Re Al Maktoum [2020] EWHC 2883 (Fam) the Divisional Court (Sir Andrew McFarlane P and Chamberlain J) held (at paras 66–72), on the assumption that the alleged hacking was a sovereign act, that this variety of act of state was not engaged. In coming to this conclusion, it was influenced by the following considerations. First, the acts alleged were directed against and had direct effects on person in the United Kingdom and within the jurisdiction of the court which, if proved, would constitute a serious interference with the process of the court. Secondly, the legality of the alleged acts fell to be judged by reference to the criminal and civil law of England, not by reference to international law, let alone contested international law. Thirdly, the alleged acts were unilateral and there was no obvious basis on which it could be said that they fell to be judged on the plane of international law. Fourthly, the acts alleged engaged the fundamental privacy rights of the mother and (derivatively) the children and the mother’s right to access to justice. In the light of these matters the Divisional Court concluded that to adjudicate on the allegations would not demonstrate any lack of respect for the principles of comity or the sovereign equality of states. On the contrary, a decision to abstain from adjudicating would undercut the sovereignty of the United Kingdom and be inconsistent with the duty of the court. Alternatively, if the principle was engaged, the Divisional Court considered (at paras 78–80) that the case fell within the established public policy exception.
On appeal, the Court of Appeal (Sir Geoffrey Vos MR, Moylan and Andrews LJJ) considered (Al Maktoum v Al Hussein [2021] EWCA Civ 129, at paras 40, 41) that the critical allegation was not the detail of the hacking but the father’s complicity in domestic UK illegal surveillance which was an entirely justiciable issue. For the same reasons, there was no infringement of comity. The central issue was not the lawfulness of the intelligence acts of a foreign state but the father’s complicity in “illegal UK phone hacking”.
Conclusion on the impact of customary international law
For the reasons explained above, even if no territorial tort exception exists in customary international law, that does not support the appellant’s interpretation of section 5 of the SIA.
For completeness, we shall address further materials which the appellant relies upon in support of its case, namely: (i) domestic case law; (ii) comparative case law; (iii) academic commentary; (iv) Notes on Clauses; and (v) El-Khouri and Osman. For the reasons set out below, none of these materials assist the appellant’s case.
Domestic case law
The appellant places reliance on the decision of Stewart J in Heiser v Islamic Republic of Iran [2019] EWHC 2074 (QB). That case concerned the enforcement in the United Kingdom of US judgments obtained by US citizens injured in terrorist attacks which took place outside the United States. Enforcement (under section 31(1)(b) of the Civil Jurisdiction and Judgments Act 1982) depended upon whether the courts of England and Wales would have had jurisdiction under section 5 of the SIA.
Stewart J held that section 5 was not concerned with where the substance of the tort is committed—“Its concern is where the act or omission causing the death, personal injury or damage occurred” (para 148). He held that, save in one case, all relevant acts and omissions occurred in Middle Eastern states and not in the United States. It was irrelevant that “either primary victims continued to suffer injury on return to the United States or that secondary victims never left the United States”.
The appellant relies on a passage in the judgment (at paras 147 to 148) in which Stewart J seemingly rejected an argument that composite acts which were initiated outside the United Kingdom but completed within it fell outside section 5. It is clear, however, that he did not decide the point. As he stated at para 160:
“I make no decision on the composite act submission eg whether firing a missile from country A into another country B is an act in both countries for the purposes of section 5. It is not necessary for me to decide that point since it does not arise on the facts of any of the cases before me.”
The case does not therefore assist the appellant.
Comparative case law
The appellant relies on US case law on the “non-commercial tort” exception contained in the FSIA as set out in section 1605(a)(5) of the United States Code. This provides that a foreign state shall not be immune where:
“money damages are sought against a foreign state for personal injury or death, or damage to or loss of property, occurring in the United States and caused by the tortious act or omission of that foreign state or of any official or employee of that foreign state while acting within the scope of his office or employment…”
As referred to above, this has been interpreted as requiring the “entire tort” to be committed in the United States. Applying this principle, the US courts have dismissed claims based on allegations materially analogous to those in this case. See, for example, Kidane v Federal Democratic Republic of Ethiopia (2017) 851 F 3d 7 and Democratic National Committee v Russian Federation (2019) 392 F Supp 3d 410. Both the injury and the act precipitating the injury must occur in the United States.
The FSIA formed part of the background context to the SIA, and, where the statutory language used is materially similar, US case law has been treated as being of persuasive authority. This, however, is not such a case. Section 1605(a)(5) requires the personal injury or property damage occurring in the United States to be “caused by thetortious act or omission” of the foreign state (emphasis added), whereas section 5 of the SIA refers to “an act or omission” (emphasis added). The use of the definite article helps to explain why the US courts have developed an “entire tort” doctrine, a principle that does not form part of UK law. The US jurisprudence also places reliance on legislative history, such as a House Report, which is particular to the United States. For all these reasons, and those given by Males LJ at paras 76 to 80 of his judgment, the US case law is of no assistance.
The appellant also relies on the decision of the Supreme Court of Canada in Kazemi, which case concerned section 6 of the Canadian State Immunity Act. According to this provision, a foreign state is not immune:
120.
The Supreme Court held that the acts constituting the tort, not just the harm, must occur in Canada. The appellant relies on the court’s reasoning that this was because it “would put the foreign state’s decisions and actions in its own territory directly under the scrutiny of Canada’s judiciary—the exact situation sovereign equality seeks to avoid”. Therefore, the “impugned events” must “take place in Canada”, at paras 69–70 and 73. In that case, however, the alternatives considered by the court were “either (1) that the injury manifest itself in Canada, even where the acts causing the death or injury occurred outside Canada, or (2) that the acts causing injury or death occur within Canada” (para 69). Faced with that choice, it is not surprising that the court chose the latter. Otherwise, there would be loss of immunity even where no causative acts or omissions occurred within the jurisdiction (as in the facts of that case). It is in that context that the comments relied upon by the appellant were made. They were not directed at and did not consider a case such as the present where there are causative acts both within and without the territory of the forum state. The wording of the Canadian Act is, moreover, materially different from section 5 of the SIA and does not refer to “an act or omission”.
121.
The appellant relies on the following passage in Dicey, Morris and Collins on the Conflict of Laws (16th ed, 2022) at para 9–050:
“Section 5 of the 1978 Act is based on the European Convention on State Immunity, Art. 11. [It] is limited to acts or omissions in the United Kingdom, and that limitation is not incompatible with the European Convention on Human Rights, Art. 6(1). [It] concerns acts or omissions in the United Kingdom, and does not involve consideration of the place where the substance of the wrong occurred. … Consequently if agents of a foreign State on an intelligence mission in England commit acts within England that cause personal injury or death to someone in England, then neither they or the foreign State will be entitled to immunity.” (Emphasis added.)
It is undoubtedly correct that section 5 is limited to where there is a causative act or omission in the United Kingdom. This passage does not, however, address the issue of whether this means all causative acts or omissions. It is to be noted that the example of an intelligence mission committing acts within England would involve initiating decision-making abroad. Furthermore, as noted at para 71 above, it is an example of a sovereign activity which would not attract immunity.
122.
The appellant also relies on the following passage from State Immunity: Selected Materials and Commentary (2004), Dickinson, Lindsay and Loonam at pp 370:
“The act or omission referred to in s 5 could be construed as being either (i) the wrongful act of (or on behalf of) the State on which the claimant’s action is based, or (ii) the immediate cause of the injury in respect of which the claimant sues. On balance, the former view is to be preferred, on the basis that it focuses on the State’s conduct within the territory in question.” (emphasis added)
Again, this passage does not address the question of whether all the wrongful acts or omissions have to occur in the United Kingdom. On the respondents’ case, their claim is based on wrongful acts and state conduct which took place within the United Kingdom.
The academic commentary does not therefore advance the appellant’s case. On the contrary it supports the respondents’ case. (See paras 49, 70 and 90–91 above.)
124.
The appellant seeks to place reliance on Notes on Clauses which were produced for Ministers at the Committee Stage of the State Immunity Bill in the House of Lords. These were not published, nor were they made available to Members of Parliament. The Court of Appeal rightly held them to be inadmissible as an aid to interpretation. As Sales J explained in R (Public and Commercial Services Union) v Minister for the Civil Service [2010] EWHC 1027 (Admin), [2010] ICR 1198 at para 55:
“Notes on clauses when not cited in debate are private documents not available to the public at large, unlike White Papers, statements reported in Hansard and published Explanatory Notes. An Act of Parliament creates law applicable to all citizens. In my judgment, it is fundamental that all materials which are relevant to the proper interpretation of such an instrument should be available to any person who wishes to inform himself about the meaning of that law. That is not the position in relation to notes on clauses and for that reason I do not consider they are a legitimate aid to construction of an Act of Parliament. …”
The appellant has a fallback argument that the Notes on Clauses should be given similar weight to academic commentary on an Act. Unpublished, undated and anonymous notes produced at the Bill stage are not, however, akin to academic commentary on enacted legislation. Their only potential relevance is as an interpretative aid and if they are not admissible for that purpose then they are irrelevant.
El-Khouri v Government of the United States of America [2025] UKSC 3; [2025] AC 845 and R v Governor of Pentonville Prison, Ex p Osman [1990] 1 WLR 277
These cases were not relied upon by the appellant but are relied upon by Lord Leggatt to support his view that section 5 of the SIA draws a clear distinction between acts and effects. We do not consider that they assist.
126.
This court in El-Khouri did state at para 51 that “the word ‘conduct’ would normally and naturally be understood as a synonym for acts done by the [person] in the specified location and not as including effects (whether intended or not) felt in that location of acts done somewhere else …”. However, the court was interpreting a different word, “conduct”, in a different statutory context, the Extradition Act 2003. At best, only limited aid can be gained from this decision in these circumstances. Further, unlike section 5 of the SIA, section 137 of the Extradition Act imposes a binary and mutually exclusive distinction between conduct that occurs “in” and conduct that occurs “outside” the territory of the state requesting extradition. In its judgment, the court repeatedly emphasised this important statutory feature when interpreting the term (see paras 24, 33, and 52). The statement in para 51 must be understood in that context. Moreover, as the court in El-Khouri expressly noted in the immediately following paragraph which is more relevant for present purposes: “[this] distinction may be difficult to draw in a situation where the conduct alleged comprises various acts some of which occurred within and some of which occurred outside the territory”.
In R v Governor of Pentonville Prison, Ex p Osman the court considered that “appropriation” occurred in the place where a telex was despatched by the individual rather than in the location where it was received, but also stated that it could “not rule out the possibility that the place where the telex is received may also be regarded as the place of appropriation, if our courts were ever to adopt the view that a crime may have a dual location” (at p 295).
As in El-Khouri, the court was faced with providing an answer to a binary question and chose one particular location. Yet in both cases, the court recognised that difficult questions might arise in other cases as to where relevant conduct should be considered to have occurred (eg when faced with acts that arguably occur in two places at once). But if the word “act” (or “conduct” and “appropriation”) necessarily has an implicit human presence requirement, there should have been no such difficulty.
The meaning of “an act … in the United Kingdom”
There is no good reason to impose a restrictive meaning on the words “an act … in the United Kingdom” in section 5 of the SIA. In particular, there is no good reason to limit it to the physical movement of a human being within this jurisdiction. Furthermore, attempts to draw a rigid distinction between an act and its effects can be controversial and can lead to artificial and unprincipled results.
130.
The natural meaning of “act” in section 5 does not distinguish between human and non-human agency. In the present context this is apparent, for example, from the terms of article 11 of the ECSI and article 12 of the UNCSI. Article 11 of the ECSI employs the linking factor “if the facts which occasioned the injury or damage occurred in the territory of the State of the forum”. That this condition is not limited to the act of a human agent is demonstrated by the inclusion of a further requirement that “the author of the injury or damage was present in that territory at the time when those facts occurred”. The second requirement would be superfluous if the first were limited to an act by a human agent necessarily present in the forum state. The position is even clearer in article 12 of the UNCSI which excludes immunity, inter alia, “if the act or omission occurred in whole or in part in the territory of that other State and if the author of the act or omission was present in that territory at the time of the act or omission”. Once again, the requirement that the author of the act be present is superfluous if the earlier reference to “the act” is limited to the conduct of a human agent present in the forum state. We consider that the express requirement of presence by the author of the act was included, in each case, to meet possible concerns in relation to jurisdiction. (See paras 101–106 above.) The inclusion of such an express requirement was necessary if that objective was to be achieved because that was not otherwise the effect of the provisions. In each case, the presence requirement was included because of the possibility that an act could occur in the forum state without a human agent necessarily being present. The requirement of presence was not otherwise implicit.
131.
Section 5 of the SIA, by contrast, does not include such an express requirement. For the reasons set out above (paras 72–78) we consider that this was a deliberate departure from the requirements of the ECSI. In our view an act includes an act done using mechanical, electrical or other automated or remote means. In these circumstances, we consider that the words “personal injury … caused by an act or omission in the United Kingdom” are capable of including acts carried out through some instrumentality, using a device or remote means, even if the person responsible for executing or carrying out the act is not physically present at the location where the act takes place. We agree with the conclusion of Males LJ in the Court of Appeal in the present case (para 34) that, as a straightforward use of language, the remote manipulation from abroad of a computer located in the United Kingdom is an act within the United Kingdom within section 5 of the SIA. The application of section 5 to the precise assumed facts of the present case is considered in section 6 of this judgment.
132.
We also endorse the following observation of Males LJ (para 34) in relation to the remote manipulation of a computer from abroad:
The boundary between an act and its effects can be controversial and such a distinction cannot always be readily drawn. To draw a rigid distinction between an act and every effect of that act can lead to absurdity. This absurdity is illustrated by the submission recited by Warby LJ at para 121 of the judgment below: counsel for the appellant submitted that when a person uses a pen to create a manuscript document the marks on the page are not part of the act of writing but only the effect of that act.
133.
Limiting section 5 of the SIA to situations in which the relevant injury or damage was caused by the act of an agent of a foreign state present in the territory of the forum state would, in our view, be an unduly restrictive reading. It fails to take account of the fact that modern technology enables acts to be carried out remotely from abroad. Nor should it be accepted that the presence of foreign state agents who do wrongful acts constitutes a greater interference with territorial sovereignty. Where a foreign state causes an explosion in this country, the infringement of sovereignty is just as great whether it is caused by agents of the foreign state present here or by a drone remotely controlled from abroad. Where a foreign state sabotages NHS computer systems in this country, the infringement of sovereignty is just as great whether it is caused by agents of the foreign state present here or by remote manipulation of those computers from abroad. Moreover, in relation to the latter example, it is not difficult to envisage significant personal injury or damage to tangible property caused by the remote manipulation of computer systems in a hospital.
In conclusion, for all these reasons, and having considered the wider context and purpose, we are satisfied that section 5 is clear and unambiguous. It means there is no sovereign immunity for a foreign state for proceedings in respect of personal injury caused by an act (or acts) in the United Kingdom, even if other causative acts take place abroad and the actor is not present in the United Kingdom.
135.
It is pleaded that the following acts took place in the United Kingdom and that they were causative of the injury to the respondents:
Transmitting executable files for installing FinSpy on the SS Computer and the MM Computer (‘the Devices’), which were at all material times located in England.
Installing FinSpy on the Devices, including by overwriting the hard disk and/or Master Boot Record with malicious code.
Executing FinSpy to the Devices’ Central Processing Units, and reading data to, and writing it from, the Devices’ Random Access Memory.
Using the Devices’ computer network interface controller to send and receive data via a wired or wireless network and telecommunications equipment within the UK.
Using the Devices’ battery power to transmit and receive data and commands, and to use other hardware components in the Devices.
Exfiltrating or causing to be exfiltrated information held on, available from and/or transmitted via the Devices.
Activating or causing to be activated the Devices’ microphones and/or cameras, and recording information with the same.
There are therefore ten classes of acts which are alleged to be causative of personal injury to the respondents, all of which took place in the United Kingdom. Those acts involved infection, accessing, exfiltration and surveillance.
Taken together we agree with the courts below that these assumed facts constitute a causative act or series of acts in the United Kingdom. In essence, these acts involved surveillance of the respondents. That surveillance was of persons in the United Kingdom by means of the hacking of computers in the United Kingdom. It took place in the United Kingdom even if it may have been initiated from abroad. It involved an interference with the territorial sovereignty of the United Kingdom, as Males LJ held (para 40), and it caused injury to the respondents.
If, for example, a person within the United Kingdom had been instructed by an official of a foreign state located abroad to carry out video-relayed surveillance on the respondents in the United Kingdom there can be little doubt that that would involve acts within the United Kingdom even though the instruction and the viewing occurred abroad. The fact that the instructed surveillance is carried out by means of the hacking of a computer rather than by a human should make no difference. In both cases the surveillance is of persons in the United Kingdom carried out through localised means.
A similar analysis applies if one concentrates on the hacking of a computer located in the United Kingdom rather than the resulting surveillance carried out. That too involves acts within the United Kingdom and in particular the gaining of access to and the manipulation of the computer.
140.
Some assistance by analogy is provided by the decision of Mr Jonathan Hirst QC in Ashton Investments Ltd v OJSC Russian Aluminium (RUSAL) [2007] 1 All ER (Comm) 857. That case involved a claim for breach of confidence, unlawful interference with business and conspiracy on the grounds that the defendants in Russia had hacked into the claimant’s computer system in London in order to obtain confidential and privileged information. It was held that the court had jurisdiction on the basis that the damage sustained resulted from an act committed within the jurisdiction. The deputy High Court judge held at para 63:
“I also consider that substantial and efficacious acts occurred in London, as well as Russia. That is where the hacking occurred and access to the server was achieved. This may have been as a result of actions taken in Russia but they were designed to make things happen in London, and they did so. Effectively the safe was opened from afar so that its contents could be removed. It would be artificial to say that the acts occurred only in Russia. On the contrary, substantial and effective acts occurred in London.”
In this case also, access to computers in the United Kingdom was achieved which was designed to and did make things happen in the United Kingdom, which acts caused injury to the respondents.
142.
The Divisional Court decision in R v Governor of Brixton Prison, Ex p Levin [1997] QB 65 is to similar effect. That was an extradition case which raised the issue of whether the alleged hacking by the applicant based in St Petersburg, Russia, of a computer in Parsipenny, USA, for the purpose of illicitly transferring funds, took place in Russia or in the USA. The court held that it was the latter. It reasoned as follows (at pp 81G–82A):
“…the operation of the keyboard by a computer operator produces a virtually instantaneous result on the magnetic disk of the computer even though it may be 10,000 miles away. It seems to us artificial to regard the act as having been done in one rather than the other place. But, in the position of having to choose on the facts of this case whether, after entering the computer in Parsipenny, the act of appropriation by inserting instructions on the disk occurred there or in St Petersburg, we would opt for Parsipenny. The fact that the applicant was physically in St Petersburg is of far less significance than the fact that he was looking at and operating on magnetic disks located in Parsipenny. The essence of what he was doing was done there.”
In this case too, the essence of what the computer operator in Bahrain was allegedly doing was done in the United Kingdom where the respondents’ computers were hacked.
The focus of section 5 is on acts or omissions which cause personal injury or property damage. In the present case the alleged cause of the personal injury was harassment. That harassment took place in the United Kingdom. It involved a course of conduct which, as pleaded, involved various acts occurring in this country.
As the respondents submit, to hold that the only causative act in a case such as this occurred abroad would lead to arbitrary consequences. They give the example of a person located in the United Kingdom who was subject to a malware attack directly facilitated or overseen by a foreign state official also located in the United Kingdom. On the appellant’s case, if the infection was activated by an IT operator located abroad, acting on the orders of the UK-based official, the state would be accorded immunity.
146.
It would also lead to arbitrary distinctions being drawn. The appellant now accepts (contrary to its case before the judge) that the murder of someone within the United Kingdom by agents present here but acting on the instructions of a foreign state would not involve immunity, with the notorious case of the polonium poisoning of Alexander Litvinenko being cited as an example (see https://www.gov.uk/government/publications/the-litvinenko-inquiry-report-into-the-death-of-alexander-litvinenko). There would, however, be immunity if the same murder was carried out by a drone operated by a state official from abroad planting and detonating a bomb. Similarly, there would be immunity if such drones entered the airspace of Heathrow causing aeroplanes to crash and multiple deaths or if NHS computer systems were hacked by operatives in a foreign state and put out of operation causing patients to die. Such acts would plainly also involve serious and intrusive interference with the territorial sovereignty of the United Kingdom. In our view, given the current state of both international and UK law, such acts are not entitled to state immunity before UK courts.
For all these reasons, we agree with the judge and the Court of Appeal that, on the respondents’ case, legally causative acts occurred in the United Kingdom so as to bring the case within the exception to immunity from jurisdiction set out in section 5 of the SIA.
For the reasons set out above we would dismiss the appeal. In these circumstances it is unnecessary to consider the issue raised by the respondents’ notice relating to access to the court and article 6 of the ECHR.
LORD LEGGATT (DISSENTING):
Introduction
It is a rule of customary international law, universally recognised, that states are immune from the jurisdiction of the courts of other states in respect of their sovereign activities. The rule is founded on the principle of sovereign equality of states, one corollary of which is that no state has jurisdiction over another (par in parem non habet imperium). Although once seen as a privilege, the grant of immunity is now understood as a right under customary international law together with a corresponding obligation on other states to respect and give effect to the immunity: see eg Jurisdictional Immunities of the State (Germany v Italy: Greece Intervening) [2012] ICJ Rep 99, para 56; Royal Embassy of Saudi Arabia (Cultural Bureau) v Costantine [2025] UKSC 9; [2025] 1 WLR 1207, paras 36–38.
150.
The United Kingdom has given effect to this international obligation through legislation. The State Immunity Act 1978, in section 1, lays down that a state is immune from the jurisdiction of the courts of the United Kingdom except as provided in sections 2 to 11. This appeal is concerned with the scope of one of these exceptions, contained in section 5. Section 5 of the 1978 Act provides:
“Personal injuries and damage to property
A state is not immune as respects proceedings in respect of–
The question on this appeal is whether this exception from state immunity applies to proceedings in which damages are claimed for psychiatric injury caused when the claimants learned that their personal computers, located in the United Kingdom, had been hacked by agents of a foreign state, located abroad. The question turns on whether the relevant acts of these agents were “act(s) … in the United Kingdom” within the meaning of section 5 of the 1978 Act. The claimants’ case, accepted by the courts below, is that they were and that, on the proper interpretation of section 5, such an act can be performed by someone who is in a foreign country. Bahrain contends that this interpretation is mistaken and that, on the correct interpretation of section 5, an act is done where the agent is located: the exception from immunity can therefore apply only if the agent of the foreign state is in the United Kingdom, which is not alleged here.
In the courts below there was also a dispute about whether the term “personal injury” in section 5(a) includes psychiatric injury. The decision that it does is not challenged on this appeal.
The claimants’ primary case is that the words “an act or omission in the United Kingdom” have always had the meaning for which they contend. But they also have an alternative case, maintained on this appeal by a respondent’s notice, that, if necessary, the court should now modify the meaning of section 5 so that it covers an act or omission of a person who is not in the United Kingdom. This is said to be both permitted and required by section 3 of the Human Rights Act 1998 to achieve compatibility with article 6 of the European Convention on Human Rights, which guarantees a right of access to a court. When addressing this alternative case, it will be necessary to take account of developments in international law and practice regarding state immunity since the 1978 Act was passed.
For the reasons which follow, I would accept Bahrain’s contention and allow the appeal. The words of section 5 can and should be read as confining the exception to personal injury and property damage caused by acts or omissions of an agent who is in the United Kingdom. That interpretation gives effect to the purpose of the 1978 Act, which was to enable the United Kingdom to ratify the European Convention on State Immunity. The claimants’ interpretation of section 5 elides the distinction between acts and their effects and creates inconsistency both with that Convention and with customary international law.
The claimants’ case
The claimants, Dr Saeed Shehabi and Mr Moosa Mohammed, are pro-democracy activists and prominent members of the Bahraini opposition movement, who live in the United Kingdom. They have each been granted refugee status and Dr Shehabi has become a British citizen, having lived in the United Kingdom since 1973. They have brought these proceedings against the Kingdom of Bahrain in the High Court of England and Wales claiming damages for the tort of harassment.
The claimants allege that in or around September 2011 their computers were infected with spyware known as “FinSpy”. This is a computer program designed to be installed on a target’s electronic device without their knowledge or consent when they open a file sent by a remote operator, usually as an infected email or attachment. Once installed, FinSpy runs automatically and allows the remote operator to access data held on the device, intercept communications conducted using the device and use any inbuilt camera and microphone to carry out live surveillance of activities in its vicinity. The claimants’ case is that the spyware installed on their computers was operated by agents of Bahrain from outside the United Kingdom via a server located in Bahrain. They claim that these agents used FinSpy to conduct covert surveillance of them over a three-year period, secretly monitoring their activities and exfiltrating data stored or available on, or transmitted by, their computers.
The claimants discovered that they had been targeted in this way in or around August 2014 from information published by WikiLeaks and by an organisation called Bahrain Watch. As a result of learning of Bahrain’s conduct, the claimants both allegedly suffered psychiatric injury, diagnosed in each case as an adjustment disorder.
Bahrain denies the allegations. But for the purpose of deciding whether Bahrain has immunity from the jurisdiction of the English courts in respect of these proceedings, it must be assumed that the allegations are true.
Section 3 of the Protection from Harassment Act 1997 provides a civil remedy for harassment. Harassment is a persistent and deliberate course of unreasonable and oppressive conduct, targeted at another person, which is calculated to and does cause that person alarm, fear or distress: see Hayes v Willoughby [2013] UKSC 17; [2013] 1 WLR 935, para 1. A feature of this case is that Bahrain’s activities were surreptitious and intended to avoid detection: it was not the activities themselves but their discovery which caused the claimants alarm, fear or distress. This may raise a legal question if the claim proceeds; but on this appeal it is to be assumed that the facts alleged, if proved, will—subject to the question of immunity—entitle the claimants to a remedy in the tort of harassment.
Statutory interpretation
Statutory language must be read in the light of the statute’s purpose
Whether Bahrain has immunity depends on whether these proceedings fall within section 5 of the State Immunity Act 1978, quoted at para 150 above. There is no doubt about the principles to be applied in interpreting this statutory provision. As Lord Hodge said in Bilta (UK) Ltd (in liquidation) v Tradition Financial Services Ltd [2025] UKSC 18; [2026] AC 140, para 20:
“The court’s approach to statutory interpretation is well established in our case law. The court derives the meaning of a legislative provision from the words which Parliament has used in that provision having regard to the context of the statute as a whole and the historical context in which the statute was enacted as the context may reveal the mischief which the provision addresses and shed light on its purpose.”
Lord Hodge also quoted a warning given by Lord Bingham of Cornhill in R (Quintavalle) v Secretary of State for Health [2003] UKHL 13; [2003] 2 AC 687, para 8, against giving a literal interpretation to a statutory provision without regard to its context and the purpose of the statute. In Lord Bingham’s words:
“Every statute other than a pure consolidating statute is, after all, enacted to make some change, or address some problem, or remove some blemish, or effect some improvement in the national life. The court’s task, within the permissible bounds of interpretation, is to give effect to Parliament’s purpose. So the controversial provisions should be read in the context of the statute as a whole, and the statute as a whole should be read in the historical context of the situation which led to its enactment.”
The State Immunity Act 1978 was enacted to bring UK law more closely in line with prevailing state practice and to enable the United Kingdom to ratify the European Convention on State Immunity signed on 16 May 1972 (and the Brussels Convention on the Immunity of State-Owned Ships 1926). To enable the United Kingdom to ratify the European Convention, it was necessary to ensure that UK national law was consistent with the obligations which the United Kingdom would undertake upon ratification. The United Kingdom duly ratified the Convention on the day when the Act came into force (22 November 1978).
There is a “strong presumption” in favour of interpreting legislation in a way that is compatible with public international law and does not place the United Kingdom in breach of an international obligation: R v Lyons [2002] UKHL 44; [2003] 1 AC 976, para 27; Assange v Swedish Prosecution Authority (Nos 1 and 2) [2012] UKSC 22; [2012] 2 AC 471, paras 122, 206; Bennion, Bailey and Norbury on Statutory Interpretation, 9th ed (2026), section 26.9. This presumption is “at its strongest” when the legislation has been passed to give effect to a treaty or may otherwise be assumed to have been drafted with the treaty in mind: see eg Boyce v The Queen [2005] 1 AC 400, para 26. In such a case it is clear law that the words of the statute are to be construed, if they reasonably can be, as consistent with the treaty obligation: Salomon v Commissioners of Customs and Excise [1967] 2 QB 116, 143; The Eschersheim [1976] 1 WLR 430, 436; Garland v British Rail Engineering Ltd [1983] 2 AC 751, 771.
Essential historical context which therefore needs to be considered in interpreting the State Immunity Act 1978 consists of, first, the customary international law of state immunity as it stood when the Act was passed and, second, the European Convention on State Immunity, which the United Kingdom wished to ratify. The Act must be interpreted in a way that does not place the United Kingdom in breach of its treaty obligations, unless it is plain that the legislation was for some reason intended to do so.
As Lord Diplock said in Alcom Ltd v Republic of Colombia [1984] AC 580, 597, the provisions of the 1978 Act “fall to be construed against the background of those principles of public international law as are generally recognised by the family of nations”. Lord Diplock explained that the feature of such customary international law most relevant to the subject matter of the Act is the distinction drawn between acts performed by a state in the exercise of sovereign authority (acta jure imperii) and acts performed by a state in the course of commercial or other activities which can also be carried out by private persons (acta jure gestionis). Under the so called “restrictive” doctrine of state immunity, which has become generally accepted, immunity applies only to the former, which I will refer to for short as “sovereign acts”, and not to the latter, which I will call “private acts”.
166.
Until 1978, the common law granted to foreign states near absolute immunity without distinguishing between sovereign and private acts. This left the United Kingdom an odd-one-out among nations (I Congreso del Partido [1983] 1 AC 244, 272, per Lord Diplock), as the trend in the decisions of national courts and the writings of international jurists had been away from absolute immunity and in favour of the restrictive doctrine. The history of the progressive adoption of the restrictive doctrine has been summarised in the judgments of this court in Benkharbouche v Embassy of the Republic of Sudan [2017] UKSC 62; [2019] AC 777, paras 8, 40–52, and Argentum Exploration Ltd v The Silver [2024] UKSC 16; [2025] AC 555, paras 17–23. In Benkharbouche, para 52, Lord Sumption (with whom the other members of the Supreme Court agreed) observed that “there has probably never been a sufficient international consensus in favour of the absolute doctrine of immunity to warrant treating it as a rule of customary international law” and that, “while there has for at least two centuries been a consensus among nations in favour of some form of state immunity, the only consensus that there has ever been about the scope of that immunity is the consensus in favour of the restrictive doctrine”. He also noted, at para 51, that: “Today, the international consensus in favour of the restrictive doctrine is almost complete”.
167.
Shortly before the 1978 Act was passed, the position under the common law changed as a result of the decisions of the Privy Council in The Philippine Admiral [1977] AC 373 and the Court of Appeal of England and Wales in Trendtex Trading Corp v Central Bank of Nigeria [1977] QB 529, both of which applied the restrictive doctrine. These decisions were later approved by the House of Lords in I Congreso del Partido [1983] 1 AC 244—a case decided after the Act came into force but arising from earlier events to which the common law applied, as the Act was not retrospective. Lord Wilberforce, who gave the leading speech in I Congreso del Partido, discussed international jurisprudence bearing on the distinction between sovereign and private acts. He said, at p 267B–D:
“The conclusion which emerges is that in considering, under the ‘restrictive’ theory whether state immunity should be granted or not, the court must consider the whole context in which the claim against the state is made, with a view to deciding whether the relevant act(s) upon which the claim is based, should, in that context, be considered as fairly within an area of activity, trading or commercial, or otherwise of a private law character, in which the state has chosen to engage, or whether the relevant act(s) should be considered as having been done outside that area, and within the sphere of governmental or sovereign activity.”
168.
As Lord Wilberforce noted, at pp 263–264, while classifying the relevant act often presents no difficulty, it is not always easy to decide whether the act is within or outside the area of sovereign activity. There were in 1978 and still are differences of state practice and opinion regarding exactly where and how the line between sovereign and private acts is to be drawn. The International Law Commission made this point when reporting on state immunity to the United Nations General Assembly in 1986:
“There is common agreement that, for acts performed in the exercise of the prérogatives de la puissance publique or ‘sovereign authority of the State’, there is undisputed immunity. Beyond or around that hard core of immunity, however, there appears to be a grey zone in which opinions and existing case law and, indeed, legislations still vary.” (Report of the International Law Commission on the work of its thirty-eighth session in Yearbook of the International Law Commission 1986, Vol II, Part 2, p 16)
For acts falling within the “grey zone”, the reasonable conclusion must be that—given the absence of a general state practice accepted as legally binding—customary international law neither requires nor precludes the grant of immunity.
170.
The distinction drawn by the restrictive doctrine between sovereign and private acts applies in tort cases as it does in other civil proceedings. One of the examples of cases in tort mentioned by Lord Wilberforce in I Congreso del Partido, at p 264D, was In re Danish State Railways in Germany (1953) 20 ILR 178, where a claim for damages for personal injuries was brought in a German court against the Kingdom of Denmark arising from the allegedly negligent driving of a bus owned by the Danish state railway company. The court held that Denmark was not entitled to immunity as the activity was of a private and civil nature. By contrast, in another German case the United Kingdom was held to be immune from suit by a haulage contractor who claimed to have suffered injury to his health after following instructions from British army officers to collect military equipment from the Soviet-occupied zone of Germany, which led to his being arrested by members of the Soviet armed forces and kept in captivity for several months. The claim was found to be concerned with sovereign activity and therefore covered by immunity: Immunity of United Kingdom from Jurisdiction (Germany) (1957) 24 ILR 207.
171.
Many decisions of national courts involving claims in tort against foreign states for personal injury or damage to property in which the distinction between sovereign and private acts was applied were cited in the Fifth report on jurisdictional immunities of States and their property, by Mr S Sucharitkul, the ILC Special Rapporteur, published in the Yearbook of the International Law Commission 1983, Vol II, Part 1, pp 41–42. A notable case mentioned in this survey is the decision of the Austrian Supreme Court in Holubek v Government of the United States (1961) 40 ILR 73. The Austrian court held that the negligent driving of a motor car owned by the US Government which was carrying mail intended for the US Embassy was to be characterised as a private act and not a sovereign act. The reasoning was that the operation of a motor vehicle and the use of public roads by a foreign state “belong to the sphere of the private activities of that state even if … such operation and use have occurred in the performance of official functions” (p 78).
What is important for present purposes is not precisely how the distinction between sovereign and private acts has been applied in tort cases but the fact that, at the time when the State Immunity Act 1978 was enacted, this distinction marked the boundary between those proceedings in which a foreign state was, or was not, entitled to immunity under customary international law. As the common law of England and Wales had by then joined the international consensus by recognising the restrictive doctrine of immunity, the distinction also defined the limit of immunity in tort cases under the common law.
The 1978 Act is not completely comprehensive. In particular, it does not apply to proceedings relating to anything done by or in relation to the armed forces of a state while present in the United Kingdom: see section 16(2). Such cases are still governed by the common law. So, in such cases, the distinction between sovereign and private acts which characterises the restrictive doctrine of immunity in public international law has continued to be applied in England and Wales. In Littrell v United States of America (No 2) [1995] 1 WLR 82 a member of the US Air Force who was serving in the United Kingdom sued the United States for damages for personal injuries caused by allegedly negligent medical treatment received at a US military hospital. Applying the test stated in I Congreso del Partido (quoted at para 167 above), the Court of Appeal upheld the claim of the United States to immunity on the ground that the acts complained of were within the sphere of sovereign activity (jure imperii).
That approach was approved and applied by the House of Lords in Holland v Lampen-Wolfe [2000] 1 WLR 1573. The claimant in that case was a military instructor at a US base in the United Kingdom who brought an action for defamation against her supervisor in respect of a memorandum he had circulated criticising her conduct. This allegedly tortious act was held by the House of Lords to be a sovereign act rather than a private act and therefore covered by state immunity.
Bahrain’s alleged acts of hacking and surveillance were sovereign acts
175.
If the common law applied, there is no doubt about what the outcome of this case would be. The acts of hacking and surveillance allegedly carried out by agents of Bahrain fall squarely within the sphere of sovereign activity. As Simler LJ said in Zu Sayn-Wittgenstein-Sayn v HM Juan Carlos de Borbón y Borbón [2022] EWCA Civ 1595; [2023] 1 WLR 1162, para 53, “covert surveillance operations” are “the classic business of a state intelligence service”. The same view has been taken in the United States. As it was put by the US Federal Court of Appeals for the Ninth Circuit in Broidy Capital Management LLC v State of Qatar (2020) 982 F 3d 582, 595:
“a foreign government’s deployment of clandestine agents to collect foreign intelligence on its behalf … is the sort of peculiarly sovereign conduct that all national governments (including our own) assert the distinctive power to perform.”
So far as this country is concerned, powers to intercept communications and target computer equipment abroad (as well as in the United Kingdom) for purposes which include the collection of data are conferred on intelligence agencies by Parts 2 and 5 of the Investigatory Powers Act 2016.
The European Convention on State Immunity and its tort exception
The European Convention on State Immunity, which was opened for signature in Basle on 16 May 1972 and entered into force on 11 June 1976, is a Council of Europe Convention, although only eight member states of the Council of Europe, including the United Kingdom, are currently parties to it. The aim of the Convention, as recorded in the Preamble, was to promote harmonisation by establishing in their mutual relations common rules relating to the scope of immunity between the contracting parties.
The Convention establishes a general rule, in article 15, that a contracting state shall be entitled to immunity from the jurisdiction of the courts of another contracting state unless the proceedings fall within articles 1 to 14. The key exception for present purposes is article 11. This provides:
178.
Two features of the wording of article 11 are important to note. First, it is apparent that article 11 was not seeking to codify customary international law on state immunity. As pointed out in the Explanatory Report on the Convention, the wording was derived, not from any prior material regarding state immunity, but from an international agreement about the recognition and enforcement of foreign judgments in civil and commercial matters: article 10(4) of the Hague Convention dated 1 February 1971. Furthermore, article 11 did not draw the distinction applicable under customary international law between sovereign and private acts. Its terms are both narrower and wider than the rule recognised by the restrictive doctrine. Article 11 is narrower because it applies only to proceedings which relate to redress for personal injury or damage to tangible property and not to other claims in tort even if the acts complained of are of a private, non-sovereign character. But article 11 is also wider than the rule recognised in customary international law because it applies to sovereign as well as private acts.
Departing from customary international law in this way did not give rise to any difficulty or involve a breach of any international obligation by states which became parties to the Convention. Sovereign states are free to make an international agreement that, in their relations with each other, they will not claim immunity in a class of case in which they would otherwise be entitled to do so: see eg Infrastructure Services Luxembourg SÀRL v Kingdom of Spain [2026] UKSC 9; [2026] 2 WLR 581 (considering the effect of the Convention on the Settlement of Investment Disputes between States and Nationals of Other States (1965)). That is what the contracting parties to the European Convention did when they agreed to article 11 in so far as it applies to sovereign acts.
The second notable feature of article 11 relevant in this appeal is that it applies only where “the author of the injury or damage was present in [the territory of the forum state] at the time when [the facts which occasioned the injury or damage] occurred”. It is thus common ground that it does not cover the present case, as it is not alleged that the agents of Bahrain who hacked the claimants’ computers were present in the United Kingdom at any time.
In considering the relationship between the European Convention and the 1978 Act, it needs to be recognised that the Convention itself allows a contracting state the option of departing from its terms within certain limits. This option is afforded by article 24 of the Convention. Article 24(1) provides:
“Notwithstanding the provisions of Article 15, any State may … declare that, in cases not falling within Articles 1 to 13, its courts shall be entitled to entertain proceedings against another Contracting State to the extent that its courts are entitled to entertain proceedings against States not party to the present Convention. Such a declaration shall be without prejudice to the immunity from jurisdiction which foreign States enjoy in respect of acts performed in the exercise of sovereign authority (acta jure imperii).”
It can be seen that article 24(1) permits a contracting state which makes a declaration under it to derogate from article 15 but not from articles 1 to 13 of the Convention. In other words, the contracting state must recognise the exemptions from immunity provided for in articles 1 to 13; but it may opt to go further and afford wider exceptions to the general rule of immunity set out in article 15. This freedom is, however, subject to two limits. First, any such wider exception must be general and not selective: that is, it must apply to all states and not just to states which are parties to the Convention. Second, any such derogation from article 15 must not prejudice the immunity from jurisdiction enjoyed by foreign states under customary international law in respect of sovereign acts. In other words, any exception to the general rule of immunity which goes further than articles 1 to 13 may extend only to private acts, and not to sovereign acts.
183.
A question was raised at the hearing about whether the reference in article 24(1) to “the immunity from jurisdiction which foreign States enjoy in respect of acts performed in the exercise of sovereign authority” should be interpreted as meaning “the immunity which foreign states do enjoy in respect of sovereign acts” or “the immunity which foreign states enjoy in so far as they enjoy such immunity”. Leading counsel for Bahrain, Tom Hickman KC, initially supported the former interpretation but later appeared to accept that the latter interpretation was to be preferred. If he did change his position, I think he was right first time. But the difference does not matter. It would only be relevant if, at the time when the European Convention was adopted and the 1978 Act was passed, there was a class of sovereign acts for which foreign states did not enjoy immunity under customary international law. But there was no such class. As discussed above, there was an international consensus that states enjoyed immunity for all sovereign acts. No instance of state practice or scholarly opinion inconsistent with that core principle has been cited. The divergence of opinion was between states which applied the restrictive doctrine and the dwindling number which adhered to an absolute doctrine of immunity. The common understanding reflected in article 24(1) was thus that any derogation from immunity for sovereign acts would involve a departure from customary international law.
The United Kingdom made a declaration under article 24 when it ratified the Convention. This declaration was in the terms provided for by article 24(1) and was therefore expressly stated to be “without prejudice to the immunity from jurisdiction which foreign States enjoy in respect of acts performed in the exercise of sovereign authority (acta jure imperii)”.
185.
In Kuwait Airways Corpn v Iraqi Airways Co [1995] 1 WLR 1147, 1158, Lord Goff of Chieveley, with whom the other law lords agreed, observed that this declaration:
“must have been intended to recognise the inapplicability in English law of the principle of sovereign immunity in cases in which the sovereign was not acting jure imperii, as had by then been recognised both in The Philippine Admiral [1977] AC 373 and in the Trendtex case [1977] QB 529, though the authoritative statement of the law by Lord Wilberforce in I Congreso del Partido [1983] 1 AC 244, 262, was not then available.”
See also NML Capital Ltd v Republic of Argentina [2011] UKSC 31; [2011] 2 AC 495, para 38 (Lord Phillips of Worth Matravers).
As noted by Lord Goff, the declaration made by the United Kingdom was needed because English law had by then accepted the restrictive doctrine of immunity to which the Convention did not give full effect. But the declaration did not authorise UK law to derogate from the principle of immunity for sovereign acts.
The Act implements the Convention without copying its language
The flexibility afforded by the declaration made under article 24(1) of the European Convention is important to keep in mind when considering how compatibility with the Convention was achieved in the State Immunity Act 1978. In La Générale des Carrières & des Mines v FG Hemisphere Associates LLC [2012] UKPC 27; [2012] 2 Lloyd’s Rep 443, para 10, Lord Mance, writing for the Judicial Committee of the Privy Council, observed that the 1978 Act “was aimed at giving broad effect to (though not following precisely the wording of) the European Convention”. Lord Mance immediately went on, at para 11, to note the limit on such divergence imposed by article 24(1) of the Convention which “left contracting states free to give effect to the restrictive principle of sovereign immunity, ‘without prejudice to the immunity from jurisdiction which foreign States enjoy in respect of acts performed in the exercise of sovereign authority (acta jure imperii)’”.
Comparison of Part I of the 1978 Act with articles 1 to 15 of the European Convention shows that the wording of the Act does not simply replicate the wording of the Convention. The provisions are arranged differently and, even where the substantive effect appears to be the same, a different drafting style is used.
This is evident when the text of section 5 of the Act (quoted at para 150 above) is compared with the wording of article 11 of the Convention (quoted at para 177 above). In Al-Adsani v United Kingdom (2001) 34 EHRR 11, para 22, the Grand Chamber of the European Court of Human Rights observed that section 5 of the 1978 Act “was enacted to implement the 1972 European Convention”. But while section 5 of the 1978 Act is clearly derived from article 11 of the Convention, the fact that it is worded differently is relied on by the claimants to support their contention that it has a materially different effect.
Section 5 refers to where “an act” is done
Turning to the text of section 5, the critical words are “an act … in the United Kingdom”. On behalf of Bahrain, two submissions are made about the meaning of these words. They are: (1) that section 5 applies only where the act or acts causing the death or personal injury or damage to property are committed entirely in the United Kingdom; and (2) that section 5 applies only where the person who does a relevant act (the agent) is in the United Kingdom when the act is done.
Section 5 does not require every causative act(s) to occur in the United Kingdom
As put forward on this appeal, Bahrain’s first contention is ambiguous. Some of Bahrain’s submissions presupposed that a single act can occur partly in one place and partly in another. It was then submitted that, to fall within section 5, the entire act must occur in the United Kingdom. I do not think that this version of the argument is coherent. An act is a type of event (involving an agent) and an event is something that happens in a particular place, in a particular way, at a particular time. It cannot occur in more than one place. Even if such a meaning were possible, it is clearly implicit in section 5 that an act either is or is not done “in the United Kingdom”. The idea that an act might occur partly in and partly outside the United Kingdom is not in contemplation.
Admittedly, injury or damage may result from a course of conduct which occurs partly in one country and partly in another. But in such a case the course of conduct can be broken down into a series of acts. They may be acts of a similar kind, as where a succession of abusive telephone calls amount to harassment; or they may be different elements of an overall scheme, as where an assassination attempt involves manufacturing an explosive device, conveying it to its destination, planting the device and then detonating it. But in either case there are multiple acts which combine to cause harm.
In the Court of Appeal, and on the alternative version of the argument put forward in this court, Bahrain submitted that section 5 applies only when, in a case involving multiple acts which combine to cause injury or damage, all those acts occur in the United Kingdom. But this is not what section 5 says. It refers to “an act”. Counsel for Bahrain rely on section 6(c) of the Interpretation Act 1978 which provides that, unless the contrary intention appears, words in the singular include the plural. But that only shows that the words “an act” should be taken to mean “an act or acts”.
Thus, in a situation where injury or damage is caused by more than one act, section 5 applies if any one or more of the causative acts occur in the United Kingdom. The words used do not require all such acts to have occurred here. Nor is it necessary to distort the language to make practical sense of the provision. The rational interpretation is that, in cases where there are multiple causative acts, the foreign state is immune in so far as proceedings are founded on acts committed outside the United Kingdom but not in so far as the proceedings are founded on acts committed in the United Kingdom.
So in the example of a course of conduct amounting to harassment of the claimant which comprises some acts committed in the United Kingdom and other acts committed abroad, the foreign state would have immunity in respect of the acts committed abroad but not in respect of the acts committed in the United Kingdom. Unless immunity was waived, the claimant could therefore rely in proceedings before a UK court only on the acts committed in the United Kingdom. If it could be shown that those acts caused personal injury for which the foreign state was legally liable, the claimant would be entitled to a remedy; but not otherwise.
A similar analysis would apply in the case of an explosive device manufactured abroad by agents of a foreign state but planted and detonated by agents of a foreign state present in the United Kingdom, causing personal injury. A UK court would lack jurisdiction by reason of immunity in relation to a claim for damages brought by a victim against the foreign state in so far as the claim was based on the manufacture and despatch of the device but not in so far as the claim was based on acts of planting and detonating the device performed by agents present in the United Kingdom.
This is not a case of either of these types. It does not involve a series of acts some of which occurred in the United Kingdom and some of which occurred in another country. Either all the acts complained of took place in the United Kingdom or none of them did. The real question is whether “an act” within the meaning of section 5 can occur somewhere other than where the agent is located. The claimants maintain that it can and that, on the assumed facts, acts which caused them personal injury were committed in the United Kingdom where their personal computers were situated. Bahrain contends that what occurred in the United Kingdom consisted of operations and processes on those computers caused by acts committed abroad and that no act was done in the United Kingdom.
An act is done where the actor is located, not where its effects are felt
If language is used with precision, as is to be expected in an Act of Parliament, the drafter can be taken to have understood the distinction between an act and its effects. An act may have effects which are distant in space or time from the act which causes them. But human beings are embodied creatures who can only act (or omit to act) where they are spatially located, even when the act is intended to and does have effects somewhere else.
To take an example discussed by Lord Keith of Kinkel in Director of Public Prosecutions v Stonehouse [1978] AC 55, 93: if a person on the Scottish bank of the River Tweed, where it forms the border between Scotland and England, fires a rifle at someone on the English bank, the act of shooting takes place in Scotland. If the bullet hits its target, injury occurs in England. But it would be inaccurate to say that the injury was caused by an act performed in England.
The position is in principle the same in a case where a computer in England is hacked by a remote operator located in Bahrain. The operator performs various acts by inputting commands to a computer in Bahrain. Those acts have effects in England which may be almost instantaneous—for example, activating the camera of the targeted computer or causing that computer to transmit data to the server in Bahrain. But the relevant acts—as opposed to their effects—all take place in Bahrain.
201.
This understanding accords with the unanimous opinion of this court in El-Khouri v Government of the United States of America [2025] UKSC 3; [2025] AC 845. That was an extradition case in which the court had to interpret a statutory condition that “the conduct occurs in the … territory” of the state requesting extradition. Of course, the words had to be construed in their context. Many of the arguments in El-Khouri depended on the context, including the fact that section 137 of the Extradition Act 2003 draws a binary distinction between conduct that occurs “in” and conduct that occurs “outside” the territory of the state requesting extradition. But before considering the significance of that context, this court began by focussing simply on the natural or ordinary meaning of the language used. As stated at para 51 of the judgment:
“the word ‘conduct’ would normally and naturally be understood as a synonym for acts done by the requested person in the specified location and not as including effects (whether intended or not) felt in that location of acts done somewhere else. A compelling reason is needed to interpret ‘conduct’ as bearing such an abnormally wide meaning.”
The only linguistic difference in the present case is that the expression used is “an act” rather than “conduct”. As discussed above, “conduct” may comprise more than one act. That matters when analysing a case involving multiple acts which occur in different places. It gave rise to a problem in interpreting section 137 of the Extradition Act because of the binary distinction mentioned above, which is difficult to apply to conduct comprising multiple acts, some of which occurred within and some of which occurred outside the relevant territory. But that problem does not arise here and is not relevant to the present question about where acts are located.
The understanding of ordinary linguistic meaning expressed in El-Khouri does not depend on how many acts are committed. It is equally applicable to a single act as to multiple acts done by a person in a specified location. Applying that understanding, the word “act”, like the word “conduct”, would normally and naturally be understood to refer to what is done by the relevant person in the specified location and not to effects felt there of an act done somewhere else. A compelling reason is needed to give the word “act” such an abnormally wide meaning.
I recognise that what I and the other members of this court in El-Khouri regarded as the normal and natural understanding of where an act which has remote effects occurs is not universally held. Some of the judges in this case, including some of my colleagues who have heard this appeal, think that, even when language is used with precision, it can sometimes without inaccuracy be said that a person located in one country is acting in another country. I will come back to this difference of opinion. It shows, as I accept, that this case cannot be decided just by appealing to the ordinary or “straightforward” meaning of the words used. It is essential to interpret those words in the light of the statutory purpose.
The courts below relied on statements made in two earlier cases which involved the hacking of computers from abroad—though neither case was concerned with state immunity.
206.
R v Governor of Brixton Prison, Ex p Levin [1997] QB 65 was, like El-Khouri, an extradition case. Mr Levin was charged in the United States with various offences the gist of which was that, using his own computer in Russia, he had gained unauthorised access to a bank’s computer in New Jersey and entered instructions which caused funds to be transferred from accounts held by others with that bank to accounts controlled by him. One issue that arose was whether, on the facts alleged, Levin would be guilty of theft under English law if the bank in New Jersey had been located in England. This in turn depended on whether he had appropriated property in the form of rights belonging to another in Russia or in the United States. A Divisional Court rejected an argument that Levin had performed an act of appropriation in Russia when he sat at his computer and typed in instructions. Beldam LJ said, at pp 81–82:
“… the operation of the keyboard by a computer operator produces a virtually instantaneous result on the magnetic disk of the computer even though it may be 10,000 miles away. It seems to us artificial to regard the act as having been done in one rather than the other place. But, in the position of having to choose on the facts of this case whether, after entering the computer in Parsippany, the act of appropriation by inserting instructions on the disk occurred there or in St Petersburg, we would opt for Parsippany. The fact that the applicant was physically in St Petersburg is of far less significance than the fact that he was looking at and operating on magnetic disks located in Parsippany. The essence of what he was doing was done there. Until the instruction is recorded on the disk, there is in fact no appropriation of the rights of [the account holder].”
207.
In an earlier case another Divisional Court had reached a contrary conclusion where instructions to transfer funds were sent by telex from Hong Kong to banks in New York. In R v Governor of Pentonville Prison, Ex p Osman [1990] 1 WLR 277 the court held that an act of appropriation took place in Hong Kong when the telex was sent. Although Osman was purportedly distinguished in Levin, it is difficult to discern a relevant distinction based on where an act was done by the person giving the unauthorised instruction. But there is nothing wrong with an analysis which regards an appropriation of rights as occurring where an instruction is received. It is not necessary to claim that an act was done there. Jurisdiction in a criminal case can quite properly be founded on events occurring in the territory of the forum which are the result of acts done abroad. That is the effect of section 2 of the Criminal Justice Act 1993 under which a person may be guilty of theft (among other offences) if a “relevant event” occurred in England and Wales. A “relevant event” is defined in section 2(1) to mean “any act or omission or other event (including any result of one or more acts or omissions) proof of which is required for conviction of the offence” (emphasis added).
208.
The second case relied on by the courts below is Ashton Investments Ltd v OJSC Russian Aluminium (RUSAL) [2006] EWHC 2545 (Comm); [2007] 1 All ER (Comm) 857. The defendants in that case were persons located in Russia who had allegedly hacked into the claimants’ computer system in London to obtain confidential and privileged information. The claimants applied for permission to serve a claim form on the defendants in Russia which included claims in tort for damages for unlawful interference with their business and for unlawful means conspiracy. To come within the jurisdictional gateway for claims in tort, the claimants had to show either that damage was sustained within England and Wales or that the damage sustained resulted from an act committed within England and Wales. Sitting as a deputy High Court judge, Mr Jonathan Hirst QC held that the first alternative was satisfied because significant damage occurred in England where the server was improperly accessed and the confidential and privileged information was viewed and downloaded. He also expressed the view, at para 63, that “substantial and efficacious acts occurred in London, as well as Russia”. He said:
“That is where the hacking occurred and access to the server was achieved. This may have been as a result of actions taken in Russia but they were designed to make things happen in London, and they did so. Effectively the safe was opened from afar so that its contents could be removed. It would be artificial to say that the acts occurred only in Russia. On the contrary, substantial and effective acts occurred in London.”
209.
In the present case the Court of Appeal, in agreement with the judge, considered the cases of Levin and Ashton to be helpful “because they demonstrate on comparable facts that to describe the act of hacking as taking place here is a natural and appropriate use of language”: [2024] EWCA Civ 1158; [2025] KB 490, para 35. Males LJ also adopted the view expressed in both Levin and Ashton that it was “artificial” to say that the act of hacking occurred only where the person who performed the relevant acts was located: see paras 34 and 39. Referring to the remote manipulation from abroad of a computer in the United Kingdom, he said, at para 34:
Treating effects as acts produces incoherence
What is meant by describing the distinction between acts performed in one country and effects occurring in another as “artificial”? I take the thinking to be that, because of the degree of control which the human agent exercises over the targeted computer, it is illogical or unprincipled to draw a legal distinction between a case of remote manipulation and one where the agent is physically present where the computer is located, inputting commands directly. The point could be expressed by saying: “the agent might as well be present” or “it is as if the agent was present” in terms of what operations the agent causes the computer to perform.
It is undoubtedly true that modern technology enables acts done in one place to have substantial and almost instantaneous effects in distant locations in ways that would have been unimaginable a generation ago. The use of such technology raises questions of legal responsibility which did not arise and would not have been in contemplation when the European Convention on State Immunity and the 1978 Act were drafted. What significance this should have for extradition law or for the service of proceedings on defendants abroad (with which Levin and Ashton were respectively concerned) it is unnecessary to consider. But I do not agree that, in determining the scope of state immunity for torts, it is illogical or unprincipled, or “artificial”, to distinguish between acts and effects; still less that such a distinction would have appeared “artificial” in 1978 when Parliament enacted the State Immunity Act. Five points may be made.
First, it is clearly true that foreign states may carry out a variety of acts on their own territory which, because of the actual and intended effects of those acts in another country, may be regarded as interfering with that other country’s territorial sovereignty. But, at the time when the European Convention was adopted and the United Kingdom decided to ratify it, it was reasonable to regard the presence of a foreign state’s agents in the territory of the forum state and the committal of wrongful acts by such agents as a greater interference with territorial sovereignty than the performance of wrongful acts by agents situated abroad. Recognising a tort exception to state immunity which was limited to acts done by agents in the territory of the forum state and which did not extend to transboundary torts drew a clear and intelligible line. The same line was later drawn in the UN Convention (see para 271 below).
If a new international convention were to be negotiated today, a different approach might be taken. In their judgment (at para 147) the majority give examples of the interference with the territorial sovereignty of the United Kingdom that would occur if a murder was carried out here by a drone operated by a foreign state official situated abroad; or if such drones entered the airspace of Heathrow causing aeroplanes to crash and multiple deaths; or if NHS computer systems were hacked by operatives in a foreign state and put out of operation causing patients to die. The points are well made. But they involve reading the 1978 Act through the prism of hindsight. No one involved in the formation of the European Convention or the 1978 Act could have had such possibilities in view. The legislation must be interpreted in the historical context of the situation which led to its enactment (see para 161 above), not on the basis of a judge’s view that a different regime would now be desirable.
214.
Second, a rule—such as that contained in article 11 of the European Convention—based on where the agent is located has advantages of clarity and certainty. It is generally simple and straightforward to apply. If that rule is abandoned in favour of a rule which treats some acts as occurring in a place other than where the agent is located, it is unclear and uncertain how the distinction between acts and effects is to be drawn. If computer hacking is said to involve acts done where the targeted computer is located, what of the rifle fired by a person standing on the Scottish bank of the Tweed at someone on the English bank? Are we to say that the act of shooting takes place in England? If not, how is this case different in principle from a computer hacking case? But if so, does the answer depend on whether the bullet hits the target? Any such approach leads down a path to absurdity exemplified by a decision of the Supreme Court of Georgia in the United States in 1893. The court reasoned that:
“if a man in the state of South Carolina criminally fires a ball into the state of Georgia, the law regards him as accompanying the ball, and as being represented by it, up to the point where it strikes.”
On this basis the act of shooting performed by a man standing in the state of South Carolina was held to have occurred in the state of Georgia because “[a person] shooting from another state goes, in a legal sense, where his bullet goes” and “the fact of his missing the object at which he aims cannot alter the legal principle”: see Simpson v State (1893) 92 Ga 41; 17 SE 984, 985.
If the term “artificial” is to be used, this reasoning might be thought to epitomise it. Rather than stretching the understanding of where an act is done beyond its breaking point, the relevant territorial connection in such cases should be recognised for what it is. The connection which justifies the exercise of jurisdiction is not that the defendant did any act in the territory of the forum but the fact that harm was caused (or an attempt was made to cause such harm) in that territory. For the purposes of deciding whether a foreign defendant can be sued in a local court or which country’s law applies to a claim in tort, the general tendency has been to move away from a test based, or based solely, on where the wrongful act was committed in favour of a test based on where the damage occurred. But no such shift has yet happened in the law governing state immunity. The court must therefore focus solely on where the act, and not its harmful effects, took place.
In oral argument leading counsel for the claimants, Timothy Otty KC, initially appeared to maintain that there is a legal distinction between cases of computer hacking, where on the claimants’ case the acts occur in the place where the victim’s computer is located, and other transboundary cases such as shooting a rifle or firing a missile across a border or sending a letter bomb or a box of poisoned chocolates by post from one country to another, where the relevant acts occur where the actor is located and not where the effects are felt. Ultimately, however, Mr Otty submitted that in all these cases the acts occur in the country of destination. This seems to collapse altogether the distinction between acts and their consequences and leaves it entirely unclear where the former are supposed to end and the latter to begin.
It also creates other conceptual difficulties. It is unclear whether on the claimants’ case a person sitting at a computer terminal in a foreign state who hacks a computer situated in England acts only in England or (as seems more plausible) also in the foreign state. If the latter, is it supposed that the same act occurs in two different places or that the agent commits two separate acts in different places one of which causes the other? Neither theory is coherent.
The short point is that no rational criterion has been identified for distinguishing between acts and effects and for deciding where acts are committed if the possibility is allowed that acts can be dissociated in space from the location of the agent who does the act. It is unreasonable to attribute to Parliament a meaning of the word “act” which is indeterminate in this way.
Third, section 5 (like article 11) is not restricted to death or personal injury or damage to tangible property sustained in the United Kingdom. The injury or damage may occur abroad. Given that the justification in principle for the departure from immunity in section 5 must be the sovereignty of the United Kingdom over its own territory, it is hard to justify such a departure when neither the victim nor the perpetrator of the wrongful act is present in the United Kingdom. An interpretation of section 5 which requires at least the perpetrator to be in the United Kingdom, even though the injury or damage may be suffered in another country, thus better accords with the underlying rationale for the exception.
220.
Fourth, when considering a case of computer hacking, it should be noted that, on the facts of Levin and Ashton, if the acts had been done by an agent of a foreign state rather than a private individual and a claim in tort had in either case been brought against the foreign state in a UK court, the foreign state would undoubtedly have had immunity. That is because it was not suggested in either case that the acts carried out had caused personal injury or damage to tangible property. The consequences alleged were financial loss and, in Ashton, unauthorised use of confidential information. It seems improbable that many cases of computer hacking will result in personal injury or that such cases, even if the relevant technology had existed, would have been cases that a person framing section 5 of the 1978 Act would have had in view. There is no principled basis for treating a foreign state as subject to the jurisdiction of the UK courts if its agents hack a computer from abroad and thereby cause psychiatric injury to someone (who may also be abroad), yet immune from suit if instead such hacking causes massive financial damage in the UK.
Fifth, any suggestion that it might have been thought illogical or unreasonable to distinguish between causative acts committed by an agent located abroad and such acts committed by an agent who is in the United Kingdom, treating the former as covered by state immunity and latter as not immune, is refuted by the fact that this distinction was indisputably drawn in the European Convention on State Immunity. It has also, as will be seen, since been adopted in the United Nations Convention on Jurisdictional Immunities of States and their Property (2004) and in various other national legislation. It cannot in these circumstances be viewed as improbable that the distinction should have commended itself to the UK Parliament when enacting the 1978 Act.
For these reasons, I consider that, on the interpretation which makes best sense of the words of section 5, even before account is taken of the purpose of the 1978 Act and the international law context, “an act or omission in the United Kingdom” means something that is done (or omitted to be done) by a person who is in the United Kingdom.
Consistency with the UK’s treaty obligations confirms this meaning
Consideration of the purpose of the 1978 Act in my view puts the matter beyond doubt. Critical to the analysis is article 24 of the European Convention, which appears to have been overlooked in the courts below.
224.
As noted earlier, a principal purpose of the 1978 Act (described in Andrew Dickinson and Alexander Thompson, The State Immunity Act 1978 (2025), para 1.2, as “[t]he main purpose”) was to enable the United Kingdom to ratify the European Convention. To achieve consistency with the obligations assumed on ratification, it was necessary to adapt UK law so as (1) to recognise an exception in the circumstances covered by article 11 of the Convention from the rule of immunity in respect of sovereign acts; but also (2) to adhere to that rule in cases of personal injury or property damage not falling within article 11. The first of these conditions flowed from the proposed agreement to article 11 itself, which applies irrespective of whether the act or omission in question is performed in the exercise of sovereign authority. The second condition followed from the proviso in article 24(1), which prohibited a state ratifying the Convention from recognising an exception to the immunity which foreign states enjoy in respect of sovereign acts that is wider than the exceptions required by articles 1 to 13 of the Convention.
It is true that the United Kingdom could, consistently with the adoption of the European Convention, have limited the application of section 5 to states which were also parties to the Convention. But, understandably, the 1978 Act did not adopt a two-tier approach of enacting different rules of immunity in UK law for those states which were, and those states which were not, parties to the Convention. With one exception (section 13(4), which concerns a procedural privilege relating to the enforcement of a judgment or arbitration award against state property), Part I of the Act treats all foreign states alike. Aligning UK law with article 11 of the Convention thus had the consequence of departing from customary international law in so far as section 5 denies immunity for sovereign acts falling within its scope committed by a state not a party to the Convention.
But the choice to depart from customary international law in this respect is not a reason to adopt an interpretation of section 5 which is inconsistent with the European Convention and the legislative purpose of enabling the United Kingdom to ratify it. The strong presumption referred to earlier (see para 163 above) applies, which requires the words of the statute to be construed as consistent with the United Kingdom’s treaty obligation if they are reasonably capable of bearing such a meaning.
As explained, it would be inconsistent with article 11 read with article 24(1) of the Convention to interpret section 5 as extending to sovereign acts performed by a person who is not present in the territory of the United Kingdom. Such inconsistency is readily avoided by giving the words “an act or omission in the United Kingdom” what I consider in any event to be their normal and rational meaning.
When, therefore, regard is had to the context and purpose of the 1978 Act, the only meaning that can reasonably be given to those words is that they refer to an act done or omitted to be done by a person present in the United Kingdom.
Why the different wording of section 5 does not widen the exception
A majority of the court has reached a different conclusion. The point which they regard as critical, and determinative of the appeal (see para 76 of their judgment), is that section 5 of the 1978 Act uses different language from article 11 of the European Convention. In particular, they emphasise that article 11 employs two distinct linking factors to the territory of the forum state: first, “the facts which occasioned the injury or damage” must have occurred in that territory; and, second, “the author of the injury or damage” must have been “present in that territory at the time when those facts occurred”. By contrast, section 5 of the 1978 Act has only one requirement: that an act or omission which caused personal injury or damage to property occurred in the United Kingdom.
230.
The Court of Appeal concluded that, in the light of these differences in wording, “the terms of article 11 are of no real help in interpreting section 5”: see para 64 of Males LJ’s judgment. The claimants took the same position on this appeal. But the majority of this court take a different view. They consider that the terms of article 11 are of help and indeed are decisive in interpreting section 5. This is because the differences in wording, in their view, lead inescapably to the conclusion that Parliament must be taken to have deliberately created an exception to immunity in UK law which is intended to be wider than the exception permitted by article 11 of the Convention. The basis for this inference is a belief that the first of the two linking factors set out in article 11 is reflected in the express terms of section 5 of the 1978 Act, but the second is not. It is said that, if the intention had been to require both linking factors in section 5, the obvious course would have been to recite both factors using the language of article 11 and that it is “inconceivable” that the second linking factor should have been omitted. The majority conclude that there was in enacting section 5 a clear and deliberate departure from the requirement in article 11 that the agent of the foreign state must be present in the territory of the forum: see para 224 above. This inference is considered to be so inexorable that it must override the presumption that Parliament intended to legislate consistently with the international obligations of the United Kingdom. There is, it is thought, simply no room to apply the principle of consistent interpretation in this case.
I disagree fundamentally with this reasoning. It is a strong thing to attribute to Parliament an intention to place the United Kingdom in breach of an international obligation. It is an even stronger thing to attribute to Parliament such an intention when, as is common ground, one of the very purposes of the 1978 Act was to enable the United Kingdom to ratify the European Convention by making UK law consistent with the obligations which the United Kingdom would assume on ratification. The majority of the court has concluded that Parliament deliberately included in an Act designed to secure compatibility with an international treaty a provision incompatible with that treaty. That is to say, Parliament included in the statute a provision which was deliberately intended to defeat a principal purpose of the statute.
That is an extreme conclusion to reach. It prompts the question: what policy goal does such an interpretation promote which could possibly have been regarded as of such critical importance as to justify placing the United Kingdom in breach of the treaty that it was proposing to ratify? None has been suggested. Indeed, no policy goal at all has been identified to explain why Parliament should be supposed to have wished to go further than the European Convention had gone in removing immunity from sovereign acts which caused personal injury or property damage.
There could in principle be language used in legislation which is so unequivocal and explicit that it is impossible to avoid giving it a meaning that defeats a purpose of the legislation and has no identifiable aim. But this case comes nowhere near to warranting such a counsel of despair. I have already explained why I consider the meaning of section 5 of the Act to be consistent with article 11 of the European Convention even without taking account of the context and purpose of the legislation and the central importance of giving effect to that purpose. The differences in wording between the two provisions do not support, let alone require, any contrary implication.
234.
I have mentioned the general feature of the 1978 Act that it does not simply copy the wording of the Convention even when enacting in UK law provisions presumably intended to have equivalent effect. This reflects the fact that the UK legislation employs a different drafting technique. As Lord Diplock observed in Fothergill v Monarch Airlines Ltd [1981] AC 251, 281–282:
“The language of [an international convention] … is meant to be understood in the same sense by the courts of all those states which ratify or accede to the Convention. Their national styles of legislative draftsmanship will vary considerably as between one another. So will the approach of their judiciaries to the interpretation of written laws …
The language of an international convention has not been chosen by an English parliamentary draftsman. It is neither couched in the conventional English legislative idiom nor designed to be construed exclusively by English judges. It is addressed to a much wider and more varied judicial audience …”
235.
It is understandable that Parliamentary counsel should have chosen to use different language in section 5 of the 1978 Act from the language of article 11. Not only is article 11 drafted in a very different style from what Lord Diplock referred to as “the conventional English legislative idiom” but it is on any view clumsily and imprecisely phrased. The formulation that “the facts which occasioned the injury or damage occurred in the territory of the state of the forum” is a solecism, as “facts” do not cause injury or damage. Nor do facts occur in any territory. Facts are true propositions. The proposition that a bomb exploded in London injuring several people, if true, is a fact. But if it is true, what caused the injuries and occurred in the United Kingdom was an event (the explosion) and not a proposition. No Parliamentary counsel would use such an inaccurate formulation. Nor would a UK lawyer refer to “the author” of the injury or damage. That expression is a literal, but not an idiomatic, equivalent of the French text which refers to “l’auteur du dommage”.
Article 11 also offends another principle of UK Parliamentary drafting, and indeed of any good writing in English, which is to use no more words than are necessary to express your meaning and, in particular, not to use roundabout phrases where single words would serve: see Sir Ernest Gowers, The Complete Plain Words (1954), p 57.
As well as being imprecise, the term “facts” in article 11 is broader in scope than “acts or omissions”. It is presumably used to mean “events” and is apt to include events other than acts of human beings. It would include events which are a result of one or more acts: for example, the explosion of a bomb. The term is broad enough even to include natural events, not caused by any human action, such as a lightning strike—though the later reference to “the author of the injury or damage” shows that only injury or damage caused by an act or omission of a human being is within the scope of article 11.
The approach adopted in drafting article 11 was to impose a very widely drawn condition which requires the occurrence in the territory of the forum state of “facts” (ie events) not confined to human acts. For example, the explosion of a letter bomb sent from abroad or injury occurring in the territory of the forum state caused by a defect in a product imported from and manufactured in another country would, at least arguably, fall within the first condition (though the significance of the definite article in referring to “the” facts is unclear). The second condition then cuts down the scope of the provision by requiring “the author of the injury or damage” to have been “present in that territory at the time when those facts occurred”.
Instead of adopting this roundabout approach, section 5 achieves a similar effect with far more precise and economical drafting. It does not use the loose and broad term “facts” and then narrow the scope of the provision with a separate presence requirement. Instead, it achieves the same end by referring simply to “an act or omission in the United Kingdom”. It was unnecessary to include an additional requirement for the actor to be present because that is inherent in the concept of an act done in a specified place.
The claim that the second of the two linking factors required by article 11 is not reflected in section 5 assumes what it is supposed to prove. It does not provide a reason for rejecting the understanding that the express terms of section 5 combine both linking factors.
The claim also rests on an erroneous assumption that the words “caused by an act or omission in the United Kingdom” are equivalent to the first condition in article 11. They are not. The first condition in article 11 refers to “the facts”, and not to “an act or omission”. If the UK statute had replicated the first condition in article 11 and omitted the second condition, there would have been at least some basis for arguing that the omission was intended to give section 5 a different legal effect. But the language of section 5 does not correspond to the first condition in article 11. Instead, it combines both conditions.
Thus, the differences in wording between section 5 of the 1978 Act and article 11 of the European Convention do not support, still less compel the court to adopt, an inconsistent interpretation of section 5.
Other provisions of the 1978 Act do not support an inconsistent interpretation
The claimants also point to other provisions of the 1978 Act which are different in scope from the corresponding provisions of the Convention. They do so to show that the Act “deliberately diverges” from the European Convention. Philippa Webb, who presented this part of the claimants’ case, identified the following sections of the 1978 Act as diverging from the Convention: sections 3(1) and (3), 4(1), 9 and 13(4).
Of these provisions, sections 4(1) and 13(4) can be put to one side because, although different in scope, it is not suggested that they conflict with the Convention. As Ms Webb accepted, they are examples of the United Kingdom exercising the freedom afforded by its declaration under article 24(1) to enact exclusions of immunity which are broader in scope than the corresponding provisions of the Convention and relate to private (non-sovereign) acts. (The divergence in section 13(4) is also specifically provided for by article 26 of the Convention as a feature of the optional regime made available to states which have made a declaration under article 24(1).)
This leaves section 3(1) and (3) and section 9. These provisions are said to be inconsistent with the Convention because they go further than the Convention in excluding immunity for sovereign acts, which is not permitted by article 24. Even if this claim were correct, it is of doubtful relevance to this appeal.
Unless perhaps the Act were shown systematically to adopt such an approach, I do not see how the fact—if it be a fact—that one provision cannot be interpreted consistently with the United Kingdom’s treaty obligations justifies giving an inconsistent interpretation to another provision. The presumption of consistency is not displaced by showing that Parliament has failed to comply with international law in some other respect.
But I am not persuaded anyway that there is such an inconsistency. I will first consider section 3(1) and (3), followed by section 9.
248.
So far as relevant, section 3 of the 1978 Act provides:
“3 Commercial transactions and contracts to be performed in United Kingdom.
an obligation of the State which by virtue of a contract (whether a commercial transaction or not) falls to be performed wholly or partly in the United Kingdom.
(3)
any loan or other transaction for the provision of finance and any guarantee or indemnity in respect of any such transaction or of any other financial obligation; and
any other transaction or activity (whether of a commercial, industrial, financial, professional or other similar character) into which a State enters or in which it engages otherwise than in the exercise of sovereign authority;
The claimants make two points about the scope of these provisions. The first relates to section 3(1)(a) when read with the definition of the term “commercial transaction” in section 3(3). The second relates to section 3(1)(b).
Section 3(1)(a) excludes from immunity proceedings relating to any commercial transaction entered into by a foreign state. There is no direct counterpart to this provision in the Convention. Although article 7 covers part of the same ground, it is limited to proceedings relating to an industrial, commercial or financial activity of an office, agency or establishment through which the state engages in that activity in the same manner as a private person on the territory of the forum state.
Section 3 can, however, be justified as seeking to give effect and set precise limits to the doctrine of restrictive immunity. There is, as noted earlier, a grey area around how the distinction between sovereign acts and private acts is drawn, and the definition of a “commercial transaction” adopted in section 3(3) of the 1978 Act can be defended as a legitimate interpretation of circumstances in which states have entered into commercial transactions with private parties which ex hypothesi are acts jure gestionis.
252.
A puzzle is created by the reference in section 3(3)(c) to “any other transaction or activity … in which [a state] engages otherwise than in the exercise of sovereign authority”. Ms Webb submitted that, as section 3(3)(a) (contracts for the supply of goods or services) and section 3(3)(b) (loans and other financial obligation) are not qualified by the words “otherwise than in the exercise of sovereign authority”, it should be inferred that those provisions apply to transactions which are sovereign in character. But that interpretation ignores the significance of the word “other” in section 3(3)(c), which implies that the transactions described in section 3(3)(a) and (b) are not regarded as sovereign in character. I agree with the analysis proposed by Hazel Fox and Philippa Webb, The Law of State Immunity, 3rd rev ed (2015), p 194, that:
“the terms ‘commercial transaction’ and ‘activity otherwise than in the exercise of sovereign authority’ are positive and negative definitions of the same concept; a comprehensive dichotomy is thus set up by the statute by which all acts not amounting to commercial transactions constitute acts in exercise of sovereign authority. This dichotomy between acts jure imperii and jure gestionis is elaborated by the addition of the descriptive words ‘commercial, industrial, financial, professional, or other similar character’ so as to embrace the widest conception of private law or commercial acts.”
This interpretation should be preferred, not least for the reason that it avoids inconsistency with customary international law and with the Convention. It also accords with the view expressed by Lord Phillips in NML Capital Ltd v Republic of Argentina [2011] UKSC 31; [2011] 2 AC 495, paras 37–39, that the broad exemption from immunity of commercial transactions in section 3(1)(a) was authorised by the declaration made by the United Kingdom under article 24(1) of the Convention.
Section 3(1)(b) excludes from immunity proceedings relating to any contractual obligation which “falls to be performed wholly or partly in the United Kingdom”. This provision corresponds to article 4 of the Convention, save that article 4 refers to any contractual obligation which “falls to be discharged in the territory of the state of the forum”. It is suggested that article 4 is narrower because it does not contain the words “wholly or partly”. I do not consider this to be a point of any substance. It seems to me that the inclusion of those words in section 3(1)(b) does no more than clarify what might otherwise be an ambiguity in article 4 and is not a material difference.
Section 9(1): arbitration
Section 9 of the 1978 Act provides as follows:
Where a State has agreed in writing to submit a dispute which has arisen, or may arise, to arbitration, the State is not immune as respects proceedings in the courts of the United Kingdom which relate to the arbitration.
This section has effect subject to any contrary provision in the arbitration agreement and does not apply to any arbitration agreement between States.”
256.
The corresponding provision of the Convention is article 12:
Where a Contracting State has agreed in writing to submit to arbitration a dispute which has arisen or may arise out of a civil or commercial matter, that State may not claim immunity from the jurisdiction of a court of another Contracting State on the territory or according to the law of which the arbitration has taken or will take place in respect of any proceedings relating to:
the validity or interpretation of the arbitration agreement;
unless the arbitration agreement otherwise provides.
Paragraph 1 shall not apply to an arbitration agreement between States.”
Section 9 of the 1978 Act is broader than article 12 of the European Convention in three respects. Section 9 is not limited, as article 12 is: (i) to disputes arising out of “a civil or commercial matter”; (ii) to arbitrations which take place on the territory or according to the law of the forum state; and (iii) to proceedings relating to the validity or interpretation of the arbitration agreement, the arbitration procedure or the setting aside of the award.
Arbitration is a consensual procedure and the principle underlying section 9 is that, if a state has agreed to submit to arbitration, it has consented to the exercise of jurisdiction by a court to the extent necessary to render the arbitration effective and has waived any entitlement to claim immunity in that regard: see Svenska Petroleum Exploration AB v Government of the Republic of Lithuania (No 2) [2006] EWCA Civ 1529; [2007] QB 886, paras 117–123; NML Capital Ltd v Republic of Argentina [2011] UKSC 31; [2011] 2 AC 495, para 89 (Lord Mance); Deutsche Bank AG v Central Bank of Venezuela [2023] EWHC 1942 (Comm); [2023] 2 Lloyd’s Rep 486, paras 19–20. An exemption based on waiver does not conflict with the principle of immunity of foreign states in respect of sovereign acts. For that reason, although the point does not arise for decision on this appeal, I am not persuaded that section 9 of the Act is inconsistent with article 24 of the Convention.
Accordingly, I do not consider that either of the other provisions of the 1978 Act on which the claimants seek to rely lends support to any suggestion that the UK Parliament was nonchalant about whether the terms of the 1978 Act complied with the United Kingdom’s treaty obligations.
For these reasons I conclude that section 5 of the 1978 Act is properly interpreted as applying only to acts done by persons who are in the United Kingdom and not to acts done by persons situated in the territory of a foreign state. Section 5 therefore does not deprive Bahrain of immunity in these proceedings.
The claimants’ alternative case: compatibility with article 6
I have mentioned in the introduction to this judgment that the claimants have an alternative case. They argue that, even if—as I have concluded—section 5 of the State Immunity Act 1978, when enacted, was similar in scope to article 11 of the European Convention, the words “an act or omission in the United Kingdom” can and must now be given a broader meaning to make the legislation compatible with article 6 of the European Convention on Human Rights.
This argument relies on section 3(1) of the Human Rights Act 1998, which provides that “[s]o far as it is possible to do so, primary legislation … must be read and given effect in a way which is compatible with the Convention rights”. This provision imposes what has been described as “a remarkably powerful interpretative obligation, which goes well beyond the normal canons of statutory construction”: In re United Nations Convention on the Rights of the Child (Incorporation) (Scotland) Bill [2021] UKSC 42; [2021] 1 WLR 5106, para 25 (Lord Reed).
263.
The nature and extent of the obligation were explained by Lord Nicholls of Birkenhead in Ghaidan v Godin-Mendoza [2004] UKHL 30; [2004] 2 AC 557. His conclusion, at para 32, was that:
“Section 3 enables language to be interpreted restrictively or expansively. But section 3 goes further than this. It is also apt to require a court to read in words which change the meaning of the enacted legislation, so as to make it Convention-compliant. In other words, the intention of Parliament in enacting section 3 was that, to an extent bounded only by what is ‘possible’, a court can modify the meaning, and hence the effect, of primary and secondary legislation.”
The limit of what is “possible” is that the courts may not adopt a meaning “inconsistent with a fundamental feature of legislation” or which is not “compatible with the underlying thrust of the legislation being construed”: Ghaidan, para 33.
264.
Article 6 (the “Convention right” on which the claimants’ argument is based) guarantees the right to a fair hearing in the determination of civil rights and obligations. Clear and constant case law of the European Court of Human Rights has held that this right includes a right of access to a court to determine a dispute and that a claim to state immunity infringes that right unless it is a proportionate means of pursuing a legitimate aim. The principal decisions are the three judgments of the Grand Chamber in Al-Adsani v United Kingdom (2001) 34 EHRR 11, Fogarty v United Kingdom (2001) 34 EHRR 12 and McElhinney v Ireland (2001) 34 EHRR 13. Those cases decide: (1) that the grant of sovereign immunity to a state in civil proceedings pursues the legitimate aim of complying with international law to promote comity and good relations between states through respect for another state’s sovereignty; and (2) that measures taken which reflect generally recognised rules of public international law on state immunity cannot in principle be regarded as imposing a disproportionate restriction on the right of access to a court.
As discussed earlier, there is no doubt that, when the State Immunity Act was enacted in 1978, affording immunity on facts such as those alleged by the claimants in this case was required by customary international law. There was undisputed immunity in tort cases in respect of sovereign acts. Although the parties to the European Convention on State Immunity had agreed to establish an exception in their relations with one another with respect to a defined class of tort cases involving personal injury or property damage, that agreement could not realistically be said to have altered customary international law. In any event it was limited to circumstances in which the foreign state agent who caused the injury was present in the territory of the forum state at the time when the injury was caused. It therefore did not extend to a case such as this involving sovereign acts committed by agents not present in the territory of the forum state.
The claimants’ alternative case requires the court to consider whether the position has since changed, so that affording immunity on these facts would no longer reflect generally recognised rules of public international law. This raises two questions. First, can it now be said that there is a “territorial tort exception” to the general rule of immunity under customary international law for sovereign acts? Second, if so, does this exception extend to a case where the acts causing personal injury were committed by persons who were not within the territory of the forum state?
It is important to keep these questions distinct. The first asks whether customary international law now permits or requires a territorial tort exception for sovereign acts at all. The second asks whether any such exception is wider than that adopted in the European and UN Conventions and extends to transboundary acts committed by agents abroad.
268.
To determine the existence and content of a rule of customary international law, it is necessary to ascertain whether there is a general practice of states which is accepted as law (opinio juris): see eg North Sea Continental Shelf (Federal Republic of Germany v Denmark; Federal Republic of Germany v The Netherlands) [1969] ICJ Rep 3, para 77; Jurisdictional Immunities of the State, para 55. To satisfy the first of these conditions, the relevant practice must be sufficiently widespread and representative, as well as consistent: see Draft conclusions on identification of customary international law, with commentaries (conclusion 8), adopted by the International Law Commission in 2018 (Yearbook of the International Law Commission 2018, Vol II, Part 2, p 100). In Jurisdictional Immunities of the State, the International Court of Justice observed, at para 55:
“In the present context, state practice of particular significance is to be found in the judgments of national courts faced with the question whether a foreign state is immune, the legislation of those states which have enacted statutes dealing with immunity, the claims to immunity advanced by states before foreign courts and the statements made by states, first in the course of extensive study of the subject by the International Law Commission and then in the context of the adoption of the United Nations Convention. Opinio Juris in this context is reflected in particular in the assertion by states claiming immunity that international law accords them a right to such immunity from the jurisdiction of other states; in the acknowledgment, by states granting immunity, that international law imposes upon them an obligation to do so; and, conversely, in the assertion by states in other cases of the right to exercise jurisdiction over foreign states.”
A major development since the enactment of the 1978 Act has been the adoption in 2004 of the United Nations Convention on Jurisdictional Immunities of States and their Property, referred to in this passage. It is therefore relevant to consider the territorial tort exception embodied in the UN Convention and its drafting history. As also mentioned in the passage quoted, a number of states have enacted statutes dealing with immunity. It is therefore relevant to consider this national legislation and, where applicable, how it has been judicially interpreted. The judgment of the International Court of Justice in Jurisdictional Immunities of the State, given in 2012, is itself another significant development.
The UN Convention was adopted by the General Assembly on 2 December 2004 and opened for signature on 17 January 2005. The treaty has not yet entered into force, as this requires ratification by 30 states and, so far, only 25 states have ratified it. The United Kingdom has signed, but not yet ratified, the UN Convention.
271.
Like article 15 of the European Convention and section 1 of the 1978 Act, the UN Convention provides, in article 5, for a general rule of state immunity subject to the other provisions of the Convention. Article 12 is a territorial tort exception for personal injuries and damage to property in the following terms:
“Personal injuries and damage to property
Unless otherwise agreed between the States concerned, a State cannot invoke immunity from jurisdiction before a court of another State which is otherwise competent in a proceeding which relates to pecuniary compensation for death or injury to the person, or damage to or loss of tangible property, caused by an act or omission which is alleged to be attributable to the State, if the act or omission occurred in whole or in part in the territory of that other State and if the author of the act or omission was present in that territory at the time of the act or omission.”
As can be seen, this provision is substantially similar in scope to article 11 of the European Convention. In particular, like article 11 of the European Convention, article 12 of the UN Convention (i) does not distinguish between sovereign and private acts and (ii) contains an express condition that the person who caused the injury or damage was present in the territory of the forum state.
The UN Convention was a long time in gestation. The process began in 1977 when the General Assembly invited the International Law Commission (“ILC”) to commence work on jurisdictional immunities of states and their property. In 1983, Special Rapporteur Sucharitkul discussed in his Fifth report what he described as “an emerging trend” in state practice and international legal opinion in favour of restricting state immunity in cases involving personal injuries or damage to property occurring in the forum state: see Yearbook of the International Law Commission 1983, Vol II, Part 1, pp 38–46.
He proposed a draft article which would limit state immunity in this area. That draft formed the basis of the provision ultimately adopted and contained the same definition of the required territorial connection with the forum state as the final version of article 12.
The proposal provoked lively discussion within the ILC: see Joanne Foakes and Roger O’Keefe, “Article 12” in The United Nations Convention on Jurisdictional Immunities of States and their Property: A Commentary (2015), pp 212–213; and Yearbook of the International Law Commission 1983, Vol I, pp 75–99. Some members opposed the draft article; others welcomed it; and some argued for an expansion of its scope. The draft article was referred to the drafting committee and, after certain drafting adjustments, was adopted on first reading in 1986.
Along with the other draft articles adopted on first reading, the draft article, then numbered 13, was circulated among states for comment. Comments submitted by various governments were published in the Yearbook of the International Law Commission 1988, Vol II, Part 1. Of those governments which made specific comments on draft article 13, several either opposed it in principle or expressed reservations about its breadth: Brazil (p 58, para 5); Bulgaria (p 59, para 10); the Byelorussian Soviet Socialist Republic (p 61, para 12); Chile (p 63, para 11); Czechoslovakia (p 64, para 8); France (p 66, para 28); the German Democratic Republic (p 69, para 21); and the USSR (p 83, para 12).
At the other extreme, Australia (p 54, para 31), the Federal Republic of Germany (p 71, para 15), Italy (p 73, para 9) and possibly Thailand (p 81, para 11) favoured deleting the requirement that the author of the relevant act or omission must be present in the territory of the forum state. In his Preliminary Report the new Special Rapporteur, Motoo Ogiso, endorsed that suggestion (p 111, para 141). In his Second Report, however, he clarified that the deletion of those words would not, in his view, make the exception “applicable to tort[s] committed abroad or other transfrontier injurious acts because of the first requirement of territorial connection: the relevant act or omission must occur ‘in whole or in part in the territory of the state of the forum’”: see Yearbook of the International Law Commission 1989, Vol II, Part 1, pp 65–66, para 20.
The ILC’s discussion of draft article 13 on second reading in 1989 revealed the same division as before between those members opposed to the provision in its entirety and those in favour of it: see Yearbook of the International Law Commission 1989, Vol II, Part 2, p 111, paras 520–522. Some other members, “while not entirely opposed to the inclusion of a provision on the matter”, considered that the draft article “required substantial redrafting” (para 523). On balance, however, “the views expressed by members generally indicated a preference to retain the text of the article as adopted on first reading” (para 524). In particular:
“The deletion of the second territorial criterion, as proposed by the Special Rapporteur, was not considered appropriate, since transboundary damage normally gave rise to international disputes which had to be settled by recourse to international law and not to the law applicable in a forum State.”
On final consideration the following year, only a single ILC member favoured deleting the draft article as a whole, but various concerns were expressed about the breadth of the provision: see Yearbook of the International Law Commission 1990, Vol II, Part 2, pp 35–36, paras 187–190. The text was referred once more to the drafting committee, which inserted the adjective “pecuniary” before “compensation” but made no other change to the wording. The provision, renumbered draft article 12, was finally adopted in this form in the draft articles submitted by the ILC (with an accompanying commentary) to the General Assembly in 1991: see Yearbook of the International Law Commission 1991, Vol II, Part 2, pp 44–46. These draft articles were subsequently adopted as the text of the Convention.
280.
Paragraph (8) of the ILC commentary on draft article 12 explained:
“The basis for the assumption and exercise of jurisdiction in cases covered by this exception is territoriality. The locus delicti commissi offers a substantial territorial connection regardless of the motivation of the act or omission, whether intentional or even malicious, or whether accidental, negligent, inadvertent, reckless or careless, and indeed irrespective of the nature of the activities involved, whether jure imperii or jure gestionis. This distinction has been maintained in the case law of some states involving motor accidents in the course of official or military duties. While immunity has been maintained for acts jure imperii, it has been rejected for acts jure gestionis. The exception proposed in article 12 makes no such distinction, subject to a qualification in the opening paragraph …”
The comment that the distinction between acts jure imperii and acts jure gestionis had been maintained in the case law of some states was accompanied by a footnote. The footnote cites a long list of cases in which courts of different countries applied the distinction between sovereign and private acts to civil claims for compensation for personal injuries or damage to property, upholding immunity where the acts were found to be sovereign acts. The only examples given of cases in which courts of any country had denied immunity for sovereign acts in any area of tort law were cases decided in the United States under its Foreign Sovereign Immunities Act (as to which, see paras 286–290 below).
282.
Para (6) of the ILC commentary described the territorial connection with the state of the forum needed for the application of this exception including the requirement that the author of the act or omission causing the injury or damage must be present in the territory of the state of the forum. As explained in para (7), the purpose of this requirement was to ensure:
“the exclusion from the application of this article of cases of transboundary injuries or trans-frontier torts or damage, such as export of explosives, fireworks or dangerous substances which could explode or cause damage through negligence, inadvertence or accident. It is also clear that cases of shooting or firing across a boundary or of spill-over across the border of shelling as a result of an armed conflict are excluded from the areas covered by article 12. The article is primarily concerned with accidents occurring routinely within the territory of the state of the forum, which in many countries may still require specific waiver of state immunity to allow suits for recovering damages to proceed, even though compensation is sought from, and would ultimately be paid by, an insurance company.”
The ILC commentary included an acknowledgment, at para (11), that:
“Some members expressed reservations about the very broad scope of the article and on the consequences that might have for state responsibility. In their view, the protection of individual victims would effectively be secured by negotiations through diplomatic channels or by insurance.”
To date, the European Convention on State Immunity is the only international agreement concerning state immunity which has entered into force.
Another regional treaty (the Inter-American Convention on Jurisdictional Immunity of States) was approved in draft by the Inter-American Juridical Committee of the Organization of American States on 21 January 1983 (ILM, vol 22, p 292). But this has never been adopted or entered into force. The draft Inter-American Convention provides, in article 6, for an exception from immunity in “proceedings for losses and damages on tort liabilities arising from the activities mentioned in article 5, paragraph one”. The activities mentioned in that paragraph are “trade or commercial activities undertaken in the state of the forum”. The second paragraph of article 5 adds that trade or commercial activities of a state “are construed to mean the performance of a particular transaction or commercial or trading act pursuant to its ordinary trade operations”. This tort exception would accordingly apply only to commercial activities and therefore only to private (non-sovereign) acts.
The first national legislation on state immunity (and the only such legislation preceding the UK State Immunity Act 1978) was the United States Foreign Sovereign Immunity Act 1976 (the “FSIA”). Under the FSIA, a foreign state is presumptively immune from the jurisdiction of a US court unless a specified exception applies. The US Supreme Court has said that the Act largely codifies the restrictive doctrine of foreign state immunity: Republic of Argentina v Weltover Inc (1992) 504 US 607, 612; Saudi Arabia v Nelson (1993) 507 US 349, 359, 363 (“the Act’s manifest purpose [was] to codify the restrictive theory of foreign sovereign immunity”).
Section 1605(a)(5), known as the “non-commercial tort exception”, exempts from immunity actions “for personal injury or death, or damage to or loss of property, occurring in the United States and caused by the tortious act or omission” of a foreign state. The exception applies to sovereign as well as private acts. Notably, in Letelier v Republic of Chile (1980) 488 F Supp 665 a US federal district court held that Chile was not entitled to immunity from civil proceedings claiming compensation for tortious acts resulting in the assassination of a former Chilean ambassador in the territory of the United States.
288.
Although section 1605(a)(5) does not in terms state that, as well as the injury, the tortious act or omission must occur in the territory of the United States, it has been so construed, in the light of its legislative history and comparison with section 1605(a)(2). The latter “commercial activity exception” refers to “an act outside the territory of the United States” that “causes a direct effect in the United States”. The absence of similar language in section 1605(a)(5) has been taken to show that the non-commercial tort exception does not encompass acts having a “direct effect” in the United States and that the entire tort must occur in the United States to remove immunity: see Persinger v Islamic Republic of Iran (1984) 729 F 2d 835, 842; Asociacion de Reclamantes v United Mexican States (1984) 735 F 2d 1517, 1524; Argentine Republic v Amerada Hess Shipping Corp (1989) 488 US 428, 441; O’Bryan v Holy See (2009) 556 F 3d 361, 382; In re Terrorist Attacks on September 11, 2001 (2013) 714 F 3d 109, 115; Jerez v Republic of Cuba (2014) 775 F 3d 419, 424; Rusesabagina v Republic of Rwanda (2023) 652 F Supp 3d 1.
So, for example, in Persinger the parents of one of the hostages held for more than a year when the US Embassy in Tehran was seized by students in 1979 claimed damages from Iran for psychiatric injury allegedly caused by their son’s detention and ill-treatment. The US Court of Appeals for the District of Columbia Circuit held that Iran was immune from suit. Section 1605(a)(5) did not apply because, although the injury had allegedly occurred in the United States, the tortious acts had not.
This interpretation has been applied to uphold immunity in recent cases involving hacking of computers in the United States by agents of a foreign state situated abroad: see Kidane v Federal Democratic Republic of Ethiopia (2017) 851 F 3d 7, 10; Broidy Capital Management, LLC v State of Qatar (2018) 2018 WL 6074570, affirmed (2020) 982 F 3d 582; Democratic National Committee v Russian Federation (2019) 392 F Supp 3d 410. The facts of Kidane are particularly close to the facts of this case. The plaintiff alleged that the Ethiopian government had installed spyware on his computer (by sending him an infected email) which was then used to spy on him from abroad. The US Court of Appeals for the District of Columbia Circuit held that such a transnational tort is outside the scope of section 1605(a)(5).
The Canadian State Immunity Act 1985, by section 6, creates an exception from immunity for proceedings that relate to personal injury or property damage “that occurs in Canada”. In Schreiber v Federal Republic of Germany and the Attorney General of Canada [2002] 3 SCR 269, 285–288, the Canadian Supreme Court rejected an argument (made, perhaps surprisingly given the scope of the FSIA, by the United States as an intervener) that the exception applies only to private acts and not to sovereign acts. The court held that section 6 does not draw this distinction.
Like the US non-commercial tort exception, section 6 of the Canadian Act also contains no express requirement that the tortious act, and not just the injury or damage, must occur in the territory of the forum. But in Kazemi Estate v Islamic Republic of Iran [2014] 3 SCR 176 the Canadian Supreme Court held that the act(s) causing injury must occur within Canada. It followed (as in the US case of Persinger) that a claim by a person situated in Canada who allegedly suffered psychiatric injury as a result of the torture and killing of his mother in Iran by agents of the Iranian government was barred by immunity.
Other national legislation relied on by the claimants includes the UK State Immunity Act itself. But unless the claimants succeed in their argument about how section 5 should be interpreted, this does not support the existence of a tort exception which extends to acts committed by a person situated abroad. Some of the other statutes to which the claimants have referred use the same wording as section 5 of the UK Act. In this category are the Singapore State Immunity Act 1979 (section 7), the South African Foreign States Immunities Act 1981 (section 6), the Malawi Immunities and Privileges Act 1984 (section 7) and the Australian Foreign States Immunities Act 1985 (section 13). No case law or other authority has been cited which suggests that any of these statutory provisions applies to an act committed by a person who is outside the territory of the forum state.
294.
Two other statutes relied on by the claimants are statutes enacted in Japan (Act on the Civil Jurisdiction of Japan with respect to Foreign States 2009, article 10) and Spain (Organic Law on Privileges and Immunities of Foreign States 2015, article 11). The Japanese legislation provides an exception for personal injury or property damage resulting from an act for which it is claimed that a foreign state should take responsibility “when all or part of the act took place in Japan and the person who performed the act was in Japan at the time it was performed” (emphasis added). The Spanish law creates a similar exception which applies where “(a) the act or omission took place wholly or partly within the territory of Spain and (b) the party materially responsible for the act or omission was present within the territory of Spain at the time when that act or omission occurred” (emphasis added). It can be seen that the wording of these provisions closely follows that of the UN Convention, which both Japan and Spain have ratified.
Legislation governing jurisdictional immunity has also been enacted in Pakistan (1981), Argentina (1995), Israel (2008), the Russian Federation (2015) and the People’s Republic of China (2023). Until their respective laws were enacted, both Russia and China adhered to the absolute doctrine of immunity, which had consequently applied in Hong Kong after 1997: see Democratic Republic of the Congo v FG Hemisphere Associates LLC [2011] HKCFA 41.
Pakistan’s State Immunity Ordinance 1981 contains no tort exception. Argentina’s Law No 24,488 on jurisdictional immunity of foreign States (article 2) and Israel’s Foreign State Immunity Law, 5769-2008 (section 5) each require the tort to have been committed in their territory. The Russian Federal Law No 297-FZ of 3 November 2015 (article 11) follows the model of the UN Convention in including an express requirement of presence on the territory of the Russian Federation at the time of the act or omission which caused harm. The Foreign State Immunity Law of the People’s Republic of China (2023) provides (in article 9) for an exception from immunity in proceedings for “compensation arising from personal injury or death or damage to movable or immovable property caused by the relevant act of the foreign state in PRC territory”. There is no reason to think that transnational torts are regarded as falling within any of these provisions.
The decision of the International Court of Justice in Jurisdictional Immunities of the State in 2012 was an important affirmation of the right to state immunity under customary international law. Although not directly relevant here, the judgment is significant for its clear rejection of the argument that state immunity does not apply in cases involving allegations of serious human rights violations or war crimes constituting breaches of jus cogens norms.
Germany brought the case against Italy following decisions of the Italian courts that Germany did not have immunity from civil claims for compensation brought by Italian nationals who had been detained by German forces during the Second World War and deported to Germany to carry out forced labour.
Three features of the case should be noted. First, as the court observed at para 54 of the judgment, any entitlement of Germany to immunity could only be based on customary international law, rather than treaty. Although Germany was a party to the European Convention, Italy was not. The UN Convention was not in force and in any case neither Germany nor Italy had signed it. (Italy has since acceded to the UN Convention, in 2013.) The court noted, at para 64, that Germany had not “been alone in suggesting that, in so far as it was intended to apply to acta jure imperii, article 12 [of the UN Convention] was not representative of customary international law” (referring to comments made by Germany, China and the United States on the draft provision).
Second, the acts of the German armed forces and other state organs which were the subject of the proceedings in the Italian courts clearly constituted sovereign acts, as Italy conceded at the hearing (para 60). Both parties agreed that states are generally entitled to immunity in respect of sovereign acts. But whereas Germany maintained that there was no relevant limitation on this immunity, Italy—as well as making the argument based on jus cogens—contended that the immunity for sovereign acts does not extend to torts or delicts occasioning death, personal injury or damage to property committed on the territory of the forum state (para 61).
Third, the acts which were the subject of the Italian proceedings occurred both within and outside Italy. Italy did not dispute that those acts which occurred outside its territory were covered by immunity. But Italy argued that “customary international law has developed to the point where a state is no longer entitled to immunity in respect of acts occasioning death, personal injury or damage to property on the territory of the forum state, even if the act in question was performed jure imperii” (para 62). In support of this argument, Italy pointed to the adoption of article 11 of the European Convention and article 12 of the UN Convention and to the fact that nine of ten states identified as having adopted legislation specifically dealing with state immunity had enacted provisions similar to those in the two Conventions (para 62).
In response, Germany maintained that, in so far as they deny a state immunity in respect of sovereign acts, neither article 11 of the European Convention nor article 12 of the UN Convention reflects customary international law. Germany also argued that, in any event, neither provision assisted Italy, because neither provision was intended to apply to the acts of armed forces (para 63).
The court took the view that it did not need to decide whether there is a “tort exception” to state immunity in customary international law that generally applies to sovereign acts. It was enough to decide—as the court did—that customary international law continues to require immunity for acts allegedly committed on the territory of the forum state by the armed forces of a foreign state in the course of conducting an armed conflict (paras 65 and 78).
304.
In Al-Adsani v United Kingdom the applicant had brought proceedings in England against the government of Kuwait, claiming damages for injuries allegedly caused by torture in Kuwait at the hands of state agents. Kuwait was held to have immunity as the acts complained of had taken place outside the United Kingdom and the proceedings therefore did not fall within section 5 of the 1978 Act. The Grand Chamber of the European Court of Human Rights rejected the applicant’s complaint that the grant of immunity violated article 6. The court considered that section 5 of the 1978 Act “complies with the relevant provisions of the [European Convention on State Immunity]” (para 57). It had not been shown that there is yet acceptance in international law of an exception for alleged torture committed outside the forum state. The 1978 Act, which grants immunity to states in respect of personal injury claims unless the injury was caused by an act or omission within the United Kingdom, “is not inconsistent with those limitations generally accepted by the community of nations as part of the doctrine of state immunity” (para 66). It thus could not be said that upholding Kuwait’s immunity amounted to an unjustified restriction on the applicant’s right of access to a court (para 67).
305.
In McElhinney v Ireland, decided at the same time as Al-Adsani, the applicant had attempted to sue the UK government in the Irish courts claiming damages for psychiatric injury for alleged assault when a British soldier fired shots at his vehicle at a checkpoint at the border of Northern Ireland with the Irish Republic. The Irish courts held that the UK government had immunity on the ground that the actions of the soldier were within the sphere of governmental or sovereign activity and thus jure imperii. Before the European Court the Irish government submitted (see para 27 of the judgment) that:
“This interpretation of international law, upholding immunity for acta jure imperii, had been followed by the courts of a number of other European countries in recent years, for example Austria, France, Germany, Italy, Spain and Switzerland. The fact that only a limited number of countries had ratified or acceded to the [European Convention] could be taken as an indication that many States were not willing to countenance all the exceptions to the doctrine of State immunity set out therein.”
In rejecting the applicant’s contention that upholding the claim to immunity violated article 6, the court observed, at para 38, that “there appears to be a trend in international and comparative law towards limiting state immunity in respect of personal injury caused by an act or omission within the forum state”, but that “this practice is by no means universal” and “may primarily refer to ‘insurable’ personal injury, that is incidents arising out of ordinary road traffic accidents, rather than matters relating to the core area of state sovereignty such as the acts of a soldier on foreign territory”. The court concluded that, “given the present state of the development of international law”, it could not be said that Ireland, in affording immunity for torts committed by acta jure imperii, “falls outside any currently accepted international standards”.
Article 11 of the European Convention is consistent with customary international law
307.
In several judgments given since McElhinney v Ireland was decided and the UN Convention was adopted, the European Court of Human Rights has treated the 1991 ILC draft articles and the UN Convention as reflecting customary international law, either by “codifying” it or by forming a new customary rule (though the court has not said which). On this basis the court has concluded that the provisions of the UN Convention apply under customary international law, “even if the state in question has not ratified that Convention, provided it has not opposed it either”: see Cudak v Lithuania (2010) 51 EHRR 15, paras 66–67; Sabeh El Leil v France (2011) 54 EHRR 14, paras 54, 57; Oleynikov v Russia (2013) 57 EHRR 15, para 66; Radunović v Montenegro (2016) 66 EHRR 19, para 69; Renouard v France [2025] ECHR 266, para 39. As pointed out by Lord Sumption in Benkharbouche, at paras 29 and 32, this approach is problematic. While some provisions of the UN Convention could be said to have codified existing customary international law, others sought to resolve differences of opinion. As the Preamble records, the Convention was intended to contribute not just to the codification but also to the “development” of international law. It is equally difficult to say that the ILC’s 1991 draft articles and the UN Convention have formed new rules of customary international law when not enough states have yet ratified it for the Convention even to have entered into force. (It is also hard to understand how the existence of such a customary rule can depend on whether the forum state has ratified or opposed the UN Convention.)
In the field of employment law, with which Benkharbouche was concerned, Lord Sumption concluded that there was no consistency of state practice capable of founding a special rule of customary international law extending beyond the immunity attaching to sovereign acts: see paras 60–62. Article 11 of the UN Convention could not be said to have established such a customary rule. The result was that the 1978 Act “can be regarded as giving effect to customary international law only so far as it distinguishes between exercises of sovereign authority and acts of a private law character, and requires immunity to be conferred on the former but not the latter”: para 63.
In this case the relevant question is not whether there is now a special rule of customary international law extending beyond the general immunity attaching to sovereign acts but whether there is now such a rule which qualifies immunity for sovereign acts in cases involving personal injuries or damage to property as reflected in article 12 of the UN Convention.
It is clear, not least from the drafting history outlined at paras 273–283 above, that there was no consistent state practice which article 12 of the UN Convention could be said to have codified. But the fact that the members of the ILC managed to reach agreement on the formulation which became article 12 of the Convention and its adoption in a treaty which a significant number of states have now ratified has undoubtedly given impetus to a developing consensus. With the sole exception of Pakistan, all the states which have enacted legislation dealing with state immunity have included a territorial tort exception in terms which either directly reflect article 12 of the UN Convention (and article 11 of the European Convention) or are capable of being interpreted consistently with it.
According to Andrew Dickinson and Alexander Thompson, The State Immunity Act 1978 (2025), para 8.11:
“the existing state legislative practice … presently falls short of a general state practice supported by opinio juris required to create a new exception to the rule of customary international law supporting immunity for acts of a sovereign character.”
That seems a fair assessment. If correct, it means that there is not yet a customary norm which requires states to recognise a territorial tort exception equivalent to article 12 of the UN Convention. So a state, such as Pakistan, which still grants immunity for sovereign acts in cases falling within article 12 is not violating customary international law.
On the other hand, I do not think that it can any longer be said to be contrary to customary international law to recognise a territorial tort exception commensurate in scope with article 12 of the UN Convention. There is now a substantial body of state practice recognising such an exception. Thus, there is no longer a widespread, representative and consistent state practice accepted as creating a legal obligation to uphold immunity for sovereign acts causing personal injury or damage to tangible property committed by agents of a foreign state present in the territory of the forum state.
I therefore take the current state of customary international law to be that a state is permitted, but not required, to afford an exception to the general rule of immunity for sovereign acts in cases falling within the scope of article 12 of the UN Convention and article 11 of the European Convention. It follows that section 5 of the 1978 Act, provided it is interpreted consistently with article 11 of the European Convention, is also now consistent with customary international law.
Customary international law does not permit any wider tort exception
The critical question is whether a territorial tort exception to state immunity for sovereign acts which is wider in scope than article 12 of the UN Convention and article 11 of the European Convention is consistent with customary international law. Specifically, is it consistent with customary international law to deny immunity for acts causing personal injury committed in the exercise of sovereign authority by an agent of a foreign state who is not present in the territory of the forum state?
The answer to this question is clear. Far from there being a widespread, representative and consistent practice of states, accepted as a legal obligation, to treat such cases as excepted from immunity, there is a complete absence of any such practice. There is not even a single instance of it.
Refusing to uphold foreign state immunity on the facts of this case would be inconsistent with the UN Convention and the European Convention. It would be inconsistent with the position that was common ground in Jurisdictional Immunities of the State (see paras 300–301 above). It would also be inconsistent with the practice of those states which have not recognised a territorial tort exception and with the national laws of all states which have enacted legislation on the subject, as those laws have been interpreted to date. Counsel for the claimants acknowledged that they had found no decided case in any country in which jurisdiction has been exercised where the agent of the foreign state performing the act which caused injury or damage was not physically present in the territory of the forum state at the relevant time.
It follows that the United Kingdom is required by international law to uphold Bahrain’s claim to immunity in this case. To do so is therefore a proportionate means of pursuing a legitimate aim and does not infringe the claimants’ right to a fair hearing guaranteed by article 6 of the European Convention on Human Rights. The argument that the meaning of section 5 of the 1978 Act must be modified to make it compatible with the Convention rights accordingly fails.
The view of the majority
The majority of the court take a different view of customary international law. They say, at paras 79 and 93 above, that, in 1978 when the State Immunity Act was passed, there was a reasonable basis for concluding that the United Kingdom could, in conformity with customary international law, enact a territorial tort exception which went beyond article 11 of the European Convention by denying immunity to a foreign state for sovereign acts committed by a person not present in the United Kingdom.
Apart from the United States statute which I have considered at paras 288–290 above, all the materials to which the majority refer post-date the enactment of the State Immunity Act. State practice has undoubtedly developed since 1978. A number of states which in 1978 did not accept any territorial tort exception to the general rule of immunity for sovereign acts have since recognised such an exception in terms consistent with article 12 of the UN Convention (itself adopted only in 2004, a quarter of a century after the UK Act was passed). Recent examples include the Russian Federation and the People’s Republic of China (see para 296 above). It is fallacious to derive from subsequent developments in state practice a conclusion about the state of customary international law at the time when section 5 of the 1978 Act was enacted.
In any case none of the subsequent developments goes further than article 11 of the European Convention and article 12 of the UN Convention by refusing immunity for sovereign acts committed by an agent of a foreign state who is not present in the territory of the forum. To do so would be contrary to those Conventions as well as customary international law. The majority acknowledge (at para 89 above) that we have not been referred to any case in which jurisdiction has been exercised when the agent of the foreign state performing the act which caused injury or damage was not present in the territory of the forum state at the relevant time. Even today, therefore, there is no reasonable basis for concluding that the United Kingdom could enact a territorial tort exception applicable to sovereign acts committed by a foreign state agent situated outside the United Kingdom without violating customary international law.
321.
At paras 99 and 101 to 106 of their judgment the majority seek to discount the significance of the presence requirement in article 11 of the European Convention and article 12 of the UN Convention by suggesting that it does not limit the scope of the territorial tort exception but “may serve the very different purpose of defining jurisdiction as opposed to immunity” (para 99). This is not a suggestion which the claimants have made in these proceedings, in my view for good reason. It is true, as I mentioned at para 178 above, that the wording of article 11 of the European Convention was derived from article 10(4) of the Hague Convention of 1 February 1971 on the Recognition and Enforcement of Foreign Judgments in Civil and Commercial Matters. But the European Convention is not concerned with the criteria for recognising and enforcing foreign judgments. It is concerned solely with state immunity. The fact that the European Convention adopts the same territorial connection to define the scope of an exception to state immunity as the Hague Convention used for a different purpose does not somehow mean that the required territorial connection can be ignored.
322.
For good or ill, both the European Convention and the UN Convention make presence in the territory of the forum state an essential pre-condition to the exception to state immunity for acts causing personal injury or damage to property. That is where the balance has been struck in those Conventions between the principle that one state has no jurisdictional competence over sovereign acts of another state and the principle that states have sovereignty over their own territory. The balance could in principle, and may in future, be struck in a different place which gives greater scope to the principle of territorial sovereignty and less scope to the principle of state immunity by removing the presence requirement in relation to acts causing personal injury or property damage. But no state practice, let alone practice accepted by states as a matter of legal obligation, currently supports such a step. If the United Kingdom ignores the presence requirement it will therefore be acting unilaterally and in breach of its international obligations under both the European Convention and customary international law.
My reasons for dissenting from the decision on this appeal can be summarised shortly. Section 5 uses territorial language directed to where the relevant act or omission is done. That points to the location of the actor, not the place where consequences are felt. This reading is consistent with the European Convention and with the present state of customary international law. The broader interpretation advanced by the claimants would create a transboundary exception for sovereign acts contrary to customary international law and the UK’s treaty obligations. The decision to adopt that interpretation and allow these proceedings against Bahrain to continue in our courts will therefore place the United Kingdom in breach of international law. There is no valid justification for reaching that result.
LORD BURROWS (DISSENTING):
I agree with Lord Leggatt’s essential reasoning and that this appeal should be allowed. I add this judgment, focussed on the central questions on this appeal, to set out what I consider to be the main points that lead me inexorably to that conclusion. Finally, I summarise why, with respect, I disagree with the majority’s judgment.
Are the words of section 5 of the State Immunity Act 1978 reasonably capable of being given an interpretation that is consistent with article 11 of the European Convention on State Immunity so that the presumption in The Eschersheim is not rebutted?
325.
The State Immunity Act 1978 (“the SIA”) section 5 provides an exception to the general rule, laid down in section 1, that a foreign state has immunity from the jurisdiction of the courts of the United Kingdom. This is commonly referred to as the “tort exception” (although that is a marginally inaccurate shorthand because, while the cause of action covered by section 5 will almost always be a tort, it could be, for example, breach of contract or a non-tortious breach of statutory duty). Section 5 reads as follows:
“Personal injuries and damage to property
A state is not immune as respects proceedings in respect of -
A principal purpose of the SIA was to implement into domestic law the European Convention on State Immunity (the “ECSI”). That treaty was signed on 16 May 1972 but was only ratified by the UK on the day that the SIA came into force (22 November 1978). The structure of the ECSI, like the SIA, was to recognise that, in general, a state has immunity from proceedings in another state. That general immunity is provided for by article 15 of the ECSI which reads:
“A Contracting State shall be entitled to immunity from the jurisdiction of the courts of another Contracting State if the proceedings do not fall within Articles 1 to 14; the court shall decline to entertain such proceedings even if the State does not appear.”
Articles 1–14 set out the exceptions to state immunity. The “tort exception” is in article 11. This reads:
The apparent difficulty is that section 5 of the SIA refers to “caused by an act or omission in the United Kingdom” whereas article 11 of the ECSI requires the “facts” occasioning the injury or damage to have occurred in the UK and the presence in the UK “at the time when those facts occurred” of the “author of the injury or damage”.
Counsel for the respondents, Timothy Otty KC, submits that, leaving aside any possible argument that the UK was here acting within a power to derogate from the ECSI (I consider that argument later on) there is an irreconcilable clash between those two provisions and that Parliament has unambiguously and deliberately chosen not to incorporate a requirement that the defendant is present in the UK. It is the words of section 5 that must therefore be applied, so that there is no “presence in the UK” requirement, even if that means that the UK is in breach of public international law as set out in the ECSI.
First, it is incorrect to regard the words of section 5 in their context as unambiguously rejecting the requirement for presence in the UK. On the contrary, an “act or omission in the United Kingdom” might naturally mean that there must be an act or omission of an agent of the state present in the UK. Put another way, a natural meaning in context would not differentiate between the actor and the act. At the very least, there is ambiguity as to whether presence in the UK is required. Say, for example, a bullet is fired from State A and hits someone in State B. Does the relevant act (or omission) occur in State A or in State B? My inclination would be to say that the act, that is the firing of the bullet, takes place in State A and that it is the effect of the act that occurs in State B. But at the very least there is room for reasonable disagreement (ie there is ambiguity) as to whether the act can be said to occur in State B.
Secondly, and closely linked to the first point, section 5 only makes sense if there is a person for whom the (foreign) state is responsible (for shorthand, an agent of the state) whose act or omission has caused the injury or damage. Without that implicit meaning, the section could not work. Necessarily, therefore, one has to interpret section 5 as if it included the words “of a person for whom that State is responsible”. If one reads those words into section 5, the ambiguity of section 5 is made doubly obvious. The last line of section 5 would then read “caused by an act or omission of a person for whom that State is responsible in the UK”. In that full formulation, the natural interpretation in context would be that the person is in the UK at the time of the act or omission. At the very least, there is plainly an ambiguity as to whether that person must be in the UK.
333.
Thirdly, as a matter of statutory interpretation, there is a presumption that, if a statute has been enacted to implement a treaty, words in the statute should be given the same meaning as those in the treaty in so far as they are “reasonably capable of bearing that meaning”. This was accepted by the House of Lords in The Eschersheim [1976] 1 WLR 430. Lord Diplock said at p 436:
“As the Act was passed to enable Her Majesty’s Government to give effect to the obligations in international law which it would assume on ratifying the Convention to which it was a signatory, the rule of statutory construction laid down in Salomon v Customs and Excise Commissioners [1967] 2 QB 116 and Post Office v Estuary Radio Ltd [1968] 2 QB 740 is applicable. If there be any difference between the language of the statutory provision and that of the corresponding provision of the Convention, the statutory language should be construed in the same sense as that of the Convention if the words of the statute are reasonably capable of bearing that meaning.”
Here it is not in dispute that a principal purpose of the SIA was to implement the ECSI. The ambiguity as to whether presence in the UK is required can be straightforwardly resolved by interpreting section 5 so as to be consistent with article 11. The words “caused by an act or omission in the United Kingdom” in section 5 are reasonably capable of being interpreted as meaning that the act or omission must be of an agent of the state present in the UK. Therefore, the presumption in The Eschersheim is not rebutted.
Fourthly, it seems clear that section 5 used different wording from article 11 because, in implementing the ECSI in the UK, it would have been inappropriate directly to use the wording of article 11. This is particularly because of its reference to “facts” occasioning the injury or damage and those “facts” occurring in the forum but also because of the reference to the “author” of the injury or damage. Neither “facts” used in this way nor the “author” are familiar terms in describing the domestic law of tort (or delict). Indeed, “facts occasioning the injury or damage” makes no sense in domestic tort law. A simple “copy out” was, therefore, not here open to the draftsperson of the domestic legislation and it seems clear that section 5 used different wording so as to implement article 11 in language familiar in our law of tort.
Fifthly, although it can be argued that it would have been a simple drafting task to have expressly included a requirement of presence in the UK, it can also be argued that, if there were to be no such requirement, it would have been equally simple to have included an express provision making clear that presence in the UK was not required.
336.
Sixthly, it has been suggested, by reference to section 16(2) of the SIA that, where presence is required in the SIA, this has been spelt out in the legislation. This is therefore a contextual interpretation point. Section 16(2) reads as follows:
“This Part of this Act does not apply to proceedings relating to anything done by or in relation to the armed forces of a State while present in the United Kingdom and, in particular, has effect subject to the Visiting Forces Act 1952.”
But, if anything, this reference to section 16(2) is a point in favour of the appellant. Section 16(2) is carving out an exception to the exception from State immunity recognised by, most obviously, section 5. However, in section 16(2) it has not been thought necessary to deal with anything done by the armed forces of the foreign state who are not present in the UK. Why is that? One obvious explanation is the recognition by Parliament that section 5 is dealing only with where the agents of the foreign state are present in the UK so that section 16(2) only needs to deal with a carve-out for armed forces of a foreign state who are present in the UK.
337.
Seventhly, there is a temporal element to consider. In 1978, drones and the internet did not exist and there was no widespread use of computers. In the most obvious situations with which section 5 was dealing (ie car accidents involving state officials), the agent of the state would necessarily have been present in the UK. It is a rational possibility that in 1978 it was not thought necessary to specify presence in the UK because, at that time, it was not easy to think of realistic situations of acts or omissions in the UK causing personal injury or property damage where the agent of the state would not be present in the UK. (Indeed, although on a conventional approach the relevant external material, contained in a briefing note to Ministers, appears to be inadmissible—see R (Public and Commercial Services Union) v Minister for the Civil Service [2010] EWHC 1027 (Admin); [2010] ICR 1198, at para 55—that was the precise approach revealed by the briefing note, held in the House of Lords’ library, that was drawn to our attention by Tom Hickman KC, counsel for the appellant). There is also the linked point that the draftsperson would be aware of the interpretative presumption that the words used should, if possible, be interpreted so as to be consistent with the UK’s obligations under public international law. In other words, the draftsperson would know that, unless made clear to the contrary, the correct interpretation would be one that adhered to article 11 of the ECSI.
Eighthly, if Mr Otty were correct that section 5 of the SIA marked a deliberate departure from article 11 of the ECSI which would therefore put the UK in breach of international law (leaving aside any possible argument that, in respect of section 5, the UK was acting within the power to derogate from the ECSI), it would be very surprising that there was no hint of this in the Parliamentary debates during the passage of the SIA. In contrast there were clear references in those debates to the departures from the ECSI in respect of non-sovereign acts (eg in respect of contracts) where it is not in dispute that the UK was free to derogate from the ECSI.
339.
Ninthly, I accept that, if one were interpreting the SIA, divorced from its public international law context and without any reference to the purpose being to implement the ECSI, the correct interpretation might be different. That is, there would be an argument that, applying the “always speaking” approach and taking account of technological developments, one would seek to avoid an interpretation that would distinguish between computer surveillance of a person in the UK by a foreign state where the surveillance is being carried out by an agent of the state in the UK and where that surveillance is being carried out without such a human presence in the UK. But, given the public international law dimension, that is not the exercise of interpretation with which we are here concerned. It is impermissible for this court, and would constitute an undermining of the rule of law, to adopt a statutory interpretation that contradicts public international law unless Parliament has enacted legislation that clearly requires that. It is beside the point that the presence requirement in article 11 of the ECSI may produce anomalies. It is not for this court to challenge the validity (whether in terms of policy or principle) of the presence requirement laid down as a matter of public international law in article 11 of the ECSI.
340.
Finally, although with the advance of technology, the requirement of presence in the UK may produce some anomalies, it may avoid other anomalies. In particular, without presence, the act or omission in the UK referred to in section 5 could comprise a minor infringement of the UK’s territorial sovereignty. That is not least because the personal injury (or property damage) need not be suffered in the UK. Let us assume, therefore, that on hypothetical facts, a claimant’s laptop computer is hacked into from abroad by a foreign state at a time when the laptop is being used by the claimant while transiting at an airport in the UK and that the claimant, who lives in France, is only aware of that hacking, causing her psychiatric illness, while in France. Without the presence of an agent in the UK, one can strongly argue that what has happened in the UK is too minor to fall within the tort exception to state immunity. Yet, applying the approach advocated by Mr Otty, in an action brought in the UK, the foreign state would not have immunity applying section 5.
For all these reasons, I conclude that, leaving aside any argument that the UK was here acting within a power to derogate from the ECSI because of its declaration under article 24(1), section 5 of the SIA is correctly interpreted as being consistent with article 11 of the ECSI so that the causative act or omission must be of an agent of the foreign state who is present in the UK.
Mr Otty, for the respondents, submitted that, in any event, the UK was free to derogate from article 11 of the ECSI and had done so, in section 5 of the SIA, by not including a presence in the UK requirement. The UK was therefore not in breach of public international law under the ECSI by rejecting the requirement of presence in the UK and the presumption in The Eschersheim was not in play.
343.
Article 24(1) of the ECSI permits a state to derogate, by expanding the exceptions to state immunity, from what is set out in the ECSI. Article 24(1) reads as follows:
The UK made a declaration under article 24(1). This means that, without being in breach of its international legal obligations under the ECSI, the UK was entitled to derogate from the provisions laid down in that Convention, by expanding the exceptions to state immunity, provided the derogation: (i) fell within the exceptions to state immunity that apply in respect of proceedings against states who are not signatories to the ECSI and therefore apply outside the ECSI ie that apply by reason of customary international law recognised in the UK; and (ii) did not infringe the last sentence of article 24(1).
Mr Otty submitted that section 5 of the SIA constituted a permitted derogation from the tort exception in article 11. It expanded the scope of that tort exception by the deliberate exclusion of the need for the author of the injury or damage to be present in the UK at the time of the relevant act or omission.
Mr Hickman argued that Mr Otty’s derogation submission should be rejected. That is because of the limits to the derogation permitted under article 24(1). He put forward a wide and a narrow submission (although, in reply, he appeared to backtrack from his wide submission).
Mr Hickman’s wide submission was that the last sentence of article 24(1) means that there can be no derogation in respect of acta jure imperii ie sovereign acts as opposed to private acts. In other words, there can be an expansion of the exceptions to state immunity in respect of private acts (and that has been done, for example, in respect of contracts by reason of section 3 of the SIA) but there can be no expansion of the exceptions to state immunity in respect of sovereign acts. It is not in dispute that in this case the acts or omissions in question were sovereign, and not private, acts of the Kingdom of Bahrain. According to Mr Hickman’s wide submission, there was therefore no power to derogate from article 11 in respect of such sovereign acts.
Mr Hickman’s narrow submission was that, even if his wide submission were to be rejected, derogation is permitted in respect of sovereign acts only if customary international law gives statutory immunity in respect of such sovereign acts and hence applies to non-ECSI states. However, in the context of tort, customary international law did not in 1978 (and does not today) give state immunity to sovereign acts without the author of the acts or omissions being present within the forum. At the very least, the burden falls on the respondents to establish that there was (or is) a wider customary international law exception and they have failed to discharge that burden.
Mr Otty argued that the last sentence of article 24(1) is not imposing any restriction on the scope of the derogation that is permitted in respect of sovereign acts. Rather the only restriction in article 24(1) is that the exception to state immunity must be recognised by customary international law and therefore applies to non-ECSI states. He argued that there is a customary international law tort exception that does not require the author of the act or omission to be present in the jurisdiction.
350.
Assuming that Mr Hickman’s wide submission is incorrect, his narrow submission and Mr Otty’s submission both require, as the correct interpretation of the derogation permitted under article 24(1), that customary international law recognises a tort exception that does not rest on the author of the act or omission being present in the forum state. The fundamental flaw in Mr Otty’s argument, and why Mr Hickman’s narrow submission is correct, is that that was not the position under customary international law in 1978 (and nor is it the position today). Mr Otty was unable to point to any case anywhere, whether in or outside the UK, that shows that customary international law has accepted a tort exception not based on presence in the forum. At the very least, the position in customary international law is unclear. Moreover, I agree with Mr Hickman that the burden of proof on this issue clearly lies with the respondents not only because they are the claimants but also because it is they who are seeking to invoke the derogation.
I therefore conclude that, having failed to discharge the burden of proof in respect of there being customary international law to the effect that there was in 1978 a tort exception not requiring presence in the forum, the respondents’ submission on derogation fails. The UK was not free to derogate from article 11 of the ECSI by rejecting a requirement of presence in the UK. The UK would therefore have been in breach of public international law if section 5 of the SIA did not require the presence of the agent of the foreign state in the UK and the presumption in The Eschersheim was applicable.
I have nothing to add to the reasoning of Lord Leggatt rejecting the alternative submission of Mr Otty that the interpretation of section 5 of the SIA that I consider to be correct is incompatible with article 6 of the European Convention on Human Rights.
It may help to clarify matters if I briefly articulate why, respectfully, I disagree with the majority’s judgment.
First, the majority takes the view that section 5 of the SIA is unambiguous. It has only one possible meaning, namely that the presence of the agent of the foreign state in the UK is not required. Section 5 was, therefore, a deliberate departure by Parliament from article 11 of the ECSI, whether or not that amounted to a breach of international law; and, even if the interpretative presumption in The Eschersheim is in play, it has been rebutted.
I have given my detailed reasons in paras 325–341 above why I cannot accept that view. The presumption in The Eschersheim has not been rebutted. Section 5 is reasonably capable of being given an interpretation that complies with article 11 of the ECSI thereby avoiding the UK being in breach of public international law.
356.
Secondly, and with great respect, the discussion of customary international law by the majority does not meet my concerns. It is not in dispute that, as is recognised in article 11 of the ECSI, there is what the majority refer to as a “territorial tort exception” to state immunity. Under article 11, that exception applies to both sovereign and non-sovereign acts. But the important point is that what is accepted as the extent of that territorial tort exception in public international law, including customary international law, is limited to where the actor is present in the forum state. In other words, to be relevant, the analysis of the majority needs to show that customary international law in 1978 (or perhaps now) recognised a territorial tort exception not based on the presence of the actor in the forum state. Yet there is nothing to show that customary international law recognised in 1978 (or recognises today) a territorial tort exception that goes beyond article 11 by not requiring the presence of the actor in the forum state. And without being able to show that customary international law has a territorial tort exception that does not require presence in the forum state, the majority has failed to explain how a departure from the presence requirement falls within the derogation permitted by article 24(1) of the ECSI. The correct analysis, therefore, is that, if section 5 of the SIA does not require presence in the UK, the UK is in breach of international law (or, at the very least, cannot be shown by the respondents, who have the burden of proof, not to be in breach of customary international law). I reiterate that, as acknowledged in the majority’s judgment at para 90, we were not directed to any case anywhere showing that customary international law has accepted a tort exception not based on presence in the forum.
For the avoidance of doubt (and in the light of what is said in the majority’s judgment at, for example, para 99) it is clear that there is no state immunity under section 5 of the SIA (or article 11 of the ECSI) if agents of the foreign state come to the UK and, while present here, carry out an assassination. That example (and, at the hearing, reference was made to the murder carried out on UK soil by the poisoning of Alexander Litvinenko) falls squarely within the presence in the UK requirement.
Contrary to the majority’s judgment, it is therefore my view that, in so far as the presence in the UK requirement is regarded as unprincipled or contrary to good policy, this court can only properly depart from that requirement, to which the UK is committed by reason of article 11 of the ECSI, if Parliament reforms section 5 of the SIA by making clear that the relevant act or omission need not be of an agent of the foreign state who is present in the UK.
