Marex Financial Ltd v Creative Finance Ltd & Anor

This is a claim for money alleged to be due on an account between the Claimant (“Marex”), a foreign exchange (“FX”) broker, and two of its clients, the Defendants, Creative Finance Limited (“Creative”) and Cosmorex Limited (“Cosmorex”).

Following the severe earthquake and catastrophic tsunami that struck Japan on 11 March 2011, the US Dollar/Japanese Yen exchange rate dropped steeply and on 16 March 2011, Marex, as it was admittedly entitled to, closed out: (i) Creative’s accumulated NZDJPY positions under which it was obliged to buy a total of 450 million New Zealand Dollars (“NZD”) in exchange for Japanese Yen (“JPY”); and (ii) Cosmorex’s accumulated EURJPY positions under which it was obliged to buy a total of 360 million Euros (“EUR”) in exchange for Japanese Yen (“JPY”). The close-out of the EURJPY positions was achieved by trades of 160 million at different times over the Currenex dealing platform and trades of two lots of 100 million at different times through two market makers, ANZ and Citibank, at the best price achievable (“at best”). The close out of the NZDJPY positions was achieved by “at best” trades also through ANZ and Citibank.

Marex debited Creative and Cosmorex sums converted to US dollars (“USD”) that reflected the average cost of the close out transactions with a mark up by way of profit in the sum of US$ 2,753,689.19. The mark up was not separately particularised but was simply included in the sums debited. The result was an overall negative ledger balance across Creative’s and Cosmorex’s accounts of US$ 7,959,383.03, which is the sum sued for, plus interest.

The defences advanced by Creative and Cosmorex allege that the manner in which the positions were closed out was irrational, alternatively grossly negligent, or negligent. In the Defendants’ submission, there was only one rational way of closing out the positions and that was by selling each of their entire positions to a market maker at a full amount “risk transfer price” which it is alleged would have produced an outcome that, far from giving rise to negative balances on their accounts, would have left those accounts in credit in respect of which they bring a Counterclaim. Creative and Cosmorex also contend that Marex has no right to charge a profit on the close-out transactions.

Marex accepts that in exercising the power to close out it was under a duty to proceed rationally, i.e. in a manner that is not arbitrary, capricious or perverse, but it denies that the way in which it closed out the Defendants’ positions was in breach of this duty. Marex also denies that it was under a contractual or tortious duty to exercise reasonable care and maintains that, if, which is denied, it owed such a duty, the way in which it carried out the close out did not involve a failure to take reasonable care. Finally, Marex contends that it had a contractual right to charge a profit on close-out transactions of such size as it should determine, alternatively that it had a right to charge a reasonable profit, which is what the profit in fact charged was.

The FX market is the largest financial market in the world and is open 24 hours a day, five days a week. It is split into three overlapping markets: London, New York and Asia (or Australasia, taking both Sydney and Tokyo into account). The London market opens at 8am and closes at 5pm. The unsynchronised adoption of Daylight Saving Time in the various markets means that opening hours of other markets, relative to London time, are not constant. Generally speaking, however, the New York market opens at 1pm and closes at 10pm (London time) and the first Asian market (Sydney) opens at 9pm (London time), with Tokyo opening 2 hours later at 11pm. There is usually therefore an overlap of one hour between the time when the New York market closes and the Sydney market opens. At the time of the close-out on Wednesday 16 March 2011, however, the New York clocks had moved back by an hour, but the Asian clocks had not, with the result that there was no overlap between the New York and Asian markets when New York closed at 9pm London time.

It is generally accepted that the London market is where the highest volume of FX trading takes place and that the Asian markets have the lowest liquidity.

In general the FX market is split into two categories, liquidity providers and liquidity takers, although with the growth of e-trading these lines have become more blurred. Traditionally, liquidity providers or “sell side” included banks and brokers, whilst liquidity takers or “buy side” included hedge funds, corporations, professional and retail clients, asset managers and central banks. Both Creative and Cosmorex are examples of professional ‘buy side’ clients.

The FX market is “over the counter” (“OTC”) rather than exchange traded. Thus, rather than there being a single source of price information or liquidity available there are numerous sources of such information eg Reuters, CME (Chicago Mercantile Exchange) and EBS (Electronic Broking System), as well as proprietary single bank electronic trading platforms such as Autobahn and CitiFXPro and multi-bank platforms such as Currenex and FXall

The four most frequently traded currencies are the USD, the EUR, the JPY and the Pound Sterling (“GBP”). Currencies are traded for other currencies and a trade is described by reference to the currency “pair” that is being exchanged, e.g. “EURUSD” or “USDJPY”. Most trades involve the USD and any currency pair not traded against that currency is called a “cross currency pair”. Historically, a “cross” trade of this kind had to be carried out via the USD but today there is a number of cross pairs that can be traded directly, including EURJPY and NZDJPY.

When a price is given for a particular currency pair, the figure quoted represents the amount of the second currency in the pair that is needed to buy one unit of the first currency in the pair. For example, if a rate of 90.0 is given for the USDJPY pair, this means that at the relevant time JPY 90.0 is needed to purchase USD 1. Although currency prices are usually quoted to four decimal places, prices for pairs involving the JPY are quoted to two decimal places only. Whatever the currency pair, the smallest unit of currency is referred to as a “pip”.

Currency trades can be either “spot” or “forward”. A spot trade is an agreement to exchange currencies on the “spot date” – usually two business days from the date of the transaction (“T + 2”) – whereas a forward trade (also known as an "outright”) is an agreement to exchange currencies at some later date. In both cases the price is fixed at the time of the transaction. However, with a forward contract, the price will take into account the gains or losses that can be made, over the maturity period, from the interest rate differential between the currencies in the traded pair.

Marex operates as a broker and does not generally take proprietary positions for its own book. Instead, it makes its money from providing clients with access to the market (liquidity). In so doing, Marex enters into direct trades with its customers on a principal to principal basis and does not act as agent bringing its customers into a contractual relationship with a third party.

Customers of Marex wanting to open or sell a position will either telephone the Marex FX desk or contact the desk using the instant message “chat” systems provided by Bloomberg (Footnote: ¹) and Reuters (Footnote: ²) or use Marex’s online trading platform (“Marex Black”). When seeking to open a position by telephone, a customer may ask for a bid-offer price on a particular currency pair for a particular notional amount or he may give an “at best” or “limit” order for a particular currency. (The “bid” price is the buying price, the “offer” price is the “sale” price and the difference between the two is the “spread”). If a customer asks for a bid-offer price for a particular currency pair and amount, the FX desk will look at prices in the market for that currency pair and amount on its single-bank or multi-bank electronic trading platform screens and will see what bid-offer prices (spreads) are available. Marex might also contact “market makers” or “liquidity providers” (major international banks) directly over the phone or over the bank’s instant messaging “chat” system. The bid-offer price on the screens is typically for an amount of between USD1 million and USD5 million (known as a “regular order”). The screens quote a wider spread for larger amounts.

In normal markets, and for liquid currency pairs, Marex takes the price obtainable in the market and then builds into it a pip or two of profit before offering it to the customer. The more illiquid the currencies, the wider the spread that will be quoted.

If the customer accepts the quoted price, the desk books the trade there and then, holding a naked position until a back-to-back order (a “fill”) is placed in the market, thereby laying off the trade with the customer. If a customer places a large order above US$ 10million the desk might cover that position (depending on in-house proprietary limits) by trading smaller “regular” amounts in the market, usually at better prices thereby making an additional profit for Marex.

Creative and Cosmorex are BVI companies owned and controlled by Mr Carlos Sevilleja Garcia. Mr Sevilleja is a self-employed businessman based in Dubai and Valencia, Spain. After serving an apprenticeship with UBS in Zurich, Mr Sevilleja started to trade FX on his own account in 1988. His principal contact at Marex was Mr Farooq Muzammal whose main role was to sell Marex’s services to new clients and maintain relationships with existing clients. Mr Sevilleja first met Mr Muzammal in 2001 when Mr Muzammal was working with IFX, a London FX broker that Mr Sevilleja was then using. When Mr Muzammal moved to Refco Capital Markets (“Refco”), Mr Sevilleja opened accounts at Refco for Creative, Cosmorex and himself and over time built up some very sizable positions with Refco. In October 2005, Refco went into liquidation and Mr Muzammal joined Marex following which he succeeded in persuading Mr Sevilleja in 2007 to agree to open accounts for Creative and Cosmorex.

Mr Muzammal is no longer employed by Marex and did not give evidence.

The two agreements entered into by each of Creative and Cosmorex with Marex were in common form and on Marex’s standard terms. These consisted of a Foreign Exchange Options Master Agreement (“the FEOMA”) executed on 22 March 2010 and a Professional Client Agreement (“the PCA”) executed on 30 March 2010.

Marex required Creative and Cosmorex to provide margin calculated at the rate of 1% of their notional positions up to a maximum of US$ 1billion, rising to 2% on positions between US$ 1 billion and US$ 2 billion. However, under a cross-margining agreement, Creative’s and Cosmorex’s two accounts were aggregated for the purposes of calculating margin so that the equity on one account could be used to the credit of the other, and vice versa. The purpose of margin was to give Marex cover for the maximum likely movement in the value of the Defendants’ positions in the course of a day. If there was insufficient equity in the Defendants’ accounts to cover the initial margin requirements and any variation margin loss (the open positions were marked-to-market daily), the Defendants were obliged to provide additional margin or face having their open positions closed out.

In the period 30 March to 23 November 2010, Creative entered into numerous trades with Marex under which it built up open positions to buy a total of NZD 450 million in exchange for JPY. In the jargon of the trade, Creative was long in 450 NZD and short in the necessary JPY to make the purchase. As for Cosmorex, it began entering into long EUR trades on 27 August 2010 and by 15 September 2010 had built up a total of EUR 360 million in that currency pair. Expressed in USD, the base currency of the Defendants’ contracts with Marex, the Defendants held a combined total position worth approximately USD 830 million as at 10 March 2011.

In the period 30 March to 24 September 2010, Marex placed price stop orders in respect of the Defendants’ positions as follows: (i) 30 March to 6 April 2010 with Citibank for 100 million NZDJPY at 60.80 then replaced by an order at 62.70 for 200 million NZDJPY; (ii) 6 April to 24 August 2010 with Deutsche Bank for between 300 million and 385 million NZDJPY at between 55.90 and 61.80; and (iii) mid August 2010 to 24 September 2010 at prices between 57.55 and 61.80 for: (a) 115 million NZDJPY with Citibank; (b) 90 million NZDJPY with HSBC; (c) 90 million NZDJPY with ANZ; and (iv) 90 million NZDJPY with Credit Suisse. All of these price stop orders would have been executed by the banks concerned obtaining the best price available in the market for quantities (“clips”) determined by them.

On 24 September 2010, all of the above-mentioned price stops were cancelled and replaced by an “equity stop” agreed with Mr Sevilleja under which Marex reserved the right to close out Creative’s and Cosmorex’s positions if the combined “equity value” of the two accounts fell below US$ 8 million. “Equity value” means the net liquidating value, namely, the total of the cash in all of the accounts, plus the mark-to- market value of the open positions.

There were also price stop arrangements in place on 16 March 2011 made through another broker, FIXI plc for positions held by the Defendants’ as follows: (i) 40 million EURJPY with a stop loss with UBS; (ii) 150 million NZDJPY, as to which there were price stops for 58 million with UBS, 50 million with RBS and 50 million with HSBC.

Mr Sevilleja was a “carry trader”. A carry trader seeks to make money from the difference between the interest rates set by the central bank in charge of one currency and the central bank in charge of another currency. To do this, the carry trader chooses a currency pair in which one currency offers a low rate of interest and the other a high one, and then builds up a forward position in that pair that is “short” (selling) the low interest currency and “long” (buying) the high interest currency. The forward exchange rate for such pairs will be lower than the equivalent spot exchange rate to take into account the profits that can be made by the trader from borrowing the low interest currency, converting it to the high interest currency and then lending this second currency out over the course of the maturity period. Therefore, on the forward contract’s maturity date, if the spot exchange rate has not changed, the trader will make a profit from the difference in value between the forward price he agreed to and the prevailing spot rate. The trader can repeat this strategy month after month by “rolling” his positions forward, taking a profit on each occasion. In addition, if the spot exchange rate increases in favour of the high interest currency (i.e. the currency in relation to which the carry trader is long), the trader benefits from an increase in the capital value of his position.

The potentially relevant provisions in the FEOMA and the PCA are:

EVENTS OF DEFAULT

NETTING

Mr Cox QC for the Defendants contended that Marex is illegitimately claiming two cumulative rights: first, a right to close out the positions by imposing trades upon the Defendants under the PCA; second, a right to “liquidate” the closed-out positions by converting them into USD under Section 8 of the FEOMA, thereby arriving at the calculation of the liquidated sum sued for. In Mr Cox’s submission, this is impermissible because both the FEOMA and the PCA grant Marex separate, self-contained and alternative methods for terminating and liquidating FX transactions entered into with the Defendants.

Section 8 of the FEOMA is prescribing not only the liquidation of terminated obligations but also the actual process of close-out itself. Thus, under clause 8.1(a), the “closing-out” of the “Currency Obligations” involves the termination of the parties’ respective delivery obligations and is accomplished by the Non-Defaulting party giving notice to the Defaulting Party. Clause 8.1(b) then provides for the “liquidation” of those Currency Obligations so as to determine what (if any) sum is payable by the Defaulting Party. Liquidation essentially involves three steps: (i) the calculation of a closing gain/loss (clause 8.1(b)(i)); (ii) the discounting of this closing gain/loss to reflect present value (clause 8.1(b)(ii)); and (iii) the netting of all closing gains or losses as between the parties to determine a net sum payable by one party to the other (8.1(b)(iii); 8.1(d))). No part of this process involves the entry into new transactions or any consideration of back-to-back transactions (if any) which Marex has entered into with market counterparties.

Unlike the FEOMA, the PCA is drafted specifically in contemplation of Marex holding back-to-back positions. Under clause 16 of the PCA, if there is an Event of Default or if Marex determines that the Defendants have not performed any of their contractual obligations, or reasonably believes that the Defendants will not be able or willing in the future to perform any of them, Marex is entitled to close out, replace or reverse any transaction (which includes Marex’s back-to-back trades in the market), buy, sell, borrow or lend or enter into any other transaction or take, or refrain from taking, such other action at such time or times and in such manner as, at its sole discretion, it considers necessary or appropriate to cover, reduce or eliminate its loss or liability under or in respect of any of the customer’s contracts, positions or commitments. However, clause 16 does not provide a formula for calculating or recovering those losses from the Defendants: this is provided by clause 15, under which the parties’ obligations one to the other are satisfied by settlement of the “Liquidation Amount” (clause 15.4(a)) which is calculated in accordance with clause15.4(b)-(c).

Marex has at no stage sought to rely upon the FEOMA formula, which (as above) requires a calculation of who, as between Marex and the Defendants, is the “winner” and the “loser” under the relevant currency positions.

Instead, Marex in substance is seeking to pass on to the Defendants the loss it suffered on its back-to-back trades. This claim cannot be brought under the FEOMA; it can only be brought under clause 15 of the PCA which in clause 15.4 (b) and (c) refers, inter alia, to “loss”

Where a contract provides for one party to indemnify another against specified “loss” then (absent provision to the contrary) this obligation is limited to loss which was not caused or contributed to by the negligence of the receiving party.

The FEOMA provisions and the CPA applicable on an Event of Default are to be read cumulatively. Marex was entitled to (and did) “close out” pursuant both to Clause 16.1(c) of the CPA and Parts IX (c) and XVI (d) of the FEOMA Schedule. Marex’s claim is therefore not based on Clause 15.4 of the FEOMA and the court is not concerned with having to construe the word “loss” in that provision.

The terms “close out” and “close-out” in both the FEOMA and the CPA are terms which are to be construed as meaning something more than a mere notional event. They each mean the actual conclusion of an equal and opposite transaction to cancel out the position that needs to be closed out and leaving a net payment (if any) to be made by one party to the other. This is the generally accepted meaning of “close out” in the FX industry and it is what is provided in Clause 9.3.8 of the CPA.

The Defendants’ suggestion that close-out under Section 8.1(a) of the FEOMA is merely accomplished by the Non-Defaulting Party giving notice to the Defaulting Party ignores the fact that, under Part IX(c) and Part XVI(d) of the FEOMA Schedule, Marex is entitled to close out without giving any notice to the Defendants. Thus, the position under the FEOMA is no different from the position under Clause 16.1(c) of the PCA, pursuant to which Marex is entitled to close out “without prior notice to” the Defendants.

It was plainly not intended under the FEOMA (or indeed under the PCA) that the close-out of the Defendants’ positions should or could only be effected by giving notice to the Defendants that Marex is terminating the relevant Currency Obligations. Indeed, Part XVI (d) of the FEOMA Schedule expressly provides that Marex does not have to take this step in order to close out the Defendants’ positions.

Thus, on the proper construction of the FEOMA, the relevance of the notice referred to in the second sentence of Section 8.1(a) is not that it constitutes the close-out itself, but that it assists in establishing the “Close-Out Date” in the context of the subsequent liquidation of the closed-out Currency Obligations under Section 8.1(b). The Defendants’ contention that the close-out process under the FEOMA involves no new transactions between Marex and the Defendants is therefore not correct. On the contrary, it is by the conclusion of new equal and opposite transactions between Marex and the Defendants that the Defendants’ existing positions would be closed out (and existing transactions terminated), as is explicitly contemplated in Clause 9.3.8 of the PCA.

The FEOMA does not refer to equal and opposite transactions but these are what the close-out process entails and Marex will inevitably base its prices for the (equal and opposite) closing trades between itself and the Defendants on the prices that it has obtained (or considers that it is able to obtain) in the market.

The Defendants’ analysis of the close-out process under the PCA is misconceived because: (i) it ignores the provisions of Clause 9.3.8 of the PCA regarding the meaning of “close-out”; (ii) Clause 15 of the PCA is concerned not with the close-out of positions in the manner described in Clause 9.3.8 of the PCA, but rather with the deemed termination and liquidation on a specially designated date of all Netting Transactions - i.e. all transactions outstanding between Marex and the Defendants - without (new) equal and opposite transactions being entered into between Marex and the Defendants; by contrast, Clause 16.1 (c) expressly confers the “additional” right “to close out … any Transaction”.

On proper analysis, Clause 15 of the PCA is a netting procedure whereby, at the election of the Non-Defaulting Party and upon designation of a Liquidation Date, all existing transactions can be treated as terminated and liquidated as of the Liquidation Date, without any equal and opposite (or closing) transactions being entered into between the parties. It is because this netting procedure does not require any equal and opposite transactions to be entered into that a calculation of the loss suffered (or gain made) by the Non-Defaulting Party as a result of the deemed termination of all Transactions (or Netting Transactions) is then required and provided for in Clause 15.4 in order to enable the final account between the parties to be settled. No such calculation is necessary, however, where there is a close-out under Clause 16.1(c) in the exercise of Marex’s separate and distinct right to close out under that clause. When equal and opposite transactions are entered into as part of such a close-out, the termination of both sets of Transactions arises automatically pursuant to Clause 9.3.8, rather than by virtue of any notice given under Clause 15.2 designating a Liquidation Date for the deemed termination and liquidation of all Transactions between the parties. What is left over after a close-out under Clause 16.1(c) is any net settlement payment due in respect of both sets of closed-out Transactions after their mutual cancellation (see Clause 9.3.8). No calculation of loss is then necessary, as required under Clause 15.4 in the event of a deemed termination and liquidation of all Transactions under Clause 15.

Once the close-out process had been completed - i.e. following the execution of back-to-back trades between Marex and the market, the execution of the closing trades (which were equal and opposite to the outstanding trades) between Marex and the Defendants, and the execution of the swaps to align the value dates of the outstanding trades and the closing trades - it was necessary for the final balance owed by the Defendants to Marex to be calculated in US dollars. This process - which was distinct from the close-out process - was undertaken by Marex pursuant to the liquidation procedure described in Section 8.1(b)(i) of the FEOMA

The closing gain or loss in USD resulting from the liquidation (including conversion into US dollars) of each of the closed out Currency Obligations (i.e. both the opening and closing trades) was specified in the Daily Equity Statement dated 17 March 2011 that was provided to the Defendants. There is no real issue between the parties as to what was involved in the liquidation under Section 8.1(b) of the FEOMA, namely: (i) the calculation of a closing gain/loss (Section 8.1(b)(i)); (ii) the discounting of this closing gain/loss to reflect present value where the value dates of the Currency Obligations were later than the Close-Out Date (Section 8.1(b)(ii)); and (iii) the netting of all closing gains or losses as between the parties to determine a net sum payable by one party to the other (Sections 8.1(b) (iii) and 8.1(d)).

In circumstances where the Currency Obligations in question formed pairs of equal (in terms of EUR and NZD amounts) and opposite (one long, one short) FX transactions, what this essentially amounted to was a calculation of the USD value of the Defendants’ JPY shortfall or surplus (in this case, a shortfall) that resulted from the close-out, discounted to reflect present value.

This was essentially what Marex did when converting the Defendants’ overall JPY debit ledger balance of JPY 5,512,562,000to the USD using the USDJPY rate of 78.60, which was the prevailing market rate at the end of the close-out process on 16 March 2011. Admittedly, in this case there was no formal discounting by Marex from the settlement dates of 22 March and 1 April 2011 back to 16 March 2011 (the Close-Out Date), but this is irrelevant because of Section 8.4 of the FEOMA which deals with payment of the net amount that is produced by the liquidation process. Under Section 8.4, the net amount payable to Marex attracts interest at overnight LIBOR from and including the Close-Out Date, so that although Marex ought to have discounted the final amount due from the Defendants back to that date, it was entitled at the same time to apply interest at the same rate, i.e. LIBOR, from that date. The absence of discounting will therefore not matter as long as Marex’s entitlement to interest is taken to accrue after 1 April 2011 and the pleaded claim for interest is from 20 April 2011.

In my judgement, Mr Choo-Choy’s submissions are correct. The term “close out” in the FEOMA connotes in my opinion the execution of an equal and opposite transaction to cancel out a position that requires to be terminated thereby establishing the net position between the parties in respect ofthe terminated position or positions. Both the expert witnesses who gave evidence before me described a close-out in substantially these terms and on this basis I readily conclude that this is how the term is generally understood in the FX market. This too is how a close-out is described in Clause 9.3.8 of the PCA, and it would be most surprising if the term meant something different in the FEOMA.

I am also of the view that the netting procedure prescribed by Clause 15 of the CPA is intended to apply only where no equal and opposite closing transactions have been executed by Marex.

Accordingly, I find that the FEOMA and the PCA are to be read side by side, so that where there is an Event of Default under either agreement, the rights conferred on Marex by the two agreements in such an eventuality are cumulative, with the result that Marex was entitled to: (i) close out the Defendant positions by executing equal and opposite transactions in the market pursuant to Clause 16.1 (c) (as further defined by in Clause 9.3.8) of the PCA and under Parts IX (c) and XVI (d) of the FEOMA Schedule; and (ii) to establish the final balance owed by the Defendants pursuant to the liquidation procedure set out in Section 8.1 (b) (i) of the FEOMA.

I further find that Marex has in fact proceeded in accordance with this contractual scheme.

In the result, I find that Marex’s claim is not made under Clause 15 of the CPA and, at least to the extent that the balance sued for represents loss on the closed out positions, is properly constituted and contractually justified (subject only to any set-off by way of counterclaim).

The close out of the Defendants’ positions involved a three stage process. First, Marex entered into a series of trades with market counterparties, selling the long NZD and EUR positions that Marex would have to buy back from Creative and Cosmorex in order to close them out. By entering into these trades, Marex “locked in” the NZDJPY and EURJPY rates achieved in the market. Second, Marex concluded trades with Creative and Cosmorex to reverse the positions that were outstanding immediately before close-out. Third, the prompt dates of the trades done in the market and with Creative and Cosmorex (22 March 2011) had to be aligned with the prompt date of Creative and Cosmorex’s original trades (1 April 2011) so that netting off could be achieved. This was done by a NZDJPY swap trade to roll forward the short NZDJPY spot position value date 22 March 2011 to the long NZDJPY forward position of value date 1 April 2011 for value 22 March 2011, and the same amount was then sold for settlement 1 April 2011. As to the EURJPY trades, the long EURJPY position with value date 18 March 2011 was rolled forward to the short spot position with value date 22 March 2011.

The price which Marex gave to Creative and Cosmorex on the close-out trades was 55.61 for the EURJPY positions and 108.02 for the NZDJPY positions, which represented mark ups of 33 pips and 22 pips respectively. This is the first element of the total profit claimed in the action. There was also a mark up of 0.7 pips and 0.35 pips respectively on the NZDJPY and EURJPY swaps and rolls which amounted to 1.248 million and 3.11 million yen respectively. Marex claims corresponding US$ amounts of $15,877.86 and $ 39,567.43 in respect of these sums. This is the second element of the total profit claimed. The third element of the profit claimed arises from the conversion to USD.

The first two of the above mark ups were fixed unilaterally by Marex. Mr Muzammal told Mr Marston, Marex’s Head of FX, that the profit on the close-out should be USD 2.5 million and these two mark ups were fixed to achieve this figure. Mr Sevilleja expected Marex to charge a mark up as part of the close out process, but no specific agreement was made between Marex and the Defendants that Marex should be entitled to do so.

Mr Cox submitted that Marex had no legal entitlement to claim any profit where the customer’s positions are closed-out. As he pointed out, neither the FEOMA nor the PCA contains any express provision entitling Marex to claim any profit in such circumstances.

Mr Choo-Choy argued that the right to claim a profit was part and parcel of the contractual right to close-out which included the essential second stage of the process by which Marex bought the long NZD and EUR positions back from Creative and Cosmorex. Since the price of such trades is not up for agreement between Marex and the customer in the way it is when trades are originally placed by the customer, the price is necessarily to be determined by Marex in the exercise of its discretion when closing out. As to the size of the mark ups, Mr Choo-Choy submitted that they were consistent with Marex’s admitted duty to act rationally when closing out. In this connection, Mr Choo-Choy maintained that the mark ups were not disproportionate compared to the bid-offer spreads experienced by the market on 16 March 2011, including during the period of the actual close-out.

In my judgement, there was no contractual right to charge a mark up on close-out trades as Marex did. True it is that a mark up is contractually charged when trades are entered into following the placing of an order by a customer, but this is because the mark up is part of the price that the customer agrees to when placing the order. When close-out trades are concluded, however, the customer is not agreeing to Marex’s proposed price. The customer has agreed by entering into the FEOMA and the PCA that if the right to close out is triggered he must sell the long positions to Marex but he has not agreed that Marex can charge a mark up, let alone any mark up Marex chooses to make, subject only to the obligation not to act arbitrarily, capriciously or perversely. In my opinion, such is the potential impact of the right to charge a discretionary mark up, the right would only arise as a contractual right if it were expressly provided for in the FEOMA or the PCA. The purpose of the right to close-out is to reduce the customer’s exposure to Marex thereby protecting Marex from the risk of the customer defaulting. That purpose is achieved if Marex charges a price that covers the difference between the cost of concluding a matching trade or trades in the market and the price to be credited to the customer on the close-out trade. In my view, that is the extent of the contractual right to charge a close-out price that arises from the FEOMA and the PCA.

Accordingly, to the extent that the sum sued for represents the mark ups on the prices achieved in executing the equal and opposite trades in the market and on the swaps and rolls, Marex’s claim fails.

As already recorded, Marex accepts that, when closing out the Defendants’ positions, it owed a duty not to act irrationally, capriciously or perversely. This concession was made in light of the decision of the Court of Appeal in Socimer International Bank Ltd v Standard Bank London Ltd [2008] 1 Lloyd’s Rep 558. There, an agreement which governed the making and performance of individual forward sales transactions of emerging markets securities provided that, where the buyer (“B”) defaulted, the defendant as seller (“A”) had the right to liquidate or retain sufficient Designated Assets and to apply the proceeds of their sale to satisfy to the extent possible any amounts payable to the Seller under the agreement. The relevant clause further provided:

The Seller may in its sole and absolute discretion sell the Designated Assets at such time, in such manner and at such price as it deems reasonable and appropriate. The value of any Designated Assets liquidated or retained and any losses, expenses or costs arising as a result of the termination or sale of the Designated Assets shall be determined on the date of termination by Seller.

Upon a default by the claimant Buyer, the defendant did not carry out such a valuation but sold what assets it could from time to time and credited the proceeds to the claimant. The claimant claimed that the defendant was obliged to have carried out such a valuation at the date of termination and that if it had done so, then, pursuant to an implied term, it should have acted with reasonable care in carrying out the valuation. The Court of Appeal, citing Abu Dhabi National Tanker Co v Product Star Shipping Ltd (No 2) [1993] 1 Lloyd’s Rep 397, Ludgate Insurance Co Ltd v Citibank NA [1998] Lloyd’s Rep IR 221, Gan Insurance Co Ltd v Tai Ping Insurance Co Ltd (No 3) [2002] Lloyd’s Rep IR 612 and Paragon Finance plc v Nash [2002] 1 WLR 685 held that no such term was to be implied into the agreement. The only restraint on the exercise of the contractual power to value the Designated Assets was a duty to act rationally, that is to say, a duty not to exercise the power arbitrarily, capriciously, or perversely.

The Defendants contend Marex did not act rationally in the manner in which they closed out the Defendants’ positions. They also contend that Marex was under a more stringent duty of having to exercise reasonable care.

The grounds relied on for the existence a duty to exercise reasonable care are: (i) Clause 18.1 of the CPA; (ii) s. 13 of the Supply of Goods and Services Act 1982 (“the 1982 Act”); (iii) an implied term; and (iv) a duty of care in tort.

Clause 18 .1 provides:

Mr Cox submitted that the effect of this clause was that Marex was accepting liability for loss that was reasonably foreseeable and arose from Marex’s “gross negligence”, wilful default or fraud. He referred to Sucden Financial Ltd v Fluxo-Cane Overseas Ltd [2010] EWHC 2133 (Comm) and Marex Financial Ltd v Fluxo-Cane Overseas Ltd [2010] EWHC 2690 (Comm). In both of these cases a broker’s customer alleged that a close-out had been negligently conducted and reliance was based on a provision similar to Clause 18.1 on the assumption that the provision predicated a duty not to cause loss by reason of gross negligence, wilful default or fraud. In Sucden, the broker argued that pursuant to the relevant clause it was not liable for the losses suffered because the customer could not prove gross negligence. Blair J held that: (i) “gross negligence” is not a term of art at common law and such a standard may amount to little more in practice than simple negligence; (ii) the concept of negligence was quite flexible enough to adapt to the circumstances in point; and (iii) in any event, the close out had not been negligently conducted. (Para 54)

In Fluxo-Cane, the clause in question provided:

Clause 15.1:General Exclusion: Neither we nor our directors, officers, employees or agents shall be liable for any direct or indirect losses, damages, costs or expenses incurred or suffered by you under this Agreement (including any Transaction or where we have declined to enter into a proposed Transaction) unless arising directly from our or their respective gross negligence, wilful default or fraud. In no circumstances shall we have any liability for consequential or special damage …

David Steel J concluded that: (i) the only relevant standard applicable to the close-out was that resulting from Clause 15.1, namely that Marex would not be liable to the customer save in respect of losses arising directly from Marex’s gross negligence, wilful default or fraud; (ii) what the epithet “gross” adds is not clear; (iii) Marex had made out its case that it had conducted the close-out in a professional and competent manner. (Paras 92 and 106).

In my judgement, construing Clause 18.1 in the context of the CPA and the FEOMA as a whole, including in particular the provisions conferring the right to close out, Clause 18.1 does not constitute an agreement by Marex that it owes a duty to take reasonable care or to avoid acting in a grossly negligent manner when closing out the Defendants’ positions. Nor does Clause 18.1 provide the foundation for the implication of such a term. I accept Mr Choo-Choy’s submission that the expression “whether arising out of negligence, breach of contract, misrepresentation or otherwise …” was used to refer to the various ways in which liability might theoretically arise; the clause says nothing as to whether a duty of care was positively intended to be owed by Marex; still less does it indicate in what particular respect or aspect of the relationship between Marex and the Defendants, or in relation to what particular activity, a duty of care might have been intended to be owed by Marex. All that the clause is providing is that if Marex is found to be in breach of a duty to take reasonable care, then the consequences set out in the clause will apply.

I also accept Mr Choo-Choy’s further submission that even if it were justifiable to infer from Clause 18.1 that some form of duty of care was intended to be owed by Marex, there is nothing in the language of Clause 18.1 that compels the conclusion that the term “negligence” was directed to or intended to cover the standard of Marex’s conduct in closing out the Defendants’ positions following the occurrence of an Event of Default, rather than in connection with the execution of any orders placed by the Defendants for best execution by Marex.

I am further of the view that in Clause 18.1, “gross negligence” means something different than “negligence”. It connotes in my opinion a want of care that is more fundamental than a failure to exercise reasonable care. The difference between the two concepts is one of degree. In reaching this conclusion I adopt the persuasive reasoning of Mance J in Red Sea Tankers Ltd & Ors v Papachristidis & Ors [1997] 2 Lloyd’s Rep 547 and Andrew Smith J in Camerata Property Inc v Credit Suisse Securities (Europe) Ltd [2011] EWHC 479 (Comm).

Section 13 of the Supply of Goods and Services Act 1982 (“the 1982 Act”) provides:

The definition of a contract for the supply of a service in Section 12(1) is “a contract under which a person (“the supplier”) agrees to carry out a service”.

Mr Cox argued that s. 13 is not confined to the performance of “primary” services to be supplied under the contract but applies also to services that are ancillary to the primary subject matter of the contract, whether that subject matter concerns the sale of goods or the provision of a service. In his submission, the closing out of positions forms part of the wider process by which Marex provided its clients with access to the FX markets. Positions that were built up also need to be closed out, whether forcibly or otherwise. It is artificial to sever the close-out from the parties’ wider relationship.

I decline to accept Mr Cox’s submissions. In my judgement, the implication under Section 13 is in respect of the particular service which the supplier has agreed to carry out pursuant to the contract, and the exercise by Marex of its right to close out the Defendants’ positions was not a “service” that Marex had agreed to carry out for the Defendants. Rather, it was a right that Marex elected to exercise in its own interests and for its own protection. This was the view of Gloster J in Euroption (paras 111, 113 and 114) and I agree with her analysis.

It was held by the Privy Council in Attorney General of Belize v Belize Telecom Ltd [2009] UKPC 10; [2009] 1 WLR 1988 that the implication of a term into a contract or other written instrument is an exercise in the construction of the instrument as a whole. The court must determine what the instrument, read as a whole against the relevant background, would reasonably be understood to mean by the reasonable addressee. That meaning is not necessarily or always what the authors or parties to the document would have intended. Tests such as the “necessary for business efficacy test” and “the officious bystander test” may be helpful in establishing what the instrument would be reasonably understood to mean, but they are not to be treated as tests different or additional to the Court’s primary inquiry. It is also not necessary that the need for the implied term should be obvious in the sense of being immediately apparent, even upon a superficial consideration of the terms of the contract and the relevant background.

In arguing for the postulated implied term, Mr Cox relied on the following features that he said arose out of the contract between Marex and the Defendants: (i) the contract vested Marex with a very wide discretion as to whether or not to close out; (ii) Marex controlled not only the decision to close out, but also the manner in which any close out was to be executed; (iii) when closing out Marex was performing a task which would directly affect the Defendants, and only indirectly Marex if the Defendants were unable to pay; (iv) it was not only foreseeable, but practically inevitable that if Marex did not exercise proper skill and care on the close out, this would cause the Defendants loss, including potentially catastrophic loss; and (v) on a close-out, there was no division of interests between Marex and the Defendants: they shared a common interest in Marex achieving a prompt and efficient close-out with minimum slippage.

Mr Cox submitted that, given these features, on a proper construction of the parties’ obligations it must have been intended that Marex would be required to exercise due skill and care when closing out the positions. Such a duty was required in order to give business efficacy to the parties’ relationship – without it the Defendants must be understood to have accepted that Marex could, through its negligence, not only “burn up” their margin but fix them with a catastrophic liability, all without the Defendants having the slightest redress. Moreover, such an implied term would be both clear and certain: an implied duty to exercise reasonable care and skill is very familiar in the law of contract; and the term would not be inconsistent with the other terms of the parties’ contract.

Mr Cox contended that Socimer should be distinguished for the following reasons: (i) in Socimer the activity to which the implied term was alleged to apply was the valuation of assets by “A” for the purposes of satisfying “B’s” existing liability which is a situation which gives rise to different considerations to where (as here) the relevant conduct does not simply concern getting in or valuing assets but a course of conduct by “A” (the close-out) which involves determining the scope of “B’s” liabilities and which, if negligently performed, may increase those liabilities by an indeterminate amount; and (ii) in the instant case, the discretionary language in Clause 16.1(c) of the PCA falls to be interpreted against a different contractual context than in Socimer because the wording of Clause 16.1(c) makes clear that Marex does not need to look to the Defendants for instructions in order to deal in the positions.

In my opinion, the agreement between Marex and the Defendants when read as a whole against the relevant background would not reasonably be understood to mean that Marex was to be subject to a duty to exercise reasonable care when exercising its right to close out. In my judgement, the compelling reasoning of the Court of Appeal in Socimer for the conclusion that there was no implied term that a valuation undertaken by the seller was subject to a duty to exercise reasonable care stands to be applied in the instant case. The contractual provisions in the two cases are of course different, but at the centre of each is a power vested in one party alone to take a course of action which may have an effect on both parties and I can see no good reason for distinguishing Socimer on the facts. Indeed, the instant case is a fortiori, since the right to close out is conferred to protect the exclusive interests of Marex and involves a very wide discretion.

In Euroption, Gloster J also rejected the contention that the power to close out was subject to a duty to exercise reasonable care. She found that the proposed term: (i) was not necessary to give business efficacy to the contract; (ii) lacked the necessary certainty, given that, when exercising the close-out power, Marex would be acting in its own interest and probably urgently, as happened in the case before her, to protect its own position; and (iii) would not have been obviously agreed to by Marex since the close-out power was manifestly intended to confer a wide discretion so as to facilitate protection of its interest and the assumption of the postulated duty would retrospectively subject its close-out conduct to a close analysis measured against every available alternative. This reasoning is directly applicable to the instant case. I agree with it and respectfully adopt it.

Mr Cox argued that whether the court approached the question in terms of assumption of liability (see eg Hedley Byrne & Co Ltd v Heller & Partners Ltd [1964] AC 465; Henderson v Merett Syndicates Ltd [1995] 2 AC 145), or the three-fold test or incremental test propounded in Caparo Industries plc v Dickman [1990] 2 AC 605, 618, the court should find that Marex owed a duty to exercise reasonable care to the Defendants in tort when exercising the power to close out.

Mr Cox submitted that if Marex chose to close out the Defendants’ positions it assumed sole control and responsibility for accomplishing that task which, if carried out carelessly, could (and in all probability would) cause the Defendants substantial (potentially catastrophic) loss. By assuming control and responsibility for the termination and reversal of the positions, Marex assumed a duty that it would exercise reasonable care.

As for the Caparo three-fold test: (1) it was abundantly foreseeable that the Defendants would suffer loss (including potentially catastrophic loss) if Marex failed to take appropriate measures to ensure that the Defendants’ positions could be closed out in an effective manner;(2)there was a proximate relationship between Marex and the Defendants: although the parties were contracting principal to principal, the Defendants were “clients” for whom Marex was providing professional “services” and the Defendants relied (and must have been expected to rely) on Marex to enter into and reverse transactions in a competent manner; further, Marex’s interests were aligned with those of its client in seeking to achieve a favourable close-out; (3) it is fair just and reasonable to impose a duty of care upon Marex and there is no reason of public policy why a duty of care should not be imposed.

As for the incremental test, the imposition of a duty of care in the present case would involve a mere amplification of the general principle that a person who assumes responsibility over the property or affairs of another will owe that person a duty of care, as is the case in respect of: (i) mortgages (a mortgagee owes a duty to exercise the power of sale with reasonable care); (ii) pledges (a pledgee owes a duty to take reasonable care to safeguard the pledgor’s propertywhile it is in his possession; and (iii) liens (a lienor must exercise reasonable care in the safekeeping of the property subject to the lien).

I do not accept Mr Cox’s submissions. In my judgement, the tortious duty of care he postulated is fundamentally at odds with Marex’s wide-ranging contractual right to protect its interests by closing out the Defendants’ positions, a right which it was entirely foreseeable might have to be exercised extremely urgently because of the state of the market. As Mr Choo-Choy submitted, there was no assumption of responsibility: Marex did not have to close out, but chose to do so for its own protection and benefit, not out of any responsibility owed to the Defendants. For this reason too, when it came to closing out: (i) the parties were not in a proximate relationship; (ii) it would not be fair and just to impose the postulated duty; and (iii) there are no good grounds for imposing the duty on an incremental basis.

Further, I agree with the view expressed by Gloster J in Euroption (Footnote: ³) that once the case for an implied statutory or contractual term fails, there is no room for the imposition of a tortious duty of care; see Tai Hing Cotton Mill Ltd v Liu Chong Hing Bank Ltd [1986] 80 at 107; Henderson v Merrett Syndicates Ltd [1995] 2 AC 295 at 316.

In addition, I accept Mr Choo-Choy’s argument rebutting Mr Cox’s incremental contention. Mortgages, pledges and liens are security interests and for that reason fall into a special category as to which the law treats the taking of possession by the security holder as a form of control or management of the owner’s property.

It is also the case that in Socimer, the Court of Appeal rejected the contention that the position of a mortgagee in equity supported the argument for an implied term imposing a duty of care.

Rix LJ said:

In my judgment, the analogy breaks down. Standard's position is governed by its commercial contract, not by the law of equity. This is the world of sophisticated investors, not that of consumer protection. These merchants in the securities of emerging markets have made an agreement which speaks of the need for a spot valuation, not of the more leisurely process of taking reasonable precautions, such as properly exposing the mortgaged property for sale, designed to get the true market price by correct process. Meanwhile, the assets involved are those of Standard, not of Socimer: and the underlying background is that where the buyer defaults, he loses both the right to complete his purchase and his downpayment. (para 122)

Lloyd LJ said:

It seems to me, with respect to her, that [the judge below] was led by that [economic] similarity into drawing, and applying, an analogy with mortgage law, while overlooking, on the one hand, the need to justify the implication on the basis of conventional contract law and, on the other hand, the fact that, in relation to a mortgage, the duties by reference to which she drew the analogy do not derive, and cannot be derived, from such a process of implication, but are imposed as a matter of general law, which does not apply in the present case because the transaction is not a mortgage. (para 154)

I do not accept Mr Cox’s argument that these observations in Socimer had no application because: (i) there the issue was whether a contractual term should be implied; and (ii) the assets to be valued belonged to the party required to undertake the valuation in question. Point (ii) is factually well made but it was by no means a determinative factor in the Court of Appeal’s reasoning. As to (i), where there is a contract between the parties, whether there is an implied term of reasonable care or a matching tortious duty involves in my judgement closely related considerations.

Accordingly, I conclude that the only duty owed to the Defendants by Marex when it closed out the Defendants’ positions was the duty to act rationally.

By way of postscript, I must deal with the Defendants’ somewhat half-hearted submission that it is not open to Marex to argue that the effect of Clause 18.1 of the CPA and Clause 16.1 (c) of the FEOMA is to exclude a duty to exercise reasonable care on a close-out because the exclusion of such a duty of care would be unreasonable under the Unfair Contract Terms Act 1977.

The requirement of reasonableness for the purposes of Clauses 16.1 (c) and 18.1 is that those terms should have been fair and reasonable to be included having regard to the circumstances which were, or ought reasonably to have been, known to or in the contemplation of the parties when the contract was made, see s. 11(1).The guidelines (Footnote: ⁴) laid down in s. 11 (2) are treated as having general application and are not exhaustive. Additional relevant factors include: (i) the way in which the relevant conditions came into being and are used generally; (ii) how far would it have been reasonable for the customer to go elsewhere to use the services of the supplier; (iii) the clause must be viewed as a whole; (iv) the reality of the consent of the customer to the supplier’s clause; see Overseas Medical Supplies Ltd v Orient Transport Services Ltd [1992] 2 Lloyd’s Rep 273 at 276-277.

In my judgement, to the extent that Clauses 16.1 (c) and 18.1 exclude a duty to take reasonable care on closing out a client’s positions, they are reasonable provisions. Mr Sevilleja did not have to use Marex as his FX broker. There were are number of other such brokers carrying on business in London and Mr Sevilleja had a relatively strong bargaining position, as evidenced by the relatively low margin rate Marex agreed to for the Defendants. Further, Mr Sevilleja was given fair notice of the terms of the FEOMA and the PCA and had an ample opportunity to consider those terms with the assistance of legal advice if desired. Further, given that it was Mr Sevilleja alone who determined the positions his companies took, it was reasonable for Marex to seek to limit its liability within the terms of Clause 18.1 and to have a wide ranging right to close-out to protect its position. Finally, substantially similar terms have been common in the FX industry for a number of years.

The earthquake and tsunami hit Japan on Friday 11 March 2011 and caused widespread substantial damage and loss of life. On Monday 14 March 2011, following the weekend close, the equity markets reacted strongly, with the Nikkei index falling 6% on the day. By the end of the following day it had lost nearly 18% of its value since the earthquake had struck. The currency markets reacted differently. Initially, the Japanese Yen weakened, which was good news for the Defendants. Later, however, the Yen appreciated, reflecting the fact that huge amounts of money were being repatriated to Japan from abroad to pay for the damage. The Yen rose five percent against the US Dollar from 10 to 17 March 2011. The Bank of Japan has a history of intervening in the currency markets to stabilise prices during times of macroeconomic shock, which meant that while the Yen was appreciating the spectre of Japanese Central Bank intervention loomed large. For this reason traders did not wish to liquidate their positions if it could be avoided, out of fear of selling the USDJPY too low.

USDJPY volatility also rose. On 11 March 2011, directly after the Japan earthquake and tsunami, USDJPY price volatility rose sharply to 14% and remained at heightened levels around a median rate of 10% until USDJPY broke its historical low on 16 March 2011 when volatility spiked again to nearly 25%. In the 10 trading days after 11 March 2011, the five day price volatility for USDJPY averaged approximately 12.5%, whereas in the 10 trading days leading up to the earthquake and tsunami, volatility had averaged closer to 5%.

At all material times, Marex’s Head of FX trading was Mr Joe Marston who had been a FX trader for thirty years. Mr Marston was vitally involved in monitoring the Defendants’ positions after 11 March 2011 and in the close-out that occurred on 16 March 2011. He left Marex’s employment on 22 February 2012 and is now employed as a Senior Trader with EDF.

Mr Marston gave evidence in the form of two witness statements on which he was extensively cross-examined. In his first witness statement he disclosed that he had entered into an agreement with Marex in lieu of his discretionary 2010/2011 bonus under which in consideration of his providing all reasonable assistance in relation to Marex’s claim against the Defendants, Marex agreed to pay him a proportion of the net recovered funds to be determined in Marex’s absolute discretion. It was submitted on behalf of the Defendants that Mr Marston did not volunteer his evidence in a free and open manner and when pressed in cross-examination he took refuge in failings of memory. I reject these criticisms. In my view, Mr Marston was a good witness who did his best to provide an accurate account of how the close-out was conducted. There were indeed a few occasions when he could not recall certain matters but I am quite satisfied that he was not disingenuously hiding behind a lapse of memory.

On Sunday evening 13 March 2011 Mr Marston discussed the Defendants’ accounts with Mr Bear, a colleague on the Marex FX trading desk. They concluded that there was little risk at this stage that the accounts would have to be closed out as the combined equity then stood at US$39 million. Later that evening, the equity value dropped to US$ 22 million and the two dealers discussed whether, if there were a close-out, the positions should be closed out via the “legs” in USD, rather than as cross currency pairs and they concluded that it would be better to close out via the cross currency pairs during overnight trading.

There was not much activity on Monday 14 March 2011 but Mr Sevilleja spoke to Mr Bob Race, another Marex FX trader, early the following day to ask for current net equity figures and exchange rates. The equity on the Defendants’ positions then stood at between US$25 and 30 million. Later that morning, Mr Muzammal told Mr Sevilleja that the equity stop should be raised to US$10 million outside of London market hours for the next week or so.

On 15 March 2011, Mr Sevilleja called Mr Muzammal six times between 12:27pm and 1:06pm to discuss the balance on the Defendants’ accounts and at 2:49 pm US$5 million was received into those accounts from Mr Sevilleja. In the evening, Mr Marston and Mr Race discussed the net equity which was at US$25 million and chatted about how a potential close-out might be executed. Mr Marston’s first thought was that two people on the desk could sell 300 million each of NZDJPY and EURJPY and then the balance of 100 million could be disposed of later. Mr Race thought that he would probably give the NZDJPY to a market maker to do 100 at a clip. He did not want to give someone the whole 450 NZDJPY and then see the price go down 50 points only for it to rally 200 points upon intervention by the Bank of Japan. Mr Marston observed: “It could be disastrous I mean you know do the best I can if the situation arises that is all.”

Between 10:00 am and 5:30 pm on 16 May 2011 the Defendants’ equity fell from US$ 28 million to US$ 18 million. Mr Marston arrived at the Marex desk between 1:00 and 2:00 pm to cover the evening shift. He joined Mr Ian Howard and Mr Lloyd Porter who were due to finish their shifts at 5:00 pm. In light of the poor liquidity in JPY pairs, Mr Marston did not want to be left alone at the trading desk at 5:00 pm waiting for Mr Race to come in for a shift starting at 10:00 pm. Mr Marston accordingly arranged for Mr Porter to stay on after 5:00 pm and asked Mr Race to arrive at 9:00 pm, rather than 10:00 pm.

The desk (principally Mr Muzammal) was in constant contact with Mr Sevilleja by telephone from 5:30 pm to 6:20 pm. At 6:10 pm the Defendants’ equity fell below US$ 10 million but before the Desk could start closing out it came back above the US$10 million level and at about 6:20 pm Marex received another US$ 3.5 million from Mr Sevilleja. By around 7:00 pm the USDJPY rate had moved slightly higher and Mr Porter was allowed to go home. Mr Marston spoke to various market makers ensuring that he had the correct chat codes should the Defendants’ positions have to be closed out and he needed to deal. At 9.00 pm New York closed and Asia opened. There were now fewer market makers available and thus less liquidity. Mr Race arrived at about 8:45 pm. Just before New York closed, the USDJPY fell through its post World War II low of 79.75. This coincided with a rise in USDJPY price volatility to 25%. After 9:00 pm (at which point the Defendants’ equity was US$14.4 million) there began a steepish drop in the USDJPY price and Mr Marston opened up a Reuters chat with Citibank (Sydney) and Mr Race did the same with ANZ (Melbourne) to check that they were ready to act if necessary. Mr Marston and Mr Race also tried to speak to other market makers over a chat system or by telephone, but none responded. In particular, Mr Marston had expected to be able to speak to Deutsche Bank having spoken to someone earlier in the evening, but no-one there picked up Mr Marston’s 9:11 pm message via Reuters headed “HiHi”.

By 9:11 pm the Defendants’ equity had fallen to US$ 9.8 million and the close-out process began. In light of exceptionally poor liquidity, Mr Marston concluded that it would take too long to close out separately the constituent legs of the positions: the rates for NZDUSD and EURUSD could move against the Defendants very quickly. The strategy was therefore to sell the EURJPY and NZDJPY as cross currency pairs.

The NZDJPY price moved between a high and a low of 57.60 and 57.10 between 9:11 and 9:12 and it moved between a high and a low of 57.46 and 56.48 between 9:12 and 9:13 pm. Likewise, the EURJPY price moved between a high and a low of 110.23 and 109.94 between 9:11 and 9:12 pm and a high and a low of 110.15 and 108.84 between 9:12 and 9:13, pm.

At 9:12 pm Mr Race started selling EURJPY through the Currenex platform and got away 30 million EUR/YEN in 19 smaller clips at an average price of 110.04. It was thought to be better to sell in small clips because liquidity was poor and spreads wide. Both Morgan Stanley and Barclays pulled their prices from Currenex in JPY and related crosses for over an hour.

Also at 9:12 pm, Mr Marston instructed ANZ to sell 100 million EURJPY “at best”, which meant ANZ were mandated to sell the entire amount of 100 million EURJPY at the best price they could get in the market. At 21:19 pm ANZ confirmed that EURJPY 100 million had been sold at an average price of 109.10.

At the same time Mr Race instructed Citibank to sell 225 million NZDJPY “at best” and this order was reported to have been “filled” at an average price of 56.59. It is not known when precisely this order was implemented although it was certainly completed before 9:43 pm.

Between 9:12 pm and 9:25 pm the Currenex and other trading platforms could not be used for trading because they were rebooting following the New York close and so Mr Marston and Mr Race continued to try to place more “at best” orders with banks including Deutsche Bank and HSBC, but without success. Then at 9:20 pm Mr Marston placed another order with ANZ, this time to sell the balance of the NZDJPY position (225 million). He chose to switch the currency pairs between Citibank and ANZ to avoid disclosing the true extent of the Defendants’ positions to any one single market maker.

Mr Race also placed an order with Citibank to sell a further EURJPY 100 million. Mr Marston decided to hold back selling the remaining 130 million EURJPY as he believed better prices would shortly be achievable on Currenex, which proved to be the case.

At 9:40 pm, ANZ reported that they had filled the NZDJPY order at an average price of 55.07and at 9:43 pm Citibank reported that they had filled the EURJPY order at 107.20.

Between 9:25 pm and 9:55 pm Mr Marston and Mr Race sold the remaining EURJPY 130 million over Currenex. The market was still illiquid but rising. Whereas Mr Race had initially sold 30 million EURJPY on Currenex by simply accepting the bid price displayed, for this set of trades Mr Marston and Mr Race made offers on the platform in clips of 10 million increasing the offer price for each subsequent clip as the market seemed to be getting stronger. Against these offers, Mr Marston got smaller “fills” in the market.

In deciding to close out using Currenex and the placing of “at best orders” with market makers, Mr Marston did not consider the possibility of seeking to sell the totality of the Defendants’ EURJPY and NZDJPY positions to a market maker (“a full amount order”). His experience told him to proceed in the manner he adopted and he was satisfied that this was the most efficient way to close out the Defendants’ positions.

Each side relied on expert evidence from experienced currency traders which addressed the question whether the close-out had been negligently conducted, not whether Marex had exercised the power to close out rationally, i.e. in a manner that was not arbitrary, capricious, or perverse.

The expert called by the Defendants was Mr Stuart Murdoch. Between 1994 and 2007, Mr Murdoch worked as a Forex and short term interest rate trader for Barclays, Goldman Sachs, Bank of America and ABN Amro. He was based in London but spent periods working in New York and Singapore. He traded a variety of currencies but specialised in “local markets”, in particular the currencies of Central and Eastern Europe, the Middle East and Africa (“CEEMEA”) and South East Asian currencies. As a trader for the banks he performed the role of market maker and proprietary trader. From April 2007 to December 2008, Mr Murdoch was a partner in a firm that advised on the raising of capital, including long dated hedging and had been involved in the establishment of a hedge fund and a bank in Georgia. Since April 2012, Mr Murdoch has been a partner in a firm that provides consulting expertise, including expert evidence, on financial disputes.

It was Mr Murdoch’s opinion that any reasonably competent professional Forex trader would have recognised the need to plan carefully for the possible close out of the Defendants’ positions and should have made pre-arrangements with a selected market maker in the form of a “spread matrix” issued by the market maker. A “spread matrix” is a document in which a market maker sets out indicative spreads for different quantities of specified currencies, including cross currency pairs. The request for a spread matrix would not disclose the size of the positions the broker is dealing with but would ask for different spreads for a range of different amounts. The arrangement should have been made by 7 September 2010 by which point the combined positions had reached a notional value of 500 million USD or at the latest in September 2010 once the equity stop was in place (“Scenario A”). The spread matrix would not have been a binding agreement but would have been a useful aide memoire which should have been checked at regular monthly intervals to insure the validity of the quotes. When it came to the close-out, it should have been used to assist in achieving a “full amount” order at a “risk transfer price” from the market maker for the constituent “legs” of the EURJPY and NZD/JPY positions.

In Mr Murdoch’s opinion, a top-level market maker who had provided a spread matrix would have given a full amount risk transfer price albeit one that reflected the risk of the price falling before it recovered. The quoted price would also reflect the advantage to the market maker of acquiring the whole position to be closed-out without the rest of the market being aware of the transaction. Such a market maker would not have refused to give a risk transfer price at 9:11 pm on 16 March 2011 and then taken advantage of the information provided by Marex for a short term gain. Instead, the market maker would have protected its franchise and sought to develop its business. The advantage to Marex would have been the speedy disposal of the entire positions to be closed, free of the risk of a subsequent market fall that could be triggered if the market got wind of the fact that Marex was having to close out very large positions, as happened on 16 March 2011 in Mr Murdoch’s opinion.

Even if the reasonably competent broker would not have obtained a spread matrix before 16 March 2011, he would not, in Mr Murdoch’s view, have closed out in the manner adopted by Marex, but instead would have requested a risk transfer price for the constituent legs of the EURJPY and NZDJPY positions from a top market maker and disposed of the entirety of the positions at the offered price, thereby avoiding the positions being sold in clips in a market that was falling “calamitously” (“Scenario B”).

Whilst selling “at best” had the advantage of virtually guaranteeing a fill, it suffered from the serious disadvantage of often achieving prices well below the investor’s expectations. What commonly happened was that a market maker with a large “at best” order would hit every bid he could see and would then use the order for himself, essentially by selling short in the market and covering those positions by buying when the price had fallen, as it inevitably would as the broker disposed of small amounts in the market. Such conduct by market makers given “at best” orders was accepted practice and was not regarded as being unscrupulous.

The turbulent market conditions after 9.11 pm on 16 March 2011 were precipitated to a significant degree by Marex’s own wrecking of the market by splitting large partial amount “at best” orders between two market makers one of whom (Citibank) was fully aware of the extent of the positions that Marex needed to liquidate because it had participated on 90% of the build up of the NZDJPY position. At the same time Marex sold 160 million EURJPY on Currenex. Those sales would have been noted by the market which would have adjusted bids progressively lower. Hence Marex, by naively splitting the orders and using “at best” orders itself chased the bids lower and significantly contributed to the ruining of its own market.

There was a particular need in this case for pre-planning because: (i) the positions were comprised of relatively volatile cross-currency pairs and were of a size and degree of leverage that meant that relatively small movements would be amplified and could rapidly erode the Defendants’ equity; (ii) the cross-currency pairs were relatively illiquid and vulnerable to the effects of “out of hours” illiquidity; (iii) the positions were large relative to their relevant markets and to Marex’s usual trading operations, so thata small negative price movement could lead rapidly to the equity margin being eroded, followed by rapid slippage; and sudden liquidation was capable of wrecking the market into which Marex would be trying to sell; (iv) the positions formed part of a wider carry trade which was susceptible to mass unwinds in the wake of macroeconomic shocks; and (v) there were obvious and serious market risks flowing out of the earthquake on 11 March 2011.

Mr Murdoch also had other criticisms of the way in which Marex conducted the close out. In his view: (i) Marex had insufficient staff handling the close-out and those trading staff it had lacked sufficient experience to execute a close-out of the sort in question; (ii) Marex’s staff demonstrated an unprofessional attitude towards their responsibilities; (iii) Marex failed to monitor carefully the “at best” orders they gave: such orders must be monitored because in effect they give to the market maker carte blanche; Marex should have asked the instructed market makers for the volumes per name they were seeing, the size of the spreads and whether they were seeing two way prices; (iv) Marex should not have placed “at best” orders with Citibank which had participated substantially in the build up of the NZDJPY positions because Citibank were thereby in a position to sell short as a proprietary trade since they would know that Marex would be disposing of NZDJPY 200 million and this would cause the price to go into freefall.

In Murdoch’s opinion, if Marex had followed Scenario A, the result would have been a positive balance on the Defendants’ accounts of between US$1,947,757.86 and US$5,581,765.58. If Scenario B had been followed there would have been a negative balance of between US$1,477,681.60 and US$495,016.01.

In coming up with these numbers for Scenario A, Mr Murdoch accepted that the Defendants’ pleaded bid adjustments to the market price at 9:11 pm on 16 March 2011, namely those adjustments that a market maker would have made in giving a risk transfer price at this time – 40 pips lower on the bid side for USDJPY; 15 pips lower on the bid side for EURUSD; and 60 pips lower on the bid side for NZDUSD -- fell within the range of adjustments he would have expected a market maker to have made based on his experience of quoting prices and spreads during similar periods of dislocation in currency markets. Mr Murdoch said that his figures were reinforced by an unidentified Chief Dealer at a global bank who was actively market making on 16 March 2011 who told Mr Murdoch that his (Mr Murdoch’s) bid side adjustments were in line with those he would have applied to amounts of similar size to the Defendant’s positions.

Mr Murdoch also stated that his figures were supported by data used in an article “The crisis in the Foreign Exchange Market” published in March 2009 by Mark Taylor, Dean of Warwick Business School and Michael Melvin of Barclays Global Investors. This data was used in the article to show how spreads changed before and after the Lehman Brothers bankruptcy (and associated Forex market crisis) announced on 15 September 2008.

Mr Murdoch’s bid adjustments used in calculating what the position would have been under Scenario B were wider than under Scenario A to reflect the fact that the order would have been ad hoc and there was no relationship between broker and market maker and hence no associated expectation of repeat business.

The expert called by Marex was Mr Steven Weller. Mr Weller has been a Forex trader since 1984. Between 1987 and 1997 he worked for Citibank in Frankfurt on secondment and later in London. Between 1997 and 2004 he was Executive Director, Global Head of Foreign Exchange spot and forward trading for Lehman Brothers London. From March 2004 to April 2009 he worked for Barclays Capital London, during which time for four years he was Managing Director, Head of Foreign Exchange – Asia Pacific, working on secondment in Singapore. He has extensive experience of all aspects of the FX market for JPY, EUR and NZD, both during and outside of London and New York trading hours.

Mr Weller agreed with Mr Murdoch that generally the overall objective of a broker on a forced close-out in normal market conditions is to liquidate the client’s entire position efficiently and as quickly as possible at the best price possible.

Mr Weller testified that it is a matter of market practice in an involuntary close-out that it is a broker’s primary, if not sole, concern to take into account the interest of his own firm and the efficient management of the firm’s risk exposure. That said, a broker has an interest to obtain as advantageous a close-out price as possible because how well he does will minimise any resulting exposure to the client and will have an impact on the size of the profit or fee he will be able to take from the client

In Mr Weller’s view, the reasonably competent broker would only be concerned about close-out pre-planning at the point when there was a real and fairly imminent risk of a forced close-out. Marex had sufficient lines in place to close out the Defendants positions in full either through market makers or through prime brokerage agreements using trading platforms like Currenex. In particular, Marex had established dealing relationships with Citibank and ANZ who could be relied on to supply large amounts of liquidity. Marex had also made arrangements to increase staffing of the desk at times of illiquidity. Further, the Marex FX team had sufficient experience to conduct the close-out. Joe Marston had worked in FX markets for over 30 years and his team of Daryl Bear, Lloyd Porter and Bob Race were experienced traders, each with substantial experience in executing FX business for clients.

Mr Weller agreed that it would be prudent for a closing out broker to consider all possibilities, including the risk transfer route, but in his opinion, in acting as it did, the Marex desk took all normal market steps to liquidate the EURJPY and NZDJPY positions and to minimise any loss on this exposure in their and ultimately their clients’ best interest. The Marex FX desk sensibly took advantage of whatever liquidity was available via the Currenex electronic portal. They also utilised the established relationships with marker makers at Citibank and ANZ. In the extreme market conditions prevailing when the right to close out occurred, the decisions made by the Marex FX team, and the measures taken to liquidate, were consistent with normal market practice. Their liquidation strategy was adapted to the prevailing market conditions and was the only way to ensure that risk could be quickly reduced. By not trading through USD by selling the constituent legs separately, the desk avoided an inadvertent increase of Marex’s and the clients’ risk from two currency pairs to three.

Given that at the time of the close-out USDJPY was trading into the unknown of historic lows, there was extreme volatility, and both EURJPY and NZDJPY crosses were gapping (Footnote: ⁵), there would have been absolutely no heroics from market makers in the prices they would show a broker and it is highly unlikely that any binding pre-agreed risk transfer price would be made available by market makers. There was also a real risk that a market maker asked for a risk transfer price on the night would: (i) change the price before it could be accepted as it saw the market falling; and/or (ii) offer an extremely low price; and/or (iii) take at least two to three minutes to come up with a price, during which time the chance to have achieved a better price through an “at best” order would be lost; or (iv) have declined to give a price and then gone on to short sell their own book.

By issuing orders in the size of the clips given to ANZ and Citibank, Marex was ensuring that above all the entire client position could be liquidated. Further, in the context of the overall size of the risk to be closed out and considering that the liquidity available to Marex via Currenex was being quickly exhausted, it was entirely reasonable for Marex to issue the orders they did to ANZ and Citibank for amounts of NZDJPY 225 million and EURJPY 100 million.

Mr Weller told the court that he had worked“at best” orders many times and would generally enter and exit his own market positions in this way with complete trust. He has never associated such orders with unscrupulous behaviour or as being open to any abuse because they are usually a reflection of a strong and trusted relationship. For these reasons he certainly never felt the need to carefully monitor execution as, amongst other things, this could get in the way of a quick efficient fill. It is also not the case, as suggested by Mr Murdoch, that market makers are more trustworthy when executing a full amount risk transfer than when executing “at best” orders.

Mr Weller disagreed with Mr Murdoch’s view that Marex wrecked the market by closing out in the manner it did. In Mr Weller’s opinion, the fall in USDJPY on 16 March 2011 was as extreme as many seasoned market professionals had ever experienced. The forced close-out was undertaken in a “perfect storm”: the Japan earthquake and tsunami had created an unstable market backdrop which was further exaggerated through reduced liquidity created by the lack of overlap between the New York market closing and Asia markets opening. When USDJPY broke through its historical low it triggered an avalanche of selling interest which, in the prevailing thin/illiquid market conditions, caused USDJPY to “fall off a cliff”.

Mr Weller also strongly dissented from Mr Murdoch’s use of the pleaded bid side adjustments in his Scenario A and Scenario B calculations. In Mr Weller’s opinion, these adjustments “have no foundation whatsoever.” The markets for USDJPY, EURJPY and NZDJPY were already collapsing by 9:11 pm on 16 March 2011. In these circumstances, any market maker quoting the spreads suggested by Mr Murdoch who were only able to sell at the prices at which Marex filled the Defendants’ positions would themselves have made a loss approaching US$ 10 million based on Mr Murdoch’s figures. If Mr Weller had been a market maker who was approached by Marex seeking to place a full amount order he would have attempted to obtain a “limit order” or an “at best” order. If Marex had insisted on a full amount transfer price, he would have made bid side adjustments based on the “prevailing market rate” and “prevailing market conditions” of 100 pips in USDJPY, 40 pips in EURUSD and 100 pips in NZDJPY. He might also have quoted a penal spread to avoid the trade altogether.

In Mr Weller’s opinion, the anonymous Chief Dealer’s statement relied on by Mr Murdoch, “is surprising”. And as for the Taylor/Melvin data, this does not provide a like-for-like comparison with the market conditions prevailing at 9:11 pm on 16 March 2011. In particular, the Lehman bankruptcy was not a currency specific event. Further, the Taylor/Melvin data relates to risk transfer trades in the range of US$ 50-100 million, whereas the risk transfer trade postulated by Mr Murdoch is for a total aggregate amount of USD 1.7 billion, a discrepancy that renders the Taylor/Melvin data of little practical use in the context of the instant case.

Marex’s key objective in closing out the Defendants’ positions should have been to close out the positions as quickly as possible, for the best price available and this being so, no rational and/or reasonably competent trader in Marex’s position would have sought to dispose of the positions in the manner adopted, namely, by a series of piecemeal “at best” orders and Currenex sales. This view is supported by the key considerations identified by Mr Murdoch in giving his view that pre-planning was necessary: (i) the positions were comprised of relatively volatile cross-currency pairs; (ii) the cross-currency pairs were relatively illiquid and vulnerable to the effects of “out of hours” illiquidity; (iii) the positions were large relative to their relevant markets and Marex’s usual trading operations; (iv) the positions formed part of a wider carry trade, which was susceptible to mass unwinds; and (v) there were obvious and serious market risks flowing out of the earthquake on 11 March 2011.

The steps that Marex actually took in the close-out were not rationally connected with the objective of closing out the positions as quickly as possible, for the best price available. In particular: (1) the decision to split the trades between “as many counterparties as possible (Currenex, Cititbank, ANZ)” in the hope of keeping the size of the positions from the market was illogical and was always likely to (and did) achieve the opposite effect; (2) placing orders on an “at best” basis whilst also peppering the market on Currenex, was always likely to (and did in fact) wreck the market into which the positions were being released; (3) placing “at best” orders at intervals -- the first as 21:12, the second eight minutes later at 21:20 -- was irrational and was likely to lead (and did lead) to much worse fills on the second tranche; (4) breaking the order into parts and selling it through various different channels was always likely to (and did in fact) take longer than making a single “full amount risk transfer” disposal to a single counterparty.

Mr Marston’s failure to consider the full amount sale route was irrational, grossly negligent and negligent. If he had considered this route he should have adopted it. A full amount sale would have been made if it had been sought because: (i) it would have secured for Marex a speedy divestment of the positions without scrambling for liquidity in an adverse market; and (ii) a market maker would have had an interest in obtaining the whole of the positions discreetly through an “off the market” sale, thereby putting himself in a position in which he could control the flow of the positions back into the market so as to maximise his profits.

Further, Marex’s failure to monitor and/or supervise the “at best” orders and the placing of “at best” orders with Citibank that had participated substantially in the build up of the NZDJPY positions was irrational.

The test of rationality is substantially different from the test of simple negligence. It is a distinction in kind, not degree. Irrationality in the context of this case connotes conduct on Marex’s part which is “so outrageous in its defiance of reason that it can properly be categorised as perverse” (see Socimer at [61]-[62] and Hayes (FC) v Willoughby [2013] UKSC 17 at [14] per Lord Sumption).

The PCA and FEOMA provisions whereby Marex was given the right to close out the Defendants’ positions confer a very wide discretion to close out without prior notice to the Defendants at such time or times and in such manner as, at Marex’s sole discretion it considers necessary or appropriate to cover, reduce or eliminate Marex’s loss or liability under or in respect of any of the Defendants’ positions. Further, none of the contractual documents dictates, prescribes or limits how the close out of the Defendants’ positions could or should be effected by Marex.

On the night of 16 March 2011 Marex faced tough choices in very difficult market circumstances and had to act quickly in order to protect itself once the US$ 10 million equity stop had been breached. In particular, Marex had to decide (amongst other things): (i) whether to close out through the cross-currency pairs or USD legs; (ii) whether to try to dispose of the positions in a single amount or split the full amount into smaller parcels; and (iii) whether to sell the positions through electronic trading platforms or market makers, or a combination of both. The decisions taken to close out directly through the cross currency pairs, to split the full amount into smaller parcels and to sell the positions through Currenex and “at best” through ANZ and Citibank were not only rational, but were well reasoned, sensible and competent, whatever the alternative routes that might have been taken.

It is far from clear whether the financial outturn would have been better if the positions had been sold at a risk transfer price. Mr Murdoch’s Scenario A and Scenario B calculations are deeply suspect for the following reasons: (1) As he himself admitted, he has never been a market maker in USDJPY, NZDJPY and EURJPY and has never been in the position of having to provide a price as market maker for positions as large as the Defendants’ positions in those currencies, let alone being asked to provide such a price or prices in the highly abnormal conditions prevailing on 16 March 2011. (2) The Taylor/Melvin data based on the Lehman bankruptcy aftermath does not provide sound support for Mr Murdoch’s calculations because the Lehman bankruptcy was not a currency-specific event and is based on quotes from a single bank in respect of risk transfer trades in the range of USD 50-100 million, whereas the full amount order postulated by Mr Murdoch would have been for a total aggregate position of about 1.7 billion USD. (3) The hearsay opinion of the Chief Dealer at a global bank relied on by Mr Murdoch should be given no or very little weight since this anonymous individual was not made available for cross-examination. (4) The contrary opinion of Mr Weller, who has been a market maker in G7/G10 and major currencies, including USDJPY, NZDJPY and EURJPY for many years, including during periods of macroeconomic shocks and during the Australasian time zone should be preferred to that of Mr Murdoch. (5) The determination of a risk transfer price at 9:11 or 9:12 pm on 16 March 2011 would have been a highly subjective exercise. (6) Mr Murdoch’s calculations are based on spreads at 9:11 pm but Marex would not have made contact with a market maker until 9:12 pm, at which time the price would have fallen further. (Marex’s first sales on Currenex were at 9:12:27 and 9:12:36 pm; the first contact with a market maker was at 9:12 pm).

In any event, for the reasons advanced by Mr Weller, there would have been a strong probability of a market maker not being willing to quote a price for a risk as large as USD 1.7 billion in aggregate. Given the volatility of prices at around 9:11 pm and 9:12 pm there was also a risk that a price that was given by a market maker would be changed before Marex had had an opportunity to accept it.

There was also a risk that if Marex asked for a risk transfer price but the bank declined to give a price, that market maker would then drive down the market and short its book before Marex had had a chance to dispose of the positions.

In contrast to a full amount order, an “at best” order was well suited to meet the objective of a close out –to liquidate the positions as quickly as possible.

Whilst it is true that an “at best order” for large positions: (i) enables a market maker, once he has filled the order with available bids in the market, to take the risk on himself at a low price, (ii) could have a tendency to drive the price down as a result of the repeated hitting of all available bids in the market; and (iii) could take longer to close out than through a full amount risk transfer order, there is nothing wrong with a market maker taking the remaining risk on to his own book at a price that he is prepared to pay and, in any event, it may be more difficult to find a suitable market maker who will give an acceptable price for the full amount rather than to close out the positions in smaller parcels on an “at best” basis as quickly as possible. Further, a market maker approached for a risk transfer price for the full amount may decline to give a price and abuse the information thereby obtained or give a price but then change it at times of high price volatility before Marex can accept it.

Marex’s use of “at best” orders in the close-out chimes with the use of price stop orders within the FX industry which customarily involve the market makers concerned executing the orders on an “at best” basis. Marex placed price stop orders in respect of the Defendants’ positions between 30 March and 24 September 2010 and FIXI plc had price stop orders in place covering positions held by the Defendants for 40 million EURJPY and for 58 million, 50 million and another 50 million NZDJPY.

One of the essential differences between the duty to act rationally and the duty to take reasonable care is that when the court considers the former it is the decision of the decision maker that is focussed upon, whereas consideration of the latter involves the court making its own judgement on the basis of objective criteria (see Socimer at para 66 and Euroption Strategic Fund Ltd v Skandinaviska Enskilda Banken AB [2012] EHHC 584 (Comm) 22, para106 (Gloster J).

The decisions made by Marex when closing out the Defendants’ positions were made pursuant to its right to close out under the FEOMA and the PCA. The power conferred by this right involves a very wide discretion as to when and how the Defendants’ positions would be closed out; its purpose is to enable Marex to protect its own position quo ad the Defendants. In exercising the power, Marex might also serve the interests of the Defendants in an indirect sense, since the greater the price achieved in obtaining equal and opposite positions in the market, the less will be the Defendants’ liability to Marex under the positions being closed out. However, the essential purpose of the power is not to protect the interest of the Defendants but to protect the interest of Marex.

It follows from the foregoing that the object of a close-out is to close-out the client’s entire position as quickly as possible at the best price possible.

I reject the contention that Marex owed a distinct duty, whether as part of the overarching duty of rationality or otherwise, to plan for the possibility of closing out the Defendants’ positions before there was a real and fairly imminent risk of a forced close-out. Instead, the question is whether the close-out strategy in fact adopted was a rational strategy. What planning Marex undertook may have a bearing on the rationality of the close-out methodology it adopted, but the proof of the pudding is in the methodology actually adopted compared with the competing postulated strategy of a full amount order.

Mr Marston did not consider whether to go down the full amount route. Instead, his experience told him that the route to adopt was to sell what could be sold on the available trading platforms and to sell the balance in individual parcels (clips) “at best” through market makers. He considered whether to close out through the legs of the positions or by selling them as cross currency pairs and decided that, given the illiquidity of the market after New York had closed, the positions should be sold as cross currency pairs. By way of preparatory and precautionary steps, he made sure on 16 March 2011 that the desk would be manned by at least two traders in the evening and he spoke to various market makers at about 7.00 pm to ensure the desk had the correct chat code. At 9.00 pm he opened Reuters chat with Citibank (Sydney) and Mr Race did the same with ANZ (Melbourne). Mr Marston and Mr Race also tried to speak to other market makers, but none responded. As Mr Weller opined, it would have been prudent to consider whether to go down the full amount route. However, this failure of Mr Marston’s does not, in my opinion, ipso facto render the manner of the close-out he adopted irrational. Instead, as I have indicated, the question is whether the full amount route was so obviously significantly superior to the route adopted by Marex that it was perverse of Marex not to have adopted it. In my judgement, for the reasons I give below, the full amount route involved significant downsides as well as upsides, to the point that it cannot be said that it was capricious, arbitrary or perverse for Marex to have closed out the positions as they did, rather than on a full amount basis.

The principal potential advantage of the full amount route was that, if such an order were achieved, there and then the Defendants’ entire positions would be disposed of at the agreed price, eliminating the possibility of the market being driven down by subsequent disposals of the positions in individual clips. However, on the basis of the evidence of Mr Weller, whom I found to be an impressive witness and whose experience was more apposite than Mr Murdoch’s, I conclude that in the particular circumstances of the market at 9:11 pm on 16 March 2011 --- high volatility (Footnote: ⁶), illiquidity, a falling NZDJPY price (USDJPY had gone through its post World War II low of 79.75 shortly after 9:00 pm) and a head of steam building up for a mass carry trade unwind -- there would have been a real possibility that any market maker approached, even one who had earlier provided a spread matrix, would have declined to give a whole amount price. And if that had been the situation, Marex would have been faced with the risk that the market maker would then go on to use the information imparted in requesting a risk transfer price to short sell its own book. Further, the approach to the market maker and its response would have taken at least a minute, possibly two or three, during which time the price for NZDJPY and EURJPY was dropping rapidly. There would also in my judgement have been a real possibility of the market maker rapidly changing its price as the market fell, thereby preventing Marex from clinching a whole amount deal.

In addition, I think that, given the circumstances of the market, any risk transfer price quoted on the night could well have been very low and would very likely not have exceeded the fill prices actually achieved on behalf of Marex. Mr Murdoch thought on both Scenario A and Scenario B the risk transfer price would have substantially exceeded the fills achieved in the market. Again, I prefer the evidence of Mr Weller on this issue and I think Marex’s submissions on this part of the case are well made.

It is further to be noted that if a market maker had quoted a very low price which Marex thought it could better through “at best” orders, Marex would have been in an acute dilemma because, if it refused to accept the price, it would run the risk of the market maker spoiling the market and short selling its own book.

I agree with Marex’s submission that an “at best” order offers the greatest assurance that the relevant position will be closed in its entirety, subject to available liquidity. This no doubt is why it is common practice for price stop arrangements to be implemented by disposals “at best”. The market maker’s own book is part of the available liquidity and it is accepted practice for a market maker to take the position onto its own book at the then prevailing price, which, if there are no other sources of liquidity, will be the price it is prepared to pay.

There is a risk that a market maker which accepts an “at best” order will abuse its position by selling short and using part or parts of the position to drive the price down. However, that risk is reduced where the broker has an established relationship with the market maker, as was the case here in respect of Citibank and ANZ. Mr Weller has disposed of positions on an “at best” basis many times and he has never associated such orders with unscrupulous behaviour or as being open to any abuse because they are usually a reflection of a strong and trusted relationship.

Where the position to be closed is large, as in this case, the closing out broker has to decide whether to instruct a market maker to sell “at best” the whole position or only part thereof. The market maker has complete discretion as to the quantities he will seek to sell at any particular time. There is therefore a risk that the market will get wind of the fact that a large position is being closed out and prices will go down accordingly. On the other hand, if to avoid this risk a market maker is given only part of a position to sell “at best”, the broker may only be able to get away the balance of the position by way of subsequent “at best” or trading platform orders, which might drive the price down, as I find happened to a not insignificant extent on 16 March 2011. There were therefore downsides in taking the “at best” and trading platform route adopted by Marex, but they were not in my judgement such in and of themselves and/or when compared with the advantages and disadvantages of a full amount strategy, as to make Marex’s strategy arbitrary, capricious or perverse.

I would add that if it had been necessary to decide whether Marex were liable in negligence for the manner in which it closed out the Defendants’ positions, I would have found against the Defendants’ on this issue. As I have observed, Marex had a wide discretion in deciding how to implement the close-out as quickly as possible and for the best possible price, and it was acting under a power conferred for its protection, not that of the Defendants. With this in mind and given the downside risks involved in adopting the full amount route at 9:11 pm on 16 March 2011 identified above, I do not think it was negligent to adopt the “at best” and trading platform strategy that Mr Marston adopted. It was also not negligent to dispose of the positions by “at best” orders for individual clips rather than for the entire positions, even though it was foreseeable that to do so could cause a steep fall in the price. This is so because: (i) Marex was entitled to protect its position by liquidating the positions as quickly as possible; and (ii) “at best” orders to dispose of the entire positions also ran the risk of causing falls in the price since the market might well realise what was afoot and drive the price down. Nor do Mr Murdoch’s other criticisms of the manner in which the close-out was executed– in particular the alleged failure to monitor the “at best” orders and using Citibank Sydney which had participated in the build up of the NZDJPY position -- constitute negligence. Marex was entitled to trust ANZ and Citibank to execute the orders in a proper and regular fashion and it was Citibank London that had participated in the build up, not Citibank Sydney.