AAA & Ors v Unilever Plc & Anor

This is my decision on two applications by the Defendants. Those applications concern a claim brought by the Claimants (‘Cs’) who are Kenyan nationals, against the First Defendant (‘D1’), a holding company registered in the United Kingdom, and against the Second Defendant (‘D2’). D1 is D2’s ultimate holding company. D2 is a company registered in Kenya. It operates a tea plantation in Kenya (‘the Plantation’). Cs were employed, or lived there. Cs were the victims of ethnic violence carried out by armed criminals on the Plantation after the Presidential election in Kenya in 2007. They claim that the risk of such violence was foreseeable by D1 and by D2, that D1 and D2 owed a duty of care to protect them from the risks of such violence, and that D1 and D2 breached that duty.

By an application dated 1 March 2016 D1 asked for Cs’ claims against it to be stayed or dismissed. D1 argues that those claims are not justiciable in England and Wales, because they require the court to adjudicate on foreign acts of state (‘FAS’). D1 also applies for a stay of the proceedings on the grounds, first, that the Cs should bring their claims in Kenya and second, that they should not be permitted to proceed with the claims in England and Wales when it is clear that the claims have only been brought for the improper purpose of attracting D2 into this jurisdiction.

By a Part 11 application dated 26 February 2016 D2 asked for

The applications give rise to several issues. The first issue is whether either claim is barred by FAS. I will consider this issue first.

The second and third group of issues concern D1’s application for a stay on other grounds, and the issues which arise under paragraph 3.1 of Practice Direction 6B. Both were considered recently by Coulson J in Lungowe v Vedanta Resources Plc [2016] EWHC 975 (TCC), and more recently still by Fraser J in Okpabi v Royal Dutch Shell [2017] EWHC 89 (TCC). The general outline of those cases was similar, in that Zambian (or as the case might be, Nigerian) claimants who were affected by the activities of a subsidiary company in Zambia/Nigeria sought to bring a claim in England and Wales both against the parent company, which was registered here, and against the Zambian/Nigerian subsidiary.

I am not persuaded that the general approach of either Judge to the second group of issues was wrong, and I will adopt it gratefully, to the extent that those cases were factually similar to this. The third group of issues was carefully considered by both Judges, with opposite results. I was told that permission to appeal against both decisions has been granted.

The third group of issues arises under paragraph 3.1 of Practice Direction 6B. This allows a claimant to serve a claim form out of the jurisdiction with the court’s permission where a claim is made against a defendant on whom a claim form has been served and there is between the claimant and the defendant a real issue which it is reasonable for the court to try, and the claimant wishes to serve the claim form on another person who is a necessary and proper party to that claim. Even if those tests are met, the court has a discretion, conferred by CPR 6.37(3), to refuse permission to the claimant to serve the claim form out of the jurisdiction if it is not satisfied that England and Wales is the proper place in which to bring the claim.

In Vedanta, at paragraph 97, Coulson J listed, in his own order, the issues which, the parties in that case agreed, had to be answered to deal with the second defendant’s application in that case (an application which was similar to the application by D2 in this case). Coulson J’s order for dealing with the issues is logical. It requires an initial focus on the merits of each claim. I will answer the disputed questions raised by D2’s application in the same order.

Leaving aside foreign act of state (‘FAS’), the key to the applications seems to me to be whether the Cs’ claims against D1 and D2 have arguable merit; both in substance, and with regard to limitation issues. In Vedanta, Coulson J held that the necessary elements of the claims had been properly pleaded and that he could not say that they lacked the relevant merit. Fraser J, on the other hand, held in Okpabi that it was not arguable that the parent company had a duty of care towards the claimants.

Cs were represented by Mr Hermer QC, Mr Weir QC and Mr Vassall-Adams QC. D1 and D2 were represented by Mr Gibson QC, Mr Heppinstall, Mr Miletic and Dr Webb. I thank all counsel for their comprehensive written submissions. I also thank them and their instructing solicitors for the meticulous preparation of the materials for this hearing. There were eight files of domestic authorities, 16 files of pleadings, evidence and expert reports, and eight files of materials referred to in the experts’ reports. The case was tenaciously argued by three advocates throughout the four-day hearing. I am grateful to all of them for their courteous and effective oral presentation of the many issues.

I am not in a position to, and do not, make any factual findings about the disputed issues in this case. The background is that the 218 Cs were, in 2007, temporary or permanent employees at, or people living in, a 13,000-hectare tea plantation owned and operated by D2 in the southern Rift Valley in Kenya. The Plantation is near Kericho town, which was then in Rift Valley Province (‘RVP’). RVP had an area of 70,000 square miles and a population of 10m people.

Cs are members of the Kisii ethnic group. The Presidential election was on 27 December 2007. After the results were announced, there was a nationwide breakdown in law and order. Criminal rioters invaded the Plantation in bands of 100 to 300 people. They were members of the Kalenjin and Luo ethnic group. They committed murders, rapes and other violent assaults, and damaged property. Cs claim they were victims of that violence and that they lost property in it. Seven Cs were killed, women were raped and many people suffered personal injuries ranging from the severe to the transient. Across Kenya, 1,333 people were killed, ‘very many more injured, thousands of public and private properties destroyed, vehicles and equipment ruined and private possessions stolen’ (page 383 of the Report of the Waki Commission – ‘the Report’).

D2 is domiciled in Kenya. It was incorporated in 1925. Until 2009 D2 was a Kenyan public company listed on the Nairobi stock exchange. In 2009 it was de-listed and now 97.65% of its shares are owned by Brooke Bond Limited, a wholly owned subsidiary of D1. Its most recent accounts (2015) show that it has £81.2m of net assets.

D1 is D2’s ultimate holding company and is registered in the United Kingdom. At the material time, D1 owned all the shares in Brooke Bond Group Limited. Brooke Bond Group Limited owned 88.2% of the shares in D2. D1 is not a trading company. It is a multi-national holding company. D1 and Unilever NV, a company domiciled in the Netherlands, own the shares in the subsidiary companies in the Unilever Group. There were about 107 subsidiary companies in 2007 (the correct figure is not clear: Mr Gibson told me in his Reply that the figure was 708). The Group operated in about 150 countries of the world in 2007. There were 174,000 employees in the subsidiaries. D2’s employees therefore represented a more significant proportion of the total employees of the Group than D2’s turnover did of the Group’s total turnover. D2’s turnover was about 2% of that of the Group.

The soil and climate in the area of the Plantation are ideal for growing tea. The tea bushes are productive all the year round. The Christmas period is a busy period for tea picking. In 2007 there were about 20,000 employees on the Plantation. The Plantation has no physical borders. It is made up of eight estates, which are not all contiguous. The overall length of the Plantation is some 75km. The boundaries of the various estates are about 80km long. In 2007 the Plantation was guarded by some 900 unarmed patrolling watchmen. There are also some police posts on the Plantation, on land leased to the authorities by D2. There is a dispute about whether despite its lack of physical boundaries, there is a limited number of obvious entry points to the Plantation such as roads.

Part of the Plantation borders on the town of Kericho. The Plantation includes areas where tea is grown, tea factories and accommodation for employees and their families, primary schools, social facilities, clinics and a hospital. There are several thousand homes there. Some 20-24,000 employees live there with their dependants; perhaps as many as 100,000 people in total. The accommodation is provided pursuant to a collective bargaining agreement, presumably because it is understood that many of those people who come to work at the Plantation come from elsewhere with their families. The Plantation’s workforce is diverse. Before 2007 there was no significant inter-ethnic violence on the Plantation.

One of the Cs’ witnesses is a distinguished Kenyan lawyer, Mr Kegoro. He was secretary to the Waki Commission (‘the Commission’). The Commission investigated the post-election violence (‘the PEV’) in 2007. The Report deals with the PEV. The Commission was unable to hold hearings in the Kericho area, but was helped by a report from the local Law Society, supported by sworn depositions from 42 witnesses (see page 155 of the Report). Much of that material seems to have been concerned with allegations of brutality by the police. Mr Kegoro is clearly well informed about the history of ethnic violence in Kenya.

He describes the different ethnic groups in paragraphs 27-8 of his witness statement. There are 40 different ethnic groups and three broad linguistic groups. The Kipsigis are a subgroup of the Kalenjin. They live in the area around Kericho. They are the majority in Kericho. They see Kericho as their homeland. They resent the presence there of members of other tribes such as the Kisii, Kikuyu, and Luo who have migrated there to work, and who they see as outsiders. Their migration into the area has led to ethnic tensions. According to some sources, the enmity between the Kipsigis and Kisii is historic and existed before the colonial state (see the Report, page 140).

The Kisii are from a neighbouring area. They come from Kisii County in what was then Nyanza Province. Kisii Town is about a two-and-a-half-hour bus ride from the Plantation. Kisii Town is 88 km from Kericho Town. The homeland of the Kisii and of the Kipsigis shares a long boundary. The Kisii have migrated to the Kericho area to work as tea pluckers. They are poor and keep homes in Kisii County. The Luo and Luhya also work as tea pluckers. Between 30-50% of the tea pluckers are Kisii. The Kalenjin are a minority among the employees on the Plantation; about 2%. They tend to be managers.

Mr Kegoro’s witness statement shows that up until 2007, with one exception, there had always been substantial violence at Kenyan elections and much is based on tribal rivalries. The 2002 election was an exception. Both main candidates were Kikuyu. There has been substantial violence between the Kalenjin and their subgroup the Kipisigis and the Kisii at every other Kenyan election. Whether there is political violence in a given election year depends on a complex interaction between ethnic and political affiliations. The picture is not always the same. There was also a referendum on a new constitution in 2005. According to Mr Kegoro, the likely ethnic alignments at the 2007 election became clear during 2005. There was ethnic violence in Kuresoi (an hour’s car journey from Kericho) in late November 2005 between Kalenjin on one side and Kikuyu and Kisii on the other. That violence continued in 2006 and 2007.

In the 2007 election the Kalenjin mostly supported Raili Odinga of the Orange Democratic Party (‘the ODM’), as did the Luo. The Kikuyu and the Kisii mostly supported Mr Odinga’s rival, Mwai Kibaki, of the Party of National Unity. Mr Kibaki is a Kikuyu. The election campaign was very close. The result could not accurately have been predicted. The campaign was accompanied by ethnic violence, including murders, and by hate speech from politicians which was designed to create and foster resentment and suspicion between ethnic groups, based on allegations about issues such as vote rigging. Threats were made between members of different ethnic groups about what would happen if the candidate supported by a rival group won.

In the event, Mr Kibaki won. The announcement of the results of the election was delayed. They were announced at 5.45pm on Sunday 30 December 2007. Bands of 100-300 young men armed with traditional weapons began to invade the Plantation that evening. Their incursions continued over the following days. 2000 homes were destroyed, 3000 were looted and many were set on fire. The local violence was not confined to the Plantation. There was mayhem in Kericho town (see page 143 of the Commission’s report). The police opened fire and killed six people; and were later criticised by the Commission for this. But most members of all tribes were peaceful and law-abiding throughout. Some Luo and Kalenjin sheltered their Kisii neighbours.

The Ds say that D2’s crisis management plan was activated. A crisis management team took charge. It promptly asked for help from the police and from the District Commissioner. The District Commissioner encouraged those who feared violence to make for ‘safe zones’ near police posts. D2’s employees were told this, and some followed that advice. Indeed, some were attacked at or near police posts and four people were killed at such posts. The criminals put up road blocks which made evacuation impossible. The Kenyan authorities advised against evacuation until they considered it was safe, and then provided escorts. D2 followed that advice. The Commission found that the police nationally failed to protect people from being killed and seriously injured and to prevent damage to property.

Paragraph 8 of the GPC alleges that Ds ‘failed to take any or any adequate steps to protect [Cs] from the foreseeable risk of attack. In particular, [Ds] failed to identify the specific risk of post-election violence, failed to put in place an adequate crisis management and/or security plan to avert or manage that risk and failed to take any, or any adequate measures, in response to [the PEV] to protect [Cs]’.

Section IV of the GPC is headed ‘Foreseeability of violence at the Plantation’. It is alleged that there has been violence at every election since 1992 and that Ds knew, or should have known, that there was a pattern of ethnic violence in Kenya which intensified at elections, that such violence was at each election particularly intense in RVP, that it involved Kalenjins attacking ‘foreign’ ethnic groups with traditional weapons, there were threats of such violence at the Plantation and violence in the surrounding areas in 1992 and 1997, and that there was a particularly high risk of such violence at the Plantation during and around the 2007 election.

The Cs rely on several factors in support of that assertion. I note that one, the suggestion (made more than once) that the largest ethnic group among D2’s employees on the Plantation was the Kalenjin, seems to be factually incorrect.

Section V of the GPC recites the events on which the claims are based. In short, unrest was reported to Ds from around 11 am on 29 December 2007. Later that day, as it seemed that the ODM’s lead had been dramatically cut, young men began barricading roads in Kericho and the surrounding area. The result was declared on 30 December at 5.39 pm. Many residents from outside the Plantation had been allowed onto the Plantation to watch the results. Cs who were present could see that the Luo and Kalenjin were upset. Cs were told, ‘You will regret this’. Shortly after the announcement, large groups of mostly Kalenjin men, some employees on the Plantation, who were recognised by Cs, began to break into, burn and loot the homes of Kisii and Kikuyu employees on the Plantation and to attack Kikuyu and Kisii they came across. The way they were armed and dressed suggested planning. I record that the defendants deny that any of the attackers were employees of the Second Defendant.

On 30 and 31 December Cs tried to contact managers on their mobile phones. Some did not answer. Those who did told Cs to hide in tea bushes. Houses owned by Kikuyu and Kisii were marked with an ‘x’ and looted and burned. People who did not escape their attackers were violently assaulted. Many women were gang-raped. Attackers pursued Cs with dogs. They were heard shouting that this was not the Kisiis’ land and that they should be killed or should leave. On 31 December D2’s lorries were used on the Chebown Estate to transport attackers onto the Estate. The attacks continued. Some Cs spent seven nights hiding in the tea bushes. Many Cs saw friends and family or other Kikuyu or Kisii being killed, raped, burned, or violently attacked with traditional weapons. All Cs feared for their lives.

The Kaptien Estate was perhaps the worst affected. 162 of the Cs lived or worked there. Some of them were at a meeting on 31 December. Some were visibly hurt. They asked for more security, and for vehicles to take them away. Managers said they could not provide that and suggested instead that Cs hide in the tea bushes. After the meeting many went to police post ‘Kaptien Number 5’. There were only three policemen there. Many Cs and the police were warned by phone that ‘…we are coming for you’. The post was attacked. Three men who had hidden in a water tank were burnt to death, witnessed, in the case of two, by their wives. Two other men were hacked to death with machetes. The attack stopped when police returned with tear gas and dispersed the attackers.

Cs plead that D1 owed a common law duty of care to them to take ‘all reasonable steps to ensure that [they] did not suffer personal injury, damage to property or economic loss due to the foreseeable risk of ethnic violence on the Plantation particularly during election periods’ (GPC, paragraph 83). This duty is said to arise from D1’s assumption of responsibility ‘for the health safety and security of employees of [D2]’. This is said to be shown by the high degree of direction and supervision by D1 of D2’s formulation and auditing of risk assessment, health and safety, and crisis management policies.

It is said to be fair just and reasonable for such a duty to be imposed because D1 knew or ought reasonably to have known of the significant risk of ethnic and political violence on the Plantation and in the Kericho district, D1 had, or ought reasonably to have had, superior expertise, knowledge, and resources in the relevant fields, D1 knew or ought reasonably to have known that D2’s plans to manage the relevant risks were inadequate, and D1 knew or ought reasonably to have foreseen that D2 would rely on D1’s superior resources. Cs rely on various corporate publications, before and after 2007, to show the ways in which D1 has assumed such responsibility.

Cs allege that D1 breached this duty in various different ways (GPC, paragraph 119). The essential allegation is that D1 ‘did not take adequate steps to enable it to foresee and respond to the risk of violence which Cs faced’, in particular by failing to ensure that D2’s relevant policies identified, and addressed, the risks of ethnic violence on the Plantation, and by failing to ensure that D2 effectively implemented such policies. Cs also allege that D1 failed to ensure that D2 ‘took adequate steps to respond to the foreseeable risk of violence, including enhanced security measures and/or arranging for vulnerable employees to return to their homelands and/or transporting vulnerable workers to safer areas’. The breach of duty is summarised as ‘failing in all the circumstances to take any or any adequate steps to prevent and/or stop the mistreatment of [the Cs]’.

Cs allege that D2 owed them (as employees or as residents on the Plantation) ‘a duty to take reasonable care for their health and safety given those relationships and its knowledge of the high risk of violence and attacks on Kisii residents on the Plantation in the period immediately before and after the election’.

Cs claim that D2 breached this duty by, among other things, failing to ‘identify the Plantation as a high risk area for ethnic and political violence… failing to realise that the 2007 election was a period of great risk, failing to realise that Cs were particularly vulnerable, failing to adopt and implement adequate policies, failing to train managers adequately, failing to take adequate steps to respond to the risk…including enhanced security measures including the organisation of additional police and or armed police officers and/or private security…failing, over the election period, to arrange for vulnerable employees to return to their homelands or to safe areas, failing to help Cs to get to safe areas in or outside the Plantation in time, failing to ensure that managers were not on leave, failing to ensure that managers could communicate effectively with each other, failing to prevent employees from being violent on the Plantation, failing to protect Cs from the risk of violence by people who lived outside the Plantation and failing in all the circumstances to take any or any adequate steps to stop the mistreatment of [Cs] during the 2007/2008 attacks’.

Should the claim against D1 be stayed on the grounds that it is barred by FAS?

The parties provided me with very detailed written and oral arguments on this issue. They also provided me with further written submissions on the decision of the Supreme Court in Belhaj v Straw [2017] UKSC 3; [2017] 2 WLR 456, which was handed down after the hearing in this case. I hope I do them no discourtesy by summarising those arguments briefly. This decision supersedes many of the parties’ initial submissions. I do not consider it necessary to set out the arguments in great detail, as in my judgment the correct approach to this part of the case is clear. I will start by summarising what, in my judgment, the majority of the Supreme Court decided about FAS in Belhaj.

Seven Supreme Court Justices sat on the appeal in Belhaj. Lords Wilson and Clarke and Baroness Hale agreed with the judgment of Lord Neuberger. Lords Mance and Sumption delivered two separate judgments. Lord Hughes agreed with the judgment of Lord Sumption. I consider that I need to do no more, for the purposes of this judgment, than to follow the reasons of the majority of the Court.

The allegations are summarised in paragraphs 3-10 of Lord Mance’s judgment. The appeals concerned the alleged complicity of the United Kingdom authorities in torts allegedly committed by various other states overseas. The defences included state immunity and appeals to FAS. Neither claim was brought against a foreign government or official. Lord Neuberger agreed with the approaches of Lords Mance and Sumption to the question of state immunity.

Lord Neuberger summarised the FAS in paragraph 118 of the judgment: ‘…the courts of the United Kingdom will not readily adjudicate upon the lawfulness or validity of sovereign acts of foreign states and it applies to claims which, while not made against the foreign state concerned, involve an allegation that the foreign state acted unlawfully…it is a principle whose precise scope is not always easy to identify’.

He considered that the cases suggested that there might be ‘four possible rules’ which have been treated as aspects of FAS.

There was a strong argument that the first rule was perhaps not part of FAS at all, or was a free-standing aspect of FAS, ‘effectively franked by international law’. That first rule was that the courts of England and Wales will recognise, and not question, the effect of the legislation of a foreign state or other laws which take effect in the territory of that state. The first rule was well established and supported by the cases, at least in relation to property (judgment, paragraph 125). The first rule was good law. It only applied to acts in the territory of the foreign state (judgment, paragraph 135).

The second rule is that the courts of England and Wales will recognise, and will not question, the effect of an act of a foreign state’s executive which takes place or takes effect in the territory of that state. That rule also had significant judicial support, but again, only in relation to property (judgment, paragraph 127). The rule is valid as it relates to an act of a state confiscating or transferring property, or controlling or transferring rights over property. He expressed some doubt about whether the rule applied to acts which were unlawful by the law of the state concerned but did not need to resolve that doubt (paragraphs 138-143)

The third rule has more than one aspect. Each gives rise to issues which it is inappropriate for the courts of England and Wales to resolve, because they involve a challenge to the lawfulness of the act of a foreign state which is of a kind that a municipal judge ‘cannot or ought not to rule on’. The courts of England and Wales will not interpret or question dealings between foreign states. Examples are making war and peace, making treaties, and annexations and cessions of territories. The courts of England and Wales will not ‘as a matter of judicial policy, determine the legality of acts of a foreign government in the conduct of foreign affairs’. Lord Neuberger said that a further facet of this rule was that international treaties and conventions, which have not been incorporated in domestic law by legislation, cannot be a source of domestic rights and duties and will not be interpreted by the courts of England and Wales (judgment, paragraph 123).

There was no doubt that the third rule applied to property and property rights. Judges should not be enthusiastic in applying it, but, equally, should not decide issues which it is not appropriate for them to decide (judgment, paragraph 144). The third rule is based on judicial restraint. It is not limited to acts in the territory of the foreign state. It is in part concerned with relationships between states. It is distinct from the first and second rules, which are almost always concerned with the acts of a single state, normally in its own territory (judgment, paragraph 146). It will nonetheless sometimes be difficult to tell whether a claim is subject to the second or to the third rule. The involvement of the executives of more than one state does not necessarily entail the application of the third rule to the exclusion of the second (judgment, practice 147).

The third rule had been applied in a number of cases, again in relation to property (judgment, paragraph 128), but also in relation to injury to the person (judgment, paragraph 130).

Lord Neuberger described a fourth possible rule in paragraph 124 of his judgment. This is that the courts would not investigate the acts of a foreign state where such an investigation would embarrass the United Kingdom Government, but only if there has been a communication to that effect from the Foreign and Commonwealth Office (‘the FCO’). He doubted whether that rule existed but if it did, it could only be invoked in exceptional circumstances (judgment, paragraph 132). He accepted (judgment, paragraph 148) that on some questions (such as the recognition of a foreign state and whether there is a state of war) the view of the executive would be conclusive for the courts. The defendants had not argued any point in reliance on the fourth rule, and the Court did not need to decide whether it existed.

In paragraphs 158-162 of his judgment, he considered whether the first and second rules applied to injury to the person. He concluded that there was a ‘very powerful argument’ that the first rule did apply to injuries to the person, but he was ‘unconvinced’ that the second rule, in so far as it applied to acts in relation to property which are unlawful by the laws of the state concerned, should apply to injuries to the person (judgment, paragraph 159). He said that the Court should decide as much (judgment, paragraph 160). So FAS does not apply where a foreign state executive has caused physical or mental harm to a claimant through an act in its territory which is unlawful according to the laws of that state.

Part of the background to Cs’ claim is that the Kenyan police forces failed to foresee the PEV and were overwhelmed by its scale and ferocity. The Kenyan police largely failed to provide Cs with any protection from the PEV. This leads D1 to submit that Cs’ claim is a direct challenge to the validity or legality of acts of the Kenyan state (cf paragraph 68 of Yukos Capital Sarl v OJSC Rosneft Oil Co (No 2) [2013] EWCA Civ 855; [2014] QB 458, cited in paragraph 14 of the D1’s skeleton argument). D1 also submits that the court will be required by the nature of the Cs claim to ‘adjudicate upon acts done abroad by virtue of sovereign authority’ (per Lord Wilberforce in Buttes Gas and Oil Co v Hammer (No 3) [1982] AC 888, at page 932).

D1 lists, in paragraphs 45 of its skeleton argument, four questions which, it submits, Cs’ claim as presently pleaded would require the court to decide. These include whether it was reasonable for D2 to rely on the police in various respects, and whether D2 could have compelled the Kenyan police to provide reinforcements. These questions are said, in turn, to require the court to decide several subsidiary questions listed in paragraph 46 of the skeleton argument about what legal duties the police owed D2 and whether those duties were discharged.

Cs submit that Ds’ arguments extend the doctrine of FAS beyond all bounds. In no previous decision has it been applied to excuse a defendant from liability because the case against it involves some factual consideration of what a foreign state has done in its own territory.

Cs submit that the core allegation against Ds is summarised at paragraph 58 of the GPC. Their case is that Ds failed to protect them against a foreseeable risk of attack. Ds failed to identify a risk of PEV, failed to have adequate crisis management and/or a security plan to avert or manage that risk and failed to take any, or any adequate measures to protect Cs in response to PEV.

The duty of care said to be owed by D1 to Cs is explained in paragraphs 83-119 of the GPC. D1 is said to be fixed with a duty of care because D1 took responsibility for ensuring that D2 adopted and operated effective systems to protect against the foreseeable risk of election violence. Paragraph 119 sets out eight allegations of negligence against D1. The allegations of negligence against D2 are pleaded in paragraph 122 of the GPC. There is nothing on the face of the pleadings, Cs submit, which requires the court to hold that the Kenyan police or other agents of the state acted unlawfully or improperly.

Ds’ case, say Cs, seems to be that it was the job of the Kenyan police to protect the Cs, D1 made every effort to get the police to help, and the Ds relied on the advice of the police about evacuating people. The various issues about the duties of the Kenyan authorities to protect the Cs from loss and damage caused by PEV which, Ds contend (Ds’ skeleton argument, paragraph 55), the court would have to decide, involve mostly factual inquiries.

Paragraph 19 of the Cs’ skeleton argument lists the things the claim is not about. There is no claim, for example, that the acts of violence were caused by state agents. Cs do not ask the court to impugn the validity of any foreign law or decree, or to judge the transactions of a foreign state. The claim raises no issue about the United Kingdom’s relations with a foreign state.

Cs submitted that their claim is not covered by FAS. In their most recent written submissions, they say that the decision of the Supreme Court ‘puts the position beyond reasonable argument’. In their skeleton argument they submitted that their claim is plainly a private law claim for damages for negligence.

Cs submit that if the Ds’ pleaded defence mirrors the contentions in its skeleton argument, the court may well have to consider what the Kenyan authorities did or did not do; but this would not require the court to make any findings about the legality of the behaviour of the Kenyan authorities (as that concept is understood in the decided cases). All that would be needed is an evaluation of what happened when the PEV erupted.

They further submit that D1’s argument extends the FAS in three other ways.

I do not consider that the court will need to decide many, if any, of the subsidiary questions set out by D1 in paragraph 46 of its skeleton argument, or to express any opinion about them. The simple question is, given that there was a Kenyan police force, whether it was reasonable for D2 to rely on that force to keep order. No doubt evidence will be given about what the police did, or did not, do in the course of the PEV. Some of that evidence will include evidence about what happened when some Cs sought refuge at police posts. There is no dispute about the scope of the duties of the Kenyan police. That has been decided by the High Court of Kenya in Murithii v Attorney General [2015] e KLR (see paragraph 49 of D1’s skeleton argument).

I accept the Cs’ submissions on this issue, and reject D1’s. This is not a case in which the court will be required to decide whether the Kenyan authorities acted lawfully or not. No ‘adjudication on the validity, legality, lawfulness acceptability or motives of state actors’ is required. The court will not have to decide whether the police were acting lawfully or not. The court will not have to decide whether the police could or should have done more, but, instead, whether, given what it could reasonably be anticipated the police might do, and given what they did do, D2’s duty of care required it to do more than D2 in fact did. This is not a case where a dispossessed owner sues in respect of his property and is met with a defence that a FAS has altered the title to that property. Such a defence would require the court to decide if title had been validly altered by that act of state. By contrast, here, at most, the court will be considering, instead, factually, what the Kenyan authorities did, or did not, do. Nor does the liability of the Ds depend on establishing that the Kenyan authorities acted unlawfully. The liability of the Ds depends, instead, on whether they owed a duty of care to the people present on the estate to protect them from unlawful violence, and, if so, whether that duty was breached.

The Cs are not arguing that the Kenyan authorities are liable for their injuries; they are claiming that the Ds are (see paragraph 18 of D1’s skeleton argument, where this point appears to be accepted). I understand that part of D2’s defence is that it was entitled to rely on the police to protect the people living on the estate, and that it did all it could to get extra police help. There is nothing in FAS which prevents D2 from relying on those two arguments.

In my judgment it is clear that FAS, as explained by Lord Neuberger in his judgment, cannot apply to Cs’ claims. Cs’ claims are not caught by any of the first three rules (D1 relies in its most recent written submissions on the third rule). The claims, quite simply, involve no challenge to the lawfulness of any act by the Kenyan state. A court trying those claims would just not be required to decide whether what the Kenyan police did was lawful. This is so, even if Cs’ claim involves showing (if only collaterally) that the Kenyan police broke their ‘constitutional promise’ to protect Cs.

The fourth rule, if it exists, can only apply in exceptional circumstances, where the executive (usually the FCO) issues a communication to that effect. There has been no such communication in this case; and D1 does not suggest that this rule is engaged.

My conclusion is that FAS does not prevent a court from deciding these claims. In case that view is wrong, I should say that I would have accepted the C’s ‘backstop’ arguments (Cs’ skeleton argument, paragraph 26).

Should the claim against D1 be stayed on other grounds?

D1 argues that the claim against it should be stayed, either, because the courts of England and Wales are not the appropriate forum for this claim, or on case management grounds, largely because, D1 submits, the claim against D1 is not viable, and it is clear that the claims have only been brought here against D1 for the improper purpose of bringing D2 into this jurisdiction. D1 does not argue that the court should decline jurisdiction (recognising that that would be contrary to article 4.1 of the re-cast Brussels Regulation (Regulation (EU) No 1215/2012) (‘the Regulation’). D1 argues, instead, that the claim should be stayed, pending the bringing of a claim against D2 in Kenya.

Vedanta made a similar application in the Vedanta case. Coulson J described the legal background to the applications in that case in paragraphs 39-46 of his judgment in that case. I gratefully adopt his analysis. He described the law in more detail in paragraphs 53-58 of his judgment. I gratefully adopt that exposition.

In short, article 4.1 of the Regulation provides that ‘Subject to this Regulation, persons domiciled in a Member State, shall, whatever their nationality, be sued in the courts of the Member State’. Regulation 5.1 provides that persons domiciled in a Member State may be sued in the courts of another Member State only by virtue of the rules set out in ‘Sections 2 to 7 of this Chapter’. The clear purpose of the Regulation is to provide rules for deciding, among different Member States, in which Member State a claim should be brought against a defendant who is domiciled in a Member State.

But in Owusu v Jackson [2005] QB 801, the Court of Justice of the European Communities held that the Regulation also applied when one of the possible jurisdictions in which a claim could be brought is a jurisdiction which is not the jurisdiction of a Member State (in that case Jamaica). The upshot was that, the defendant being domiciled in England and Wales, the claim had to be brought here, and not in Jamaica even though the claim had close links with Jamaica. As Coulson J said in Vedanta (at paragraph 54) ‘Owusu is authority for the proposition that forum non conveniens arguments are irrelevant [to the claim against the English-domiciled defendant], given the terms of what is now Article 4’. In paragraph 56 of his judgment he listed some of the cases in which the courts of England and Wales have since decided that Owusu prevents any consideration of forum non conveniens when a defendant is domiciled in England and Wales.

Coulson J noted (judgment, paragraph 59) the defendants’ arguments that the claim against Vedanta was a device designed simply to ensure that all the claims were brought in the United Kingdom, and that that claim was hopeless. The defendants argued that that meant that the claim was not viable, and would not, realistically, come to trial in the United Kingdom. He referred to his analysis of the claim against Vedanta, and to his conclusions that it raised a real issue (and was not bound to fail), and that it was not ‘weak or very unlikely to succeed’. His view was that it was ‘difficult to say that a claim which raises a real issue to be tried is, at the same time, a device’ (judgment, paragraph 63).

Coulson J considered Vedanta’s application for a stay against that background. In paragraphs 65-72 of his judgment he gave his reasons for refusing a stay on forum non conveniens grounds. He held that he was bound to reach that conclusion by the decision in Owusu and by the domestic authorities he referred to in paragraph 56 of his judgment. I can see no reason to differ from his analysis of this issue. I gratefully adopt his reasoning. Applying it here, I conclude that I should not order a stay of the claim against D1 on the grounds of forum non conveniens.

Coulson J recognised in paragraph 84 of his judgment that, despite Owusu, the court retained a discretion to stay a claim on case management grounds (see Pacific International Sports Club v Surkis [2009] EWHC 1839 (Ch)). He accepted that that discretion could be exercised in ‘rare and compelling circumstances’ (see per Lord Bingham in Reichold Norway ASA v Goldman Sachs International [2000] 1 WLR 173). He refused to exercise that discretion because he had decided that there was a real issue between the claimants and the domiciled defendant, and that the sole purpose of the proceedings was not to act as a hook for bringing the second defendant into the jurisdiction. He could not conclude, therefore, that the claim against the domiciled defendant was not viable, or that it was unrealistic that there would ever be a trial of the claim against that defendant.

If I had concluded that Cs’ claim against D1 raised a real issue to be tried, I would unhesitatingly have adopted Coulson J’s reasoning for refusing a stay on case management grounds.

I have, however, decided (see below) that C’s claim does not raise a real issue to be tried. But I do not consider that that is a reason for staying the claim. Many authorities were cited and analysed in both parties’ skeleton arguments and in Ds’ note on the court’s power to stay on case management grounds. None is directly in point. The cases show that this residual discretion must not be exercised in such a way as to circumvent Owusu by, so to speak, the back door. The cases do not suggest that the merits of the claim are relevant to the question whether this discretion should be exercised; the court has other powers to deal with weak claims. The cases give no guidance about the ‘exceptionally strong grounds’ that are required to found the exercise of the discretion (per Lawrence Collins J, as he then was, in Mazur Media Limited v Mazur Media GmbH [2004] EWHC 1566 (Ch); [2004] 1 WLR 2966). They do not suggest that the discretion can be exercised where there are no proceedings pending in a foreign jurisdiction against the non-domiciled defendant. In my judgment, the effect of a stay would be to circumvent Owusu. It would force Cs to bring proceedings in Kenya and to prevent the litigation of a claim here which Cs are entitled to bring here by virtue of regulation 4.1 of the Regulation, as interpreted in Owusu.

I therefore refuse to stay the claim against D1 on case management grounds. D1’s remedy in respect of the weakness of that claim is to apply to strike it out.

Is there a real issue between Cs and D1?

Cs submit that if the claim depends on substantial controversial points of law in a developing area, then it would not normally be appropriate to rule the claim out at this stage. The higher courts have repeatedly emphasised that it is unwise to strike out a claim at the very outset, particularly in the field of negligence, because whether it is fair just and reasonable to impose a duty of care will very much depend on precisely what facts are ultimately found at trial. It is undesirable to make decisions on questions of law before the facts have been found. There are also disputed issues, for example, about the relationship between D1 and D2 which are not suitable for any decision before disclosure has been given. The Cs submit that they will not be able to present their best case about the relationship between D1 and D2 until after disclosure. They have prepared, in support of that submission, a schedule of the documents for which they have asked and which Ds have not provided, or which they have not looked for, or they have not been able to find.

Ds submit that I can nonetheless be confident that the claim against D1 is fanciful, and should not go any further. D1 submits that the duty which Cs rely on is a ‘novel and unarguable duty of demonstrably breath-taking scope’. In that situation, submits D1, it is open to the court to conclude that the claim against D1 simply is not arguable, despite the caveats made by Cs.

There are three issues.

I bear in mind that I am not making a final decision about these issues after a trial. I have to decide, rather, whether Cs’ claim raises a real issue, or conversely, whether the claim is bound to fail. That is the approach which I have adopted, but I do not repeat those formulae for each issue in this part of my judgment.

Which law applies to Cs’ claim against D2?

Cs accept that the law of Kenya applies to their claims (see paragraph 29 of Mr Leader’s fourth witness statement; and this was conceded at the hearing). In my judgment that concession was rightly made, and I say no more about it. Cs remind me that questions of foreign law are questions of fact to be decided on the evidence. There are reports from experts. They disagree on some issues. Cs submit that it is not for me to substitute my view for the views of the experts, especially in the absence of any cross-examination, unless I were to decide that the view on an issue is ‘manifestly absurd on its face’ and incapable of clarification.

Cs submit that whether Kenyan law or the law of England and Wales applies to substantive merits makes no difference, as the experts agree that on these issues, the law of England and Wales would be treated as very persuasive in Kenya. Although the duty said to be owed by D2 to Cs, as pleaded, is supported by more than one source, the parties also agree, for the purposes of this application only, that my decision can be based on the duty imposed on D2 by the Occupiers’ Liability Act (Cap 34) (‘the OLA (K)’). That duty would include, the experts agree, a duty to take reasonable care to protect the Cs against the foreseeable criminal acts of third parties (see SOMA Properties Limited v HAYM Civil Appeal No 74 of 2005).

Are the criteria in Caparo v Dickman satisfied?

Both sides agree that Kenyan law would follow the law of England and Wales on the question whether a duty of care should be imposed on D1. The court would apply the three-stage test in Caparo Industries plc v Dickman. There are three ingredients. The damage suffered should be foreseeable. There ‘should exist between the party owing the duty and the party to whom it is owed a relationship characterised by the law as one of ‘proximity’ or ‘neighbourhood’ and … the situation should be one in which the court considers it fair, just and reasonable that the law should impose a duty of a given scope upon the one party for the benefit of the other”: per Lord Bridge of Harwich, at p 618-619.

I consider, first, whether Cs’ claim raises a real issue that, on the balance of probabilities, the loss and damage suffered by Cs was foreseeable (conversely, whether such a claim is bound to fail). I am not trying this issue. But I agree with Cs (skeleton argument, paragraph 97) that I must consider it. It would not be enough, in my judgment, for Cs to show at trial that PEV was foreseeable in Kenya, or indeed in RVP. I would have had no hesitation in deciding that they could do that. The issue, however, is different. It is whether the loss and damage suffered by Cs, caused as it was by PEV of the scale and ferocity of the PEV which broke out on the Plantation, was foreseeable.

The Cs list, in paragraph 99 of their skeleton argument, 19 separate matters which, they say, support this aspect of their claim. I was taken to the relevant materials in more detail during the hearing. The Cs rely on four types of evidence: general evidence about previous election violence, evidence from former managers working for D2 about tensions on the Plantation which they reported to more senior managers of D2, evidence from Cs about an increase in tensions, and independent reporting from risk management companies about increasing tension in the run-up to the election and a concomitant risk of violence.

The Cs rely on Mr Kegoro’s report and on his witness statement. In short, there has been violence at every election since 1991. 2002 was an exception because both candidates were Kikuyu. The pattern at elections after 2007 is not relevant here. The violence is most intense when the rival candidates have been from different ethnic groups.

In paragraph 23 of his witness statement Mr Kegoro gives his view that the risk of PEV was foreseeable and was foreseen by the Kenyan authorities. He lists nine findings of the Waki Commission in paragraph 24 which support that view. There was election violence in 1992 and 1997, including in RVP by Kipsigis on ‘outsider’ ethnic groups in the Rift Valley. The Southern Rift Valley is a hotspot for violence between the Kipsigis and the Kisii. There was a gradual and well-publicised build-up of tensions before the PEV. There was widespread violence in the run up to the election including in two different places which were an hour’s car journey away from the Plantation. Violence had been simmering between the Kipsigis and the Kisii in Kuresoi (an hour away from Kericho by car) since the 2005 referendum. He explains in paragraphs 56-59 of this witness statement why it became clear, once Mr Kibaki appointed Mr Nyachae, a Kisii, to his Cabinet, that the Kisii and Kalenjin would support opposite sides in the 2007 election. The tension between them was increased when, on 22 September 2007, Mr Ruto, a Kalenjin, and leader of the opposition party, was violently attacked in Kisii County, an attack which was widely believed to have been orchestrated by Mr Nyachae. This triggered revenge attacks by Kalenjin on Kisii, for example in Sondu, about 45 km west of Kericho.

There was longstanding hostility between the Kipsigis and the Kisii in the Kericho area. The authorities recognised the possibility of violence in the Kericho area and convened peace committee meetings before the election to prevent violence. There were telephone warnings from perpetrators to victims in the South Rift by phone and leaflet, which suggested planning. A large number of people seem to have known in advance that there would be violence. Violence started in Sotik, 35 km away from Kericho, on 28 December, before the results were announced.

It is right that the Report found that the Kenyan security service (the NSIS) accurately predicted the violence before it started. The NSIS provided regular reports to state agencies. As early as September 2007 they warned of the potential for election violence in specified areas (see the Report, page 362). The NSIS reports even named suspects and recommended that other agencies make contingency plans. They warned that security organs might be overwhelmed.

But the Report makes clear that this information did not reach the police on the ground. The Commission was very critical of the police. ‘In terms of preparedness for dealing with [PEV], the Police were simply too far off the mark. There was, and still is, a lack of understanding and reluctance to accept that a comprehensive exercise in planning and preparation for the elections were not only required but constitute a basic responsibility of the Police’ (Report, page 369). It identified ‘fundamental failures’ in planning, and in intelligence gathering which ‘set the scene for police preparedness, and response, nationwide’. Their approach reflected a ‘misplaced arrogance that they would always be able to control what came up’ (ibid, page 370).

The police ‘quite simply failed, and failed badly, in this fundamental requirement of their mandate’ (ibid, page 372). ‘Many citizens seemed to have known at least a few days in advance that there would be targeting by tribal and/or political affiliation although the police seemed unaware or chose not to take the information seriously. In any event the police were unprepared and failed to take the necessary pro-active steps to reduce the opportunity for violence to eventuate’ (ibid, page 375). This was a ‘failure to perform in what is a critical and fundamental element of policing. As a result, many more people lost their lives, were injured, were displaced, their homes and personal property destroyed and ability to continue their livelihoods severed than would otherwise have been the case’ (ibid, page 376). The police seemed to lack any independent capacity to gather or to use intelligence (ibid, page 376).

Mr Kegoro refers to five further factors in paragraph 25 of his witness statement. They are the coverage by international media of the tribal violence in 1992, 1997 and before the 2007 election; the fact that the election was the closest since 1991; the fact that it was clear from 2005 that the Kalenjin and the Kisii would support rival candidates in the 2007 election; the incessant claims of vote-rigging and doubts about the impartiality of the Electoral Commission; and ‘truly alarming levels of hate speech’ in the campaign, which were spread by every means conceivable.

Ds submit that the violence at the 2007 election was unique in that it was the only election in which large bands of several hundred young men invaded the Plantation and attacked residents and employees. Not all the victims of this violence were Kisii: some were Kikuyu. The Commission found that the violence in 2007 was ‘unprecedented. It was by far the most deadly and the most destructive violence ever experienced in Kenya’. The police were overwhelmed by the ‘massive’ numbers of attackers and the apparent co-ordination of their attacks. The authorities’ failure ‘to act on intelligence and other early warning signs contributed to the escalation of the violence’.

The response of the authorities in Kericho described in Mr Kegoro’s witness statement should be set in the context of the Report. The account of government officials in the area was that they did not know that there would be PEV and did not prepare for it. The ‘worst case scenario for them’ was violence in the lead-up to the elections in areas they considered to be prone to violence such as some parts of Kericho town. ‘There were no contingency arrangements in relation to the violence that ended up engulfing the multinational estates in Kericho and, therefore, no provision to deal with the violence’ (see the Report, page 148-9). When the violence broke out in Kericho town, the police were not prepared for it, and responded with extreme force.

The ‘peace committee meetings’ were ‘the response by the authorities in Kericho to the violence that they thought might happen’. The meetings were organised by the Electoral Commission and were ‘its preferred intervention in dealing with the possibility of election violence’. They were organised all over the country, and they seem to have been prompted by widespread hate speech (Report, pages 148-149).

Ds point out that in the run-up to the 2007 election during December 2007 the Kenyan Government said publicly that it would be responsible for security during the election. The Ds’ case is that the PEV which happened on the Plantation was not reasonably foreseeable by D2 or by D1. They also argue (see further below), that it was reasonable of D2 to rely on the police for protection when it did break out.

I note the approach to the foreseeability of criminal acts by third parties taken by the Court of Appeal of Kenya in Soma Properties Limited v Haym [2015] KLR (see further, paragraph 106, below). In essence, criminal acts by third parties causing loss and damage to a visitor on a defendant’s land may well be foreseeable if something comparable has happened on the defendant’s land in the past. There is no evidence, and no pleaded case, that anything comparable had happened on D2’s land before. In my judgment Cs have not shown that there is a real issue that the losses and damage which the Cs did suffer as a result of the acts of the criminals who invaded the Plantation were foreseeable by D1, resulting as they did from PEV of the scale and ferocity which broke out on the Plantation, and from the almost total failure of the police to keep law and order there (as elsewhere in Kenya).

Most of the material on which Cs rely is capable of showing that some PEV was foreseeable in RVP, and indeed, in Kericho. What it cannot show, in my judgment, is that D1 should have foreseen that large armed bands of Kipsigis would invade the Plantation in their hundreds, and burn homes, loot those, and also hunt down and attack Cs. Crucially, as I have said, nothing remotely comparable had ever happened before on D2’s land. D1 had no means of foreseeing what would happen other than what was communicated to it by D2, or by Control Risks, or similar organisations. The material I have seen from Control Risks was very general. While it may well be that the Kenyan police were at fault in not foreseeing what was to happen, I find it inconceivable that a court would hold after a trial that D2, let alone D1, which was based thousands of miles away, should have foreseen the loss and damage caused to Cs by these terrible events. The fact that D2 has changed its approach for the future in the light of what happened in 2007 does not show that it had should have foreseen those events before they happened, when such events had not happened before. Cs plead that threats were made in 1997, but they resulted in nothing on this scale. Nor does the fact that general threats were made to some Cs in 2007, and communicated by them to managers of D2, mean that there is a real issue that D2 should have foreseen the damage suffered by Cs. These were no doubt attempts to intimidate members of the minority tribes on the Plantation, but there were not specific threats to the safety of individuals.

It is not pleaded that it was foreseeable that the Kenyan police would fail to protect Cs from the criminal acts of the Kipsigis who invaded the Plantation. But the premise of the nature and scope of the duty which is pleaded is that it was foreseeable, not only that Cs would suffer as they did, but that law and order would break down generally, and police would fail to protect Cs when it did break down. I consider that any such argument is bound to fail.

In case that is wrong, I should consider whether Cs claims that there is a sufficient relationship of proximity between them and D1, and that it is fair just and reasonable (‘FJR’) to impose a duty of care on D1 on the facts of this case are bound to fail. A substantial focus of the argument was the question of proximity. An aspect of that is whether there is a sufficient relationship of proximity between an indirect parent company registered in a different country and visitors to the property of its indirect subsidiary. Is it FJR to impose a duty on the foreign parent to keep them safe from the criminal acts of third parties? A further aspect is whether it is FJR to impose a duty to keep visitors safe in the second country when there has been a total breakdown in law and order or whether as Ds contend, they were reasonably entitled to rely on the police to provide that protection. A third aspect concerns the relationship between the scope of the duties pleaded, and the conduct which is said to constitute a breach of those duties.

On the first question, both sides have referred to Chandler v Cape plc [2012] EWCA Civ 525; [2012] 1 WLR 3111 (and, by analogy, to Connelly v RTZ [1998] AC 854), to Thomson v Renwick Group Plc [2014] EWCA Civ 635 and to Lubbe v Cape Plc [2000] 1 WLR 1545). In Chandler, the claimant had been employed by a subsidiary of the defendant as a brick loader. He worked next to a factory which had until 1956 been operated by the defendant, and was then operated by the subsidiary. The factory had open sides. An asbestos product was made in the factory. There was asbestos dust in the area where he worked. Many years later, he became ill because he had been exposed to asbestos dust.

The Court of Appeal noted that slowly but surely, the company which employed the claimant became integrated into the defendant’s group of companies (see paragraph 10 of the judgment). But it kept its separate identity (paragraph 11). The defendant shared its expertise with the subsidiary in various ways (paragraphs 14-27); especially in relation to the health and safety of employees of group companies. The Judge had found that the subsidiary had inherited its working practices from the defendant and that ‘As and when it felt it appropriate the defendant did control what [the subsidiary] was doing’ (judgment, paragraph 30). The Judge held that the defendant owed a duty of care to the claimant. There had been ‘a systemic failure’ of which the defendant had been fully aware’ (judgment, paragraph 57). The Court of Appeal upheld that decision.

The Court of Appeal held that a duty to intervene to prevent damage to another would arise where there was a relationship between the parties which gave rise to an imposition or assumption of responsibility on the part of the defendant. The Court acknowledged that a subsidiary and its parent company were two separate entities and there was no imposition of responsibility just because the defendant was the subsidiary’s parent. But it was not necessary for the parent company to have absolute control of its subsidiary before a duty of care could exist. At paragraph 66 the Court said that ‘The parent company is not likely to accept responsibility towards its subsidiary’s employees in all respects but only for example in relation to what might be called high level advice or strategy’.

The Court held that the law might impose on the parent responsibility for the health and safety of the employees of a subsidiary, where (as on the facts of Chandler), ‘(1) the businesses of the two companies were in a relevant respect the same and (2) the parent has, or ought to have, superior knowledge on some relevant aspect of health and safety in the particular industry; (3) the subsidiary’s system of work is unsafe as the parent company knew or ought to have known, and (4) the parent knew, or ought to have foreseen that the subsidiary or its employees would rely on its using that superior knowledge for the employees’ protection’ (judgment, paragraph 80). It was not necessary to show that the parent was in the habit of intervening in the health and safety practices of the subsidiary. Element (4) could be shown if the parent had the practice of intervening in trading operations such as production and funding issues.

The defendant had known about the works where the claimant had been employed. In the light of that knowledge and of its better knowledge about the nature and management of asbestos risks, it was appropriate to find that it had assumed a duty of care either to advise the subsidiary on what steps it had to take to provide employees with a safe system of work or to ensure that those steps were taken. The Court held that a direct duty of care had been owed by the defendant to employees of the subsidiary company. The defendant had breached its duty by not advising on precautionary measures. That had caused injury to the claimant.

What Chandler shows is that a parent company may, on appropriate facts, and even though it does not exercise complete control over the operations of its subsidiary, be responsible for the health and safety of employees of the subsidiary, in particular where the parent knows that the system of work operated by the subsidiary is unsafe. This case is a long way from the facts of Chandler. The close geographical links between the parent company and subsidiary in that case are not present. D2 is not a direct subsidiary of D1. In this case, the risks against which such a duty would require D1 to provide protection are not foreseeable risks of personal injury caused by a dangerous activity in which D1 knows D2 is engaged. I say more about these and other distinguishing features below.

Nonetheless, the relationship between D1 and D2 differs from that between the parent and subsidiary in Okpabi. There is no equivalent of Unilever Executive (‘UEx’) in Royal Dutch Shell. Cs submit that the publicly available documents show that D1 is not just a holding company but plays a more active role in managing D2’s affairs. Cs draw attention to D1’s memorandum of association, among other things. I was also shown many documents in the hearing in which D1 laid down rules about the policies and procedures (including for health, safety and risk management) which D2 (and all its other subsidiaries across the world) should adopt, and documents about monitoring and auditing those policies and procedures. On paper, D1 has assumed apparent control of the content and auditing of such policies. It is also true that there has been no discovery, and Cs point out that Ds’ response to requests for discovery has not been especially co-operative. At this stage, I must assume that Cs will make out their pleaded claim that the four indicia in Chandler are met. On the material I have seen, I would have some hesitation in concluding that what I have seen shows the sort of control and superior knowledge which the Court of Appeal described in Chandler. It is formal control exercised by UEx at a high level of abstraction, and over the content and auditing of general policies and procedures; although that is to some extent fleshed out by evidence from ex-managers. I would also hesitate before finding that D1 had, or had the means of acquiring, better knowledge about the risks in Kenya than did D2, or that D2 would have relied on D1 in that regard. I bear in mind that if such a duty is imposed on D1, it would apply to foreseeable risks of harm to employees and third parties arising in any of the countries of the world in which D1 has subsidiaries, and that those subsidiaries have some 174,000 employees. In the Chandler case the duty extended to the employees of one subsidiary operating in the same country. Nonetheless, I am not trying the claim (see Sabbagh v Khoursy [2014] EWHC 3233 (Comm), cited in paragraph 118 of Vedanta). I am deciding whether the pleaded case, and the material I have been shown, raises a real issue. That this issue is finely balanced is shown by the different decisions made by Coulson J and Fraser J on broadly similar material, applying the same law. With some hesitation, I conclude, on the Chandler aspect of the case, that it does. By that I mean that in theory, a claim against D1 as the parent of D2 might succeed, based on the documents by which D1 has sought to exercise control over the management of D2 and of D2’s various policies. But I must take into account the other unusual features of this claim before I can decide whether or not the claim as a whole raises a real issue between Cs and D1.

A further feature of this claim is that it seeks to impose liability on Ds for the criminal acts of third parties. In Chandler, by contrast, the personal injury suffered by the claimant was caused by the tortious conduct of the defendant and of the subsidiary, that is, by the way in which the subsidiary conducted its dangerous operations and by the way in which the defendant permitted that to happen. The opinion of Mr Havelock, Ds’ expert, is that the Kenyan courts are slow to impose such liability. Mr Ohaga, Cs’ expert disagrees (see paragraph 53 of his report). Two Kenyan cases Mr Ohaga cites involve an employee employed in an intrinsically dangerous job. They are not obviously analogous. He also relies on Everett, an English case about occupier’s liability, and Pike, a case to which the law of India applied and in which it was conceded that the point was arguable. Everett was a case about a nightclub. The Court of Appeal held the defendant had a duty of care towards the guest of a member of the club to protect him from attack by other visitors, but that the duty was only to do what was reasonable in the circumstances. The claim failed.

In Soma Properties Limited v Haym the Court of Appeal of Kenya held, having reviewed various authorities from other jurisdictions, that in principle an occupier of premises could owe a duty of care to protect visitors from the criminal acts of third parties. The defendant occupied a large shopping centre. One tenant was a bank, and there had been thefts and robberies in the past. The defendant had 35 security guards from a security company, armed with truncheons, and an alarm system. Two armed police officers were on duty in the centre. They and other police officers were summoned by the alarm. There was a gun battle between robbers and the police after a robbery failed. The plaintiff’s daughter was shot, and later died.

Whether a risk of such acts was foreseeable was said to depend on such features as ‘the nature of the business, frequency and similarity of prior incidents of crime on the premises and the neighbourhood’ (per Ouko JA). The Court of Appeal held that there was a foreseeable risk ‘in view of the previous robberies and thefts in the premises’.

A further feature of the claim is that it seeks to impose liability on D1 for loss and damage suffered in another country caused by criminal acts in circumstances where law and order has broken down. I have already decided that the claim that Cs’ loss and damage was foreseeable is bound to fail, as is the implied claim that it was foreseeable that law and order would break down. The parties disagreed in oral submissions at the hearing whether it was lawful in Kenya for private security guards to use firearms. There was no evidence about this. In any event, it is not pleaded that D1 or D2 could or should have employed private security guards equipped with firearms. The duty as pleaded, on the facts of this case, in effect requires D1 to act as a surrogate police force to maintain law and order. In the course of his review of the foreign authorities, Ouko JA observed in Soma that ‘…security matters are dynamic and…it is a difficult thing to impose on an individual a duty to ensure citizens are safe, a duty, viewed by many, including the dissenting Judges in the above decisions, as a national duty of the State’.

I turn to the pleading. I appreciate that D2, not D1, is the occupier of the Plantation. But I consider that the duty imposed on the occupier by the relevant Kenyan statute, the OLA (K), is informative. It is a duty ‘in respect of dangers due to the state of the premises or to things done or omitted to be done on them’. The duty is ‘to take such care as in all the circumstances of the case is reasonable to see that the visitor will be reasonably safe’. The Court of Appeal in Soma concluded that the defendant had taken ‘reasonable measures intended to ensure that the visitors were reasonably safe’. To demand, as the trial judge had, that the defendant ‘put in place “adequate measures”, was to raise the threshold beyond that set by statute’. I consider it vanishingly unlikely that a court would impose on D1 a duty the practical effect of which is more onerous than the scope of the duty imposed by the OLA (K) on D2, the actual occupier of the Plantation. I should deal here with the suggestion made by Mr Ohaga in his discussion of Soma (report, paragraph 73) that the services of two armed police officers were ‘privately procured’ by the appellant in that case, and that the decision of the Court of Appeal was based in part on that. I do not consider that it is clear from the summary of the pleadings, or from the summaries of the evidence in the judgments, that the two armed police officers were privately procured, as opposed to merely being on patrol in the shopping centre at the material time. The word ‘deployed’ is used more than once in the judgments, but it is not clear who was responsible for that deployment. The summary of Mr Singh’s report in the judgment of Ouko JA suggests that the two armed police officers were ‘on duty’ at the centre on the day of the robbery.

In this case, the duty pleaded is a duty to take ‘all reasonable steps to ensure that [Cs] did not suffer personal injury, damage to property or economic loss due to the foreseeable risk of ethnic violence on the Plantation particularly during election periods. I appreciate that D1 is not the occupier of the Plantation, but the pleaded duty is wider than the duty imposed by the OLA (K). It is pleaded as a duty to take all reasonable steps, not to ensure that Cs were reasonably safe, but to ensure that Cs did not suffer the loss and damage which they did suffer. The pleaded particulars of the breach of that duty in sum are that D1 did not take adequate steps to enable it to foresee and respond to the risk of violence which Cs faced, in particular by failing to ensure that D2’s relevant policies identified, and addressed, the risks of ethnic violence on the Plantation, and by failing to ensure that D2 effectively implemented such policies. Cs also allege that D1 failed to ensure that D2 ‘took adequate steps to respond to the foreseeable risk of violence, including enhanced security measures and/or arranging for vulnerable employees to return to their homelands amended/or transporting vulnerable workers to safer areas’. The breach of duty is summarised as ‘failing in all the circumstances to take any or any adequate steps to prevent and/or stop the mistreatment of [Cs]’. These particulars show that there is a mis-match between the scope of the pleaded duty and the conduct which is said to be a breach of it. The pleaded breach is not a failure to take reasonable steps, but a failure to secure a result.

It is not pleaded that Cs have any features which singled them out for the violent attacks they suffered, other than that they were Kisii. I bear in mind that to secure that result, since the election was very close, and its outcome was not known until the last moment, D1 would have had to arrange, in advance, for D2 to evacuate all the potential victims of PEV from the Plantation (Kisii, Kikuyu and perhaps also Luo and Luhya), that is, thousands of employees and their families, before the result was announced, that is, before law and order broke down. Any claim that D1 could or should have arranged for D2 to procure extra police protection faces the difficulty that, in the event, none would have been available, as is shown by the response to D2’s attempts to get help from police once the violence broke out; but I recognise that if the question whether police help could or should have been procured in advance stood on its own, that would be an issue for trial. Any claim that D1 should have arranged for D2 to get extra private security faces the difficulty that the Kipsigis were armed, and there is no pleaded claim that it is lawful for private security personnel to use firearms. It is not remotely probable that private security personnel equipped with sticks or truncheons could have protected Cs from the large mobs of armed Kipsigis. D2’s watchmen were not able to.

I would not have decided that Cs’ claim raised no real issue against D1 if the only obstacle in its path had been the Chandler point. But I do consider that when the scope of the duty is analysed by reference to what is pleaded expressly, and by what is implied in the pleading, it does not raise such an issue. I consider that a claim that it is FJR to impose on D1 a duty either to anticipate and protect Cs against a breakdown of law and order, or to keep law and order when it had broken down and to ensure that Cs did not suffer as they did from the criminal acts of the Kipsigis, is bound to fail.

Is the claim barred by limitation?

I have read the relevant passages in the reports of the two experts. Section 4(2) of the Limitation of Actions Act 1968 (Cap 22) (‘the LAA’) prescribes a limitation period of three years for cause of action in tort. Time runs from the date when the cause of action accrues. The effect of section 27(1) of the LAA is that where the damages claimed include damages for personal injuries the three-year limitation period does not afford a defence to the claim if the requirements of section 27(2) are met. Those requirements are fulfilled in relation to a cause of action if the plaintiff proves that ‘material facts relating to that cause of action were or included facts of a decisive character which were at all times outside the knowledge (actual or constructive) of the plaintiff until’ the material date. That date is described in section 27(2)(a) and (b). The parties agree that the effect of those provisions is that, for the claims not to be barred by limitation, material facts relating to the cause of action must have been outside the actual or constructive knowledge of Cs until one year before the claims were made (that is, in the cases where there was no standstill agreement, 1 August 2014, or in those cases where there is such an agreement, 4 October 2013 or 29 April 2014, depending on which standstill agreement applies) (see Mr Havelock’s first report, paragraph 238).

‘Material facts relating to a cause of action’ are defined in section 30(1) of the LAA as ‘a reference to one or more the following - (a) the fact that personal injuries resulting from the negligence… or breach of duty constituting that cause of action; (b) the nature or extent of the personal injuries resulting from that negligence…or breach of duty; (c) the fact that the personal injuries so resulting were attributable to that negligence…or breach of duty, or the extent to which any of those personal injuries were so attributable’. Section 30(1)(a) does not make sense. It seems that some words are missing, perhaps ‘have been suffered’, or that ‘of’ should replace the first ‘that’. The parties agreed.

Section 30(2) provides that for the purposes of section 27, ‘any of the material facts relating to a cause of action shall be taken, at any particular time, to have been facts of a decisive character if they were facts which a reasonable person, knowing those facts and having obtained appropriate advice with respect to them, would have regarded at that time as determining, in relation to that cause of action that (apart from section 4(2) of this Act) an action would have a reasonable prospect of succeeding and of resulting in the award of damages sufficient to justify the bringing of the action’.

Section 30(3) provides that for the purposes of section 27, a fact ‘shall be taken at any particular time to have been outside the knowledge (actual or constructive) of a person if but only if – (a) he did not know that fact, and (b) in so far as that fact was capable of being ascertained by him, he had taken all such steps (if any) as it was reasonable for him to have taken before that time for the purpose of ascertaining it; and (c) in so far as there existed, and were known to him, circumstances from which, with appropriate advice, that fact might have been ascertained or inferred, he had taken all such steps (if any) as it was reasonable for him to have taken before that time for the purpose of obtaining appropriate advice with respect to those circumstances’. ‘Appropriate advice’ is defined in section 30(5) in relation to any fact or circumstances, as ‘the advice of a competent person qualified in their spheres, to advise on the medical, legal or other aspects of that fact or those circumstances as the case may be’.

Section 39(1) provides that a period of limitation does not run if there is a contract not to plead limitation or the person relying on it is estopped from doing so. This confers protection on potential claimants who have made a standstill agreement with a potential defendant.

The policy of these provisions is to grant a short extension to the three-year limitation period for claims in tort in relatively narrow circumstances when a claim includes a claim for damages for personal injuries. The scheme of the provisions is that the extension is available where the ‘material facts’ about the cause of action (as defined in section 30(1) of the LAA) include facts of a ‘decisive character’ (as defined in section 30(2)), and the material facts of a decisive character were outside the actual or constructive knowledge of the plaintiff (as described section 30(3)).

The definition of ‘facts of a decisive character’ makes limited allowances for the ignorance of a plaintiff and assumes that he has taken appropriate advice about them. The exhaustive prescription of the circumstances in which a plaintiff will be taken not to know a material fact of a decisive character assumes that a plaintiff has taken all reasonable steps to find out such facts which he was able to discover and that, to the extent that he could with appropriate advice have found out about, or inferred such facts, he has taken all reasonable steps to get appropriate advice.

I have read the material about limitation in the reports of both experts. They do not agree about some of the key issues of interpretation which arise from the LAA. It seems to me reasonably clear that Cs knew, when they suffered the injuries which they suffered, that the people who were the direct cause of those injuries were the people who attacked them. Having regard to Mr Ohaga’s views, however, I consider that it is at least arguable the material facts of a decisive character, in relation to the putative causes of action against D1 and D2, who did not inflict the injuries and losses, include that the Cs’ personal injuries and other losses were attributable to breaches of duty by D1 and D2, and that Cs could make a claim against them in respect of those injuries and losses. This issue only arises if I am wrong about the strength of Cs’ claim against D1. I assume, for the purposes of this stage of the analysis (as seems to me to be probable) that Cs were advised that this claim had reasonable prospects of success.

The critical question is whether those facts of a decisive character were known to Cs, or within their constructive knowledge, at any stage before the critical dates I have referred to above, or whether, as Cs contend, they only actually knew those facts of a decisive character when their solicitors (‘LD’) advised them that they had a claim against D1 which had reasonable prospects of success.

Mr Ohaga’s view on constructive knowledge is that the question is whether Cs took reasonable steps to ascertain the facts before the limitation period expired. Relevant to that question, he says, are the facts that Cs are unskilled, uneducated and poor, that their poverty was increased by moving away from the Plantation, by losing all their possessions, by not being paid for the periods when they were away from the Plantation, and by their loss of earnings because of their injuries, the serious psychological trauma which they allege they have suffered, their fear that legal action in Kenya would expose them to unacceptable risks, the fact that they have no money to pay for a private lawyer in Kenya, that they sought advice re getting compensation from the government, that they had not appreciated that any action was viable until LD advised them that a case could be brought in England, that they were specifically warned when they returned to the Plantation not to raise any issues about their lost property and that they feared losing their jobs if they complained about the PEV.

I was taken to many passages in the reports of the parties’ experts which are relevant to this question. The parties’ experts do not agree about many of them. I am not in a position to decide those questions finally. Cs’ evidence includes evidence that Cs are very poor and cannot afford to bring a claim in Kenya. I reject D1’s argument, which is based on Mr Havelock’s evidence, that what Mr Nderitu describes as the ‘harambee’ system would be remotely likely to generate enough money to fund a claim such as this. CFAs are unlawful in Kenya. I also reject D1’s suggestion that Cs could or should, nonetheless, have engaged Kenyan lawyers who might be prepared to disobey that prohibition. There is evidence that the Kenyan legal system is inefficient. Files regularly get lost; Mr Havelock was effectively forced to concede this when faced with a list of his own judgments in cases in which he had referred to files being lost. The fact that, in his view, the loss of those files did not affect the outcome of those cases is neither here nor there. The Kenyan system has also had significant historic and continuing problems with corruption, which Mr Havelock acknowledges. There is a body of evidence about the physical risks to people who have sought to give evidence about the PEV (in particular in relation to the abortive ICC prosecution), and about witness tampering and intimidation. I consider those topics in more detail at paragraphs 154-170, below. As I found in my judgment on the application for a private hearing, this case is seen in Kenya as a case brought by Kisii against Kipsigis. Cs, having suffered significant unlawful violence at the hands of the Kipsigis, have, in my judgment, a reasonable fear that if they were to litigate this case openly in Kenya, they would be subject to further violence at their hands. In the light of the evidence about inefficiency and corruption in the Kenyan legal system, they would also have had a reasonable fear that anonymity orders, if made, would not be an adequate protection.

These factors, in combination, lead me to conclude that it is, at the very least, arguable, that material facts of a decisive character were not known to Cs until they were advised by LD that they had a claim against D1 which they could litigate in the courts of England and Wales. I also consider it arguable that, given the factors to which I have referred, Cs did not act unreasonably in not attempting to investigate possible legal remedies in Kenya, having attempted unsuccessfully to get compensation from the Government for their losses. It follows that, had I considered that there was merit in their claim against D1, I would not have held that it was barred by limitation.

In case my decision on the limitation point is wrong, I should consider whether the Kenyan law of limitation (not just the primary limitation period) should be dis-applied pursuant to section 2 of the Foreign Limitation Periods Act 1984 (‘the 1984 Act’). The scheme of section 1 of the 1984 Act is that if foreign law applies to a claim, so does the foreign law of limitation. Section 2(1) requires the court to dis-apply section 1 if Cs show that to apply it would conflict with ‘public policy’. The short question raised by section 2(2) is whether the application of the Kenyan law of limitation would cause ‘undue hardship’ to Cs. ‘Undue hardship’ is not caused by the mere fact that the foreign limitation period is shorter than the period which would apply under the Limitation Act 1980 (‘the 1980 Act’) (Harley v Smith [2010] EWCA Civ 78). In fact, the primary limitation periods provided for by the two enactments are the same; what differs is the scope of the court’s power to extend that period. I accept Cs’ submission that it is at least arguable that, pursuant to section 33 of the 1980 Act, they might obtain the necessary extension of time to bring their claim. Cs have referred me to the reasoning of the Court of Appeal in Bank St Petersburg v Arkhangelsky [2014] EWCA Civ 593, at paragraphs 19 and 24, and that of Hildyard J in his supplemental judgment in the same case [2013] EWHC 3674 (Ch) at paragraph 17. I have to consider all the circumstances of the case. For reasons which resemble my reasons for holding that it is arguable that the limitation period should be extended under Kenyan law, I accept Cs’ submission that it is arguable that the application of the Kenyan law of limitation would cause Cs undue hardship.

The claim against D2 is based on more than one duty alleged to be owed by D2 to Cs. As I have said, the parties have agreed that for the purposes of this application I can approach it as a claim under OLA (K). I have summarised the pleadings, and referred to the relevant provisions of OLA (K), and to the decision in Soma. Crucially, I have already decided that the claim that the loss and damage suffered by Cs were foreseeable by D1 is bound to fail. For reasons similar to those I have already given, it is my view that the claim that the loss and damage suffered by Cs were foreseeable by D2 and that it was foreseeable by D2 that the Kenyan police would fail to protect Cs from the unlawful attacks by the Kipsigis does not have a real prospect of success.

Had I been required to do so, I would also have decided that the argument that the OLA (K) imposes a duty on D2 to protect Cs from those attacks does not have real prospects of success. The Chandler point and related arguments are not relevant to this analysis, of course, but my reasons for this decision are similar to my reasons for holding that it would not be FJR to impose a duty on D1 to protect Cs from the loss and damage they suffered at the hands of the Kipsigis. Again, the duty pleaded against D2 (in short) is a duty to take reasonable steps, but the particulars of breach do not match the duty pleaded, as the breaches alleged are failures, in the main, to secure a result. The duty imposed directly on D2 by OLA (K) is a narrower duty, in short, than the duty relied on by Cs, as can be seen from the particulars of breach of that duty relied on by Cs.

If I had had to decide the issue of limitation, for reasons broadly similar to the reasons I have given for deciding that the claim against D1 was not barred by limitation, I would have decided that the claim against D2 was not barred by limitation, either. The claim against D2 is a more straightforward claim than the claim against D1, but the background factors are the same. If that decision was wrong, I would have dis-applied the Kenyan law of limitation pursuant to section 2 of the 1984 Act, on grounds similar to the grounds for dis-applying it in relation to Cs’ claim against D1.

This issue only arises if my reasoning on the other issues I have decided is wrong. I approach it, therefore, on the assumption, made for the purposes of this stage of analysis only, that I have decided all the logically prior issues in favour of Cs. D2 relied on the decision of the Court of Appeal in Erste Bank AG (London) v JSC (MV Red October) [2015] EWCA Civ 279. The Court of Appeal emphasised in this case that the question whether it is reasonable to try a claim in England and Wales is a distinct question from the question whether a claimant has a claim with real prospects of success. It must be separately considered by the court, and it is to be considered in isolation from any claim against the non-domiciled defendants (judgment paragraph 38).

In Red October the claimant, the London branch of an Austrian bank, had participated in a syndicated loan agreement with D1, which was based in Russia. D2, D1’s parent company, guaranteed D1’s obligations under that agreement. Both contracts provided that they were governed by English law and for arbitration in London, or, at the lender’s option, for the exclusive jurisdiction of the English courts. D3 was a state corporation domiciled in Russia and D5 was its subsidiary. The claimant alleged that the defendants had conspired together to put the assets of D1 and D2 beyond the reach of their creditors by transferring them to D7. D1 defaulted under the loan agreement. D1 and D2 were then made subject to insolvency procedures in Russia (as was D7). The claimant claimed in those procedures as a creditor and took part in them in several other ways (judgment, paragraph 12). The guarantee was invalid under Russian law. The claimant made claims in debt and damages in England against D1 and D2. It also made various tort claims against the other defendants, and for relief under section 423 of the Insolvency Act 1986. The claimant had obtained summary judgment against D1 and D2 in its debt and contract claims. The judgment debt was outstanding. The claimant then applied for permission to serve the other defendants out of the jurisdiction. D3 and D5 challenged the jurisdiction of the English court.

At paragraph 30 of the judgment the Court of Appeal listed the issues which were relevant to the question whether the claimant had shown that there was an issue which it was reasonable for the English court to try. Those factors are absent in this case, nor are any analogous factors present (see also paragraph 34 of the judgment, which lists the submissions of D3 and D5 on this issue). That question (judgment, paragraph 48) is a ‘finely nuanced, soft-edged’ one. The question is not whether it is reasonable for the claimant to issue proceedings in England and Wales but whether it is reasonable for the court to try the claim.

In Vedanta, Coulson J observed, at paragraph 129, that it was not until the decision in Red October that the question of reasonableness was given any real or separate prominence. Coulson J considered the effect of this decision in paragraphs 130-133 of his judgment. His conclusion was that the decision in Red October has to be seen in the context of the facts of that case, and, in particular, of the Court’s decision that there was no real issue between the claimant and D1 and D2. I agree with Coulson J that the importance of the question whether there is an issue which is reasonable for the English court to try will depend on the facts of the case. On the facts of Red October, it was obviously decisive. This case is very different. If I assume that there is a real issue between Cs and D1, I can see no reason why, if that claim is viewed in isolation from the claim against D2, it would not be reasonable for the English court to try it. As Coulson J said in Vedanta, a further important distinction between his case and this case, on the one hand, and Red October, on the other, is that in these two cases, the claim against the domiciled defendant goes ahead ‘as a matter of United Kingdom and European law’, whereas the foundation for any jurisdiction over the claims in Red October was a contractual agreement. Moreover, the claim against D1 has factual connections with England, and is hotly contested. The claim in Red October had no connections with England other than the jurisdiction clause in the two agreements. Further, judgment had been entered on the claims against D1 and D2 in Red October, so that there was nothing at all to try in relation to those claims.

In Vedanta, at paragraph 139, Coulson J quoted paragraph 96 of Carr J’s judgment in Sabbagh. She said

“The “necessary or proper party” test is at least as broad as the court's power to add or substitute a party under CPR 19.2 (2) (see United Film Distribution Ltd v Chhabria [2001] EWCA Civ 416 at paragraphs 36 to 38 and Altimo Holdings v Krygyz Mobil Tel Ltd [2011] UKPC 7 at paragraph 87.) For present purposes, it is therefore sufficient for Sana to show that there is a serious issue involving HH which is connected to the matters in dispute in the proceedings, and it is desirable to add HH so that the court can resolve that issue.”

On the assumptions that there is a real issue between Cs and D1 and that it is reasonable for the English court to try that issue, I would have no hesitation in deciding that D2 would be a necessary and a proper party to that claim. The claims against D1 and D2 are closely connected. Their resolution would require only one investigation. The claims are based on the same core facts and on similar legal principles, albeit that the claim against D1 involves additional elements. D1 and D2 would be sued in respect of the same losses. The causation arguments would be similar. If D2 were domiciled in the United Kingdom, it could unquestionably be joined to the claim against D1 (cf paragraph 141 of Vedanta, where Coulson J took similar points into account). D2 submitted in its skeleton argument that ‘no properly advised claimant would waste their resources (or the Court’s) on a claim against both these Defendants if both were within the jurisdiction’. In my judgment this is not a good reason for holding that this part of the test is not met. The test is not about whether two overlapping claims should be litigated at the same time. Its sole focus is whether the other party is a ‘necessary or proper party’ to the claim. That threshold is plainly crossed here.

Cs must show that England and Wales is ‘the proper place’ to bring the claim against D2 (3.1(3) of Practice Direction 6B). Coulson J summarised the relevant principles at paragraph 148 of his judgment in Vedanta. He cited paragraph 88 of the judgment of Lord Collins in Altimo Holdings v Krygyz Mobil Tel Ltd [2011] UKPC 7; [2012] 1 WLR 1804. Lord Collins said, ‘…the task of the court is to identify the forum in which the case can be suitably tried for the interests of the parties and the ends of justice’. That passage reflects something Lord Goff said in Spiliada Maritime Corporation v Cansulex Limited [1987] AC 460 at p 476, a statement, which, in turn, he said, derived from the judgment of Lord Kinnear in Sim v Robinow (1892) 19 R 665, at p 668. As Coulson J observed in Vedanta, this test involves two separate issues.

In Connelly v RTZ, the plaintiff was domiciled in Scotland but had been employed by a subsidiary of the defendant at a uranium mine in Namibia. His case was that his laryngeal cancer was caused by his breathing in carcinogens in the mine. He did not sue the Namibian subsidiary, but brought a claim against the parent company, which was registered in the United Kingdom, which, he alleged, had devised and implemented the subsidiary’s policy on health and safety. He then joined an English-registered subsidiary which had also been involved in implementing that policy.

Lord Goff said of the first issue that the defendant had to show that ‘there is another available forum which is clearly or distinctly more appropriate than the English forum in which jurisdiction had been founded by the plaintiff as of right. The court’s task on application for a stay is to look for ‘connecting factors which indicate that it is with the other forum that the action has its most real and substantial connection’ (at p 871F). The plaintiff in that case conceded that the claim’s ‘most substantial and real connection’ was with Namibia.

The second issue was described by Lord Goff in Connelly v RTZ quoting (at p 866C) from the judgment of Lord Bingham MR (as he then was) in the Court of Appeal:

“But faced with a stark choice between one jurisdiction, albeit not the most appropriate in which there could in fact be a trial, and another jurisdiction, the most appropriate, in which there never could, in my judgment, the interests of justice would tend to weigh, and weigh strongly, in favour of that forum in which the plaintiff could assert his rights.”

The House of Lords in Connelly upheld the decision of the Court of Appeal. Lord Goff described this passage as setting out ‘the decisive consideration’ in the reasoning of the Court of Appeal. As Lord Goff put it at p 873A: ‘Only if the plaintiff can establish that substantial justice cannot be done in the appropriate forum, will the court refuse to grant the stay…’.

The burden of satisfying the first part of the test is on the defendant, and of the second, on the claimant.

In Connelly the plaintiff’s case was that it was impossible for his claim to be presented without very substantial help with funding. That funding was available in England and Wales, but not in Namibia. Lord Goff said that the plaintiff had sued the two defendants as of right, because they were both registered in the United Kingdom. No doubt he had done so because he was domiciled here and because he was entitled to legal aid. But that did not of itself expose him to criticism. If he was going to sue the defendants, this was an appropriate jurisdiction in which to sue them. Lord Goff started from the position that ‘at least as a general rule, the court will not refuse to grant a stay simply because’ he cannot fund the claim in the appropriate jurisdiction. Many countries cannot afford legal aid. But exceptionally, the fact that the plaintiff could fund the litigation here and could not do so in the appropriate jurisdiction might ‘be a relevant factor’. The question was whether the plaintiff could show that substantial justice would not, in the circumstances of the case, be done if the plaintiff has to bring the claim in a forum where he would get no financial help (p 873H). This was not a case where the plaintiff was simply trying to get ‘a Rolls Royce presentation of his case, as opposed to more rudimentary representation in the appropriate forum’. There was ‘every reason to believe that this case calls for highly professional representation by both lawyers and scientific experts’. Lord Goff was satisfied that ‘this is a case in which, having regard to the nature of the litigation, substantial justice cannot be done in the appropriate forum, but can be done in this jurisdiction where the resources are available’.

In Lubbe v Cape Plc the House of Lords considered two appeals. Some 3000 claims had been brought in England by plaintiffs almost all of whom were resident in South Africa. The claims were brought in two tranches. They were brought against the defendant, a company which was registered in England but had had various South African subsidiaries which mined and sold asbestos and things made of asbestos. No claims were brought against the subsidiary companies. The basis of the claims against the defendant was that it knew that asbestos was dangerous but did not take appropriate steps to ensure that the working practices of its subsidiaries were safe. This was said to be a breach of a duty of care owed by the defendant to people employed by the subsidiaries, or living in the area of their operations. By the time the claims were brought, the defendant had stopped trading in South Africa and owned no assets there. This meant that the courts of South Africa had no jurisdiction over the defendant or its subsidiaries when the first tranche of claims were brought. The courts of South Africa only had jurisdiction over the defendant or its subsidiaries to the extent that it undertook, and could validly undertake, to submit to their jurisdiction (see p 1563F-G per Lord Hope). The defendant applied to stay the claims.

There were two applications for a stay. The first application concerned the first claims, brought by four plaintiffs, one of whom was resident in the United Kingdom. A stay was granted at first instance but overturned on appeal. The Court of Appeal on that appeal (‘the first Court of Appeal’) differed from the judge about whether South Africa was the appropriate forum for the claims and held, in any event, that that forum had not been available to the plaintiffs until undertakings to submit to the jurisdiction of the South African courts were given by the defendant. Once leave to appeal from the Court of Appeal to the House of Lords was refused, some 3000 further claims were brought against the defendant. A judge stayed those claims. His decision was upheld by a different division of the Court of Appeal (‘the second Court of Appeal’). The second Court of Appeal held that the factors suggesting that South Africa was the appropriate forum were ‘overwhelming’ and that that forum was available to the plaintiffs. The House of Lords vacated its refusal of leave to appeal in the first group of cases on 30 March 2000. The House of Lords then heard appeals against the decisions of the first and second Court of Appeal.

On the first part of the test, Lord Bingham said that the case (against the parent company) was in two parts. The first part, concerning the liability of the parent company, depended on the actions of the parent, and many of the documents were likely to be in England. The second part of the case involved the plaintiffs’ personal injuries claims. The plaintiffs would have to be examined by doctors, and their working conditions would have to be investigated. Lord Bingham said that the conclusion of the first judge that South Africa was the appropriate forum was ‘tenable…but not inevitable’ on the case as then presented. He questioned whether the first Court of Appeal was justified in overturning the judge’s view on that question, but held that its own assessment of the factors was not unreasonable or wrong, either. On the case as then presented, ‘…there was room for the view that South Africa was not shown to be a clearly more appropriate forum’. It is ‘a field in which different conclusions can be reached by different tribunals without either being susceptible to legal challenge’. The jurisdiction to stay was ‘liable to be perverted if parties litigate the issue at different levels of the judicial hierarchy in the hope of persuading a higher court to strike a different balance in the factors pointing for and against a foreign forum’ (at p 1556E-F).

Lord Bingham had a different view about the second tranche of claims. He said that ‘The emergence of over 3,000 new plaintiffs…had an obvious and significant effect on the balance of the proceedings…the personal injuries assumed very much greater significance.’ This would require ‘a careful, detailed and cumbersome factual inquiry, and, at least potentially, a very large body of expert evidence’. That tipped the balance ‘very clearly in favour of South Africa’. He held that the judge and the Court of Appeal were ‘plainly correct’ to hold that South Africa was clearly the more appropriate forum for the claims.

Lord Bingham said in Lubbe that if a forum other than the courts of England and Wales was the appropriate forum, it was only if the plaintiff could show that ‘substantial justice will not be done that a stay will be refused’. It was not enough for the plaintiff to show that he would enjoy procedural advantages, bigger damages, or more generous rules about limitation (and see Connelly at p 872F-H per Lord Goff). ‘This is not an easy condition to satisfy’. Nor was it generally enough to show that legal aid (or other help with funding) is available in England and Wales but not in the appropriate forum (at p 1554G-H). It may, however, exceptionally, be a relevant factor (at p 1555D) (echoing Lord Goff in Connelly).

Legal aid for personal injury claims had recently been withdrawn in South Africa, but conditional fee agreements were lawful. The plaintiffs argued that it was not likely that anyone would be able to fund a group action of the size and complexity of the current claims in South Africa. The claims could not be prosecuted effectively without legal representation. If the proceedings in England, where funding was available, were stayed, that would deny the plaintiffs any reasonable prospect of pursuing their claims to trial. The House of Lords, in short, accepted that argument.

In Vedanta, Coulson J held that if the claims against the subsidiary were considered on their own, ‘it was plain and obvious’ that England was not the appropriate place to bring the proceedings (judgment, paragraph 152). He said that the claimants were all Zambians, the claims all involved personal injuries or damage to land, the place of the commission of the alleged tort was Zambia, the mine was owned and operated by a Zambian company regulated by Zambian law and supervised by Zambian regulatory authorities, and the applicable law was Zambian law (judgment, paragraph 153). He described, at paragraph 154 of his judgment, the ‘raft of logistical matters’ which suggested that Zambia was the appropriate place for any trial. He said that ‘The claimants have no real answer to these points’ (judgment, paragraph 156).

He then considered what the position was if the claim against the parent company was taken into account. He then referred to three cases in which the exercise of this discretion had been considered where there was a parallel, related claim running against a domiciled defendant. Those cases show that where there is such a claim, that will be a very powerful, if not decisive, factor, because it is undesirable for proceedings to be duplicated in two different jurisdictions. He cited paragraph 16 of the judgment of Leggatt J in OJSC VTB Bank v Parline Limited [2013] EWHC 3538 (Comm). Leggatt J said, ‘The real question, in other words, is whether the factors which connect the claim against the second defendant with Russia carry weight in circumstances where to require the claim to be pursued in Russia would result in duplication of cost and the risk of inconsistent judgments – the same factors which make the second defendant a necessary or proper party’. Coulson J decided that, subject to two arguments to the contrary from the defendants which he rejected, the existence of the parallel claim against the domiciled defendant did in effect conclude the issue.

This conclusion is apparently inconsistent with the views of the House of Lords in Connelly and Lubbe about the appropriate forum for claims against English-domiciled defendant parent companies in relation to the acts of subsidiaries in other jurisdictions (defendants which were not, in either case, sued in those jurisdictions, or joined in the claims against the parent companies). This is a consequence, it seems to me, of the Regulation, as interpreted in Owusu. The Brussels Convention came into force in the United Kingdom on 1 January 1987. It was in force when both Connelly and Lubbe were decided. It is not referred to in Connelly. The House of Lords heard argument about the effect of article 2 in Lubbe (see per Lord Bingham at p 1561-1562B). The Appellate Committee did not need to decide its effect because a stay was refused on other grounds. It seems that its effect in a case where the appropriate forum is not a Member State might not have been appreciated until the decision in Owusu.

I repeat that these issues only arise if my conclusions about the merits of the claims against D1 and D2 are wrong. So for the purposes of this stage of the analysis, I have to assume that both those claims have the required merits. For reasons which are similar to those given by Coulson J for his equivalent conclusions in Vedanta, my view is that, absent the claim against D1, Kenya would clearly be the appropriate forum for the Cs’ claim against D2. The claimants all live in Kenya. The claims all involve personal injuries to Cs, and claims for loss of property in Kenya. The relevant torts were committed in Kenya. Cs were employed by, or visitors to land owned by, D2, a company registered in Kenya. The law of Kenya applies to their claims. Similar logistical factors to those which were present in Vedanta support the view that Kenya would be the appropriate forum.

However, like Coulson J, I conclude that the existence of the claim against D1, which Cs can bring as of right in the courts of England and Wales, because D1 is domiciled in England, means that the courts of England and Wales are the appropriate forum. I reject Ds’ argument that the decision of the Court of Appeal in Red October, which, Ds argue, was made after the decisions in the cases on which Coulson J relied, can affect the analysis in any way. The facts of Red October are very different from the facts of this case, as I have already explained. I do not deduce from Red October any principle which undermines the approach taken by Leggatt J in OJSC VTB.

Jong v HSBC Private Bank (Monaco) SA [2015] EWCA Civ 1057, in which the Court of Appeal briefly considered OJSC VTB, does not suggest that I should adopt any different approach from Coulson J in Vedanta. The claim in that case against the non-domiciled defendant was governed by an exclusive jurisdiction clause, and the first instance judge had to decide what weight to give to that clause as against the claimant’s right, conferred by the Regulation, to sue the other two defendants in England. The claims against the domiciled defendants (alleged failures by them in dealing with the claimant’s complaints about the performance of the non-domiciled defendant) were wholly parasitic on the claim against the non-domiciled defendant. On conventional principles, the judge not having misdirected himself in law, the Court of Appeal upheld his approach to weighing the competing factors. In any event, the Court of Appeal did not say that Leggatt J’s in reasoning in OJSC VTB was incorrect.

Nor do I accept Ds’ argument that to permit the claim to be brought in England would be to allow ‘the tail to wag the dog’. On the assumption that both claims are well founded, the claim against D1 cannot be said to be such as to make D1 ‘a relatively minor player in the dispute’ (cf Pacific International Sports Club Limited v Soccer Marketing International Limited [2009] EWHC 1839 (Ch) per Blackburne J at paragraph 112), a case relied on by Ds. If the claim against D1 is well founded, it was responsible for devising the policies which, on Cs’ case, should have made all the difference, and for their implementation by D2; and one would expect witnesses and documents in respect of that claim to be in the United Kingdom. Moreover, Mr Hermer submitted that the claim against D1 could succeed in circumstances where the claim against D2 failed. I accept that that is logically conceivable.

It was common ground in the Pacific International case that the appropriate forum for the claim in that case was Ukraine (see paragraph 40 of the judgment of the Court of Appeal; [2010] EWCA Civ 753). As the Court of Appeal observed (judgment, paragraph 4), ‘The principal parties, their claims, their defences and the events that have generated this litigation have no connection with England, its laws or judicial system’. The only connection with England was that one of the defendants, SMI, was registered here. SMI was dissolved before the claim was started, and was restored to the register by the claimant so that it could be sued. It had no assets. The connections of the case with Ukraine are summarised at paragraph 10 of the judgment. There was no dispute, on the appeal, that the claim against SMI could proceed in England. I get no help, on the facts of this case, from the approach of Blackburne J in Pacific International.

This issue only arises on a double contingency; broadly, that my conclusions about the merits and about the appropriate forum are wrong. Like Coulson J in Vedanta, I was shown much material on this issue. In case those conclusions are wrong, I should reach, and express a view on it.

There are two main strands in Cs’ case on this issue. They are that there are significant impediments to access to justice in Kenya and that Cs will face significant risks to their lives if they litigate in Kenya. The first strand had two elements in Cs’ skeleton argument: the lack of funding in Kenya for these claims and a suggestion that there were significant delays and adjournments in Kenya. The second element was not relied on by Cs in oral argument.

Cs’ evidence on these issues is mainly in the witness statement of Mr Nderitu. He is an advocate of the High Court of Kenya. He is also Chair of the International Commission of Jurists in Kenya and was appointed by the Government of Kenya to its Task Force on Judicial Reforms. Ds’ evidence is from Mr Havelock, who is a former High Court Judge in Kenya, from Mr Ojiambo, an experienced Advocate of the High Court of Kenya and the senior partner in a commercial firm in Nairobi (he has much experience of complex litigation since he qualified in 1975), and from Mr McCourt, who has been qualified as an Advocate since 1987.

Mr Hermer was careful to emphasise in his oral submissions that it was not part of his case to suggest that the legal system in Kenya was intrinsically unjust. His point was a different one. It was that Cs would not get justice because of the particular features of their claim. He made the point which he made in Vedanta (see paragraph 175 of Coulson J’s judgment) that the House of Lords held that the claimants in Lubbe could not get justice in South Africa, despite the fact two members of the Appellate Committee were lawyers from South Africa, as were at least two of the advocates in the case. As Coulson J pointed out, ‘South Africa is the largest economy in southern Africa. It is a country where CFAs are lawful. In addition, it has one of the most developed legal systems in the world’.

Mr Hermer drew attention to various features of this claim which made it an unusual, complex claim for the Kenyan courts, and one for which no obvious funding was available. There are over 200 claimants. The nature of claims meant that they had to be handled with sensitivity. Some Cs had suffered psychological trauma and some had been raped. Group claims had been heard in Kenya but not on this scale. The case will be expensive to run; factual investigations in Kenya will be necessary, and significant disclosure of documents will be required. Cs will need to instruct a range of medical experts to provide reports on their injuries. Cs would not be able to pay privately to make these claims. They are all very poor.

A Legal Aid Act has recently been enacted (it came into force on 10 May 2016). There is, at present, no budgetary provision for it and no functioning system of legal aid in Kenya (see paragraphs 27-29 of Mr Nderitu’s witness statement and paragraph 22 of Mr Kitonga’s witness statement: he is one of the 27 Advocates in Kenya with the rank of Senior Counsel). CFAs are unlawful. No NGO would be able to fund these claims. They do not have unlimited money, and concentrate on funding cases against the Government (see paragraph 22 of witness statement of Ms Nyaundi, the chief executive officer to the Kenya Commission of Human Rights). There is no evidence that any lawyers would be willing to do the case pro bono. If Cs lost the case, they would be liable for Ds’ costs. There is no costs protection for claimants (cf the qualified one-way costs-shifting regime that operates here).

Ds submitted that even though conditional fee agreements are unlawful, ‘de facto CFAs are rife’. Mr McCourt confirmed this (witness statement, paragraphs 18 and 19). Mr Gibson drew attention to the witness statement of Mr Ojiambo. The Civil Procedure Rules provide for group actions. He has, in the past 15 years, acted for defendants in more than ten claims brought by large groups of plaintiffs; in one, there were over 600 claims for personal injuries and about employment issues. He is in the early stages of acting in another. He describes his experience of group actions, which, he says, are becoming more and more common in Kenya. In appropriate cases, there will be split trials on liability and quantum and test claims will be chosen. Mr McCourt describes acting in a claim for 23 claimants making a tort claim against a railway company. The claimants were represented by a consortium of firms. He says that there is no reason why a consortium of firms should not act in this case. Expert medical reports, in his experience, cost the equivalent of between £37 and £75. Mr Ojiambo also says that there are experts who are qualified to give evidence about psychiatric injuries in several cities in Kenya. He has used them himself. It is common practice for them not to charge fees for their work until the end of a case.

Mr Nderitu gives powerful evidence about the intimidation of witnesses and attempts to influence witnesses which caused the ICC to abandon its prosecution of four leading political figures for their alleged roles in organising and fomenting the PEV. One important witness was brutally murdered. There was an attempt to reveal the identity of a key witness, the first to give evidence. The ICC prosecutor alleged that 16 witnesses in one prosecution recanted their evidence because of intimidation and bribery. The ICC has brought charges against three people for such offences. Security officials twice tried to dissuade Mr Nderitu from meeting victims. He suspected that his phone was tapped and his office was under surveillance. The prosecution also alleged that three crucial witnesses in the other prosecution were offered bribes to withdraw their evidence. Ds accepted in the course of the procedural hearing that Cs’ claim would be seen as a claim brought by Kisii against Kalenjin. No one has been prosecuted in Kenya for offences committed in the 2007 PEV.

It is common ground that the Kenyan courts are able to make anonymity orders. Mr Nderitu, however, describes a systemic problem with insecure file management in Kenya. He gives several examples of high-profile cases in which files have gone missing or been tampered with. In one criminal case a file went missing from the court strong room. The court had ordered it to be kept there because it had already gone missing once before. The court’s judgment, quoted in paragraph 67 d of Mr Nderitu’s witness statement, refers to endemic problems and to the temptation for registry staff to engage in corrupt practices. Mr Havelock’s response to this, in his second witness statement, was to say that the cases referred to by Mr Nderitu were all criminal cases and so ‘not necessarily relevant to this case’, and to say that if he had concerns about the security of a file in a particular case, he would order it to be kept in the court strong room, or lock it in a cabinet or cupboard in his chambers, and take the key home.

On 29 November 2016, LD sent Ds’ solicitors a table of seven cases in which Mr Havelock had referred, in public judgments, to court files going missing. Mr Havelock’s third witness statement deals with these cases. He said (at paragraph 8) that paper files do sometimes go missing. The seven cases were interlocutory rulings, not final judgments, and the disappearance of files did not affect the outcome. In some cases, he would order some files to be kept in the court strong room, and in others, would keep the file in a locked cupboard in his chambers. A party can apply for a file to be kept in the court’s strong room. Sometimes missing files re-appear, which suggests that they have been misfiled. One of the seven cases was such a case.

Mr Nderitu also describes problems with judicial corruption in Kenya. In 2003, the Anti-Corruption Committee removed many judges, including five out of nine Justices of Appeal and half of the 36 High Court Judges. There was an inquiry from May 2009 which led to the Ouko Report. Mr Nderitu was involved in the Task Force. The Ouko Report (July 2010) said that ‘…corruption remains one of the greatest challenges to the judiciary’. As recently as March 2016, the Chairman of the Kenya Judges and Magistrates Vetting Board admitted publicly that the corruption considered in the most recent round of vetting was ‘only the tip of the iceberg’ of corruption in the Kenyan judiciary. Mr Nderitu refers to the current case of Judge Tunoi, who went through two rounds of vetting and is accused of taking a bribe of about £1.3m in order to decide a case in favour of a politician. There are allegations that two other Supreme Court Judges bought luxury homes in Dubai around the time when Judge Tunoi is alleged to have been bribed. A High Court Judge has been suspended since 2013 over allegations that he mishandled a case about corruption.

There are allegations by a former Chief Justice that appointments to the Judicial Services Commission are ‘riddled with corruption’. He has dismissed one official, and she, and another, face criminal corruption charges. In January 2016, he called for a ‘lifestyle audit’ of judicial officers on the basis that many enjoy lifestyles which their salaries cannot support.

Mr Havelock accepts ‘that there are ongoing issues in relation to judicial conduct in Kenya which are being dealt with by the Kenyan Government and senior judiciary and indeed I am participating in that process’. He is confident that judges appointed or vetted since 2010 will be able to provide justice for Cs. Cs’ argument, as I understand it, however, is not that the judges hearing the claim would be corrupt, but that there is a risk, because of endemic corruption among judicial officers, that, if an anonymity order were made, their identities might, nonetheless, be disclosed.

I cannot make a factual assessment, on the balance of probabilities, about whether or not Cs will get substantial justice in Kenya (Cherney v Deripaska [2009] EWCA Civ 849 paragraphs 27-29). What I can do, rather, is to assess the risk that they will not get substantial justice. I accept that these claims could in theory be litigated in Kenya in the sense that procedures exist for litigating group actions, and that there is sufficient local expertise to enable the OLA (K) claim against D2 to be brought, and to enable suitable medical experts to advise on Cs’ injuries. I accept that no local firm is big enough to litigate this claim on its own, and that it would therefore be necessary for Cs to be represented by a consortium of firms. This would not be ideal, but it would not deprive Cs of substantial justice.

However, I am persuaded that there is cogent evidence of two linked factors, and that there is a third, which, in combination with the first two, mean that there is a real risk that Cs will not get substantial justice in Kenya. Cs are the victims of terrible violence in 2007 which was based on a combination of their ethnicity and perceived political affiliation. Their case is not seen in Kenya as a claim by Cs against their employer or against a foreign multi-national. It is seen as a case brought by Kisii against Kalenjin. The evidence shows that ethnic and political rivalries still cause violence in Kenya. The evidence also shows that, apart from some civil claims against state organs such as the police, no-one has been held to account for the crimes committed in the PEV in 2007. In particular, significant intimidation and bribery led to the abandonment of two prosecutions by the ICC of four politicians. Cs are able to identify some of their attackers who are still employed by D2. I am persuaded that if Cs were to litigate this claim in Kenya there is a real risk that, unless anonymity orders were made, they would be exposed to further violence from Kalenjin. But even if such orders were made, there is no provision, and no precedent for, a confidentiality club in Kenya.

I am also persuaded that there is a continuing problem with judicial corruption in Kenya, as Mr Havelock accepts. I take into account Mr Havelock’s evidence that he knows the judges who would be likely to hear this claim and his views about their impeccable integrity. But once it is common ground among the experts that there is a continuing problem with judicial corruption, in other words, once it is accepted that the system in Kenya is not free from corruption, it cannot be said that there is no risk that corruption of judicial officers might affect the conduct of the case. It is also not disputed that files go missing in the Kenyan system; Mr Nderitu quotes a Kenyan judge directly associating missing files with corruption. I consider that there is cogent evidence that there is a real risk that if anonymity orders were made in any litigation in Kenya, that those orders could be breached or circumvented by those who wished to discourage Cs from bringing this claim, and were prepared to use corruption, and violence or intimidation to do so.

The third factor is the evidence about funding litigation like this. I must be careful not to approach this on the basis that Cs are necessarily entitled, in Kenya, to representation of the quality which they have undoubtedly had so far in England. They are not entitled to ‘a Rolls Royce presentation of [their] case, as opposed to more rudimentary representation in the appropriate forum’. Ds argued that the claim could be litigated in Kenya without the need for extensive expert evidence and suggested that the claim against D1 should not be brought at all, as it was a superfluous complication to a straightforward claim against D2 under OLA (K).

Nonetheless, I consider (on the assumption that I am wrong about the merits of their claims) that Cs are entitled to sue both Ds. I also consider that, while it is likely some corners could be cut, the claims will be expensive and complex to prepare. I consider that, on the evidence, there is a real risk that Cs would not be able to afford to bring these claims in Kenya. I do not consider that they should be required to make unlawful arrangements for conditional fee agreements. There is no functioning legal aid system. There is no evidence which satisfies me that money could be found to enable Cs to bring and prosecute these claims even as far as a trial on liability of a small number of test cases.

For these reasons, my conclusion is that there is a real risk that Cs could not obtain substantial justice in Kenya.

For the reasons I have given at some length, I have reached three broad conclusions.