Braganza v BP Shipping Limited and another

Between 01.00 and 07.00 on 11 May 2009, Mr Renford Braganza, Chief Engineer on BP’s oil tanker the British Unity, then in the mid-North Atlantic, disappeared. No-one knows for certain what happened to him. But his employers formed the opinion that the most likely explanation for his disappearance was that he had committed suicide by throwing himself overboard. This would mean that his widow was not entitled to the death benefits provided for in his contract of employment. Clause 7.6.3 of that contract provided relevantly as follows:

It is not the task of this or any other court determining a claim under such a contract to decide what actually happened to Mr Braganza. The task of the court is to decide whether his employer was entitled to form the opinion which it did. The issue of general principle in this appeal, therefore, is the test to be applied by the court in deciding that question.

Mr Braganza was an able and well qualified Chief Engineer. Like all the crew of the MV British Unity, he was an Indian national. He was a Roman Catholic and married with two children. In July 2008, the family had moved from India to Toronto in Canada and he had taken extended leave for this purpose. After returning to work with BP, he joined the British Unity in Gibraltar in February 2009. The vessel’s main engine had broken down in August 2008 and been repaired but the damaged cylinder liners had not been replaced. This major work was done in April 2009 in Ferrol, Spain. Shortly after leaving Ferrol, the cooling water jacket of one of the cylinders began to leak. The vessel then proceeded via Falmouth to Brofjorden in Sweden, where a cargo of unleaded gasoline was loaded and also two spare cooling water jackets, which were stored in an alleyway on the main deck, close to a hatch providing access to the engine room below.

After sailing from Brofjorden, Mr Braganza received an email on 5 May 2009 from an engineering superintendent asking him to carry out a scavenge inspection/ring inspection of the engine in about six days’ time and suggesting that it would be prudent to replace the cooling water jacket at the same time as the engine was stopped for that purpose. Mr Braganza replied that he would do the inspection in about six days’ time and try to carry out the jacket replacement at the earliest opportunity. To do this, it would first be necessary to lower the cooling water jacket through the hatch into the engine room below. Ms Belinda Bucknall QC, for Mrs Braganza, stresses the “highly weather-sensitive nature” of the cooling water jacket exchange. (Indeed, Mr Williamson, the engineering superintendent who had supervised the major works in Ferrol had intended to telephone Mr Braganza on the morning of 11 May to suggest that he think about the operation “because obviously the lifting of a one and half tonne cylinder head in any sort of swell would be risky”.) The main engine would have to be shut down for several hours, leaving the vessel at the mercy of wind and wave.

The vessel was originally bound for Jebel Ali in the United Arab Emirates, but on 7 May, while in the middle of the Bay of Biscay, it was ordered to proceed to New York instead. It therefore altered course and headed across the Atlantic. On 9 May, the weather worsened during the day and the log recorded at 20.00 a rough sea and high swell and shipping water on deck. On 10 May the weather began to improve. At 19.00, the Master, the Chief Officer, Mr Braganza and the second engineer met to discuss the plans for the next day. These included “lowering liner jacket to engine room from main deck” and “M/E scavenge inspection and liner jacket renewal”. It was agreed that the Master and Chief Officer would check the weather in the morning.

At about 23.30 that same night, the Master and Mr Braganza met for about an hour. They discussed the weather conditions for the next day, with the assistance of a weather routing report which the Master had sent to Mr Braganza’s computer. During the meeting, Mr Braganza emailed an engineering superintendent to report that he intended to stop the main engines the next day and do a scavenge inspection and that, weather permitting, he would like to change the liner jacket at the same time. He was advised to go ahead “if weather and schedule permits”. The Master left Mr Braganza’s cabin at about 00.30 on 11 May. That was the last anyone saw of the Chief Engineer. At 01.00 he sent a routine email to the second engineer. At 07.00 it was noticed that his cabin door was open, as it habitually was unless he was sleeping. His bed looked as if it had been slept in. He did not breakfast as usual in the officers’ mess.

The Chief Officer carried out a risk assessment around 08.00 and the Master agreed that the job could be carried out. The Judge concluded that it was unlikely that there was no pitching or rolling but that it was not such as to make the planned operations unsafe. Lowering the cooling water jackets into the engine room began shortly after 08.00 and was safely completed by about 09.30. Mr Braganza was not there. Soon afterwards, the Master announced that he was missing. A search was made on board but he was not found. The vessel was turned around and a search and rescue operation conducted but he was not found. The cooling water jacket replacement was postponed until some days later.

On arrival in New York, an investigation was carried out on behalf of the Isle of Man where the British Unity was registered. The ship’s crew were interviewed under caution. Its conclusion was that Mr Braganza was lost overboard, presumed drowned, but no finding was made as to the reason for this. BP then set up its own inquiry, in accordance with its own procedures, to “investigate the relevant circumstances leading up to the loss of Mr Braganza, identify if possible the root causes of the incident and identify any changes required to the BP Shipping Safety Management System”. The five person team took about four months to make their extensive inquiries. Their eventual report was dated 17 September 2009. It considered five “possible scenarios”. The team were able to discount three explanations – hiding or being hidden on board, collection by another vessel and fall from vessel due to horseplay, altercation or foul play. That left an accidental fall from the vessel, which could not be discounted, and suicide.

Under the heading “suicide”, the team made six “bullet points” which led them to consider that suicide was a possibility. There was much criticism of each of these points in the courts below and the conclusions of each are briefly summarised below:

Mr Braganza’s “behaviour was reported to be notably different on this voyage than on previous voyages”: (a) he was quiet and withdrawn, (b) there were no clean officer’s uniforms in his cabin, and (c) his attention to detail in record keeping had slipped.

The shoes and sandals he usually wore on board were found in his cabin after his disappearance.

“Several e-mail messages received from his immediate family ... suggest [he] had some family and/or financial difficulties that were causing him concern”.

He was not aware, before joining the ship, of its status and reputation and was reported to be unhappy about this.

He considered himself eligible for the 2008 bonus which had been paid to him, but the employer had later advised him that it would be withdrawn.

There were indications that the watertight door from the accommodation block that opens onto the upper deck on the starboard side may have been opened during the hours of darkness on the morning of 11 May. No member of the crew reported opening this door.

The team’s report went through several drafts. The first version did not mention suicide. After exchanges with BP Legal, the final version concluded that “Having regard for all the evidence the investigation team considers the most likely scenario to be that the C/E jumped overboard intentionally and therefore took his own life”. The judge rejected the suggestion that there had been a “fundamental shift” between the first and final versions. The team had always felt that suicide was the most likely explanation. However, they could not rule out that he had gone on deck for a work related reason and that his fall had been accidental. The judge commented that “What can perhaps be said is that the team’s initial reluctance to identify suicide as ... the likely cause of death suggests that such a conclusion was not clearly proven in the minds of the team” (para 88).

The report was produced by and for BP Shipping Limited, who owned and managed the vessel. It was then forwarded to Mr Sullivan, General Manager of BP Maritime Services (Singapore) Pte Ltd, a Singapore Company which provides management services to the shipping company and employed the officers on board the vessel. Mr Sullivan made no further enquiries of his own. On the basis of the team’s report, he concluded that there had been “wilful default” within the meaning of Clause 7.6.3 of the contract of employment and thus that death in service benefits were not payable to Mrs Braganza. BP Legal informed her solicitors of this on 13 November 2009.

Mrs Braganza brought a claim in contract against the employer for death benefits amounting to US$230,265. She also brought a claim in tort under the Fatal Accidents Act 1976 and the Law Reform (Miscellaneous Provisions) Act 1934 for damages quantified as US$1,325,945, alleging that the death was caused by negligence on the part of both the shipping company and the employer. BP produced a supplemental report in answer to the claims (see paras 83 to 85 of Lord Neuberger’s judgment). Mr Sullivan maintained his decision. The action was tried over 8 days in 2012 before Teare J, sitting as a judge of the Admiralty Court: [2012] EWHC 1423 (Comm).

Teare J was unable to make a finding as to the cause of Mr Braganza’s death (para 60): there was a real possibility, but it was not more likely than not, that he had fallen overboard (para 57); but the evidence was not sufficiently cogent to warrant a finding of suicide on the balance of probabilities (para 58). He rejected the claim in tort on the ground that there had been no breach of duty in deciding to carry out the two operations that day (para 64) and that, even if there had been a breach of duty, it had not caused the death (para 65). Mrs Braganza has not appealed against that conclusion.

Teare J upheld the contractual claim. It was common ground between the parties that the opinion formed by the employer had to be “reasonable” (para 76). This meant reasonable in the sense in which that expression is used in Associated Provincial Picture Houses Ltd v Wednesbury Corporation. [1948] 1 KB 223, which had been applied to the exercise of a contractual discretion (by a P & I Club) in The Vainqueur José [1979] 1 Lloyds Rep 557 (para 91). As Clause 7.6.3 was in the nature of an exception or exclusion clause, the burden of proving that the opinion was a reasonable one lay with the employer (para 93). The investigation team did not direct themselves that “before making a finding of suicide there should be cogent evidence commensurate with or proportionate to the seriousness of a finding of suicide” (para 94). They, and therefore Mr Sullivan, failed to take into account the real possibility that Mr Braganza had gone out on deck in order to check the weather to see whether it was safe to carry out the planned work (para 96). They had stated in the first draft of their report that they “could not find any work related reason for him to be outside during this time”. His interest in the weather furnished such a reason and they should have taken it into account (para 97). This mattered because they had concluded that an accidental fall could not be discounted (para 98). Thus they were not properly directed in law and failed to take into account a relevant matter when forming their opinion (para 99). BP had conceded that if the opinion was not reasonable, the contractual claim should succeed (rather than the matter be sent back to them for reconsideration) (para 93). Having reached that conclusion, nevertheless Teare J very properly went on to consider the reasons advanced against the team’s conclusion and formed the views outlined in para 9 above on the six “bullet points”.

The employer appealed: [2013] EWCA Civ 230. Longmore LJ (with whom Rimer and Tomlinson LJJ agreed) thought it “not entirely clear” whether the judge had considered that the failure of the team, and Mr Sullivan, to direct themselves as to the need for cogent evidence before making a finding of suicide was in itself enough to render their opinions unreasonable (para 9). But, as Mocatta J had pointed out in The Vainqueur José, at 577, “it would be a mistake to expect [of a lay body] the same expert, professional and almost microscopic investigation of the problems, both factual and legal, that is demanded of a suit in a Court of Law” (para 13). It could not be the law that a non-lawyer such as Mr Sullivan had to give himself directions before forming his opinion (para 14). It would be impossible for him to give himself such a direction without taking legal advice of a kind which cannot have been contemplated by the requirements of the death benefit clause (para 20).

As to the failure to appreciate that there might be work-related reasons for Mr Braganza to go on deck, that failure could not make the employer’s opinion unreasonable in the absence of a mechanism explaining how he could accidentally fall overboard (para 22). Thus Longmore LJ was unable to agree with either of Teare J’s reasons for saying that the opinion was unreasonable (para 23). He therefore went on to consider and reject the wider ranging attack mounted upon the six “bullet points” (see para 9 above). The conclusion of suicide was a reasonable one in all the circumstances (para 34).

This case raises two inter-linked questions of principle, one general and one particular. The particular issue is the proper approach of a contractual fact-finder who is considering whether a person may have committed suicide. Does the fact-finder have to bear in mind the need for cogent evidence before forming the opinion that a person has committed suicide? The general issue is what it means to say that the decision of a contractual fact-finder must be a reasonable one. There are many statements in the reported cases to the effect that the principles are well-settled and well-understood, but this case illustrates that all is not as clear or as well understood as it might be.

Contractual terms in which one party to the contract is given the power to exercise a discretion, or to form an opinion as to relevant facts, are extremely common. It is not for the courts to re-write the parties’ bargain for them, still less to substitute themselves for the contractually agreed decision-maker. Nevertheless, the party who is charged with making decisions which affect the rights of both parties to the contract has a clear conflict of interest. That conflict is heightened where there is a significant imbalance of power between the contracting parties as there often will be in an employment contract. The courts have therefore sought to ensure that such contractual powers are not abused. They have done so by implying a term as to the manner in which such powers may be exercised, a term which may vary according to the terms of the contract and the context in which the decision-making power is given.

There is an obvious parallel between cases where a contract assigns a decision-making function to one of the parties and cases where a statute (or the royal prerogative) assigns a decision-making function to a public authority. In neither case is the court the primary decision-maker. The primary decision-maker is the contracting party or the public authority. It is right, therefore, that the standard of review generally adopted by the courts to the decisions of a contracting party should be no more demanding than the standard of review adopted in the judicial review of administrative action. The question is whether it should be any less demanding.

The decided cases reveal an understandable reluctance to adopt the fully developed rigour of the principles of judicial review of administrative action in a contractual context. But at the same time they have struggled to articulate precisely what the difference might be. In Abu Dhabi National Tanker Co v Product Star Shipping Ltd (The “Product Star”) (No 2) [1993] 1 Lloyd’s Rep 397, 404, after contrasting the position in judicial review, Leggatt LJ explained that:

That was in the context of a ship-owner’s decision as to whether a port to which a vessel was directed was dangerous. In Paragon Finance plc v Nash [2001] EWCA Civ 1466, [2002] 1 WLR 685, the court had to consider whether there was any implied term limiting the power of a mortgagee to set interest rates under a variable rate mortgage. Dyson LJ had no difficulty in holding (at paras 32 to 36) that it was necessary, in order to give effect to the reasonable expectations of the parties, to imply a term that the power would not be exercised dishonestly, for an improper purpose, capriciously or arbitrarily. He went on to discuss whether there should also be a term that the power would not be exercised unreasonably. He concluded that there had been a “somewhat reluctant” extension of the implied term to include “unreasonableness that is analogous to Wednesbury unreasonableness” (paras 37 to 42).

These authorities, together with Ludgate Insurance Co Ltd v Citibank NA [1998] Lloyd’s Rep IR 221, 239-240, and Gan Insurance Co Ltd v Tai Ping Insurance Co Ltd (No 2) [2001] EWCA Civ 1047, [2001] 2 All ER (Comm) 299, at paras 64, 67, 73, are helpfully summarised by Rix LJ in Socimer International Bank Ltd v Standard Bank London Ltd [2008] EWCA Civ 116, [2008] Bus LR 1304. In his conclusion, at para 66, he substitutes the more modern term “irrationality” for unreasonableness:

The same point was made (albeit in a completely different context, and so obiter) by Lord Sumption in Hayes v Willoughby [2013] UKSC 17, [2013] 1 WLR 935, at para 14:

The problem with this formulation, which is highlighted in this case, is that it is not a precise rendition of the test of the reasonableness of an administrative decision which was adopted by Lord Greene MR in Associated Provincial Pictures Houses Ltd v Wednesbury Corporation [1948] 1 KB 223, 233-234. His test has two limbs:

The parties in this case disagree as to whether the term to be implied into this contract includes both limbs. Mrs Braganza argues that the employer must “keep within the four corners of the matters which they ought to consider”, while the employer argues that its decision may only be impugned if it is a decision which no reasonable employer could have reached.

Mrs Braganza can pray in aid the approach of Mocatta J in The Vainqueur José. He held that the common law principles applicable to the exercise of a contractual discretion include fairness, reasonableness, bona fides and absence of misdirection in law (p 574). He later quoted (p 575), without reservation, Lord Greene’s summary of the public law concept of reasonableness. There is nothing on Mocatta J’s judgment to suggest that only the second of those elements is applicable to the exercise of a contractual discretion. He did (at 574) contrast the contractual principles with the principles applicable to the exercise of a statutory discretion by Ministers of the Crown, but on the basis that, in addition, the Minister’s decision had to be consistent with the objects and other provisions of the statute in question, citing Laker Airways Ltd v Department of Trade [1977] QB 643.

On that point, on the other hand, in Hayes v Willoughby, just before the passage quoted in para 23 above, Lord Sumption stated that rationality “has … in recent years played an increasingly significant role in the law relating to contractual discretions, where the law’s object is also to limit the decision-maker to some relevant contractual purpose”: [2013] 1 WLR 935, para 14. This is consistent with his earlier observations in British Telecommunications Plc v Telefónica O2 UK Ltd [2014] UKSC 42, [2014] Bus LR 765, at para 37:

There are signs, therefore, that the contractual implied term is drawing closer and closer to the principles applicable in judicial review. The contractual cases do not in terms discuss whether both limbs of the Wednesbury test apply. However, in Gan Insurance, where the issue was the limits, if any, to the reinsurers’ power to withhold approval to the insured’s agreement to settle a claim, Mance LJ first commented that “what was proscribed was unreasonableness in the sense of conduct or a decision to which no reasonable person having the relevant discretion could have subscribed” (para 64); but he concluded that “any withholding of approval by reinsurers should take place in good faith after consideration of and on the basis of the facts giving rise to the particular claim and not with reference to considerations wholly extraneous to the subject matter of the particular reinsurance ….” (para 67).

If it is part of a rational decision-making process to exclude extraneous considerations, it is in my view also part of a rational decision-making process to take into account those considerations which are obviously relevant to the decision in question. It is of the essence of “Wednesbury reasonableness” (or “GCHQ rationality”) review to consider the rationality of the decision-making process rather than to concentrate upon the outcome. Concentrating on the outcome runs the risk that the court will substitute its own decision for that of the primary decision-maker.

It is clear, however, that unless the court can imply a term that the outcome be objectively reasonable – for example, a reasonable price or a reasonable term – the court will only imply a term that the decision-making process be lawful and rational in the public law sense, that the decision is made rationally (as well as in good faith) and consistently with its contractual purpose. For my part, I would include both limbs of the Wednesbury formulation in the rationality test. Indeed, I understand Lord Neuberger (at para 103 of his judgment) and I to be agreed as to the nature of the test.

But whatever term may be implied will depend upon the terms and the context of the particular contract involved. I would add to that Mocatta J’s observation in The Vainqueur José, that “it would be a mistake to expect [of a lay body] the same expert, professional and almost microscopic investigation of the problems, both factual and legal, that is demanded of a suit in a court of law” (577). Nor would “some slight misdirection” matter, at least if it were clear that, had the legal position been properly appreciated, the decision would have been the same. It may very well be that the same high standards of decision-making ought not to be expected of most contractual decision-makers as are expected of the modern state.

However, it is unnecessary to reach a final conclusion on the precise extent to which an implied contractual term may differ from the principles applicable to judicial review of administrative action. Given that the question may arise in so many different contractual contexts, it may well be that no precise answer can be given. The particular context of this case is an employment contract, which, as Lord Hodge explains, is of a different character from an ordinary commercial contract. Any decision-making function entrusted to the employer has to be exercised in accordance with the implied obligation of trust and confidence. This must be borne in mind in considering how the contractual decision-maker should approach the question of whether a person has committed suicide.

Teare J directed himself, in relation to his own decision as to the cause of Mr Braganza’s disappearance, that “before a finding of suicide is made there must be evidence of sufficient cogency commensurate with or proportionate to the seriousness of the finding” (para 46), citing the observation of Watkins LJ in R v West London Coroner, Ex p Gray [1988] QB 467, 477-478, that suicide is “still a drastic action which often leaves in its wake serious social, economic and other consequences”. He also directed himself, following the House of Lords’ decision in The Popi M [1985] 1 WLR 948, 955-956, that, where two improbable causes are suggested, he was not bound to make a finding one way or another. I agree with Lord Neuberger, at para 100 of his judgment, that it is also perfectly proper for the employer to conclude that he or she is unable to form an opinion as to the cause of death. But the question is how he or she should go about making a positive finding of suicide.

Longmore LJ pointed out that the direction based on Gray “might itself be said to be a little outdated since the decisions in In re H (Minors) (Sexual Abuse: Standard of Proof) [1996] AC 563 and In re B (Children) (Care Proceedings: Standard of Proof) [2009] AC 11, which have emphasised that in civil cases there is only one standard of proof, viz. the balance of probabilities” (para 15). Those cases make it clear that there is not a sliding scale of probability to be applied, commensurate with the seriousness of the subject-matter or the consequences of the decision. The only question is whether something is more likely than not to have happened. Lord Hoffmann put it thus in In re B at para 15:

Some things are inherently a great deal less likely than others. The more unlikely something is, the more cogent must be the evidence required to persuade the decision-maker that it has indeed happened. As Lord Nicholls explained in In re H, at 586,

However, Longmore LJ also took the view that the employer did not have to approach the matter in this way. I respectfully disagree. The employer is entrusted with making a decision which has serious consequences for the family of a deceased employee. It deprives them of what would otherwise be a contractual right. There is no reason why the employer should not approach that decision in the same way that any other decision-maker should do. On the contrary, in view of the special nature of the employment relationship, there is every reason why they should do so. Employers can reasonably be expected to inform themselves of the principles which are relevant to the decisions which they have to make. Employment law is complicated and demanding in many legal systems, but employers are expected to know it. They can also reasonably be expected to know how they should approach making the important decisions which they are required or empowered to make under the terms of the employment contract. In my view, a decision that an employee has committed suicide is not a rational or reasonable decision, in the terms discussed above, unless the employer has had it clearly in mind that suicide is such an improbability that cogent evidence is required to form the positive opinion that it has taken place.

The employer now accepts that it is for him to show that the decision which it reached was a reasonable decision in the sense which is required by the contract.

In my view, Mr Sullivan should not simply have accepted the view of the investigation team that suicide was the most likely explanation for Mr Braganza’s disappearance. The team had been conducting their investigation for purposes which were different from the purpose of his decision. Their purpose was to see whether BP’s systems could be improved. They could and did make recommendations about the steps to be taken to support officers who might be experiencing financial or emotional problems. Those recommendations were equally valid and sensible whether or not Mr Braganza had in fact committed suicide.

Mr Sullivan’s task was quite different. He had to consider whether he was in a position to make a positive finding that Mr Braganza had committed suicide. He should have asked himself whether the evidence was sufficiently cogent to overcome the inherent improbability of such a thing. In my view that can be expected of any employer making a decision under a provision such as this. But it could certainly be expected of BP, which clearly had access to in-house legal expertise to guide it in the decision-making process.

In this case, there were no positive indications of suicide. There was no suicide note, no evidence of suicidal thoughts (apart perhaps from his wife’s reference six weeks earlier to his seeming “so afraid of life”), no evidence of over-whelming personal or financial pressures of the sort which would be likely to lead a mature professional man to take his own life, no evidence of psychiatric problems or a depressive personality. The “bullet points” are at most straws in the wind. The two most significant are the emails and the record-keeping deficiencies. The cogency of the emails from Mrs Braganza is much diminished by the failure to ask her about them. The team’s failure to do so is completely understandable, given the task which had been set for them. But the employer’s failure to do so is much less understandable. Nor do the record-keeping deficiencies appear to have been explored in any depth.

Against those straws in the wind is the evidence that Mr Braganza’s behaviour had appeared entirely normal to the Master and the other officers with whom he was in contact the night before. There was also a good deal of evidence of his concern about the weather, which would have constituted a good work-related reason for him to go on deck that morning. A further relevant factor which ought to have been in the mind of this employer is that Mr Braganza was a Roman Catholic. There are cultures in which suicide is an acceptable, even an honourable, solution to certain problems or dilemmas. But his was not one of them. For him, suicide was a mortal sin. This increases its inherent improbability in his case and the corresponding need for cogent evidence to support a positive finding.

Although I would not have phrased the correct approach exactly as Teare J phrased it, in my view he was right to conclude (para 95) that the investigation team’s report and conclusion could not be regarded as sufficiently cogent evidence to justify Mr Sullivan, and hence BP, in forming the positive opinion that he had committed suicide. No-one suggests that his decision was “arbitrary, capricious or perverse”, but in my view it was unreasonable in the Wednesbury sense, having been formed without taking relevant matters into account.

I would therefore allow this appeal, with the result that Mrs Braganza’s claim in contract (for the comparatively modest sum of US$230,265, albeit with interest) succeeds.

I agree with Lady Hale that this appeal should be allowed.

For the sake of brevity I do not set out again the facts which Lady Hale and Lord Neuberger have summarised at paras 1 to 16 and 66 to 96 respectively. For the sake of simplicity I refer to the two BP companies involved in this appeal as “BP” without drawing distinctions between them, except in the next paragraph in which I identify Mr Braganza’s employer.

Mr Renford Braganza’s contract of employment entitled him to compensation in the form of a death in service benefit if he died in the employment of BP Maritime Services (Singapore) Pte Ltd, the second defendants (Clause 7.3.1). Any sum payable as compensation was to be paid to his nominated beneficiary, in this case his widow, the claimant and appellant. That entitlement was subject to Clause 7.6.3 which provided (so far as material)

BP denied Mrs Braganza the compensation because it was of the opinion that Mr Braganza had committed suicide. That opinion, which Mr Sullivan reached on behalf of BP, was based on a report of the investigation team and his subsequent confirmation of his decision was based on the second report of that team after Mrs Braganza challenged the decision. As Lady Hale has shown in para 9, the team’s conclusion that suicide was the most likely explanation for Mr Braganza’s disappearance was based on six points which it set out as bullet points in its report.

The task which the investigation team and Mr Sullivan faced was, as Lord Neuberger says, to decide how an unlikely event, which undoubtedly occurred, was actually caused. As a result of the detailed investigations there were only two realistic possibilities: accident or suicide.

I am struck by the paucity and the insubstantial nature of the evidence from which BP inferred that Mr Braganza committed suicide. While the six points must be considered in aggregate, the only ones which seem to me to be capable of bearing any weight are (a) his lack of timely record keeping on his last voyage, (b) the evidence in Mrs Braganza’s emails of his financial worries since his family had settled in Canada and (c) his concerns about the state of repair of his ship and the workload which fell on him as chief engineer as a result. Evidence of some moodiness during the voyage and irritation over the refusal of a bonus added little to the picture. I agree with Lady Hale’s description of the six points on which the investigation team relied as “straws in the wind”: para 40. Unsurprisingly, the team could not rule out the possibility that Mr Braganza had gone on deck for some work related reason and that he had fallen into the sea by accident.

Against the conclusion which BP reached is the inherent improbability of suicide. Mr Braganza had no history of depression or mental disorder. There was no evidence that he had spoken to anyone about suicide. On 13 February 2009, before commencing his last voyage, he had undergone a medical examination and was pronounced physically and psychologically fit. His colleagues who saw him in the 36 hours before his disappearance expressed no concerns about his appearance or behaviour. The ship’s master, who was the last person to see him alive, also had no concerns about his demeanour at 0030 hours when they discussed the arrangements for the planned inspection and repairs which were scheduled for the following morning. Mr Braganza left no suicide note or message that indicated any intention to kill himself. Further, he and his wife were devout Roman Catholics, for whom suicide is a mortal sin.

Accordingly, I readily understand Teare J’s conclusion, in the part of his judgment in which he considered Mrs Braganza’s claim in tort, that the evidence before the court was not sufficiently cogent to warrant a finding of suicide on the balance of probabilities. But the issue in this case is not what a court would decide if determining the matter at first instance. As Teare J recognised later in his judgment when he addressed Mrs Braganza’s claim in contract, the decision maker was BP. The principal issue is: “when is the court entitled to intervene in the exercise of a contractual discretion?”.

As Lady Hale has shown (paras 18-31), the court is not entitled to substitute its own view of what is a reasonable decision for that of the person who is charged with making the decision; it conducts a rationality review. In Clark v Nomura International Plc [2000] IRLR 766, a case concerning an employee’s entitlement to a discretionary bonus, Burton J stated (at para 40):

Like Lady Hale, with whom Lord Neuberger agrees on this matter (para 103), I think that it is difficult to treat as rational the product of a process of reasoning if that process is flawed by the taking into consideration of an irrelevant matter or the failure to consider a relevant matter. While the courts have not as yet spoken with one voice, I agree that, in reviewing at least some contractual discretionary decisions, the court should address both limbs of Lord Greene’s test in Associated Provincial Picture Houses Ltd v Wednesbury Corporation [1948] 1 KB 223, 233-234.

In my view it is clearly appropriate to do so in contracts of employment which have specialties that do not normally exist in commercial contracts. In Johnson v Unisys Ltd [2003] 1 AC 518 (at para 20) Lord Steyn stated:

The personal relationship which employment involves may justify a more intense scrutiny of the employer’s decision-making process than would be appropriate in some commercial contracts.

The scope for such scrutiny differs according to the nature of the decision which an employer makes. In this case Clause 7 gave the employee a prima facie entitlement to the death benefit unless BP could satisfy itself of a factual circumstance which excluded the benefit. The nature of the exercise which BP had to undertake in deciding the cause of death was very different from the assessment of whether an employee was entitled to a discretionary bonus, which is an exercise that involves a qualitative judgement of the employee’s performance.

In cases such as Clark v Nomura International Plc, Keen v Commerzbank AG and Horkulak v Cantor Fitzgerald International [2005] ICR 402 the courts have reviewed contractual decisions on the grant of performance-related bonuses where there were no specific criteria of performance or established formulae for calculating a bonus. In such cases the employee is entitled to a bona fide and rational exercise by the employer of its discretion. The courts are charged with enforcing that entitlement but there is little scope for intensive scrutiny of the decision-making process. The courts are in a much better position to review the good faith and rationality of the decision-making process where the issue is whether or not a state of fact existed, such as whether an employee’s wilful act caused his death. The decision of the employer is not a judicial determination and the court cannot expect judicial reasoning. But I see no reason why an employer’s decision-making should be subject to scrutiny that is any less intense than that which the court applies to the decision of a public authority which is charged with making a finding of fact. A large company such as BP is in a position to support its officials with legal and other advisory services and should be able to face such scrutiny.

The investigation team and Mr Sullivan appear not to have considered the real possibility of accident if Mr Braganza had a work-related reason to be on deck after dawn, namely to check the weather conditions which would affect the planned replacement of the cooling water jacket, and had acted carelessly while standing at or even on the deck railings. While there was evidence that Mr Braganza was generally very conscious of safety, people do unexpected things and unforeseen accidents occur; accidents are often difficult to predict, as Teare J stated at para 56.

The focus of the team’s second report was on responding to Mrs Braganza’s suggestion that her husband might have been washed overboard or slipped while inspecting engine spares on the deck and fallen over or between the railings as a result of the rolling of the ship. The team found no evidence in support of such an accident. The weather conditions did not support such hypotheses. The railings were in good repair and adequate to prevent someone falling on the deck and sliding overboard. Those railings would also protect someone from falling overboard unless they acted carelessly. But if Mr Braganza had behaved carelessly at the railings, there would in all probability have been no evidence of the cause of his fall which could be set against the exiguous evidence pointing to suicide.

Given the improbability of suicide in this case, I agree with Lady Hale (para 36 above) there had to be cogent evidence to overcome that improbability. Not only do I not see that evidence but also I do not detect any consideration of both the possibility of Mr Braganza having acted carelessly while at the railings and that there would in all probability be no evidence of such behaviour. On those bases the appeal succeeds.

There is a further point. I do not rely on it in reaching my decision because it was not argued in this appeal and it may merit further argument in a suitable case. I think that the employer’s obligation of trust and confidence may provide a further reason for requiring cogent evidence. A finding that an employee has committed suicide carries a stigma for the spouse of the deceased employee. Teare J in his judgment (at para 46) cited Watkins LJ in R v West London Coroner, Ex p Gray [1988] QB 467 (at pp 477-478) where he stated that suicide is

Whether or not the obligation of trust and confidence imposes this constraint on an employer, I am satisfied that there was not sufficient evidence of suicide in this case to outweigh its inherent improbability and Teare J was justified in so finding. Absent cogent evidence to support it, Mr Sullivan and thus BP should not have made a finding of suicide.

I do not question the good faith of the investigation team, who carried out a thorough investigation under a different remit, or of Mr Sullivan, who made the positive finding on behalf of BP. Nor do I think that BP acted unfairly in the manner it carried out the task. But, like Lady Hale and Teare J, I do not think that the report of the investigation team gave Mr Sullivan the evidential basis for forming the positive opinion that Mr Braganza had committed suicide.

I would therefore allow the appeal.

The ultimate issue on this appeal is whether Niloufer Braganza, the widow of Renford Braganza, is disentitled from obtaining a death-in-service payment from his employer (whom I will refer to as BP) following his death at sea, on the ground that his death was caused by suicide. In agreement with the Court of Appeal, and respectfully differing from the majority of the court and from Teare J, I consider that Mrs Braganza is disentitled from obtaining the benefit in question.

In order to explain my reasons for reaching this conclusion, I will set out the significant facts as I see them, although I am conscious that this means a degree of overlap with Lady Hale’s judgment. Mr Braganza had been employed by BP since 2004 as Chief Engineer. His terms of employment were set out in a contract, of which only parts of clause 7 (“Clause 7”) are relevant for present purposes. The effect of clause 7.3 was that, in the event of his death “whilst in the employment of [BP]”, Mr Braganza’s “nominated beneficiary” (in this case Mrs Braganza, the claimant in these proceedings) would be entitled to compensation equal to “three times [his] annual salary”. However, this entitlement was subject to certain exceptions. One of those exceptions was set out in clause 7.6.3 which stipulated that no such compensation would be payable if, inter alia, “in the opinion of [BP] or its insurers, the death … resulted from … [his] wilful act, default or misconduct whether at sea or ashore …”.

On 9 February 2009, Mr Braganza joined the MV “British Unity”, a 183 metre long, 46,803 mt deadweight oil tanker, at Gibraltar. Unusually, and contrary to standard practice, he was not given a briefing, and in particular he was not told that the vessel needed substantial repairs. Major work was carried out while the vessel was in Ferrol, Spain between 18 and 22 April 2009, an operation for which Mr Braganza was responsible, albeit under supervision. After leaving Ferrol, further problems came to light and the vessel berthed in Falmouth for work and then went on to Brofjorden in Sweden for further work. The vessel was then loaded with unleaded gasoline, and set sail for Jebel Ali, but was then ordered to divert to New York. On 11 May 2009, Mr Braganza was lost overboard when the vessel was in the mid-North Atlantic, en route for New York, between 01.00 (when he sent an email) and 07.00 (when his cabin was seen to be empty).

Mr Braganza was an able and experienced engineer, a staunch Roman Catholic and a strong family man, with a wife and two children, and he telephoned his family when he was at sea every two or three days. He had been found to be physically and psychologically fit when examined in February 2009.

During the 36 or so hours before his disappearance, Mr Braganza was described as “normal” by those who saw him. On the evening of 9 May 2009, he was the caller in a game of bingo played by a number of the crew members, and, according to the Chief Officer, he made the party “a lot of fun”. He had discussions for around ninety minutes on 10 May with the Master and the second engineer about the replacement of a cooling water jacket (“CWJ”) which cooled one of the six cylinders of the vessel’s diesel engine, which involved the engines being turned off and which required reasonably clement weather. Mr Braganza thought that this needed doing, and it was agreed that it would be carried out next day, weather permitting. After meeting crew members and watching a comedy film with them, he discussed the proposed operation and the likely weather conditions in his cabin with the Master at about 23.30 for rather over an hour. At 23.24, while the Master was with him, he emailed an engineering superintendent about the proposed operation, which was approved in an email nine minutes later “[i]f weather and schedule permits”. Before the Master left his cabin, Mr Braganza checked the Indian Premier League cricket results on his computer. At 01.00 on 11 May, Mr Braganza emailed the second engineer concerning entries in the log book. He was not seen alive again.

While he was Chief Engineer on the MV “British Unity”, Mr Braganza was described by some of those on board as being withdrawn and “staring into space”, in contrast with his happy character noted on previous voyages - although he was, as mentioned, in good humour on 9 May. During the carrying out of the works at Ferrol in April, Mr Braganza’s skills did not impress the BP superintendent, who told him how he might improve them.

On 24 March 2009, Mr Braganza was advised by the Captain of the “British Unity” that he had been awarded a bonus of $1688, but on 18 April, the Master was told to deduct this sum from Mr Braganza’s salary as he had had a break in service in 2008, which disentitled him from having a bonus. This was communicated on the same day to Mr Braganza, who immediately challenged the deduction, but his challenge was rejected by BP on 7 May.

As Chief Engineer, Mr Braganza was responsible for writing up the daily orders he gave his staff in the Chief Engineer’s Order Book. The last entry which he made was on 18 April 2009, when the vessel was still at berth, and when he was not under any obligation to make any entries. However, from 22 April 2009, when the vessel left Ferrol, he was obliged to make such entries, but he made none at any time thereafter.

It is also appropriate to refer to emails which were sent to Mr Braganza by his wife after he had joined the “British Unity”. According to the evidence of Mrs Braganza to Teare J, Mr Braganza was worried about the cost of living in Toronto, where he had moved with his family from India in July 2008. The two had had telephone discussions on 13, 14 and 15 February, and in an email on the last of those dates, his wife had emailed Mr Braganza saying that she “thank[ed] God for what we have and what the last seven months have done to us”, and encouraging him “to make a decision to let go of anything that is holding you down”. Again according to the evidence of Mrs Braganza, Mr Braganza was also concerned about the amount of work that needed to be done to the vessel, and on 17 February, his wife emailed him saying “please do not break your head about getting this opportunity on Unity. … Just relax, … think this is like any other ship. … So please do not stress yourself thinking of unnecessary stuff”. However, it appears that Mr Braganza continued to worry about money and the repairs.

On 27 March, in what strikes me as a particularly relevant email, Mrs Braganza asked her husband not to worry, saying that “everything will work out just fine” and that “I really cannot figure out what has shaken you out so much that you seem to be so afraid of life”, adding “If we keep crying over spilt milk we will not be able to go on with life”. And, on 6 May, she emailed him, begging him to “be happy for whatever we have” and expressing concern that “In Toronto … you seem[ed] sad” and “Now on the ship too you do not seem happy”. She also said that “you are thinking too much about the future and that is eating you”. She added that she wished he “could share [her] enthusiasm”, asking “what are you frightened about?”.

After Mr Braganza had disappeared, BP appointed a five person investigation team (“the team”), pursuant to its “Group Defined Practice”, which applies to any fatality on board a vessel. The team consisted of a Vice-President in the Engineering, Refining and Logistics Technology department, a Health Safety Security and Environment Manager, a Marine Incident Investigator, a Marine Standards Superintendent, and a Fleet Offshore Marine Services Superintendent.

The members of the team boarded the vessel on 18 May 2009 when it arrived in New York, and carried out an investigation, and produced their detailed Fatal Incident Report (“the First Report”) on 17 September 2009. According to the uncontradicted evidence, the First Report had been produced after carrying out careful inspections, interviewing 54 witnesses, and looking at “hundreds of documents”.

The “terms of reference” were set out at the start of the First Report. They were to “investigate the relevant circumstances leading up to the loss of” Mr Braganza, to “identify if possible the root causes of the incident”, “[a]s a result [to] identify any changes or improvements” in BP’s safety systems, and to “identify any common factors” with another incident on a different vessel. The First Report then set out some other formal information.

The First Report then summarised its conclusions. After referring to the team’s “in depth investigation”, the summary stated that the team could not “conclude for certain the root cause of the incident”. Then, after ruling out certain other possibilities, the Findings concluded by saying that the team “considers the most likely scenario to be that the C/E jumped overboard intentionally and therefore took his own life”. The First Report then set out certain recommendations for the future.

The First Report next recited the history leading up to the tragic death of Mr Braganza in some detail, including the need for the vessel’s repairs, the experience of Mr Braganza, and such details of the “incident” as the team could identify, including the response to his disappearance from the vessel. The First Report then explained the investigatory steps which the team had taken, including interviewing those on board, interviewing other people who might have relevant information, photographing and logging documents, inspecting the vessel, collecting other relevant information and analysing all data. The First Report went on to consider and reject the possibilities of “Hiding on board”, “Collec[tion] by a 3^rd party vessel” or “Fall from vessel due to horseplay, altercation or foul play”.

The First Report then addressed “Accidental fall from vessel”, and while the team found “[n]o evidence of sub-standard or unsafe structures”, they considered that slipping overboard “cannot be discounted” although there was no evidence to support it. The report then considered the possibility of suicide, and mentioned the matters referred to in paras 69-74 above, namely Mr Braganza’s withdrawn character, his concerns about his finances, his worry about the vessel’s poor repair, the emails from his wife, the criticism he received at Ferrol, and the withdrawal of his bonus. The First Report then stated that, although Mr Braganza’s state of mind “cannot be known”, it was “possible” that he jumped overboard. Under the heading of “Findings”, the First Report then repeated the conclusions contained at the start, including this: “the most likely scenario to be that the C/E jumped overboard intentionally and therefore took his own life”.

Mr Sullivan, BP’s General Manager, was appointed by BP to address the cause of Mr Braganza’s death for the purpose of clause 7. He considered the First Report and accepted the conclusion reached by the team. Accordingly, he decided that the death in service benefits under clause 7.3 were not payable to Mrs Braganza, and this was duly communicated to her.

Mrs Braganza then challenged the team’s rejection of the possibility that Mr Braganza’s death was caused by an accident, on the basis that, for the purpose of seeing whether the CWJ could be replaced, he could well have gone on deck to see the state of the weather and fallen overboard. Accordingly, Mrs Braganza contended that the right conclusion was that her late husband had died as a result of an accident. She provided a statement to the team, which then reconsidered the matter. On 12 July 2011, the team produced a Supplemental Report (“the Second Report”) in which they adhered to their conclusion.

The Second Report referred to many of the findings of the First Report. It also observed that the state of the weather during the night of the 10 May and the morning of the 11 May was “ever improving” and could “be considered good weather mid-Atlantic”, which rendered it improbable that Mr Braganza had been washed overboard or had fallen on deck due to a roll. It stated that, as “a senior and respected Chief Engineer, and reported as being ‘very safety conscious’ by his ship-mates”, Mr Braganza “is highly unlikely to have ventured out on deck during the hours of darkness and certainly not without informing the officer of the watch on the bridge”. The Second Report went on to point out that there would have been no reason for his checking on the weather, because that assessment “would in any event be carried out from the bridge by the Master and/or the [Chief Officer] who was responsible for the task”. The report also said that the safety standards on board “impressed those of the team visiting a BP ship for the first time”. A little later, the team stated that “it would be extremely unlikely that a person could trip, slip or fall in such a manner so as to fall overboard while carrying out normal shipboard duties and in the weather conditions which were known to prevail at that time”. The Second Report also stated that the team were “not aware of any previous cases within BP where weather conditions or rolling of the vessel had caused anyone to go overboard accidentally” (emphasis in the original). The report also stated that the team had “considered the statement of Mrs Braganza dated May 2011” and that “the contents do not affect the team’s conclusions”.

The Second Report went on to say that the team considered “that the scenario of C/E Braganza going outside to check on equipment during the night or early morning and suffering a fall that took him over the side of the vessel, [to have been] unlikely in the extreme”. The report then turned to the possibility of suicide and repeated the points which gave some support for that notion, which were summarised in these terms: “there seemed a lot of independent and reported (although circumstantial) evidence indicating a depressed state of mind combined with personal problems”. The lack of a suicide note, it said, “cannot be taken as a firm indication that suicide was not the cause of his disappearance”.

The Second Report therefore explained that the team “assess[ed] suicide as the most likely scenario, although equally the investigation team cannot firmly conclude what happened to C/E Braganza on the night of the incident” (emphasis in the original). The team also explained that they “did not reach this conclusion [sc that Mr Braganza committed suicide] on the basis of exclusion ie because we could not find any other cause as being likely or possible”. They added that “there was no single piece of evidence on which [the team] concluded suicide”, and that it was “very much a collection of a number of issues and the general feeling based on this evidence that C/E Braganza intended to take his own life”.

Following Mrs Braganza’s contentions and the provision of the Second Report, Mr Sullivan reconsidered his opinion in the light of those contentions and that report, and remained of the view that Mr Braganza’s death resulted from suicide. As he explained in the subsequent court proceedings, he “took the view that there was a reasonable basis on which to conclude that deliberate suicide was the most likely scenario”, which was a conclusion reached on the basis that “there was positive evidence that made deliberate suicide the most probable cause of death”. He also said that he did not arrive at this conclusion “by a process of elimination”.

Mrs Braganza issued proceedings against BP claiming (i) damages under the Fatal Accidents Act 1976 on the ground that BP was in breach of its duty of care to Mr Braganza (“the first claim”), and (ii) that BP’s opinion as to the cause of Mr Braganza’s death, based as it was on the First and Second Reports (“the Reports”) was flawed and ought to be declared invalid (“the second claim”). The trial lasted eight days before Teare J, who heard a number of witnesses primarily directed to the first claim. Having heard argument, he gave a reserved judgment in favour of BP on the first claim and in favour of Mrs Braganza on the second claim - [2012] EWHC 1423 (Comm). The judgment included some detailed findings of fact and conclusions in relation to Mr Braganza’s death, because of the first claim, namely that under the 1976 Act.

In paras 54 and 57 of his judgment, the judge concluded that, while “there was a real possibility that Mr Braganza went out on deck … to assess the weather himself”, he was “not … persuaded that it is probable that he did so”, and he concluded that while there was “a real, not a fanciful, possibility that Mr Braganza fell overboard”, it was “not … probable that he did so”. In para 58, the judge said that there was also “a real, not a fanciful, possibility that he committed suicide”.

The judge therefore concluded that Mrs Braganza’s claim under the 1976 Act failed, as the onus was on her to show that her husband’s death was accidental – para 61. Although it was not necessary to do so, the judge went on in paras 64-65 to hold that, even if she had succeeded in establishing an accident, Mrs Braganza’s claim under the 1976 Act would have failed on the additional grounds that she had failed to prove breach of duty by BP or causation.

The judge then turned to the second claim, namely Mrs Braganza’s challenge to BP’s opinion under clause 7.6.3 that Mr Braganza committed suicide, and concluded that Mr Sullivan had not properly directed himself before forming the opinion that Mr Braganza had committed suicide. In particular, in para 94, Teare J said that the team:

Teare J expanded on this in paras 96-97 in these terms:

The judge then went on to reject various other criticisms of the Reports, namely the lack of any psychiatric expertise (irrelevant as BP did not suggest that Mr Braganza suffered from a psychiatric illness), lack of engineering expertise (irrelevant to the issues which the team had to decide), and the fact that the team divided up to hear evidence (there was no requirement that they heard all the evidence together). The judge also considered and accepted some of Mrs Braganza’s criticisms about the findings made by the team (and the more important examples are given by Lady Hale in para 9(1), (2) and (6) above), but he did not base his decision in favour of Mrs Braganza on these criticisms (para 120).

One further criticism raised on behalf of Mrs Braganza before the judge against the opinion formed by Mr Sullivan is worth mentioning, namely that neither the team nor Mr Sullivan had interviewed Mrs Braganza. The Judge thought that “in principle … fairness did … require” her to be interviewed, although he accepted Mr Sullivan’s explanation that he did not want to “intrude unnecessarily into the family’s grief”. However, given that Mrs Braganza had the opportunity to comment on the First Report and did not offer any comment in relation to the matters the team relied on to justify its finding of suicide, the Judge rejected her case on this point.

In these circumstances, the Judge effectively set aside the opinion of Mr Sullivan on behalf of BP that Mr Braganza had committed suicide, and determined that Mrs Braganza was entitled to the death in service benefit under clause 7.3, and gave judgment in her favour in the sum of $230,265 with interest.

BP appealed against the judgment, and the Court of Appeal, for reasons given in a judgment given by Longmore LJ, with which Rimer and Tomlinson LJJ agreed, allowed its appeal – [2013] EWCA Civ 230. Longmore LJ disagreed with the notion that Mr Sullivan (or the team) should have directed himself (or themselves) that “there should be cogent evidence commensurate with or proportionate to the seriousness of a finding of suicide”, on the ground that it was unreasonable to expect him (or them) to take such an approach (paras 19-20). Longmore LJ then went to explain that that was not necessarily the end of the matter, as “[t]he relevant opinion still has to be reasonable”. However, he disagreed with Teare J’s view that the team “did not take into account the fact that Mr Braganza might have had legitimate (non-suicidal) reasons for being on deck”. Longmore LJ also said that, even if the team had failed to take that possibility into account, it would not render its conclusion unreasonable, given “the absence of a mechanism explaining how he could accidentally fall overboard” – para 22.

The Court of Appeal also considered the various minor items of evidence relied on by the team and referred to in para 92 above. They considered that the Judge had adopted too rigid a test in relation to some of those matters, but as the Judge had not relied on them to justify his conclusion, that aspect took matters no further.

Miss Belinda Bucknall QC, on behalf of Mrs Braganza, raised three lines of argument in support of her contention that this court should set aside the order of the Court of Appeal and restore Teare J’s order. First, she contended that the Court of Appeal did not appear to have appreciated that it should not have interfered with the Judge’s conclusion, given that he was the trial judge, who had heard the witnesses and reached a fact-sensitive conclusion. In my view, that contention is not well-founded, because it confuses the judge’s role when he was deciding what in para 87 above I have called the first claim, namely whether BP had been in breach of duty, with his role when considering the second claim, namely whether BP’s opinion under clause 7.6.3, based on the two Reports, should be set aside. When ruling on the first claim, the Judge was indeed the primary finder of fact, and it would only be if he had gone seriously wrong (eg if he had come to a conclusion which no reasonable judge could have come to, or made a demonstrable error such as ignoring a significant relevant piece of evidence) on a factual issue that an appeal court could properly have reversed his conclusion. However, the position on the second claim was entirely different: far from being the primary finder of fact, Teare J was carrying out a reviewing function of the two Reports, prepared by the team, and it was the team, or more precisely Mr Sullivan (who relied on the two Reports in forming his opinion), not the Judge, who represented the primary finder of fact.

Miss Bucknall’s second line of attack was based on a number of alleged errors in the two Reports, which are briefly referred to in para 92 above. Although I agree with some of them, I do not accept all the criticisms which were accepted by the Judge. However, it is unnecessary to discuss this aspect further, as I agree with the Judge and the Court of Appeal that the mistakes could not possibly justify challenging the conclusion reached in the two Reports. It appears to me to be fanciful to think that, if all of them had been pointed out to the team and corrected, it could conceivably have resulted in the team reaching any different conclusion from that which they did. Just as minor errors in a full and careful judgment do not justify interfering with a judgment, so it is with minor errors in a report which may have legal effect. To hold otherwise would be wrong in principle, and it would serve to encourage unmeritorious appeals and to discourage full, considered and informative reports and judgments. Of course, different considerations apply where the error of fact might well have affected the outcome.

The third line of attack raised by Miss Bucknall was very similar to the Judge’s reasoning for rejecting Mr Sullivan’s reliance on the two reports as set out in para 94 of his judgment, namely that the basis for determining that Mr Braganza committed suicide was too flimsy to justify such a conclusion, at least by Mr Sullivan. That argument merits much fuller consideration, but, before turning to consider it, it is convenient to address the appropriate approach which the court should take to a contention such as that raised by Mrs Braganza’s second claim.

In some circumstances, it may be that the cause of an officer’s death is very difficult to determine. In such a case, the exercise envisaged in clause 7.6.3 does not require the person charged with carrying out the enquiry to form an opinion. It is perfectly proper for that person to conclude that he or she is unable to form an opinion as to the cause of the death in question. Clause 7.6.3 provides for an exception to BP’s liability to pay compensation under clause 7.3, and if it is impossible for the person who is charged with forming an “opinion” under clause 7.6.3 to come to a conclusion, then BP must pay the compensation.

A second point, which requires a little more exegesis, arises from the fact that the person who is given the primary duty of determining the cause of death under clause 7.6.3 is BP or its insurers (and for present purposes one can ignore the reference to the insurers). If a provision such as clause 7.6.3 does not specify how the issue is to be determined, it would be for a judge to decide it as the primary fact-finder. Save where for some reason BP cannot or refuses to consider the cause of death, the courts only have a role to play once BP forms an opinion on the issue and an interested party seeks to challenge the opinion – as happened in this case.

There was some discussion as to the standard which the court should expect of the decision-maker or opinion-former in such circumstances. That question was fully considered by Rix LJ in Socimer International Bank Ltd v Standard Bank Ltd [2008] EWCA Civ 116, paras 60-66. He began by describing the issue as arising “When a contract allocates only to one party a power to make decisions under the contract which may have an effect on both parties”, and, after considering a number of previous authorities, he concluded:

Like Lady Hale, I consider that there is considerable force in the notion that this approach is, and at any rate should be, the same as the approach which domestic courts adopt to a decision of the executive, as described in the judicial observations which she quotes in paras 23 and 24 above (as indeed is reflected by the reference to the Wednesbury decision in the passage I have quoted from Rix LJ). I do not think that there is any inconsistency of approach between Lady Hale and Lord Hodge or myself in this connection.

However, a third point I should mention does concern a point of difference between us. It is best expressed by reference to Lord Hodge’s statement that “contracts of employment … have specialties that do not normally exist in commercial contracts”, which he discusses in para 54-57. It appears to me questionable whether the special implied mutual duty of trust and confidence survived the death of Mr Braganza, but I accept that there is a powerful case for contending that it did, at least for present purposes. However, I do not think it necessary to decide the point, because I fail to see how it assists Mrs Braganza’s case. Once it is accepted that BP had to carry out the investigation with “honesty, good faith, and genuineness” and had to avoid “arbitrariness, capriciousness, perversity and irrationality”, I do not see what trust and confidence add. I find it difficult to accept that trust and confidence would require more than what in a normal commercial context would be expected, either of BP when carrying out the investigation, or of the court when scrutinising the investigation and its results. Either the investigation was properly carried out or it was not, either there was enough evidence to justify the conclusion reached by Mr Sullivan or there was not, and either the reasoning which led to the conclusion was defensible or it was not. Accordingly, as I see it, the duty of trust and confidence is simply irrelevant to deciding that question.

A fourth point is also worth mentioning. A court considering a decision such as that reached in this case by the team or by Mr Sullivan should bear in mind the fact that it is performing a reviewing function, and, as I have already mentioned, not an originating fact-finding function. The court’s approach should therefore be similar to that of an appellate court reviewing a trial judge’s decision. In this case, the team, and then Mr Sullivan on behalf of BP, were entrusted with forming a view on a point of fact, namely how Mr Braganza died. As Lord Hoffmann pointed out in Biogen Inc v Medeva PLC [1997] RPC 1, 45, albeit in connection with a trial judge’s decision, this is “‘a kind of jury question’ … and should be treated with appropriate respect by an appellate court”. And that is, in my view, at least as true where the parties have agreed by whom the issue should be determined.

Lord Hoffmann then explained that “[t]he need for appellate caution in reversing the trial judge’s evaluation of the facts is … because specific findings of fact, even by the most meticulous judge, are inherently an incomplete statement of the impression which was made upon him by the primary evidence”. He continued: “[the judge’s] expressed findings are always surrounded by a penumbra of imprecision as to emphasis, relative weight, minor qualification and nuance … of which time and language do not permit exact expression, but which may play an important part in the judge’s overall valuation”. I would add that (i) it cannot be right to expect a higher standard from non-lawyer decision makers than from a judge, and (ii) the respect to be shown to the decision maker’s determination may often be greater when it is that of a person or people with relevant expertise or experience than when it is that of a judge.

One would have to be unusually stony-hearted not to hope that a way could be found to ensure that, having suffered the terrible blow of losing her husband, Mrs Braganza could be spared the additional blows of an inquiry concluding that he had killed himself and the deprivation of a death-in-service benefit. However, it is the most fundamental duty of a judge to apply the law, even if it sometimes leads to hard consequences in the circumstances of a particular case.

In my view, the position in this case is, in summary, as follows. As the First Report made clear (see para 77 above), BP appointed a team of experienced people from different disciplines specifically to form a view as to how Mr Braganza had died, the team carried out what appears to have been a very thorough investigation (see para 78 above), and produced a full and meticulous report in which they expressed themselves in moderate and considered terms, and in which they concluded that, while Mr Braganza could have suffered an accident, that was very unlikely, and that the probable cause of his death was suicide (see paras 79-81 above). They then carefully reconsidered that conclusion following a request from Mrs Braganza, and, sadly for her, confirmed it in a further carefully considered report (paras 83-85 above). When Mr Sullivan was instructed by BP to consider the cause of Mr Braganza’s death for the purpose of clause 7, it was entirely reasonable for him, in the light of that background to consider the First Report and adopt its conclusions, and, when the First Report was challenged, to readdress the matter and to consider the Second Report and adopt its conclusions.

However, it is said that Mr Sullivan’s opinion is flawed because of a combination of factors. When analysed, it appears to me that those factors may be summarised as follows:

It was inappropriate for Mr Sullivan simply to rely on the Reports, as they were prepared for a different purpose;

A finding that a person committed suicide amounts to an inherently improbable, serious or damaging conclusion which required more cogent evidence than was available;

Mr Braganza exhibited no signs that he was depressed or had suicidal intentions during the 24 or 36 hours prior to his death;

The consequence of Mr Sullivan concluding that Mr Braganza’s death was caused by suicide was so severe, namely, the loss of a death-in-service benefit, that it was not justified on the evidence;

Mr Sullivan ought at least have directed himself as to the inherent unlikelihood of Mr Braganza having committed suicide;

The team, and therefore Mr Sullivan, failed to take into account the fact that Mr Braganza had good reason to go on deck in the early morning;

Mr Sullivan’s failure to ask Mrs Braganza about the emails impugned his opinion.

As to argument (a), it was, in my view, neither unreasonable nor inappropriate for Mr Sullivan, when forming his opinion for the purposes of clause 7.3, to adopt the conclusion reached in a carefully considered report prepared by a group, such as the team, on behalf of BP. I am unpersuaded by the argument that the team was concerned with a different issue from Mr Sullivan. The team was specifically instructed to investigate the “relevant circumstances” and “root causes” of Mr Braganza’s disappearance, and they clearly could only conclude that it was suicide if that was, in their view, the likely cause, ie on the balance of probabilities, after fairly considering all the evidence. Precisely the same applies to the forming of Mr Sullivan’s opinion as to the cause of Mr Braganza’s death. Further, the fact that the team was investigating the issue for a purpose different from Mr Sullivan is nothing to the point. Even if one assumes (which I do not accept) that the sole purpose of the team was to make recommendations as to what steps should be taken by BP to improve safety on board their vessels, it was still essential for them to assess, if they could, just how Mr Braganza had died.

If the team had concluded in their report that Mr Braganza had fallen overboard as a result of an accident or that it was impossible to say how Mr Braganza had died, it could not seriously be suggested that Mr Sullivan could not have adopted that conclusion. Indeed, to many people, it would have been a little surprising if, in the absence of new facts not known to the team or plain error on the part of the team, Mr Sullivan had reached a different view from that taken by the team. Of course, Mr Sullivan was not entitled unthinkingly to adopt the view of the team: he had to form his own opinion. However, once he was satisfied that the team had conducted a very thorough investigation, and had carefully considered all the evidence and had reached a conclusion with which he considered that he agreed, it would in my view be little short of absurd to hold that he was nonetheless obliged in law to carry out his own separate investigation.

As to argument (b) in para 109, support for the proposition that the more serious the allegation, the more cogent the evidence required to prove it can be found in the opinion of Lord Nicholls in In re H (Minors) (Sexual Abuse: Standard of Proof) [1996] AC 563. At p 586, he said “the more serious the allegation the less likely it is that the event occurred and, hence, the stronger should be the evidence”. However, that point also can only be taken so far, as Lord Hoffmann and Baroness Hale each said in In re B (Children) (Care Proceedings: Standard of Proof) [2008] UKHL 35, [2009] AC 11, paras 1-15 and 64-70 respectively. As Baroness Hale pointed out in para 72, “there is no logical or necessary connection between seriousness and probability”. She then said that “serious allegations are [not] made in a vacuum”, and went on to give examples of statements which might seem intrinsically unlikely if made without a particular context, but appear quite likely if made in the context of certain assumed facts. In the end, the decision-maker has to come to a conclusion on the particular facts of the case.

The point is also apparent from the opinion of Lord Carswell (with whom the other members of the committee agreed) in In re D (Secretary of State for Northern Ireland intervening) [2008] UKHL 33, [2008] 1 WLR 1499. At para 29, he said, in connection with the serious finding in civil family proceedings that an adult had been guilty of very serious sexual assaults on young girls, “[t]he panel had to be satisfied on the balance of probabilities that he was, but it did not follow that specially cogent evidence was required”. In reaching that decision, the House of Lords accepted the appellant panel’s argument identified at para 22 that the court below was wrong to say that “the ‘flexible approach to the civil standard of proof’ required ‘more cogent evidence than would be conventionally required’ and that ‘a more compelling quality of evidence’ was needed”.

In the present case, it seems to me clear that there was a combination of reasons which can fairly be said to be sufficiently cogent to justify Mr Sullivan’s opinion, based on the two Reports, that Mr Braganza had taken the unusual and tragic course of committing suicide. First, this is a case where it is clear that Mr Braganza died at sea, and the only two plausible possible causes (as the Judge and the team both considered) were accident or suicide. So this is not a case where the issue is whether an unlikely event occurred: the issue is how an unlikely event, which undoubtedly occurred, was actually caused.

Secondly, when one turns to consider that issue, it is clear that there were various factors which either concerned, or could well have concerned, Mr Braganza. In particular, the emails from his wife strongly suggest that he was depressed about his financial affairs and that he was oppressed by the vessel’s state of repair (paras 73 and 74 above). Furthermore, he could well have been upset by the withdrawal of his bonus (especially if he had money problems) (para 71), and by the criticism of his supervision of the works at Ferrol (para 70). It does seem pretty clear from the emails from his wife that, to a significant degree, he let at least some of these matters cause him serious concern. Although there was evidence that he was on good form in the 36 hours immediately before his disappearance, there were a number of comments about his uncharacteristically moody state for most of the three weeks he was on board the British Unity – see paras 69 and 70 above. And there is the curious fact that he did not make entries as he should have done in the Order Book – see para 72 above.

Thirdly, the only serious alternative to suicide, as both the Judge and the team separately concluded, was an accident, and for the reasons summarised in paras 80 and 83-84 above, that would seem to have been very unlikely. The weather was relatively calm, Mr Braganza was very safety-conscious, he was unlikely to have gone on deck without warning someone, he had no reason to go on deck, the vessel was very well protected, and nobody in the BP fleet had ever fallen overboard from this class of vessel. On the team’s analysis, it was, at least in my view, plainly open to them, and therefore to Mr Sullivan, to conclude that suicide was more likely than an accident.

Of course, as already mentioned, it was open to the team to conclude (as the Judge did) that it was not possible to reach a conclusion as to how Mr Braganza had died. However, that is not at all the same thing as saying that this was the only conclusion that the team could have reached. If the team or Mr Sullivan had proceeded on the basis that they were bound to come to a conclusion as to how Mr Braganza had died, or if they had arrived at their conclusion by a process of simply eliminating all other possible causes, they would have been guilty of an error of law which would, I think, have vitiated the conclusion which the team reached. However, that is not how they approached the issue: on the contrary – see paras 85 and 86 above.

Argument (c) in para 109 above is that there was no evidence of depression or suicidal behaviour in the day or two before Mr Braganza’s disappearance. In the first place, there was evidence to support the possibility of suicide as already discussed. Quite apart from that, we do not, alas, need expert evidence to tell us that many suicides occur “out of the blue” so far as loving relatives are concerned. The very fact that some suicides occur at all is attributable to the fact that there are no signs that it will happen. Of course, a number of suicides are preceded by aberrant behaviour or warnings, but I am quite unpersuaded that the absence of any such matters is of much significance.

As to argument (d), namely the consequence of Mr Sullivan’s forming the opinion that the Mr Braganza killed himself was so great that the decision-making process was flawed, of course, cogent evidence was needed: it always is. And, as The House of Lords has emphasised in a number of different cases, apart from the criminal standard of proof, there is but one standard of proof, namely on the balance of probabilities, ie more likely than not – see what Lord Nicholls said in In re H [1996] AC 563, 586, affirmed and explained by Lord Hoffmann and Lady Hale in In re B, [2009] AC 11, paras 13 and 68-70 respectively, and applied by Lord Carswell in In re D [2008] 1 WLR 1499, para 28.

I accept that it may be right to take into account the seriousness of the consequences of a particular finding when deciding whether the evidence justified it, but, as Lord Carswell explained in In re D, para 28, “[t]he seriousness of consequences is another facet of the … [the seriousness of the allegation]”. In other words, in this case it is illogical to suggest that, because Mr Sullivan’s opinion, unlike the decision of the team, would result in Mrs Braganza not receiving the death-in-service benefit, he should have been more reluctant than the team to conclude that Mr Braganza had committed suicide. The only relevance that the non-receipt of the benefit could have had to Mr Sullivan’s opinion would be that, because the benefit would be lost if Mr Braganza committed suicide, it could be said to render it less likely that he would have done so. However, that applies equally to the team’s decision.

As to the argument (e) in para 109 above, I do not think that it would be appropriate to hold it against either the team or Mr Sullivan that they did not expressly say that it is inherently unlikely that a person would commit suicide. There was evidence that could support the notion that Mr Braganza was in an unhappy state emotionally, and the only alternative to suicide was an accident, which the team rejected as very unlikely indeed. Particularly in those circumstances, it would, in my view, involve setting an unrealistically, and therefore an undesirably, high standard on investigators or writers of reports, whose investigations and reports are intended to have legal effect, to hold that the investigator or writer had to mention in terms that suicide was inherently unlikely.

Even if the Reports had been judgments, I very much doubt that they could have been fairly criticised for failing expressly to say that suicide is an inherently unlikely act. However, even if that is wrong, any such criticism is impossible to sustain in connection with the Reports. As Mocatta J said in The Vainqueur Jose [1979] 1 Lloyd’s Rep 557, 577:

So far as argument (f) in para 109 is concerned, Miss Bucknall, in the course of her submissions, emphasised on a number of occasions that Mr Braganza could have gone on deck in daylight to check the weather. This echoes the Judge’s observation in para 96 of his judgment that the team “and hence Mr. Sullivan, failed to take into account the real possibility that Mr Braganza went out onto deck in order to check the weather to see whether it was safe to carry out the planned work”. However, the simple answer to that contention is that (i) in the First Report, the team simply thought that an accident was very unlikely for the reasons they gave, even if Mr Braganza had gone on deck, and (ii) in the Second Report, the team reconsidered the possibility of an accident and rejected it as very unlikely, because (a) Mr Braganza would probably not have checked the weather, (b) if he had gone on deck to do so he would have told the officer of the watch, and (c) perhaps most importantly, because the possibility of his falling overboard by accident was very unlikely indeed – see paras 80 and 83-84 above.

Nor would I accept that argument (f) in para 109 above provides a good reason for doubting Mr Sullivan’s opinion, namely that he did not ask Mrs Braganza about the emails before forming his opinion. If, as the Judge found, the team was entitled to assume that Mrs Braganza had had the opportunity to make representations about the emails to them once she got the First Report, then Mr Sullivan was equally entitled to make that assumption. Indeed, she was free to make any points which she wanted, and apparently did so in her note of May 2009, which was referred to in the Second Report (para 83 above). Mr Sullivan knew that she had complained about the First Report, as it was her complaint which resulted in the Second Report, and, having read the Second Report, he knew that she had been free to make any point she wanted to the team.

In any event, although Mr Sullivan referred in his witness statement to the fact that he had not thought it appropriate to approach Mrs Braganza about the emails, he was not cross-examined about this. It is therefore difficult to see how he could be criticised on this ground. Quite apart from this, Mrs Braganza gave evidence before Teare J that her husband “was not happy because their money was running out”, “[t]hey had sold everything”, that “life [was] tough in Toronto”, and that Mr Braganza “was frightened”: para 17. Although the Judge rejected BP’s case that this evidence understated Mrs Braganza’s degree of concern, it does not seem to me that such evidence would have undermined a conclusion that he had killed himself: it confirms what is plain from the emails.

In my view, in the light of the discussion in paras 107-125 above, it is not fairly open to a court to decide that the conclusion reached by the team in the First and Second Reports, and therefore the opinion formed by Mr Sullivan, fell foul of the test laid down by Rix LJ in the passage cited in para 102 above. In my view, neither the conclusion reached by the team nor the consequential opinion formed by Mr Sullivan can be characterised as “arbitrar[y], capricious …, pervers[e] [or] irrational …”, to use Rix LJ’s words. The two Reports are, as I have indicated, impressive both in the extent of the investigations on which they were based and the care with which they were compiled, and the conclusion they reached was carefully and rationally explained, and Mr Sullivan cannot be criticised for relying on them.

I would accordingly dismiss this appeal.