Rhine Shipping DMCC v Vitol SA

This appeal is concerned with the effect of ‘internal hedging’. The Respondent (‘Vitol’) is a large and well known trader in oil and petroleum products. The Appellant (‘Rhine’) was the disponent owner of the vessel MT DIJILAH (‘the Vessel’), which Vitol voyage chartered to carry cargoes from West Africa to China. The Judge held Rhine liable to Vitol on its counterclaim for breach of the charterparty causing a 6 day delay in the Vessel reaching and completing loading at the second loading port, Djeno (in Congo). The delay increased the purchase price which Vitol had to pay its seller because the purchase price was linked to ‘Brent Dated’ prices published by the Platts index on dates referable to the bill of lading date, and because there was a rise in the published index price during the delay in the bill(s) of lading being issued. The Judge determined a number of issues against which there is no appeal. The single ground of appeal is that he erred in deciding that Vitol’s ‘internal hedging’ did not fall to be taken into account in reducing damages.

The charterparty, dated 27 March 2020, was for carriage of a minimum of 260,000 m.t. of crude oil from 1/2 port(s) West Africa (Ghana-Angola range) to 1/3 port(s) Far East Singapore-Japan range. The first cargo loaded thereunder at Offshore Cape Three Points (in Ghana) was some 128,000 mt. As a result of a court arrest of bunkers and stores on the Vessel at the suit of third parties pursuing claims against the bareboat charterer, with whom Rhine was associated, the Vessel was detained at Takoradi anchorage off the first loading port on 30 April 2020 following loading. On 7 May 2020, following the posting of security, the Vessel was permitted to proceed to the second loadport, Djeno, where she completed loading a cargo of some 920,000 bbls of Djeno crude oil on 12 May 2020 (‘the Djeno cargo’). Three bills of lading were issued for the Djeno cargo, each dated 12 May 2020. The Judge held that Rhine was liable to Vitol under the terms of the charter for the delay to the completion of loading of the Djeno cargo caused by the detention in Ghana; and that but for such detention the Djeno cargo would have completed loading on 6 May 2020, and the bills of lading would have been issued on that date. There is no appeal from those findings.

On 3 March 2020 Vitol agreed to sell 920,000 bbls of Djeno crude oil to its associated group company Vitol Asia Pte Ltd (‘Vitol Asia’) CIF Qingdao China with an estimated arrival period of 20-31 May 2020 (‘the sale contract’). The sale price was agreed as the average of the settlement quotations for July-20 ICE Brent (‘the ICE Index’), for quotations between 27 – 28 May 2020, plus US$2.60 per bbl.

On 26 March 2020 Vitol purchased the physical cargo to fulfil this short position by a contract of purchase of 920,000 bbls of Djeno crude from TOTSA Total Oil Trading SA (‘TOTSA’) on terms FOB Djeno (‘the TOTSA contract’). The price payable by Vitol under the TOTSA contract was determined by reference to the bill of lading date for the Djeno cargo and was to be calculated as the average of the prices published in the Platts Crude Oil Marketwise index under the heading Brent Dated, on the five publication dates following the bill of lading date, minus a discount of US$4 per bbl. Such published prices, described within Vitol simply as Brent Dated for short, are prices for one of the benchmark grades of crude oil which are traded for fixed prices, based on a basket of North Sea grades of crude oil. Others are West Texas Intermediate (‘WTI’) representing cargo exported from Houston Texas (or certain terminals inland from Houston); and ‘Dubai/Oman’ representing Middle East export. Brent Dated represents the price for delivery within 10-30 days. Platts is not the only index to publish Brent Dated prices, its main rival being Argus. Where delivery is anticipated to be more than 30 days ahead, the relevant index used for Brent crude is that quoted on the regulated exchange ICE, and known as Brent Futures. This was the index used for the sale contract price in this case.

The pricing terms of the TOTSA contract and the sale contract each involved variables by reference to the Dated Brent and Brent Futures indices respectively around the time of loading and anticipated delivery. These two variables could be hedged in order to lock in the pricing profit and remove the risk of market movement on those indices.

The appeal is concerned only with the first aspect of such hedging, namely the hedging of the TOTSA purchase price, which is the only hedging which could potentially affect the loss caused by the delay in the bill of lading being issued at Djeno, for which Rhine was contractually liable under the charterparty. That delay had no effect on the Vitol Asia sale price which was calculated by reference to the ICE Index on defined dates which were unaffected by any delay in arrival at the discharge port.

The floating element of the TOTSA purchase price, dependent on the Platts Brent Dated price on the five published days after the bill of lading date, could be fixed by Vitol purchasing a swap for a fixed price (‘x’), under which Vitol would pay x to the counterparty, and the counterparty would pay to Vitol the Platts Brent dated price on those days.

It is common in the oil industry to hedge pricing risks in these and other ways, although not necessarily to hedge them completely. Trading houses will have their own risk appetite towards taking a speculative position on pricing risk.

The Judge described Vitol’s internal risk management system, known as Vista, and its application to the hedging of the TOTSA contract price, in the following paragraphs of his judgment, which it is necessary to set out at some length. His account was, he said, largely uncontroversial, and was based on the evidence of Mr Jordan Smith, a commercial analyst at Vitol; and the expert evidence, given on behalf of Rhine by Mr Beckett, and on behalf of Vitol by Ms Bossley:

It is important to identify the case advanced by Rhine at the trial as to the effect of the internal hedging because, as I shall explain, the argument advanced on the appeal is significantly different.

Rhine argued that it was the internal hedge between portfolio 645 and portfolio 101 (as I shall call them for short), which fell to be taken into account to reduce the damages claim because had that been an external hedge it would have conferred the benefit which had to be taken into account. The benefit was the US$2,871,971 recorded in the Vista entry for portfolio 645 as a result of the closing of the rolling of the swaps, the rolling also being recorded as an ‘internal’ hedge with portfolio 101 on the dates of the rolling. Rhine argued that although the portfolios were internal to Vitol, the effect was the same as if the hedging had been external because two divisions within a company could be treated as separate entities for these purposes, notwithstanding their lack of separate legal personality, relying on the decisions of Donaldson J and the Court of Appeal in Bremer Handelsgesellschaft mbH v Toepfer [1978] 1 Lloyd’s Rep 643 and [1980] 2 Lloyd’s Rep 43. This was the case supported by its expert Mr Beckett who confirmed that it was not relevant to investigate what happened to the risk after its transfer to portfolio 101.

In a meticulous and well-reasoned judgment, the Judge rejected this argument. He started by accepting the submission that had the hedge been an external one with a third party, the benefit would have had to be taken into account, relying on the judgments of Christopher Clarke J in Choil Trading SA v Sahara Energy Resources Ltd [2010] EWHC 374 (Comm) at [161] and Blair J in Glencore Energy UK Ltd v Transworld Oil Ltd [2010] EWHC 141 (Comm) [2010] 1 CLC 284 at [78]. That conclusion was not challenged on the appeal by Vitol and is in my view correct. The basis for the Judge’s conclusion was that an external rolling of external swaps would have been a step taken in mitigation of the delay caused by Rhine’s breach, and the resultant benefit from a third party would have fallen to be taken into account as a benefit received as a result of such reasonable mitigation. The conclusion would also follow, in my view, from an application of the principles of avoided loss, as summarised in Swynson Ltd v Lowick Rose LLP (in liquidation) [2017] UKSC 32 [2018] AC 313 by Lord Sumption JSC at [11]: the rolling of the swaps with a third party and the resultant benefit received would have been the direct result of the breach and would not have fallen within the scope of the collateral benefit exception characterised by the latin tag res inter alios acta. As Mustill LJ observed in Hussey v Eels [1990] 2 QB 227 at p. 232D-E the principle of avoided loss and the principle of credit for benefits procured by mitigation are separate principles, but in relation to causation they raise the same issues of fact and lead to the same conclusion.

The Judge rejected, however, the argument that the internal rolling of the swaps between portfolios 645 and 101 could be treated as having any effect on Vitol’s profit or loss. The swaps, and their rolling, merely transferred the risk of loss or gain between Vitol portfolios and did not make good any loss to Vitol as a company. In reaching this conclusion he explained why Bremer Handelsgesellschaft v Toepfer supported it, rather than, as Rhine had contended, providing support for the opposite conclusion: in short it confirmed that a company cannot contract with itself. This aspect of his reasoning is in my view unimpeachable. Indeed this aspect of his reasoning was not challenged by Mr Berry KC in advancing Vitol’s argument on the appeal.

The Judge went on to consider an alternative way in which the point was pursued. The Judge found that the likely position was that the risk transferred into portfolio 101 by the rolling of the swaps was offset within the Vista system by opposite risks which had ultimately originated with other physical trades (however complex the mix of other transactions which gave rise to the match of this particular risk). Rhine therefore contended that those other physical transactions, or the parts of them which together represented the matching risk, had reduced or mitigated the risk in question and therefore Vitol’s loss.

The Judge rejected the argument on the basis that the avoided loss on such other physical transactions was not a benefit derived from mitigation; and it was res inter alios acta, applying what Lord Sumption said in Swynson at [11]. The other transactions were not entered into for the purposes of hedging the TOTSA purchase and Vitol’s decision to enter into them was in no way caused by the price risks to which Vitol was exposed on the TOTSA transaction, still less by Rhine’s breach of contract giving rise to the loss on the TOTSA purchase price. The existence of separate transactions within Vitol, which offset the risks represented by rolling the swaps into portfolio 101, would be separate and unconnected, neither entered into for the purposes of hedging the other, or for mitigating risks or potential loss on the other; rather they were each based on different physical trades in the course of ordinary trading operations. They were in no sense steps taken in mitigation of Rhine’s breach (as would have been the case had Vitol taken an external hedge with a third party in the form of the rolled swaps). The “offsetting” risks were paper hedges derived from somewhere in Vitol’s book reflecting its ordinary trading operations. To the extent that they gave rise to a benefit they did so entirely independently of the circumstances giving rise to the loss on the Djeno cargo transaction. Any such benefit had nothing to do with the charter, the delay to the Vessel or the reason why it had been delayed. It was not derived from steps taken by Vitol in consequence of the breach of charter. Nor did it derive from a contract for the benefit of Rhine.

Again, I find this reasoning compelling. Indeed this aspect of the reasoning was also not the subject of challenge by Mr Berry on the appeal.

I would therefore have no hesitation in dismissing any appeal based on the case advanced by Rhine before the Judge. Indeed Mr Berry did not seek to argue the points which had been advanced before the Judge, but confined himself to a new argument.

Permission to appeal was granted by the Judge for the following ground (and expressly on the basis that Rhine would not challenge any findings of fact):

Rhine’s replacement skeleton argument for the appeal was signed by Mr Young KC, Rhine’s leading counsel at the trial. A week before the hearing of the appeal, Rhine served an application to amend the grounds of appeal and to rely on a supplementary skeleton argument, each signed by Mr Berry who appeared for Rhine on the appeal.

The supplementary skeleton advanced a new argument. The benefit which it is said should have been brought into account was the avoided cost of a “book hedge” on Vitol’s net book risk which would, absent breach, have been required to hedge the risks which were in fact met by the oppositional rolling of the swaps. As articulated by Mr Berry in his attractively simple oral submissions it proceeded by the following steps:

Mr Berry accepted that the argument had not been put “in exactly this way” at the trial, although he sought to downplay its novelty. It is, however, entirely new in three respects, namely conceptually; in relation to the factual and expert issues it raises; and in relation to the legal issues it raises. It is conceptually new because it addresses the counterfactual scenario of what would have happened in and as a result of the Vista system, and in Vitol’s book hedging decisions, had there been no breach by Rhine; whereas the case at trial depended wholly upon what had in fact happened in the breach scenario, which was what gave rise to the paper entry in portfolio 645 of the internal rolling of the swaps and their notional monetary consequences had they been external. The benefit is now said to arise from an avoided loss on a transaction which would have occurred absent breach, rather than, as at trial, a benefit from a transaction or transactions which ‘occurred’ as a result of the breach. It raises new factual issues, as to how the Vista system works in netting off pricing risk between different transactions of different kinds or linked to different indices or linked to indices on different dates; whether Vitol always hedged all its pricing risk on a whole book basis so as to run no speculative risk whatever on any pricing risk; what the terms of the avoided hedge would have been; when it would have been taken out; and what loss it would have produced. It raises new legal issues as to whether the avoided book hedge can be characterised as a step taken in mitigation of loss caused by Rhine’s breach (being the absence of a decision to take out a hedge by someone in all probability ignorant of such breach) or as avoided loss, which in either case is not res inter alios acta. I return to explore some of these aspects more fully below.

The application for permission to amend the grounds of appeal raises the same issues as arise on the application to advance the new argument, and stands or falls together with it, although this was not immediately apparent from the draft amended grounds attached to the application notice. This was because those draft amended grounds contained the first sentence of the ground for which the Judge had given permission; and proposed an amendment to it by referring to the rolling of the swaps rather than the swaps themselves; and by referring to paragraphs [61]-[64] of the Judgment in place of paragraph [127] so as to correct the cross reference to the rolling of the swaps. These amendments in isolation appeared unobjectionable because it had always been Rhine’s case that it was the rolling of the swaps, not the original swaps themselves, which had given rise to the profit which had to be brought into account, and Vitol would always have understood that. However the vice in the proposed amendment, to which Vitol objected, lay in its intended use to lay the foundation for the new argument advanced in the supplementary skeleton, leave to advance which was also opposed by Vitol. Mr Berry submitted that he did not really need permission to amend the grounds, the amendment being simply by way of clarification of something already understood. That might have had some initial attraction had the original ground been limited to the first sentence, as was the impression given by that sentence alone being included in the draft amended grounds attached to the application notice. The second sentence did not appear, suggesting that the proposed amendment to the grounds was intended to delete it notwithstanding that it was not included with strike-out. However the existence of the second sentence in the ground for which the Judge gave permission puts an entirely different complexion on whether permission should be granted, because the second sentence limited the scope of the first: the first sentence had it stood alone might have been wide enough to cover almost any argument, including the new argument raised in the supplementary skeleton, but the leave given by the Judge, including the second sentence, was directed to the arguments which were addressed to him and in my view are not wide enough to encompass the new argument. This is not surprising since the new argument had not been identified or advanced at the time permission was sought from, and granted by, the Judge. Mr Berry told us that the omission of the second sentence was an inadvertent error and that it was intended that the second sentence should remain in the proposed amended grounds. It follows that there is still no draft amended ground which would encompass the new argument. However that is a matter of form rather than substance because the new argument has been clearly identified in the supplemental skeleton argument. We allowed Mr Berry to develop it, de bene esse, and Mr Toms to respond to it, as well as hearing argument on both sides as to whether it should be permitted. If I had concluded that the new argument could properly be raised for the first time on the appeal, I would have favoured giving Rhine a further opportunity to formalise the position by redrafting the proposed amendment to the grounds.

The applicable principles were not in dispute. In Singh v Dass [2019] EWCA Civ 360 Haddon-Cave LJ summarised the guidance in the authorities which this court generally applies in deciding whether a new point may be advanced on appeal:

There are aspects of such guidance which I would emphasise as of importance to the present case. First, the caution about allowing a point to be raised for the first time on appeal, irrespective of whether it affects the evidence or argument which would have been advanced at trial, reflects a policy both of fairness and of the efficient use of court resources which was encapsulated by May LJ in Jones v MBNA [2000] EWCA Civ 514 (unreported 30 June 2000) at [52]:

As Snowden LJ said in Notting Hill Finance Ltd v Sheikh [2019] EWCA Civ 1337 [2019] 4 WLR 146 at [26] (as Snowden J sitting in this court), there is no general rule that a case needs to be exceptional before a new point will be allowed to be taken on appeal. The court must examine each application on its own facts in the light of the guidance to be found in the authorities. Nevertheless the considerations of fairness and efficiency adverted to by May LJ remain factors to be taken into account.

Secondly, it will usually be fatal to the new point being permitted to be raised for the first time on an appeal if there is a real possibility that, had the point been taken at trial, evidence could have been adduced which might have affected the outcome. As Nourse LJ put it in Pittalis v Grant [1989] QB 605 at p. 611:

The reference to the mere possibility of other evidence affecting the outcome of the point is also reflected in similar language in the classic dictum of Lord Herschell in The Tasmania at p. 225:

“It appears to me that under these circumstances a Court of Appeal ought only to decide in favour of an appellant on a ground there put forward for the first time, if it be satisfied beyond doubt, first, that it has before it all the facts bearing upon the new contention, as completely as would have been the case if the controversy had arisen at the trial; and next, that no satisfactory explanation could have been offered by those whose conduct is impugned if an opportunity for explanation had been afforded them when in the witness box.” (my emphasis)

To similar effect is the judgment of Gibson LJ in Jones v MBNA at [38]. The court must be confident that there is no realistic possibility that had the point been raised below there would have been evidence which the other party could or would have adduced which might have had any possibility of affecting the outcome. Moreover where there has been no disclosure relevant to the new way in which the appellant seeks to put its case, the court cannot reasonably expect the other party to be specific about the evidence it would have adduced had the point been raised at the trial. As Arden LJ observed in Crane t/a Indigital Satelite Services v Sky In-Home Ltd [2008] EWCA Civ 978 at [21]:

Thirdly, it will usually be fatal to the application if the new point requires findings of fact which the Judge has not made. The appellant may be able to point to evidence at the trial which would support a finding of fact which is necessary for the new point to succeed. But unless that evidence was expressly agreed, in the absence of a finding by the judge accepting that evidence, there will not ordinarily be the necessary evidential foundation for the new point to succeed.

Fourthly, even if there is evidence at the trial which is agreed or unchallenged, it will often be unfair to allow it to be relied on in support of a new point taken for the first time on appeal if it was not directed to the points then in issue and was not necessary or relevant to their resolution. In conducting litigation the parties can be expected to address the facts which matter for the way in which the claim has been framed, and indeed the overriding objective requires them to do so. As Lord Herschell put it in The Tasmania at p. 225:

Fifthly, care needs to be taken in interpreting answers given by a witness at trial when the appellant is seeking to rely on them for the new point. If the answers were given by reference to a different point in issue at the trial, there is a real risk of unfairness in assuming that they convey the same meaning by reference to the new point. Some of Mr Berry’s submissions illustrated this unfairness. He relied on evidence given by both Mr Smith and Ms Bossley that they had seen no evidence that Vitol externally hedged “this risk”. The questions were asked, and the answers given, in relation to a specific external hedge of the pricing risks created first by the TOTSA contract itself and then by the effect of delay on the pricing risk, i.e. an equivalent external swap to that initially entered as a notional paper swap internally in portfolio 645 and then as entries for its rolling. The question and answers were not addressed to a book hedge of whatever the risk might have looked like at the end of a process of netting off within the Vista system, still less to a hypothetical hedging of a book risk which would have existed but for Rhine’s breach. Mr Berry sought to treat this evidence as an admission that Vitol would not, in the counterfactual scenario, have left unhedged any book risk which was the oppositional equivalent to the pricing risk arising from the delay caused by Rhine. It was no such thing, and neither witness was addressing their mind to the question of what book hedge Vitol would have purchased in the counterfactual hypothesis of no breach. The answers they gave were not addressing that question.

I have reached the clear conclusion that Rhine should not be permitted to raise the new point on appeal because determination of the point would require further factual findings, not made by the Judge, on which it is in my view likely (and at the very lowest possible) that further evidence, which might affect the outcome, could and would have been adduced at trial by Vitol. This applies to each of three elements which are essential to the success of the new point, any one of which would be fatal to the application, namely:

Paragraph 10 of the supplementary skeleton argument containing the new point asserts that “in the non-breach counterfactual there would have been no rolling. It necessarily follows that the ‘opposite risks’ deriving from the complex mix of other transactions would not have been offset by the rolling. Those opposite risks would necessarily have been identified by Vista as an unmatched and un-offset risk of a fall in the market causing a net exposure to a fall in the market price of 920,000 physical oil over 7 days.” The ‘fall in the market price’ must be intended to refer more specifically to a fall in the published prices for Brent Direct in Platts. The reference to a fall “over 7 days” must be intended to be a reference more specifically to the fall between those published prices on 13, 14, 15, 18 and 19 May 2020 and those published on 7, 11, 12, 13 and 14 May 2020. To talk simply of a fall in the market over 7 days is an oversimplification of the risk, which affects the nature and value of an oppositional pricing risk or a book hedge of it.

That more specific formulation of the oppositional pricing risk is indeed the basis for Rhine contending that the avoided loss on the book hedge is equivalent in money terms to the profit which would have been made by rolling the swaps had the roll been an external hedge. It is implicit in this that the oppositional risk asserted would have been of a fall in the Brent Crude price published on the Platts Index for the specific publication dates covered by the internal rolled swaps.

It will be immediately apparent that it is a matter of speculation whether there would have been such an exact oppositional risk in Vitol’s physical trading book; and given the spread of dates and amounts involved, being fractions of a 920,000 bbl cargo at different dates, it seems improbable. Indeed it was put to Mr Smith by Mr Young on behalf of Rhine in cross examination that such a perfect match was unlikely.

The new point addresses the question at a higher level of generality by assuming that once complex offsetting had taken place within the Vista system the remaining risk would be the exact equivalent. This, however, is also far from having been established by the evidence about the Vista system put forward at trial. The evidence was that the Vista system was intended to allow many different risks to be pooled and set-off across the book of each sub-division and then division, for the purposes of an assessment of overall risk to Vitol. The evidence did not address whether that latter assessment involved assessing risk across divisions or whether, for example, there was one ‘book’ for the crude trading division, and a separate ‘book’ for other divisions, with book risk being assessed separately for each division. An answer of Mr Smith referring to setting off gasoline pricing risks suggests that it may have been the former, but this was simply not addressed in the evidence. It appears from the evidence of what was contained in the portfolios examined, namely 645 and 101, that a variety of different risks would be collected within a single portfolio. The pricing risk in issue in this case arose by reference to market prices published by Platts for Brent Direct. The portfolios 645 and 101 contained pricing risks by reference to other indices, notably Brent Futures on the ICE index. They included foreign exchange risk. They may have contained others, because the evidence was that they contained thousands of entries, most of which were not before the court at the trial. Nor were the entries in any other portfolios. If there was to be a book wide netting off to assess overall risk, whether division by division or on Vitol’s whole book, there would have to have been aspects of the system which correlated a pricing risk of market movement as expressed in different indices and over different periods of time; and there would have to have been a system which allowed a correlation between all the risks of different kinds which were included within the portfolios. It is simply not known what these contained by way of different kinds of risk other than through the small window of the evidence about portfolios 645 and 101, which was not complete even in relation to those portfolios.

Moreover it was not explored in the evidence at the trial how exactly the Vista system went about netting off risk. It was not clear how and to what extent there was human intervention by commercial analysts (which the evidence established was what determined that portfolio 645 was created for the matching of the purchase and sale contracts, and the internal hedge of the swaps and their rolling between 645 and 101); and how or to what extent the netting off was an algorithmic function of the code written into the Vista system in netting off between risks.

None of this was explored in the evidence because it was Rhine’s case, specifically confirmed by Mr Beckett in his evidence, that it was not necessary to examine what happened in the Vista system beyond the roll-over of the swaps from portfolio 645 into portfolio 101. Indeed Mr Young expressly confirmed this in his oral opening submissions, as recorded at paragraph [175] of the Judgment. However if it were necessary to identify what a pricing risk looked like at the end of the Vista system process for the purpose of assessing overall book risk, it would have been necessary to understand how the system worked so as to be able to identify what the unhedged ‘oppositional risk’ would have looked like. Without such evidence it is simply impossible to draw the conclusion, upon which the new point depends, that but for the breach there would have been an oppositional pricing risk equivalent to the risk of a loss from a rise in the market arising from Rhine’s breach by reference to specific published prices for specific amounts on specific dates.

The new point would also have raised questions of the margin of accuracy of the system. Vista was intended to assess overall risk, but it is unlikely that it could do so with 100% accuracy to the nearest cent given the multiplicity of risks and factors involved in different aspects of Vitol’s trading in different products by reference to different indices and different currencies. The bespoke algorithms and/or decisions by commercial analysts might well have had to make value judgments about correlations. The sums involved in Rhine’s breach and the rolled swaps were tiny compared to Vitol’s overall trading, the latter being estimated in one press article before the Judge as involving revenue in 2020 of about US$3 billion and trading in oil alone of an average of 7.1 million bbls per day (described in one Financial Times article from June 2020 as the equivalent of the combined consumption of the UK, Germany, France, Italy and Spain). Unless the Vista system was accurate to a degree of tolerance of less than this fractional proportion, it might not produce any ‘oppositional pricing risk’ at all; and unless it was 100% accurate to the nearest cent it would not produce an ‘oppositional pricing risk’ which was equivalent to the effect of Rhine’s breach. It seems improbable that Vitol would claim 100% accuracy for the system, whether by reference to algorithmic code or human judgment or the combination of both. But all that matters for present purposes is that it is at least realistically possible that evidence could and would have been addressed to this question had the new point been raised below.

The new issue also raises questions about the timing of any assessment of book risk and of the decision whether to hedge. Mr Smith’s evidence in his statement about hedging the aggregate position of its book, which he characterised as a ‘book hedge’, was that “These hedges (to the extent risk is laid off externally) are dynamically managed as that book of trades adjusts over time”. The Vista system was dynamic in that the netting off processes would result in a constantly shifting position by reference to the continual trading activity (including derivatives trading) across Vitol’s entire business activity (or division, if they were treated individually). What might appear as an overall exposure to the risk of a fall in the market at one moment might appear as exposure to a risk of a rise in the market the next, or a further increase or reduction of exposure in either direction. The evidence did not address exactly when and how often a decision would be made to effect external book hedges. It did not even identify who took them (Mr Smith explained that it would not be him or any of the commercial analysts). Moreover a chart attached to Mr Beckett’s supplementary report suggested that there might at any given time be a pool of internal and external hedges awaiting allocation, but without any detail of for how long they would remain unallocated or when and how a net book position was evaluated for external book hedge decisions to be made. Without evidence about these matters it becomes even less possible to say whether there would have been an oppositional pricing risk which required a decision on book hedging.

For these reasons Mr Berry’s simplistic theoretical approach to the hedging of the book as a whole may well be falsified in practice when the particular workings of the Vista system are explored and understood. That is a matter on which there was no evidence at the trial and no relevant findings by the Judge. Rhine’s supplementary skeleton argument seems to recognise this at paragraph 15 by accepting that ‘the precise terms of the external hedge are impossible to assess precisely’. That is because the terms of the oppositional risk which it is asserted would have been externally hedged (in the counterfactual) could not be ascertained precisely. In truth, on the evidence adduced at trial, the nature and terms of such oppositional risk could not be identified at all.

Mr Berry argued that Vitol would not have run a speculative position of a fall in the price of 920,000 bbl of physical oil over 7 days. Even assuming that this generalised characterisation of the oppositional risk is sound, which it is not, the conclusion that it would have been externally hedged is not supported by the evidence at trial or the Judge’s findings (neither of which, of course were specifically directed to this question which was simply not in issue).

Mr Berry submitted that this conclusion followed (a) as a matter of inference from the Judge’s findings at [144] and [178]; and (b) from strands of evidence said to establish that Vitol would hedge against a rise or fall in the market wherever possible. Neither aspect of the argument is sound.

As to the judgment, paragraph [144] concluded that the likely position within Vista in relation to the notional paper hedge sold by portfolio 101 to portfolio 645 in respect of the rolling of the swaps was that it was offset by risks elsewhere in the system, albeit by a complex mix of other transactions both physical and paper. That was premised, as is clear from paragraph [143], on Mr Smith’s evidence that he had not identified an external hedge “in relation to this risk (i.e. the pricing on the TOTSA contract)”; and that in the absence of any positive decision to externally hedge or speculate “in relation to a particular risk”, the inference was that the risk was offset within the Vista system. The premise was an absence of an external hedge, or decision to speculate, specifically on the TOTSA contract risk which arose as a result of the breach. It was not whether there was or would have been a decision externally to hedge or speculate in relation to other net risks within the system at any given time, and a fortiori risks which did not exist (because the argument addresses the counterfactual, not the breach scenario). Paragraph [144] cannot be treated as containing any conclusions about unidentified oppositional risks in the non-breach scenario and whether they would have been externally hedged or a speculative position run in relation to them or the book in general. These questions were simply not in the arena at the trial by reference to the arguments which Rhine was advancing, and the Judge was making no findings about them. He had said at [142]:

Nor can the inference be derived from paragraph [178] of the judgment. The only point which the Judge was making there was that Rhine was seeking to read too much into Vitol’s statement that it sought to “[hedge] directional price risk where possible” (see below) by treating it as a statement that it externally hedged every transaction individually. The paragraph tells one nothing about the Judge’s views on whether a book hedge would have been made on an oppositional risk not identified at the hearing in a counterfactual situation not considered at the hearing.

Indeed the finding for which Mr Berry contends would be inconsistent with the Judge’s express finding at paragraph [51] that “To the extent that Vitol is long or short on a net basis any particular position across its portfolio, Vitol may choose to use external hedges (i.e. with third parties) in order to protect its position as a whole or it may choose to maintain that market position.” So too with paragraph [139] quoted above.

Nor did the strands of evidence relied on, whether individually or cumulatively, come anywhere close to establishing that Vitol would not have run a speculative position in relation to the unidentified oppositional pricing risk but rather would have externally hedged it. The evidence at trial was not addressed to this issue, which would no doubt have been explored more fully and directly had the new point been squarely raised. But such as it was it undermines rather than supports this aspect of the new point.

The evidence relied on by Mr Berry was as follows:

By contrast there was also evidence which positively suggested that Vitol did take speculative pricing risks:

These snippets of evidence, in a trial which did not address the factual issue raised by the new point, illustrate that determination of this aspect of the new point would require further factual findings, not made by the Judge, on which it is likely, and at the lowest possible, that further evidence, which might affect the outcome, could and would have been adduced at trial by Vitol.

For the reasons I have explained, it is impossible to identify the nature and terms of the oppositional pricing risk which it is said would have been the subject of a book hedge in the counterfactual scenario of no breach. Paragraph 15 of Rhine’s supplementary skeleton argument accepts that the terms of such a book hedge cannot be identified, but asserts that they would likely have given rise to the same profit as was recorded internally by the rolled swaps in portfolio 645. There is no basis for such an assertion once it is accepted, as it is, that the hedge would not have been in the same terms as the internal rolling of the swaps. Mr Berry submitted that this court could be confident that it would have involved at least an equivalent loss to the profit booked for the roll-over of the swaps because any margin for error would be accounted for by the fact that an external hedge would have involved transaction costs and so given rise to a greater loss than the profit booked in portfolio 645. However since the nature and terms of the avoided book hedge cannot be identified there is simply no basis for such an assertion. Unless the book hedge would have been an exact match for the rolled swaps booked in portfolio 645, which it is accepted it would not have been, it is simply impossible to say what loss it would have produced and therefore what loss was (on this hypothesis) avoided by the breach. Mr Berry described the profit booked on the rolled swaps as a proxy for the ‘best estimate’ of the loss on the avoided book hedge. I would not accept that that is so without any evidence addressed to the issue: a miss may be as good as a mile. There are no findings, nor any agreed/unchallenged evidence directly on the point, and the proposition which Mr Berry asserts does not follow inexorably from any such findings as were made or any such evidence adduced at the trial.

Mr Berry argued that the new point could have been advanced in closing submissions because it was within the scope of the pleadings, in which case it would have been too late for Vitol to adduce any further factual issues to which the new point gives rise. This is not sufficient to meet the point that Rhine does not have findings by the Judge or agreed/unchallenged evidence on the three crucial aspects of the point I have identified which are each necessary for the point to succeed. But the submission is in any event unsound in my view.

It is true that the Defence to Counterclaim pleaded the hedging point in very general terms at paragraphs 25(3) and (4):

However, even the width of this pleading, couched in terms of hedging arrangements Vitol had, is not sufficient to encompass an argument that what falls to be taken into account is a hedging arrangement which Vitol did not have, which is the avoided book hedge upon which the new point rests.

Moreover the scope of a party’s case is usually more precisely defined prior to the trial by lists of issues and skeleton arguments, which the opposing party is entitled to treat as confining the limits of the case being advanced by its opponent; and an attempt to raise in closing submissions a point outside the scope of such narrowed issues should not be allowed if it would have affected the approach to the evidence: see Clements-Siddall v Dunbobbin Hotels Ltd [2023] EWCA Civ 1300 [2024] PIQR P4 at [78] and [80]-[92]. We were not provided with any such list of issues or the skeleton arguments for trial in this case, but it is reasonable to conclude from the fact that the point was not advanced at the trial that it was outside the scope of the arguments contained in Rhine’s skeleton argument. Rhine’s case was also narrowed in the expert evidence served on its behalf prior to trial in the form of Mr Beckett’s report, which relied solely on the internal hedges recorded in portfolios 645 and 101 and made clear that no reliance was being placed on what happened to the risk within the Vista system after the swaps were rolled from portfolio 645 into portfolio 101. I therefore have little doubt that had Rhine sought to raise the point for the first time in closing arguments, it would have been met with the objections advanced in this court, and that the point would not have been permitted to be advanced, either at all by reason of its very late introduction, or at least without affording Vitol an opportunity to address it by additional evidence.

There remains to be considered Vitol’s argument that the new point would be bound to fail for the additional reason that an avoided book hedge loss would be res inter alios acta. I do not see how it could be characterised as a step taken in mitigation of loss since what is posited is (a) the absence of any decision to make a book hedge (b) in the counterfactual no breach situation in which there would be no knowledge of the breach. An analysis treating it as an avoided loss engages consideration of whether it is a collateral benefit in accordance with the principles in Swynson Ltd v Lowick Rose.

Mr Berry submitted that the test propounded by Lord Sumption was merely one of causation, save in the category of cases, not here relevant, where the party is treated as having made the gain from his own resources, as with insurance. A causation analysis is not, however one of ‘but for causation’; and the authorities suggest that what amounts to causation in law raises special considerations in this context. The hearing before the Supreme Court in Swynson took place at the same time, and before the same constitution, as the appeal in Globalia Business Travel SAU v Fulton Shipping Inc (The New Flamenco) [2017] UKSC 43 [2017] 1 WLR 2581. That was an example of a case in which a gain made as a direct result of what an arbitrator had found was a step taken in mitigation of loss caused by the breach was held to be a benefit which did not fall to be taken into account, because there was insufficient causation in law for these purposes, notwithstanding that it would have satisfied a causation in fact analysis in other contexts. The authorities which I referred to at length and summarised at first instance in that case ([2014] EWHC 1547 (Comm) [2014] 2 Lloyd's Rep. 230 at [9] to [62]) illustrate that causation issues in this context are not straightforward.

My summary was set out in the Supreme Court judgment at [16] without disapproval, but without express approval save to the extent of the Court expressing a preference for my reasoning generally (at [28]) and express endorsement of proposition 8 (at [30]).

None of these authorities was cited to us in this appeal. My purpose in referring to them, and my attempt in The New Flamenco to summarise the principles established by them, is not to treat that summary as an authoritative statement of the law. Rather it is to illustrate that questions of causation in this area raise difficult issues involving taking into account a wide variety of circumstances, being matters of fact and degree which are highly sensitive to the particular circumstances of individual cases. For that reason, and although I consider that the collateral benefit principles would place formidable difficulties in the way of Rhine’s new argument succeeding even if the avoided loss were made out on the facts, I would prefer not to express a final concluded view on the question in circumstances where the relevant process of causation through the Vista system has not been the subject matter of the necessary evidence and findings of fact. It is unnecessary to decide the issue because Rhine should not be allowed to raise the new point for the reasons I have identified.

For these reasons I would dismiss the appeal.

I agree.

I also agree.