WS Tankship II BV v The Kwangju Bank Ltd & Anor

These are claims by beneficiaries under what are called refund or advance payment guarantees. The claimants are companies which form part of a Dutch shipping group, and the defendants are the Korean bank and guarantee company which issued the guarantees. The case arises out of contracts for the building of bitumen tankers which the claimants ordered from a yard in Korea (in all orders were placed for eight such vessels). During construction, the claimants paid various advance payments to the builders, and took a number of guarantees issued by the defendants payable in the event that they became entitled to a refund. The contracts were eventually terminated by the claimants, leading to claims for a refund. Payment has been made by the defendants in part only. A considerable number of defences have been advanced, including a contention that the guarantees were restricted to particular advance payments only, whereas the claimants maintain that the guarantees covered all advance payments. The nature of the guarantees is itself in dispute, it being asserted by the defendant (or at least the first defendant) that they are contracts of suretyship, whilst the claimants assert that they are instruments payable on demand. The defendants’ counterclaim for the return of the payments they have already made, on the basis (among others) that the payments were made under a mistake of fact or law.

The claimants in the three actions before the court are three “WS Tankship” companies, each of which was the buyer of a ship pursuant to a newbuilding contract with the builder. Each is a separate subsidiary of the Dutch shipping company, Vroon Group B.V. (I shall refer to them compendiously as “Vroon”). Each of these Tankship companies would, had things gone to plan, have been single purpose ship-owning companies of each of the ships purchased under the contracts.

The first defendant, The Kwangju Bank Ltd (“Kwangju Bank”), is a regional bank operating in South West Korea, and is part of a nationwide banking group, namely the Woori Bank group. Kwangju Bank is concerned only with the first set of proceedings. The second defendant, Seoul Guarantee Insurance Company (“SGIC”), is a company whose business consists (so far as relevant) in the issuance of guarantees, the evidence being that until 2004 this was limited to the domestic context. SGIC is a defendant in all three sets of proceedings.

As well as the parties, I must describe the shipbuilders, who have played a significant part in the trial of the action. At the material time, GEO Marine Engineering & Shipbuilding Co. Ltd. (“GEO”) operated a small shipyard based at Masan in South Korea. As a company, it was a newcomer to the business of shipbuilding.

There are three sets of proceedings corresponding to the three shipbuilding contracts in respect of which the disputed guarantees were issued (it has been ordered that the actions be heard together). They concern guarantees in respect of (1) the hull identified as GMS-103 against both defendants, (2) the hull identified as GMS-104 against Seoul Guarantee Insurance Company only, and (3) the hull identified as GMS-105 against Seoul Guarantee Insurance Company only.

Most of the relevant evidence came from witnesses of fact. For the claimants, oral evidence was given by Mr Herman Marks, who is the director of Vroon with overall responsibility for the projects. Vroon’s other witnesses were Mr Sturle Erichsen of Associated Shipbroking Monaco, who was Vroon’s broker for the projects, Ms Alexandra Kranenkamp, an administrator in Vroon’s Newbuildings Department, and Mr Nicolaas Spiljard, Vroon’s Newbuildings Manager. Reliance was also placed on the witness statement of Ms Lieke Pasquier-Melse, a lawyer from Vroon’s legal department, who for practical reasons was unable to give evidence in person. All these witnesses (except for Mr Erichsen) are based in the Netherlands.

Oral evidence was also called by the defendants. Both defendants relied on evidence from the shipbuilders, in the form of Mr Bong Jung Lee, who was the representative director of GEO, and effectively its senior officer, and Mr Yong Mok Ha, the assistant manager who at the relevant time was involved in handling the issue of refund guarantees. Kwangju Bank called Mr Han-Beom Ryu who was the manager of its Baekundong branch during the period at issue. SGIC called Mr Hee Moon Lee who was assistant manager of its Masan branch, Mr Pil Ku Noh who was section chief of the branch until June 2008, and Mr Myeong Jun Doh who was manager of the branch from June 2008. None of these witnesses speaks English with any degree of fluency (I mention this only because the contractual documentation and the guarantees were all in the English language, and an issue at trial has been the extent to which the documentation was fully understood by the defendants at the time).

There is a comparatively small amount of contemporaneous documentary evidence with which to compare the oral evidence. In fact, comparatively little documentary evidence has featured in the trial, though this is due in good part to the focused efforts of the parties to produce and work with a core bundle of trial documents, for which I am grateful.

Vroon’s case has been straightforward. It submits that its claim is made on first demand refund guarantees, upon which payment must be made in full according to the terms of the guarantees in question, and without regard to the underlying contractual issues between shipbuilder and buyer. Upon termination of the shipbuilding contracts, a number of instalments were paid by the defendants under the guarantees. But a number were not. Vroon claims:

As will be seen, a feature of this case is that there was a series of refund guarantees issued in respect of each of the relevant shipbuilding contracts. In this regard, in its opening submissions, Vroon has submitted (and I quote) that as “a matter of logic, C’s primary case in each action is that the full amount can be recovered pursuant to the RG which was granted first and that it is not necessary for it to rely upon subsequent RGs. In any event, it is accepted that it cannot recover more than the amount of the instalments paid (plus interest). Each guarantee provides that it will become null and void upon receipt by C of the sum guaranteed, so there is no scope for double recovery”.

The defendants have raised a substantial number of defences, and it is necessary to outline these before consideration of the facts. I have adopted the order in which they are set out in the first defendant Kwangju Bank’s submissions, and note that it takes a number of points which SGIC does not argue (or does not actively argue). In summary:

I begin by noting that the court has been provided with an agreed chronology. This has been useful, because with a number of different contracts, and events spanning some three years or more, the story is quite complicated. Much of the story is however not relevant to the issues in the case. The facts as I find them are as follows. In about 2006, GEO Marine Engineering & Shipbuilding Co. Ltd was formed from the merger of a shipyard and a design and management company. At that time, the shipbuilding market was buoyant, there being (in Mr Marks’ words) an exceptional demand for vessels. This position prevailed (according to his evidence) until the third quarter of 2007. GEO took over a small shipyard at Masan and subsequently at Mokpo in South Korea. As far as shipbuilding was concerned, it was essentially a start up operation. So far as this trial has been concerned, the most important officer of the company was Mr Bong Jung Lee, the senior officer of GEO, who is by profession a shipbuilding engineer. In a presentation prepared in February 2007 for marketing purposes, the company described itself as providing “A new paradigm in the small and medium shipbuilding industry”. At that time, the company had 38 employees.

At the early stage of their relations, a substantial amount (though by no means all) of the communication between buyers and builder came through the brokers, that is Mr Sturle Erichsen of Associated Shipbroking Monaco for Vroon (who gave evidence at trial), and GEO’s broker who is Mr Ung Seop Jeong (usually referred to in the documents as Mr US Chung). Mr US Chung is the managing director of US Maritime Korea, and he did not give evidence. From about the end of 2007 however, a considerable number of communications with Vroon were made on behalf of GEO by Ms Inn Su (Susan) Lee, who is Mr Lee’s sister, and who speaks English. Ms Lee did not give evidence at trial.

Vroon Group B.V. (which owns the three claimant companies) is the well-known Dutch shipping group based in the Netherlands. For some time, it had been interested in entering the bitumen tanker market. The introduction to GEO as potential builders came through the shipbrokers. What was envisaged (and indeed eventually ordered) were small tankers of 6,000 DWT. GEO was identified as able to build such vessels at a competitive price, and on 17 October 2006, a Letter of Intent was signed between GEO and Vroon Tankers B.V. for the building and purchase of four vessels (with an option at that stage on two more) at a price of US$15,150,000 per vessel, which (as Mr Herman Marks put it) was a good price. These would be the first ships that GEO had built. Mr Marks was the director of Vroon responsible for the GEO project, and he remained involved at a supervisory level throughout, and was the key witness called by the claimants at the trial.

In the usual way, the price was to be payable by instalments in advance, and the buyer wanted a guarantee from a suitable institution that the payments would be repaid should the contract fail prior to delivery of the vessel. The Letter of Intent stated that the builder would provide a refund guarantee issued by an insurance company in Korea acceptable to the buyer. On 12 October 2006, GEO proposed a company called Green Fire and Marine Insurance Co Ltd. Mr Bong Jung Lee said in his oral evidence that he thought Vroon had agreed this company, though this was not the case. On 31 October 2006, GEO’s broker provided a standard form of refund guarantee in the English language that I infer came from the insurance company. However Vroon’s bankers had difficulty in finding sufficient information about Green, and in discussions over the next few months Vroon was not prepared to agree to it as issuer. The evidence of Mr Bong-Jung Lee shows that despite considerable efforts, the procurement of guarantees from an alternative issuer was to prove to be a matter of difficulty for GEO, as a small company, whose business was newly established.

On 13 December 2006, four shipbuilding contracts in materially identical terms (save as to delivery and instalment dates) were signed by Mr Marks on behalf of Vroon as buyers, and Mr Bong Jung Lee on behalf of GEO as builders. The contracts were entered into by four Vroon nominee companies, namely:-

So far as material the contracts provided as follows:

It is common ground that, by these provisions, the contract contemplated a letter of guarantee (in other words a single letter of guarantee) issued by an insurance company acceptable to the buyer for the refund of all the pre-delivery instalments in respect of the contract in question in the event that the buyer was entitled to terminate prior to delivery. This has been called a “blanket” guarantee at trial. The evidence on behalf of both Vroon and GEO is that it was required and expected. (The draft refund guarantee refers to four pre-delivery instalments, though in the event five were agreed, and nothing turns on this difference.)

On the same day (that is 13 December 2006), a Memorandum of Agreement was signed giving Vroon the option to order up to four further vessels. Eventually, Vroon would exercise options in respect of all these vessels, namely GMS-106 – GMS-109. To anticipate events, the final exercise of the option by Vroon was on 21 November 2007. Neither defendant was concerned with the refund guarantees in respect of these vessels, which were issued by a different company.

Difficulties that GEO was experiencing in purchasing the main engines caused it to seek an extension of a few months in the delivery dates, which was done by way of Addendum No 1 to each of the contracts (each Addendum is dated 9 February 2007). The buyers agreed to an extension on the proviso that the 2^nd instalments for GMS 103, 104 and 105 were also postponed from 30 April 2007 to 30 June 2007.

Further, since the first instalments were payable on receipt of the required refund guarantee, the lack of accord as to the identity of the issuer meant that GEO was not getting paid the first instalments, which caused it cash flow problems. The problems are shown by an email between the brokers of 25 January 2007, to the effect that, “If they can’t [i.e. if Vroon could not accept Green] then we will have to forget the deal. We can not have both parties waiting for so long it is not good for either the shipyard or Vroon”. It is argued on behalf of the defendants that GEO was beginning to take a hard line, and that Vroon had to make concessions because it needed the project to succeed and because it wanted the vessels at the competitive price it had secured. Consequently, it is contended that the balance of bargaining power was with GEO as builder. On the evidence, I do not unreservedly accept these points. Speaking in general terms, Mr Marks said that the “negotiation power was with the shipyard, more than with the buyer”—further, GEO’s price was competitive, and it is obvious that Vroon made considerable efforts to try to get the vessels built. However, these contracts were also important to GEO as the first contracts that the new venture had entered into, and I am satisfied that both parties had a strong interest in keeping the deal going. It is clear that the parties worked on a cooperative basis through much of the relationship. It is not the case that GEO was able (or sought) to dictate terms, as is shown by the fact that the buyers would not accept its proposed guarantor. So far as it matters, I accept Mr Marks’ oral evidence to the effect that he was willing to provide assistance, rather than concessions.

To mitigate the effect of the delay in getting the refund guarantees in place, and to enable GEO to purchase supplies and to enter into contracts for the purchase of supplies necessary for construction of the ships, Vroon agreed at the end of February 2007 to make a payment of half of the first instalment for each of GMS-102 and GMS-103. This was in effect a down payment of 5% of the purchase price of each of the vessels notwithstanding that the refund guarantees were not in place. The payment was structured as two loans of US$785,500 by Ship Finance B.V. (another Vroon company) to the builder, to be repaid out of the first instalment for each project. That relating to GMS-103 was entered into on about 5 March 2007. The position is made plain in the loan agreement. It records that, “In accordance with Article X of the Contract, the Borrower has to provide a refund guarantee to the Buyer as security for the Buyer’s payment of pre-delivery instalments. The Borrower has difficulty providing such refund guarantee and in order not to delay the building schedule of the Vessel, the Borrower wants to borrow sufficient funds from the Lender in order to purchase supplies or enter into contracts for the purchase of supplies necessary for the construction of the Vessel”. The drafting reflects the fact that at this time, GEO had anticipated receiving a refund guarantee from Dongbu Insurance by 20 March 2007, and consequently receiving the first instalment under the contracts then (the loan was interest free up to this date).

This was the first of three occasions on which Vroon extended financial assistance to GEO. It gives rise to an issue between the parties, because the defendants argue that these loans were not disclosed to them by the buyers, and that this non-disclosure discharges the defendants’ liabilities under the guarantees. As a matter of fact the loans were not disclosed by the buyers, who say that they were under no obligation to do so, though they say that in terms of the defendants’ knowledge the factual position is not as clear as they suggest. I will have to come back to this question when dealing with the non-disclosure defence.

It appears that quite soon after the loan agreements were signed, GEO ran into difficulties obtaining the guarantees from Dongbu Insurance. On 19 March 2007, it sent a message via the brokers about further delays on the part of Dongbu. This is also the first reference so far as Vroon is concerned to Kwangju Bank. The email stated among other things that, “If Dongbu fails issuing R/G until end of May 2007, we have a plan to submit R/G to Vroon issued by bank(s) (Woori bank or/or Kwangju bank) for each 20% of contract price (first + second instalment) within end of May 2007”. As is pointed out by Kwangju Bank in its submissions, this suggestion as to the extent of the refund guarantee was not something which had been discussed by the parties previously. Further emails followed, which included a reference to an extension of time being required due to GEO having to “deposit 100% of RG amount into above bank for RG issuance”.

This sets the scene for an important factual dispute. It is not in dispute that at the outset GEO understood that the contracts stipulated for and Vroon required what was described at trial as a “blanket” or “full cover” guarantee covering all instalments under the relevant contract, as opposed to “discrete” guarantees covering individual instalments—Mr Bong Jung Lee accepted this in his evidence. The defendants contend however that at the end of March 2007, Vroon and GEO agreed to vary the shipbuilding contracts so that, rather than deliver a guarantee for the refund of all the instalments to be paid under the contracts, GEO would instead deliver, in the first instance, a guarantee only for the refund of the 1^st instalment (plus interest) to be followed by further separate refund guarantees for the refund of subsequent instalments. There were eventually twelve of these in respect of vessels GMS 102, 103, 104 and 105. The variation argument was pursued with particular force by SGIC, but Kwangju Bank also advanced it, and made detailed submissions in closing. Variation of the shipbuilding contracts would not in itself provide a defence to the defendants. But the thrust of their contentions has been that such variation explains the fact that a series of guarantees were issued, rather than one per contract. It is linked therefore to the defences based on what is said to be the continuing common intention of the parties, and the estoppel defences which are based on the contention that each of Vroon, Kwangju, GEO and SGIC intended and understood that the refund guarantees provided would be discrete guarantees in relation to discrete advance payments, and acted on that assumption. A further point to note is that the alleged variation arises at the beginning of the process by which the guarantees were issued. In fact, the first guarantee was in respect of GMS-102, which vessel GEO ultimately delivered, albeit not completed. But, as I shall explain, this guarantee was the template for all the further guarantees.

Vroon says that it makes no substantive difference to the liability of the defendants whether or not such a variation was agreed between Vroon and GEO. However, for the moment I am concerned with the question whether the contracts were varied or not. Vroon submits that the evidence does not begin to support the existence of any such agreement. The defendants (in the words of the closing submissions of SGIC) say that the evidence for a contractual variation is overwhelming. But Vroon also contends that there has been a lack of specificity in the defendants’ case (particularly that of SGIC) as to how the alleged variation of the shipbuilding contracts came to be agreed. Kwangju Bank’s pleaded case is that it relies on the communications, written and oral, between Mr Chung and Mr Erichsen on 28, 29 and 30 March, and on the emails sent by Mr Kim of the builder and by Mr Marks of the claimant on 28, 29 and 30 March 2007. That in substance was its case at trial. SGIC (to quote from its opening submissions) says that want of particularity in this regard “misses the point: the variation of the shipbuilding contracts was agreed by a combination of oral consent and then conduct consistent with it”. SGIC relies on Kwangju Bank’s case, though saying that even if there was no variation at that time, the ship building contracts must have been varied at or immediately prior to the provision of the SGIC guarantees on 25 May 2007. Whenever it happened, the shipbuilding contracts must, it submits, have been varied.

In resolving this dispute, it is necessary to analyse events over the last few days of March 2007. There are a number of relevant emails, the order of which is not always easy to follow (partly because of the time difference between Korea and Europe). I shall take the order from the agreed chronology. On 28 March 2007 Mr US Chung emailed Mr Erichsen as follows—“Since Dongbu can still not confirm possible date for issuing R/G, GEO will not wait any longer and by all means they managed that Kwang Joo Bank, whose main office in Kwangjoo city covering Mokpa city, will issue the R/G for 10% within this week. Please see the bank’s 2005 Annual report.” The email concludes, “Please kindly get Vroon’s acceptance on the above in order this project to be actually materialized”.

As I have said, Kwangju Bank is part of the nationwide Woori Bank group. Mr Erichsen recommended it as a guarantor in enthusiastic terms to Mr Marks by email of 28 March 2007. Mr Marks was not in the office that day, but Mr Erichsen sent an optimistic email to Mr US Chung to the effect that he was “confident that this will be in order tomorrow”. Mr Chung chased Mr Erichsen by email sent next morning saying, “GEO with Kwangju bank has a plan to have the R/G issued today 29th March. Please contact Mr. Marks by phone and confirm us of his agreement through my mobile phone bcos I will be outside the office today”.

Mr Marks responded by email to Mr Erichsen on 29 March 2007 saying, “Confirm this guarantor is acceptable to Buyers”. He says that he was not doing more than confirming the suitability of Kwangju Bank as guarantor. Kwangju Bank disputes this, contending that the buyer had been asked to accept a proposal that contained two inter-dependent components (i) Kwangju Bank as guarantor and (ii) a guarantee for 10% of the price. As a result, to accept one was to accept the other. On the other hand, as Vroon points out, an agreement to “discrete” refund guarantees would represent a significant change to the security arrangements between GEO and Vroon, which would be exposed to the risk that the builder would find itself unable to procure subsequent guarantees. The email makes no mention of new arrangements, confirming only that “guarantor is acceptable”. On balance I accept what Mr Marks says in this regard.

That however is only part of the picture. As a matter of background, GEO was a new customer to Kwangju Bank, and Mr Han-Beom Ryu (who was the manager of the Baekundong branch and the officer called by Kwangju Bank to give evidence at trial) was only introduced to the company in January 2007. An account was opened at the branch at the end of February 2007, which received the sum of US$1,571,000 a few days later. It seems originally that the parties envisaged a banking relationship. At some point, the question of the issuance of guarantees came up. Meetings took place over 29-30 March 2007 at the Baekundong branch involving Mr Ryu together with Mr Bong Jung Lee and Mr Yong Mok Ha of GEO. Each of these people gave evidence, and they were good witnesses, and I generally accept their evidence so far as it went (though not as I shall explain that of Mr Lee and Mr Ha as to the position much later). At this point, the discussions concerned a refund guarantee in respect of GMS-102. It is not in dispute that Kwangju Bank was only asked by GEO to produce a refund guarantee for the refund of the 1^st instalment only, that is to say, US$1,571,000.00, plus interest. The transaction in the form I have described is recorded in Kwangju Bank’s own internal documents and also the applications made by GEO. The bank’s security was (initially) the cash deposit, and its annual charge was a fee based on 0.8% of US$1,721,600, which took account of possible exposure to a claim for interest. It is worth emphasising that at no time was there any contact between Vroon and Kwangju Bank.

According to Mr Lee’s first witness statement, on 28 March 2007 (that is the day before the meeting at the bank) GEO asked its broker Mr US Chung to obtain Vroon’s acceptance of a refund guarantee issued by Kwangju Bank covering just 10 percent. At the meeting with the bank on 29 March 2007, as they had not received Vroon’s response, he says that he spoke on the telephone with Mr Jang-Ho Kim (who was GEO’s managing director). Mr Kim was then at GEO’s office in Busan. Mr Lee says that he asked him to obtain Vroon’s answer. He says that Mr Kim called him later that day, saying that he had spoken with Mr US Chung who confirmed that Vroon agreed to what GEO was proposing.

In his first witness statement, Mr Ha says that he remembers Mr Lee calling Mr Kim who came on the telephone, and asking him to make contact with Vroon through the broker in order to find out whether Vroon would allow GEO to break down the refund guarantee into separate instalments. He says that he heard Mr Lee getting confirmation from Mr Kim on the phone that Vroon agreed to this. On that basis, they went ahead and arranged for Kwangju Bank to issue the refund guarantee in respect of the first instalment under contract GMS-102.

Broadly, both witnesses confirmed that account in cross-examination. However, what they may have believed was agreed by Vroon, and what was actually said by those in a position to agree may not be the same. There was no direct contact between Vroon and GEO. The chain of communication was from Mr Lee in the bank’s Baekundong branch to Mr Kim in GEO’s offices in Busan, and from Mr Kim to Mr US Chung, and from Mr Chung to Vroon or its broker. In fact, in cross-examination Mr Lee said that he did not know whether Mr Chung spoke to Mr Erichsen or directly to a Vroon representative. Neither Mr Kim nor Mr US Chung gave evidence. Mr Ha heard what Mr Lee said on the phone to Mr Kim, but not what Mr Kim was saying.

On the other side, Mr Marks denied that he had reached such an agreement. He said in cross-examination that there was never a discussion between him or GEO, and that he didn’t think there had been a discussion between the brokers either to go to a system of discrete refund guarantees “or anything that will come close to it or look like it”. It was put to him that he was not telling the truth in that regard. However I see no reason to doubt that that was his understanding at the time. Mr Erichsen’s witness statement says that so far as he was aware, following the initial agreement between the parties in October 2006, there was no later agreement that a separate refund guarantee would be issued for each instalment. In cross-examination, he said that he could not recall calling Mr US Chung on 29 March 2007, but accepted that there could have been a conversation between them in which he told Mr Chung that Vroon agreed to what GEO was proposing. So far as oral conversations are concerned, this was the high point of the defendants’ case on variation. The parties differ as to what agreement this conversation (if it happened) extended to, that is, identity of guarantor only, or refund guarantee limited to the first instalment of GMS-102 also.

There were further emails on 29 March 2007, which must be examined in that regard. Mr Kim emailed Mr US Chung as follows:

“Thank you for your good cooperation.

1.

With regard to subject, we are going to be issued RG from Kwangju Bank today (when receiving Vroon’s confirmation not later than tomorrow).

2.

In the meanwhile, we would like to ask Vroon that we want to pay back the loan with interest around 20th April at the time for receiving 2nd instalment by deduction from Vroon’s 2nd instalment.

Because we have to spend our money to get RG from the bank by deposit so much”.

Kwangju Bank says that paragraph 1 was a reference back to the proposal made on 28 March 2007 (this was the email with the reference to “the R/G for 10%”), and this must be right.

Mr Kim was clearly anxious for confirmation from Vroon, because the same day he emailed Vroon direct with reference to this email with its two numbered paragraphs as follows:

“I’m sorry to send you directly this urgent e-mail without Mr U.S. Chung/USC maritime because he is work out now. I have received that Vroon have confirmed item 1 below, but we have to receive to confirm item 2 below by reason below.

Therefore you are kindly requested to make a confirmation the item 2 below and reply to us and Mr U.S. Chung”

Mr Marks responded shortly afterwards “We accept your proposal and look forward to receiving the refund guarantees soonest”. Although Kwangju Bank contests this, it appears to me that he was thereby confirming paragraph 2, which had to do with repayment of the loan, rather than paragraph 1 which had to do with the refund guarantees. I reject the case that Vroon and GEO agreed (in variation of the shipbuilding contract) that, rather than deliver a guarantee for the refund of all the instalments to be paid under the contract, GEO would instead deliver, in the first instance, a guarantee only for the refund of the 1^st instalment (plus interest) to be followed by further separate refund guarantees for the refund of subsequent instalments.

The way the guarantee was issued by Kwangju Bank is described in Mr Ryu’s witness statement so far as he was aware of it. He said that this was the first such guarantee he had issued. The draft refund guarantee wording from Exhibit A to the shipbuilding contract was used as the text of the guarantee. Issuance was not by the branch however, but by the bank’s international department in Seoul. Communications from the branch were handled by the assistant manager (who did not give evidence), and there was no evidence from anyone in the international department. Mr Ryu says that the procedure was that the international department received the request from the branch, and would confirm whether issuance was properly approved, which could be verified from the computer systems which would show that issue was approved on the basis of a cash deposit. Though this was an international instrument subject to a foreign law, there is no evidence that advice was sought as to its terms. It may be noted however that there is a line at the end of the instrument as issued which is not in the Exhibit A form, namely, “AND GUARANTEE IS SUBJECT TO THE UNIFROM RULES DEMEND GUARANTEE ICC PUBLICATION 458”. This is clearly a reference to the first edition of the ICC Uniform Rules for Demand Guarantees which was published in 1992 (the successor Uniform Rules for Demand Guarantees 758 was published in 2010), and must have been added by the international department. It suggests that the international department thought that the bank was issuing a demand guarantee. The guarantee was sent by SWIFT message to the buyers’ bank, which was ABN AMRO Bank, Amsterdam on 30 March 2007.

Also on 30 March 2007, Mr Kim sent by email to Mr Marks a PDF copy of the guarantee. Mr Kim said in the email:

“We are pleased to submit you the copy of RG for the first instalment for GMS-102 attached herewith.

Please check the same at your bank.

Please be informed that we are going to submit you the copy of RG for the first instalment for GMS-103 within 10^th April.

And we are also going to submit you the copy of RG for the second instalment for GMS-102 on around 15^th April.”

As Kwangju Bank says, this implied that there would be a further refund guarantee in respect of the second instalment of GMS-102, which under the contract was due on 15 April 2007. However, I do not accept its further proposition that it is only consistent with the parties having agreed to vary the shipbuilding contract the day before. Nor do I accept that Vroon knew as at late March that if it did not agree to a variation of the refund guarantee structure, there would not be any project at all. I am not satisfied on the evidence that it was asked to agree to a variation of the refund guarantee structure.

Mr Marks did not reply to this email (it appears that he was on holiday at or about this time). What in fact happened was this. Kwangju Bank had correctly used the draft refund guarantee wording from Exhibit A to the shipbuilding contract. (I set out below the terms of the guarantee in its final form.) In the second paragraph of Exhibit A, there is a blank space to be filled in after the words “…after demand not exceeding…”. In the third paragraph, after a reference to automatic increase, there is a blank to be filled in after the words “…in any eventuality the amount of this guarantee shall not exceed…”. Kwangju Bank had inserted into these paragraphs the same figure, namely US$1,571,000 (which was the amount of the 1^st instalment). Had this figure remained in the third paragraph of the guarantee, then this would have been a strong indication that despite the reference to “automatic increase”, the intention was that the issuer’s liability was capped.

But this was not what happened. The evidence shows that a Mr Frans Waterman (who was in Vroon’s Treasury department and who did not give evidence at trial) received the guarantee. On the same day (30 March 2007) he emailed Mr Erichsen as follows:

“Sorry to inform you but the text of the RG is not acceptable

give and take a few typo’s there is one paragraph on page 2/3

QTE

“THE AMOUNT OF THIS GUARANTEE WILL BE AUTOMATICALLY INCREASED UPON THE BUILDERS’S RECEIPT (OF) THE THE SECOND, THIRD, FOURTH AND FIFTH INSTALMENTS, PLUS INTEREST THEREON AS PROVIDED IN THE CONTRACT”

UNQTE

So far so good

But in the next line

QTY

“BUT IN ANY EVENTUALLY THE AMOUNT OF THIS GUARANTEE SHALL NOT EXCEED THE TOTAL SUM OF USD ONE MILLION FIVE HUNDRED SEVENTY ONE THOUSAND (SAP U.S. DOLLARS 1,571,000)”

UNQTE

That cannot be the case once the second etc. instalments are actual paid; the maximum amount is the total paid instalments + interest thereon.

Also the last alinea “AND GUARANTEE UIS SUBJECT TO THE UNIFORM RULES..” is not mentioned in the contract and I do not know the content of these rules and therefore we suggest to skip it.

Attached the scan including remarks..also where we think a part of the text is missing

Waiting for your reply.”

This email states in unequivocal terms that the guarantee is not acceptable because once the second (etc.) instalments were paid, the maximum amount of the guarantee would be the total paid instalments, and not US$1,571,000. The message was passed to GEO. Mr Kim did not however query it or ask for a further explanation. On 31 March 2007 (a Saturday) he responded to Mr Erichsen, saying, “First of all we are very sorry to you and Vroon for sending incorrect document. I have sent it without checking because of late night. Now the bank dosen’t work due to Saturday. We will contact with the bank and take the R/G again and resubmit to you.” That message was passed on to Mr Waterman.

On Monday 2 April 2007, Mr Ha of GEO sent Kwangju Bank a fax which (though not copying Mr Waterman’s email) copied the annotated version of the refund guarantee showing the changes which needed to be made. He asked the bank to “review and correct”. Mr Ryu says he was told about the fax by the assistant manager, and asked him to discuss it with the international department. Thus, he was not involved in dealing with the request for the amendment. There is no evidence from the international department about what it may have understood by the changes. But the same day, Kwangju Bank issued amendments to the guarantee as asked, which were again sent by SWIFT message to ABN AMRO Bank. The number in the third paragraph of the instrument after the automatic increase wording was amended to read US$12,568,000 which was the amount of all five instalments prior to the delivery instalment for the vessel.

One would have thought that this change would have alerted either GEO or the international department of the bank to the possibility that the guarantee as amended was in terms covering this greater sum. But I accept the evidence that it did not alert them. Mr Ha’s evidence was that he told the assistant branch manager of the bank’s Baekundong branch that the reason for changing the amount in the third paragraph of the instrument was that if successive guarantees were issued in stages, then the total guaranteed amount would become greater and that this was why the number had to be amended, and if it was passed to the international department, presumably this explanation was accepted. Mr Ryu says, “At the time there was a limit on the amount of guarantee which the Bank could provide and the amount of US$12,568,000 was far beyond the limit on which the Bank could provide a refund guarantee to GEO”. There was no change in the security arrangements under which the guarantee had been issued.

On 2 April 2007, Mr Kim sent a PDF copy of the corrections to Mr Marks by email saying, “As we promised last Saturday, we are pleased to submit you the copy of the Correction for RG for the first instalment for GMS-102 which corrected on the first RG letter sent on 30th March by the bank attached herewith”. On 4 April 2007, the 1^st instalment for GMS-102 was paid by Vroon in the sum of US$1,571,000.

I have rejected the case of Kwangju Bank as to variation. SGIC does not identify how the alleged variation was agreed, arguing that it can be inferred from subsequent facts. (I have set out its position in paragraph 26 above.) But whilst (as Vroon puts it) reliance can be placed on conduct, as well as on written and oral communications, it must still be shown how that conduct involved the making, or acceptance, of an offer, not just that it was (in the broad sense) consistent with such an agreement having been reached. A contractual variation is different from a common assumption, which is the distinction between this aspect of the defence and that concerned with estoppel by convention. I agree with the claimants that the court cannot infer an agreement here from the parties’ subsequent conduct, though it may be relevant for other purposes. The defendants are financial institutions. They were not parties to the ship building contracts. Mr Lee recognised in his oral evidence that the alleged variation would be important for both Vroon and GEO. As he put it, there were potentially “big problems” if it had subsequently proved impossible to obtain further refund guarantees. No addendum was drawn up recording a variation as happened in some other instances (though it is right to say that the contracts did not require variations to be in any particular form). Nor is there an email exchange showing clearly a proposal for a new structure for the issuance of refund guarantees, and its acceptance by Vroon. The emails demonstrate an acceptance of Kwangju Bank as guarantor, but no more. Despite GEO’s understanding at the time, I am not persuaded that the ship building contracts were amended as alleged.

To put the factual findings into perspective, it is useful to summarise an aspect of the defendants’ defences that is particularly fact sensitive. It is Kwangju Banks’ case that as at late March and up to 18 April 2007 Kwangju Bank and the buyers had a continuing common intention that Kwangju Bank’s refund guarantee was to cover only the 1st instalment of the price of GMS-103 and that subsequent instalments would be covered by separate refund guarantees. It is not in dispute that Kwangju Bank was under the impression that it was being asked by GEO to produce a refund guarantee for one instalment only. The dispute is as to Vroon’s understanding in this regard. Kwangju Bank asks the court to find that from about 19 March (the date of the email first mentioning the bank to Vroon which I have set out above) up to 18 April 2007 (the day after Kwangju Bank issued the second and final of its two guarantees) it was Mr Marks’ intention that Kwangju Bank’s refund guarantee was to cover only the 1^st instalment (plus interest) to be paid under the shipbuilding contract and that subsequent instalments would be covered by separate guarantees. This is central to the defences both in rectification and estoppel.

A different period is relevant in the case of SGIC (the first SGIC guarantees were issued on 25 May 2007 and the last one was issued on 17 October 2008). However it is convenient to state at this point that SGIC’s case, as it is put in its closing submissions, is that each of Vroon, Kwangju Bank, GEO and SGIC intended and understood that the refund guarantees provided would be discrete guarantees in relation to discrete advance payments, and this again gives rise to defences both in rectification and estoppel.

It is also necessary to state that a significant plank of the case of SGIC (in closing, Mr Stephen Cogley QC, counsel for SGIC, described it as his bull point) is that Mr Herman Marks (Vroon’s director) was not an honest witness, and that he knew all along that the buyers had agreed to accept refund guarantees provided limited to single instalments of the advance payments. Kwangju Bank also strongly criticises Mr Marks’ evidence, describing it as a succession of reconstructions of events. I am going to have to reach conclusions in this respect.

I can deal with one point at the outset. I have already said sufficient as regards the defendants’ variation argument to reject the suggestion that it was Mr Marks’ intention from about 19 March 2007 that Kwangju Bank’s refund guarantee was to cover only the 1^st instalment. For essentially the same reasons as set out above, I am satisfied that he had no such intention for the period which I have so far covered, that is the period up to the issuance of the first guarantee. I shall now consider the facts as to the period following.

At this point in time, the only guarantee in place was for GMS-102 as described above. Shortly afterwards, GEO asked Kwangju Bank to issue a further refund guarantee to cover the 1^st instalment under the contract for GMS-103. This time it proposed that the security which it would provide to cover both this new guarantee and the guarantee already issued for GMS-102 would be in the form of 30% cash deposit and 70% by way of counter-guarantee in Kwangju Bank’s favour from SGIC. Because the security was not all in cash, Mr Ryu explains that this required Head Office approval. Again, the documentation recorded that the new refund guarantee was only for the 1^st instalment of the price of GMS-103. Mr Ryu explains that this was based on what the bank had been told by Mr Bong Jung Lee and Mr Yong Mok Ha of GEO. It accords with their understanding of the position.

On 16 April 2007, Mr Ha of GEO sent to Kwangju Bank draft terms of the refund guarantee to be issued in relation to GMS-103. The bank’s Head Office gave its approval on 17 April 2007, and a refund guarantee was issued (presumably by the international department) the same day. Again, the guarantee was sent by SWIFT message to the buyers’ bank, ABN AMRO Bank, Amsterdam. The same sentence was included at the end as had been in the un-amended guarantee relating to GMS – 102 as regards the ICC Uniform Rules for Demand Guarantees. This was spotted by Mr Ha of GEO, who on 18 April 2007 asked Kwangju Bank to delete it, which the bank did by amendment the same day. A PDF copy was sent by GEO to Mr US Chung who forwarded it to Mr Erichsen who forwarded it to Mr Marks by email saying, “GEO has confirmed R/G for 1^st instalment of the 2^nd vessel has been issued today by Kwangju bank as per attachment”.

It is convenient at this point to set out the terms of the guarantee. All subsequent guarantees issued by SGIC were in the same form (with a slight difference of wording as regard the number of instalments). By the instrument, Kwangju Bank guaranteed as follows:

Also on 18 April 2007, Mr Kim emailed Vroon directly as follows:

“We have been issued RG for the 1st instalment of GMS-103 by Kwangju Bank today and will be forwarded to you soon.

Our RG system is consisted by 2 steps (guarantors). First step is guaranteed 70% by Seoul Guarantee Insurance, the 2nd step is 30% by Kwangju Bank and RG is issued by the Kwangju Bank.

RG for the 1st instalment of GMS-104 and 105 will be issued around 11th May. And the next stage will be 2nd instalment of GMS-102. After this stage, we will send you the massage for confirmation before 2 weeks of issuing RG...”

There was then a reference to repayment of the loan, which is not relevant for present purposes. Lastly on that day, GEO sent a copy of the refund guarantee direct to Vroon together with monthly cashflows.

As regards the facts at around this time, it is also relevant to note that a delegation from GEO was attending at Vroon’s offices for meetings between 17 and 19 April 2007. Mr Marks said that the meetings were focused on the finalisation of vessel design, specifications, makers’ lists, and construction of the shipyard. But part of the meetings concerned commercial matters (and Mr Marks said he would have attended the commercial discussions) and part of these focused on the cash flow statements which were sent with the last email sent on 18 April 2007 mentioned above. Further, as Kwangju Bank emphasised in its submissions, the 2^nd instalment in relation to GMS-102 was contractually due on 15 April 2007, though it was not paid on that date (it was not paid until 29 October).

To complete the factual narrative as regards this period, on 19 and 20 April 2007, ABN AMRO notified Vroon of the issue of the guarantee and the amendment respectively. The first instalment in respect of GMS-103 was paid by Vroon on 20 April 2007 in the sum of US$1,571,000.

So far as Vroon is concerned, the main evidence as to continuing common intention and also what was intended and understood as regards the scope of the guarantee issued on 17 April 2007 came from Mr Marks. (The evidence of Ms Kranenkamp and Mr Spiljard was described by the defendants as of marginal relevance.) In its closing, Kwangju Bank submitted that the starting point was paragraph 68 of his witness statement, where he says “Looking back at the messages which I was receiving and sending at this time, I can see that I was being told by Geo that the sequence of events would be further refund guarantees, followed by invoices, followed by payment of the second instalment, and I think that ... I simply took this at face value”. However, I think that Vroon was correct to submit that read in full this passage in his witness statement shows that he is talking about the summer of 2007. Though there are parts of his oral evidence that can be read differently, on the whole I think he stuck to this in cross-examination. It is common ground that for the purposes of both rectification and estoppel, the position at the time of the issuance of the guarantee is what matters, since from that time Kwangju Bank was liable on the instrument as issued. There is a dispute whether that should be treated as 17 or 18 April. The bank says the relevant “cut-off date” for these purposes is 18 April 2007 because that was the day on which the buyer first saw the refund guarantee (when GEO sent it to them), whereas Vroon says that only the period prior to 18 April 2007 can be relevant. This difference of view is primarily because of the parties’ desire respectively to include and exclude from the factual enquiry Mr Kim’s email to Vroon of 18 April 2007 which mentions “our RG system”.

I do not accept the contention made in Vroon’s closing submissions that the email came after the Kwangju Bank RG was issued and hence too late for the purpose of establishing a common intention prior to its issue. It seems to me to be part of the factual picture contemporaneous with issuance, and relevant in that regard. But it is not determinative. Mr Marks said in his witness statement that he could not remember reading the email, though the evidence showed that it was forwarded to him, and he accepted in cross-examination that he probably had read it. He also accepted that the implication of the email was that a further refund guarantee would be provided in relation to the 2^nd instalment under GMS-102. The nub of Mr Marks’ evidence was that he did not recall looking at the message very closely and had not noticed the reference to refund guarantees for the second instalments. I shall set out my appraisal of Mr Marks later in this judgment, but for the moment state that I accept his evidence about this email. I do not find (as the defendants say I should) that it was his intention at this time that that the refund guarantee was only to cover the 1^st instalment, subsequent instalments to be covered by separate guarantees.

Mr Ryu agreed that, after 18 April 2007 until 30 June 2009, Kwangju Bank had no cause to have any dealings with GEO, and there is no evidence that it did so. It effectively dropped out of the picture.

The first of SGIC’s witnesses was Mr Hee Moon Lee who was assistant manager of the Masan branch of SGIC and whose role was marketing. In his witness statement he describes the process by which the company came to be involved. SGIC is in the business of issuing guarantees, and at the behest of the International Relations Division, he made a marketing approach to GEO in December 2006. There was a general meeting in March 2007. Mr Ha was the point of contact with GEO. Mr HM Lee recalls that he spoke to him between March and April 2007, at which time GEO was seeking a full cover refund guarantee for each contract which would cover 80% of the advance payments. However this proposal did not get through SGIC’s filter conditions, because of GEO’s lack of financial standing. Mr HM Lee says that he communicated this to Mr Ha on the phone probably sometime in early April 2007.

Sometime between 3 and 17 April (Mr HM Lee thinks), Mr Ha asked what SGIC’s attitude would be to providing refund guarantees on an incremental basis in relation to advance payments when such payments were made. He was interested in obtaining the business, and regarded this as a way forward. He made it clear that SGIC would consider each request as and when made, and that the security requirements were 30%. He says that he was told by Mr Ha that Vroon would not pay GEO until guarantees for the second instalments were provided.

In the case of SGIC there has not been the same focus on whatever contemporary documents there may have been passing between Vroon and Geo as to the issue of particular instruments. But GEO’s applications, internal SGIC evaluation requests, reports, and agreements are in the core bundle and have not been the subject of challenge. Mr Pil Ku Noh was section chief of SGIC’s Masan branch until June 2008, and was responsible for the issue of eight of the ten guarantees that SGIC eventually provided. These guarantees were the first that he had issued. The reason, he says, that there are so many guarantees is that they only related to one instalment each, and he believes that Vroon knew this. He says that he was provided with refund guarantees by GEO, and thought that all he had to do was to fill in certain of the details. There are in-house lawyers at SGIC, but they were not asked to review the guarantees. His recollection is that at the beginning of April 2007, he could see that GEO would not be financially strong enough to risk providing full refund guarantees in respect of the four shipbuilding contracts. His next involvement was in mid-April 2007, when Mr Ha asked if SGIC would issue refund guarantees in respect of each advance payment. He was not involved in the decision making process, but was instructed that SGIC was prepared in principle to do so, on the basis of a 30% cash deposit by way of security (later reduced to 20%). At around this time, SGIC also provided a counter-guarantee to Kwangju Bank in respect of a large proportion of its liability.

Each guarantee issued by SGIC, Mr Noh explains, was based on the standard form sent by GEO sometime in March 2007. Once the financial situation had been evaluated, security provided and rates agreed, he would then modify the standard form so that it covered the instalment in question. The decision had been made by the “examination committee” at SGIC’s head office. As regards the first one, he says he spoke with Mr Ha who told him that the part he now knows is described as the automatic extension provision related to the accumulated liability of SGIC under all refund guarantees including those which were due to be issued in the future, as distinct from the amount to be inserted in the top right hand corner of the refund guarantee which represented the liability of SGIC under one single refund guarantee. So far as he was aware, he says, everybody acted on the basis that there was a refund guarantee for each instalment, and he says that he has no doubt that GEO and Vroon fully understood this as well. However, neither he nor any other SGIC officer had any contact with Vroon at any time. Mr Noh’s understanding derived from what he was told by Mr Ha (and Kwangju Bank).

At this point, the refund guarantees covering GMS-104 and GMS-105 should have been issued by 11 May 2007, and Vroon had been chasing for them and would not pay the first instalments in respect of these contracts until it had them. It seems from the record that the first time Vroon was aware of SGIC was on 21 May 2007, when Mr US Chung emailed Mr Erichsen stating that GEO was expecting the refund guarantees for the remaining two ships (GMS-104 and GMS-105) to come from “Korea Export Insurance Corporation”, which is probably a reference to the second defendant, Seoul Guarantee Insurance Company. This company was satisfactory as an issuer so far as Vroon was concerned. For completeness, I should say that there was nothing to suggest (as SGIC stated in its opening submissions) that the shipbuilding contracts must have been varied at or immediately prior to the provision of the SGIC guarantees in May 2007.

The guarantees covering GMS-104 and GMS-105 were in fact issued by SGIC on 25 May 2007. By email of that day, GEO forwarded a copy “of RG for the first instalment for Hull No GMS-104/5” via the brokers to Vroon. These instruments were in the form of physical bonds issued by SGIC. Each is headed “Advance Payment Bond”, and uses the same words as those issued by Kwangju Bank which I have set out in full above (additionally the words “Guarantee Amount USD 1,571,000” appear in the top right hand corner). Unlike Kwangju Bank, SGIC does not contend that these instruments were not first demand instruments.

On 1 June 2007, the first instalments for GMS-104 and GMS-105 were paid (in each case US$795,056.92 went to Ship Finance B.V. on behalf of GEO in repayment of the loans and US$775,943.08 went to GEO).

By this time, as I have said, Kwangju Bank was no longer actively involved (in its opening it referred to subsequent exchanges as evidential support for its case as to the common intention of the parties at the time of its 17 April 2007 guarantee). After that time, there were frequent on-going exchanges between Vroon and GEO as the project got underway. GEO aside, Mr Marks says that Vroon’s business was very active at this time, with an extensive programme of acquisition of new vessels, as well as perhaps fifty four newbuildings in various parts of the world. He was heavily engaged overseeing three projects involving eleven vessels in addition to those being built by GEO, and spent a considerable time travelling, when he would read emails on his Blackberry. He makes these points in order to give what he sees as the proper context to the various exchanges of emails over the period between June and November 2007 on which reliance is placed by the defendants. He says that he was confused at this stage, and assumed that GEO was right when it said that further refund guarantees would need to be provided before the second instalments on the contracts fell due. It follows that over this period both GEO and Vroon were both working on that assumption. To anticipate however, Mr Marks says that he finally checked the wording of the guarantees in October, and realised that each covered all instalments, and told GEO accordingly. The defendants say that he is making up the story as to his confusion, and that he realised all along that the agreement was that there would be one refund guarantee per instalment.

I now must look at the emails. On 8 June 2007, Mr Erichsen emailed Mr US Chung stating, “As per the agreement between Vroon and GEO, the second instalments for the first 4 ships is due by end June. Pls discuss with the yard if they will be ready with refund guarantees by then and from where they expect the guarantees will be issued. Vroon will ask about this when they see the yard on Tuesday…”. This was a reference to a forthcoming visit by Mr Marks, who was in Korea for a naming ceremony unrelated to GEO. Mr Marks maintained in cross-examination that he was not the source of this message, and Mr Erichsen seemed to agree with that. GEO’s response on 11 June was "…we are preparing RG for the second instalment for the first 4 (four) ships by the Seoul Guarantee Insurance or Kwangju Bank around 27th July”.

In fact what happened is that GEO did not produce refund guarantees, but requested further financial accommodation from Vroon. On 24 July 2007 Mr Kim sent Mr Marks an email with the subject line “R/G for 2nd instalment”. It read:

“Regarding of the issuing R/G matter, we like to ask one thing favour of you to bellowing.

As you can assume R/G that it should be deposit 30% of the amount to the guarantor from total R/G amount for issuing.

At this present GEO that it is preparing to issued the R/G for 2nd instalment of GMS 102~105 (4 ships). But we are in faced with lack of money for issuing R/G, at this moment, due to we have had allocated a certain company capital move to facilities for block assembly yard.

With that, we are concerning that the best way of the solution is loan from you. Therefore, we would like to ask you to borrow around US$2,000,000, as our loan agreement before…”

Mr Kim went on to express GEO’s great appreciation if such a loan could be made. A similar message was sent by Mr US Chung on the same day referring to GEO’s purchase of a steel block fabrication factory in Daebul for approximately US$3m, saying, “they would like to ask for your help to put USD2mil into Escrow Account in Kwangju Bank that is prerequisite for next Refund Guarantees to be issued”.

Vroon agreed in principle to provide such a loan the next day, and as the defendants say, this shows that Vroon was prepared to afford the builder substantial assistance. On 25 July 2007 Mr Marks emailed Mr Kim in what was plainly a considered reply, which shows that though Vroon was prepared to make a loan, it wanted GEO to concentrate on its order and not on selling vessels to other buyers. At that point, two further vessels had been ordered pursuant to the option, and Vroon expressed its intention of ordering a further two. He said:

“Many thanks for your kind message which we duly noted. I had earlier received a similar request from Mr Chung on your behalf. We appreciate all your efforts to ensure the successful design, construction and delivery of the six vessels we now have on order with you. As you know we intend to have further talks with you to extend our order beyond the current six vessels and also to include larger design Asphalt/Bitumen vessels. We also hope that for the time being you continue to focus your design and engineering efforts on completing the design for our vessels. We earlier asked you not to spend marketing efforts on selling the vessels to other clients untill the design is completed and the new Mokpo facility is up and running. We would appreciate if that is still the case.

We would be willing in some form or another to assist financially at this point in time if in one way or another you could offer some form of security for the loan. We would also like to make sure that no further loans are necessary and that this loan would suffice to get the yard up and running and also to make sure that this will be sufficient to assist the banks in issuing/increasing the refund guarantees for the six vessels as well as for the two optional vessels (that we intend to declare in the 4th quarter). For your bankers I enclose a financial summary on Vroon Group B.V. for 2006. We would appreciate if our financial data are kept private and confidential.”

Mr Marks followed this with a further message to Mr Kim on 2 August 2007 as follows:

“As I mentioned this morning over the phone we would be willing to help you with providing a loan for USD 2 million.

In order to do this we would like to know the following.

1.

Can you offer some security for the loan of USD 2 million (e.g. assignement of the contract of the main engines or something else, like a personal garantee pledge of shares, or something else??)

2.

A confirmation that you do not plan to build for other owners for the time being (until the design for the vessels is completed and the yard is up and running).

3.

A payment plan for the next installment on the four vessels. When do you expect us to pay the 2nd 10% downpayment for all 4 ships and when will we receive the refund garantees for these payments (we will pay in accordance with the contract terms)

4.

When do you expect the refund garantees for vessel 5 and 6 and who will issue these? (KEIC?)

Look forward to hear from you soonest.

As discussed I will visit on August 20th 2007, at the meeting I would like to discuss further the option vessels 9 and 10.

Mr Kim responded the same day, and the relevant loan agreement was then drawn up and signed on 16 August 2007. By this agreement, Vroon loaned GEO US$2m. The agreement records that the loan is to fund investment in a steel building block facility to expedite the construction of GMS-102. It was repayable on launching or on 31 August 2008 whichever was sooner. It is part of the case of SGIC that the fact that Vroon did not disclose this loan to it discharges SGIC from responsibilities under its guarantees.

Following the emails as set out in the agreed chronology, Mr Marks visited Korea at the end of August 2007, and was entertained by Mr Bong-Jung Lee on 23 August. An email of 24 August from Mr US Chung to Mr Erichsen reports on a meeting the next day between representatives of Vroon and GEO, at which (among other things) the refund guarantees were discussed.

Mr Marks emailed the brokers on 26 August, his prime purpose seemingly being to ask them to accept a reduced rate of 1% commission on proposed vessels 9 and 10. As regards the first six vessels, he said “there is a lot to be done yet in order to make sure that these six ships get build (a.o. to increase the refund garantees so we can pay the 2^nd 10% instalment on the first 4 vessels and receive the RG’s for vessel 5 and 6)”. He accepted in cross examination that he was clearly working on the basis that the refund guarantees that were in existence at that point in time did not provide sufficient cover in relation to the payments to be made on the first four vessels. However he did not accept that these guarantees were confined to the instalments that had already been paid. His evidence was that he was confused at this time. He was mistaken, he said, but did not act on the mistake, having been, as he put it “passive in this whole process”.

Ms Susan Lee (Mr Bong-Jung Lee’s sister) emailed back on 28 August 2007 referring to the slow progress in obtaining refund guarantees from the “Seoul Insurance company”, obviously the second defendants. On 29 August, she emailed Mr Marks as follows:

“ … we would like to suggest you to Refund Garantees under SSANGYOUNG insurance company with 70% of whole amount of ships price for Vessel No 5 & 6. We will notice to the amount on the Letter of Guarantee.it is not percentage of the value. It is mean that we will guarantee the amount on this letter. The R/G for remain 30% will be issued before the time of steel cutting for Vessel No. 5. …”

The same day Mr Erichsen emailed Mr US Chung saying, “We got the following msg in from Geo. Neither myself or Mr Marks can understand what the Yard is trying to say. Can you pls look into this.” Mr US Chung explained on 30 August that, “… I have checked with GEO as follows. Originally GEO expects and understands Busan Bank to issue R/G under their name after taking Ssangyong’s portion of 70%, which is normal practice here, but this bank advised GEO that they can only issue R/G for their own portion of 30%. Then, alternatively, Ssangyong will issue 70% of whole amount for 5th/6th vessels now, but for remaining 30% R/G which Busan Bank is going to issue can be issued by Ssangyong around 15th June 2009 (just before steel cutting of 5th vessel), because at that time Ssangyong’s guarantee ceiling can be increased. The above is what I have heard and understood from GEO.” Then Mr US Chung continues, “GEO is asking for Vroon’s approval on the following: Each separate guarantee from Busan Bank (30%) and Ssangyong (70%); or one guarantee from Ssangyong. Meantime, R/Gs for the first four vessels are on the process through Seoul Guarantee Insurance Co.”.

On 17 September 2007 Mr Erichsen emailed Mr US Chung to the effect that Vroon was waiting for the next batch of guarantees to be issued. A chasing email was sent to Ms Susan Lee by Mr Marks on 24 September 2007, and followed up by Mr Erichsen on 27 September 2007.

Mr Marks’ evidence was to the effect that he began to focus properly on the terms of the refund guarantees in October. On 11 October 2007, Mr Marks emailed Ms Lee principally as regards the refund guarantees for vessels five and six. He said also, “we understand that for vessels 1 to 4 the current RG’s will remain in force and that you will invoice us for the second 10% instalment (and increase the current RG amounts)”. By vessels 1 to 4, he was referring to GMS 102 to 105. The email says that he will call to discuss, it seems that they were talking on the phone about that time.

SGIC made the following submissions as to these exchanges (and a later email exchange of 10 September 2008 from Mr Marks to Ms Lee confirming that the steel cutting instalment, that is the third instalment, in respect of GMS-105 would be secured under the refund guarantee). SGIC’s case is that it was agreed between Vroon and GEO that discrete RGs were “contractually” acceptable, and that position never changed. If it did, nobody told SGIC: likewise, even if there was no formal variation, the common assumption and conventional behaviour of the parties prescribed and permitted the provision of discrete RGs, and again nobody disabused SGIC of this understanding. However, an “alternative scenario” is suggested as follows (quoting from the closing submissions of SGIC): “… it is possible that what actually happened was that it dawned upon Susan Lee and Mr Marks that irrespective of the basis upon which the RGs had been provided at that stage, it might be possible to somehow exploit the automatic extension provisions in the future. However, the understanding between Geo and Mr Marks was that, nevertheless, the position being uncertain (particularly in view of the fact that Kwangju and SGIC had only agreed to provide discrete RGs) Geo would continue to obtain discrete RGs so that these could be “locked away for a rainy day”. Thus the obligation was still to obtain discrete RGs. This scenario explains, for example, why Vroon ultimately made payments which were not matched by discrete RGs toward the “end” of the relationship [Footnote: Another explanation is that Vroon simply took a risk being “too far in” and being desirous of obtaining the vessels], but neither Mr Marks nor Susan Lee/Geo were sure of whether any claim based on the automatic extension provisions would “stick”. This explains why Geo continued to obtain further RGs, and Vroon accepted them without demur. It also explains why Mr Marks was not sure whether the pre-payments (on the assumption that they constituted advance payments) would be covered under any RG …”.

SGIC’s suggestion that it dawned on Ms Lee and Mr Marks that it might be possible to exploit the automatic extension provisions and obtain further guarantees to be kept for a rainy day, is in my view inconsistent with the evidence, and I reject it. Although Ms Lee did not give evidence, her brother (Mr Bong-Jung Lee) did do so. I have dealt with his evidence elsewhere in this judgment, and have generally accepted it. I accept that he believed that Vroon had agreed that separate refund guarantees would be obtained in respect of each instalment. Whether his belief was correct or not is a different matter, but it effectively rules out SGIC’s alternative scenario. As to Mr Marks, his understanding of the position has to be seen in the light of what happened next.

At this point in time, notwithstanding the various discussions as to further guarantees which I have set out above, there was still only a single refund guarantee for each of the first four vessels—those issued by Kwangju Bank in respect of vessels GMS 102 and 103, and those issued by SGIC in respect of vessels GMS 104 and 105. Following these exchanges between Ms Lee and Mr Marks, invoices for the second instalments of GMS 102 – 105 were issued and sent to him by Ms Lee on 12 October 2007 (they appear to have been received at Vroon on 15 October 2007).

Mr Marks says that he suspects that he returned to the office about this time, and asked someone to check the wording of the guarantees. On 17 October 2007, Mr Marks emailed Ms Lee. This is an important document, upon which Vroon places considerable reliance. The email says:

“I am back into the office and look into the Refund Guarantee tesks [i.e. texts] for the vessel no 102-105. I noted that the RG’s will increased autimatically we pay and this seems fine to me. I will arrange for payment of the 4 invoices on Monday October 29th. Currently our Treasury Mngr is travelling as well as Mrs Alexandra Kranekamp the newbuilding project administrator. In any case this matter is resolved.

I now look forward to receive the RG’s for the vessel 5 and 6

Floris and Mr Nick Spiljaard are both travelling and will return tomorrow and revert to you on the site office matters.”

Vroon’s case is that having checked, Mr Marks realised that the wording of the guarantees increased automatically. His earlier confusion was thereby dispelled, he informed GEO accordingly, and authorised payment of the second instalments on 29 October 2007. (As to that date, Vroon suggests that payment 14 days from receipt of the invoices was in compliance with the relevant term of the contracts. Payment instructions to ABN AMRO were given on 26 October 2007.) As he put it in cross-examination, “There was refund guarantees being mentioned for second instalments, third instalments and that whole issue was seemingly confusing me until quite late in the process. It was only late in October I really sat down and looked at this and said all these things are in place.”

As I have mentioned, it is said on behalf of SGIC that Mr Marks was not an honest witness, and that he knew all along that the buyers had agreed to accept refund guarantees provided limited to single instalments of the advance payments. Much the same point is made on behalf of Kwangju Bank (though it does not directly impugn his honesty), the bank submitting that his recollection of events was poor. It follows that the defendants’ case is that this account of events culminating in the 17 October 2007 email must be untruthful, or at least reconstructed after the event. It comes down to a short but important point so far as this case is concerned. Should the court accept Mr Marks’ evidence in this regard? Or had he agreed to accept discrete refund guarantees, and sent what must have been a knowingly misleading email on 17 October 2007?

My findings on this issue are as follows. I accept the defendants’ submission that there are criticisms that can be made of Mr Marks’ evidence. He was, as they say, defensive in dealing with his understanding of emails from GEO and in seeking to explain why Vroon did not query the further guarantees which (as I shall describe) were subsequently received from SGIC—he was extensively cross-examined on these matters, but not unreasonably so given Vroon’s claim. On the other hand, I can accept the force of his point that he receives many emails every day, and does not necessarily absorb or respond to all of them. As he put it, an email may look different when seen in hard copy after the event than it appeared at the time to the recipient—particularly one who travels as much as he does.

In reaching a conclusion as to the veracity and reliability of Mr Marks’ evidence, a significant point is that the email of 17 October 2007 is a contemporaneous document. Of course, an email is not necessarily a good indication of a person’s state of mind, and can be deliberately drafted so as to create an impression which is misleading or incomplete, or to create a record. But the parties were not in dispute at the time, in fact they were cooperating, and there is no reason to think that this was the case here. Having observed him give evidence, I do not think that Mr Marks would behave in this way. Further, the email did not come out of the blue, but came after earlier email of 11 October 2007 that I have quoted above. Finally, the version of events set out in the email fits with the facts. It says that Mr Marks is back in the office, and has looked into the text of the refund guarantees. He has noted—correctly—that the guarantees increase automatically on payment by Vroon. He says that he would arrange for payment of the four GEO invoices on 29 October 2007, which he did. By authorising payment without further guarantees, he acted in a way that was contrary to what the defendants say was the common intention or understanding of the parties in this case. I do not think he sent a misleading email. I consider the email shows what he thought at the time. For these reasons, I reject the contention that he has given false evidence, or reconstructed events ex post facto. I broadly accept his evidence as to what happened.

As to GEO’s understanding of the email of 17 October 2007, it was clear from the evidence of Mr Bong-Jung Lee that he considered it carefully. He appeared to accept that he realised that Mr Marks was saying that the existing guarantees would automatically increase and that it followed that no further guarantees were necessary, but said that he did not think that this was what Mr Marks could have meant. He said that he thought that it was something which Vroon ought to have raised with the defendants rather than GEO. Mr Yong Mok Ha (the other GEO witness) said that the content of this email had not been communicated to him. GEO did not go back to Mr Marks for an explanation.

What this episode shows, in my judgment, are the misunderstandings that have given rise to this litigation. Communications between Vroon and GEO were not ideal, because the key people in GEO did not speak English. As Mr Marks said, GEO was not asking him to agree to something, or even asking him a question. If so, he might have engaged with it. GEO assumed that all parties were proceeding on the basis that there would be separate guarantees for each instalment. SGIC (though not Kwangju Bank) now accepts that by their terms, the amount of the guarantees increased automatically. But the bank and SGIC took it on trust that their liability on each was limited to a single instalment, without seemingly focusing on the language of the instruments that they were issuing, because that is what GEO told them. Mr Marks can be criticised for allowing himself to be confused over the summer of 2007, but when he finally focused, at a time when there was a single guarantee in respect of each contract, he realised that the value of the instruments increased automatically and told GEO accordingly. GEO however assumed that this could not be his meaning, and continued as before. It obtained the issue of further refund guarantees by SGIC, as I shall now explain.

The second instalment was paid to GEO in respect of GMS-102 to GMS-105 (i.e. all four vessels) in four payments of US$1,571,000 made on 29 October 2007, pursuant to instructions given on 26 October 2007.

Also on 29 October 2007, Ms Lee emailed Mr Marks enclosing pdf copies of further refund guarantees issued by SGIC in respect of GMS-102 and GMS-103. On 2 November 2007, Ms Lee emailed Mr Marks enclosing pdf copies of further refund guarantees issued by SGIC in respect of GMS-104 and GMS-105. SGIC contends in its closing submissions that “no earthly reason has been suggested for why Geo should be acting in this way, or why Vroon stood by and allowed Geo to act in this way”. Its case is that Mr Marks must have known that GEO was proceeding on the assumption that it was providing discrete guarantees, but chose to keep silent. This together with the fact that SGIC continued to be under the impression that it was issuing guarantees limited to individual instalments founds SGIC’s contention that its liability on each of these guarantees is limited to amount of the second instalment of the contract in question, namely US$1,571,000.

In fact, there was no response from Mr Marks. His evidence was that he didn’t react because he had made the payments, and the matter was “finished”. He accepted in cross-examination however that the message did not get through to GEO. His general explanation for a lack of response on his part to the provision of these further guarantees was that they were not, as he put it, in piles on his desk, but came intermittently with lots of other things in between. This was not something, he said, that he was occupied with. It was much later (his evidence was in late 2009) that he asked Vroon’s legal department about it, and was told that it was “not a problem”. He denied that he knew that SGIC was acting under a misapprehension after October 2007. As a factual matter, it is common ground that Vroon had no contact with SGIC. To quote its closing submissions, Mr Marks never disabused SGIC of any intention, post 17 October 2007, to either (a) rely on the Kwangju refund guarantees as blanket guarantees, or (b) treat any of the subsequently provided guarantees as blanket guarantees. In principle however, there would be no reason for Vroon as beneficiary of the guarantees, to have contact with SGIC as issuer. I do not think that Mr Marks can be criticised for not making contact with SGIC when the further guarantees were provided. The fact that no response was made to GEO is more surprising, though Mr Marks explained it on the basis that he had already made the position clear. I do not accept the “alternative scenario” put to him in cross-examination, namely that he kept quiet because he wanted to keep the guarantees for a rainy day—I have set out my reasons above in relation to the closing submissions of SGIC.

There are further email exchanges on which SGIC places reliance. On 20 November 2007, Mr Erichsen emailed Mr US Chung saying, “Is there any further news on the R/G for the next ships? Vroon pushing me hard on this. Also understand the Steel cutting on the first ship moved back to mid Dec. Will GEO be able to increase the R/G so that Vroon can pay the next instalment on the first ship?” It is the last sentence that SGIC relies on, saying that this can only have related to the third instalments as the second had been paid by this date and the refund guarantees for the second instalments provided. Mr Mark’s reaction in cross examination about this email was that it had been sent broker to broker. Vroon says that it shows that the confusion about the need for further RGs continued in the mind of Mr Erichsen even after Mr Marks’ email of 17 October 2007. The impression Mr Erichsen gave in his evidence is that the refund guarantees were not a matter of primary concern to him as broker. In any case, the email is mainly dealing with guarantees for the next ships, those in respect of which Vroon had exercised its option, and for which there were no guarantees in place.

A further email was sent by Mr Erichsen to Mr US Chung on 21 November 2007. It read:

“RE: GEO – VROON (GMS -108/109)

I spoke to Vroon this morning any they are infuriated and disappointed that Refund guarantees had been issued for Courtney Fynn and Novo ship. After All the assistance Vroon has Showed to GEO over the last 8-10 months with issuing loans and assistance on design and technical matters, the yards conduct is far from acceptable on this matter.

As per the meeting on the 23rd August President B.J. Lee confirmed to Mr Marks that the first RG’s to be issued were for hull the second instalments on the first 4 ships then for hull 106/107/. It was never even a question that Novoship or Courtney Fynn would be getting their R/G’s issued in full prior to the outstanding RG’s for Vroon.

During this meeting GEO also confirmed they would give us and offer for ship no 9 and 10.

Vroon informed us that their intention is to declare the option for Hull 108 and 109 today, we will get an official msg from them later today.

Upon doing so we expect to have an offer for the next two ships within today as promised by President Lee we would have last month.

Pls also inform the shipyard that Mr Marks and myself will be coming to Korea over the next 2 weeks to meet with GEO!”

In fact, the options for vessels GMS 108 and 109 (which were the seventh and eighth vessels ordered by Vroon from GEO) were declared that day by email from Mr Marks. Whilst SGIC suggested that the reference to the other buyers getting their refund guarantees “issued in full” meant they were “blanket” guarantees as opposed to the “discrete” guarantees which Vroon was getting, Mr Erichsen did not accept this reading. Vroon’s case is “issued in full” is a reference to the fact that these other buyers had received guarantees for all of their vessels. It is not clear, but this in my view is the better reading. Although these emails from Mr Erichsen provide some further support for the case of SGIC, when seen in context, I do not think they alter the factual conclusions set out above.

In most respects, subsequent events have not been the subject of the same degree of focus in this trial, and I can deal with them more briefly. The 3^rd instalment (steel cutting) for GMS-102 was paid by Vroon to GEO on 28 December 2007 in the sum of US$3,142,000. A further refund guarantee was issued by SGIC the same day.

The 3^rd instalment (steel cutting) for GMS-103 was paid by Vroon to GEO on 20 May 2008 in the sum of US$3,142,000. A further refund guarantee was issued by SGIC on 30 May 2008.

The 4^th instalment (keel laying) for GMS-102 was paid by Vroon to GEO on 23 June 2008 in the sum of US$3,142,000. A further refund guarantee was issued by SGIC on 3 July 2008. This guarantee, and the last one issued by SGIC in October 2008 was handled by Mr Myeong Jung Doh, who by then had taken over from Mr Noh at SGIC’s Masan branch. His evidence (which I accept) was that he had no previous experience of refund guarantees, seeing his task as essentially an administrative one. He used the form that had been used previously. At no time was there any contact between SGIC and Vroon.

In mid-July 2008, GEO requested that Vroon provide assistance by making a prepayment in respect of the cost of the extra items which had been requested by it, which would usually only be paid as part of the final instalment. These amounted to US$2,125,000 per vessel for GMS-102 and GMS-103. Mr Marks tried to persuade GEO to have the prepayments covered under the refund guarantees for each vessel. Ms Lee responded to the effect this could not be done. Ultimately, Mr Lee provided a personal guarantee. Loan agreements were entered into on 24 July 2008 pursuant to which Vroon loaned GEO the sum of US$2,125,000 in respect of each vessel. SGIC contends that it is discharged from liability under its guarantees because Vroon did not disclose these loan agreements to it.

The 3^rd instalment for GMS-104 was paid by Vroon to GEO in the sum of US$3,142,000 on 21 August 2008. A further refund guarantee was provided by SGIC in respect of this contract on 17 October 2008. This was the last guarantee which SGIC issued.

In fact, instalments continued to be paid by Vroon as follows. The 3^rd instalment for GMS-105 in the sum US$3,112,000 (which took account of a rebate of $30,000 for early payment) was paid on 12 September 2008. (Ms Lee confirmed by email of 10 September 2008 that this payment was fully secured under “refund guarantee” though there was no separate instrument issued at this time.) The 4^th instalment (keel laying) for GMS-103 in the sum of US$3,142,000 was paid on 7 October 2008. The 4^th instalment for GMS-104 in the sum of US$3,142,000 was paid on 5 January 2009. The 5^th instalment for GMS-102 was paid (subject to a deduction in respect of outstanding loan amount from August 2007) on 14 April 2009.

As Vroon submits, even where a further refund guarantee was provided, the instalment payment had already been made before it was received. On no view however, were these last instalments the subject of any “discrete” refund guarantee. Here again, there is a significant dispute of fact. In his witness statement, Mr Bong-Jung Lee says that although there was never any formal agreement with Vroon that GEO would not continue to issue further guarantees, ultimately an understanding was reached with Vroon that the 4^th and 5^th instalments would be unsecured by any refund guarantee. However, in cross-examination this evidence was considerably qualified. Mr Lee indicated that his witness statement reflected simply what he had been told by Mr Yong- Mok Ha, but in his cross-examination, Mr Ha accepted that he was not a party to any such agreement, and was not directly involved. In his cross examination, Mr Marks rejected the suggestion that any such agreement had been reached. He said that Vroon felt secure that the payments made were covered under the refund guarantees. In my view, the evidence from GEO fell considerably short of establishing any such understanding or agreement to the contrary, and I prefer the evidence of Mr Marks in this regard, which is consistent with the factual position as it actually was.

By the beginning of 2009, the market had decisively turned, and I infer that GEO was in considerable difficulty. It indicated that it wanted to agree the return of the refund guarantees issued in respect of GMS-102, on the basis that it would “deliver” (in the sense of transferring title to) that hull to Vroon once it had been launched. Vroon would have the part-completed hull as security for the instalments it had prepaid. This was agreed by way of Addendum No.3 dated 19 March 2009. A bill of sale was executed dated 20 March 2009. Once that had been agreed, the refund guarantees for GMS-102 were returned by Vroon. It took delivery of GMS-102 on 5 November 2009, but she was incomplete and required further work at a different shipyard. In the event, this delivery was the only one which GEO made.

In mid-2009, GEO requested that Vroon make prepayments in respect of forthcoming instalments. Two prepayments of US$500,000 each against the 6^th and final (delivery) instalment for GMS-102 were made on 20 May 2009. Two prepayments of US$500,000 each against the 5^th instalment (launching) for GMS-103 were made on 27 May 2009 and 17 June 2009. Originally the latter payments were included in the claim in the first action, but this was abandoned shortly before trial.

By late June/early July 2009 there was increasing concern on the part of Vroon about the extent of its exposure to default by GEO in respect of GMS-103. Mr Lee told me that a lot of people started leaving the company as of October 2009. By November 2009, it seemed to Vroon that work had stopped on its vessels, and I find that this was so.

On 27 January 2010, Vroon cancelled the GMS-103 contract on grounds of delay and called upon GEO to refund the instalment payments made. This was followed up on 3 February 2010 by demands on Kwangju Bank and SGIC. Both financial institutions adopted the position that they were only responsible for repaying the particular instalments for which their refund guarantees had been issued. Payments were made by them on this basis on 5 and 24 March respectively. As noted at the beginning of this judgment, Kwangju Bank (but not SGIC) argues that it has a defence on the grounds that the buyers’ termination of the shipbuilding contract in relation to GMS-103 (upon which the return of the advance payments was premised) was premature.

A similar pattern was followed for GMS-104 and GMS-105 (for which SGIC was the sole refund guarantor). The GMS-104 contract was cancelled on 29 April 2010, a demand was made on SGIC on 6 May 2010, and a part-payment of the sum demanded was made by SGIC on 16 June 2010. The GMS-105 contract was cancelled on 30 July 2010, a demand was made on SGIC on 9 August 2010, and a part-payment of the sum demanded was made by SGIC on 7 September 2010.

As I have indicated above, there have been a considerable number of defences raised in this case, particularly by Kwangju Bank. I shall deal with them in the following order: (1) the secondary liability/demand guarantee point (Kwangju Bank only), (2) the construction defence as to the limit of the guarantee (Kwangju Bank only), (3) the premature termination of the shipbuilding contract defence (Kwangju Bank only/secondary liability only), (4) the variation of the shipbuilding contract defence, (5) the non-disclosure defence, (6) the Statute of Frauds defence (Kwangju Bank only/secondary liability only), (7) rectification, (8) estoppel, (9) the counterclaim for the return of sums already paid on the guarantees on restitutionary grounds.

The secondary liability/demand guarantee point

This is the first of two issues of construction that arise between Vroon and Kwangju Bank. The bank argues that the refund guarantee applicable in its case is a “see to it” guarantee imposing secondary, not primary, liability, in other words that it was a true guarantee rather than an instrument payable on first demand. Strictly speaking this is a premise upon which various defences are adumbrated rather than a defence in itself. There were two guarantees issued by Kwangju Bank, and both are relevant in the factual account. However only the second is the subject of a claim (the vessel which was the subject of the first guarantee was delivered). The guarantees issued by SGIC were in the same terms, but it does not dispute the primary nature of the instruments (it does however adopt the position of Kwangju Bank should I find in favour of the bank on the issue).

The terms of the instruments which the defendants issued are set out in full above. Although those issued by Kwangju Bank are headed “Letter of Guarantee” whereas those issued by SGIC are headed “Advance Payment Bond”, the terms are the same. I have explained above why this was so. The instruments were in the form required by the shipbuilding contracts, the form originally having been provided by the builders’ broker. There are also some immaterial differences in terms of method of issue. Kwangju Bank’s guarantees were issued bank to bank through the SWIFT system, whereas those of SGIC (whose business consists so far as relevant in the issuance of guarantees) were physical bonds.

Whether Kwangju Bank is correct to submit that its liability is conditional on GEO being liable under the shipbuilding contract depends upon the nature of the instrument. There is a fundamental difference between a guarantee and a demand instrument. A guarantee imposes secondary liability, primary liability remaining with the principal debtor. A demand guarantee—also called a “first” demand or “on demand” guarantee or bond, the nomenclature does not matter—imposes primary liability on the issuer. Liability flows from the instrument, not the underlying contractual arrangements which give rise to the issue of the instrument. These are often international instruments, and the English law distinction between a guarantee and an indemnity may not be helpful. A better (though not perfect) analogy is with an irrevocable documentary credit, which like a demand guarantee, is an autonomous undertaking by the issuer. In the case of a demand guarantee, liability arises upon a demand that conforms with the requirements of the guarantee, including as to any supporting documents, and made within the validity period of the instrument. Such instruments cover a variety of situations, including tender guarantees, advance payment guarantees, performance guarantees, and so on. They have proved to be of great value in international commerce. Particularly when issued by financial institutions, the beneficiary has a promise to pay from a creditworthy promisor which is independent from that of its contractual counterparty. These points are uncontroversial, and emerge from the authorities beginning with early cases such as Edward Owen Engineering Ltd v Barclays Bank International Ltd [1978] QB 159, and the more recent Court of Appeal authority cited by the parties, Marubeni Hong Kong and South China Ltd v Government of Mongolia [2005] 1 WLR 2497, Gold Coast Ltd v Caja de Ahorros del Mediterraneo [2002] 1 Lloyd’s Rep 617, and IIG Capital LLC v Van Der Merwe [2008] 2 Lloyd’s Rep 187.

The nature of the instrument in question is a matter of construction of the instrument as a whole without any preconceptions as to what it is (Gold Coast at [15], IIG at [7]). As Tuckey LJ pointed out in Gold Coast at [21], the description of an instrument as a “guarantee” is simply a label. The question is as to its effect. At [16], he referred to a passage in Paget’s Law of Banking (11^th edn) where the authors say that an instrument which (i) relates to an underlying transaction between parties in different jurisdictions, (ii) is issued by a bank (iii) contains an undertaking to pay “on demand”, and (iv) does not contain clauses excluding or limiting the defences available to a guarantor, will almost always be construed as a demand guarantee. There was force in Kwangju Bank’s submission that subsequent authority shows that these indicia are not determinative. For example, Associated British Ports v Ferryways NV [2009] 1 Lloyd’s Rep 595 at [12] shows that the fourth factor may be neutral. The mere use of the words “on demand” is also not determinative, since a secondary guarantee often requires a demand on the guarantor (IIG at [25]) (which may be relevant for example for limitation purposes). On the other hand, the significance of the fact that issue is by a financial institution is emphasised in Marubeni at [28]. It appears to me that whether or not determinative, the list of indicia is at least a useful checklist.

Kwangju Bank relies on the following points in support of its contention that its guarantee created secondary liability. (1) The word “If” at the beginning of the second paragraph: the provision reads, “If, in connection with the terms of the contract, the buyer shall become entitled to a refund of the advance payments…”. Therefore, it is contended, the provision imposes liability on a contingency. (2) Under the second paragraph Kwangju Bank guarantees the “repayment” of the advance payments to the buyer. GEO is the only party which can repay monies to the buyer, hence Kwangju Bank is guaranteeing the performance of the GEO’s repayment obligation, not promising its own primary obligation. The following words “the same” are a reference back, and Kwangju Bank is promising to refund those advance payments if GEO is obliged to do so and fails to perform that obligation. (3) The reference in the fifth paragraph to payment of any interest being by way of liquidated damages due to cancellation of the contract shows that liability is secondary, since if the instrument imposed a primary liability, this paragraph would be redundant. This provision mirrors the third paragraph of Article X.5 of the shipbuilding contract, and as such, it specifically ties in the obligation to pay interest under the guarantee with the obligation to pay interest under the shipbuilding contract. (4) The provision in the sixth paragraph is another positive indication that the liability is secondary because, if the claim/demand for refund from GEO is challenged in arbitration, the liability of Kwangju Bank is dependent on the determination in the arbitration of the underlying obligations between the buyer and GEO. Payment is stated to be “the sum adjudged to be due to you by the Builder pursuant to the award made under such arbitration”, which may differ from the amount demanded by the buyer. (5) The provision in the seventh paragraph that upon receipt by the buyer of the sum guaranteed or upon its acceptance of the vessel the refund guarantee shall “become null and void” is language consistent with a secondary rather than a primary obligation. (6) Finally, the refund guarantee repeatedly uses the words “guarantee” “guarantees” etc. It is also headed “Letter of Guarantee”. Vroon takes issue with each of these points.

My conclusions are as follows. An argument similar to (1) was rejected by Beatson J in Meritz Fire & Marine Insurance v. Jan de Nul NV [2011] BLR 320 at [74] and [75]. In my view, the fact that the guarantee states that, “If … the buyer shall become entitled to a refund of the advance payments … we hereby irrevocably guarantee the repayment of the same …” does not in context point to Kwangju Bank’s liability being secondary. The guarantee provides in the fourth paragraph for payment “on demand” against “a signed statement certifying that the buyer’s demand has been made in accordance with Article X of the contract and that the builder has failed to make the refund”. It is (as Vroon submits) this statement which conditions the bank’s obligation to pay, rather than the buyer’s contractual entitlement to a refund of the advance payments, something which in the normal course, the bank will not be in a position to evaluate. In my judgment, this also answers points (2) and (3).

The fifth paragraph of the guarantee (point (3)) provides that payment of any interest is by way of liquidated damages due to cancellation of the contract and not by way of compensation for use of money. This was a point raised by Kwangju Bank in its closing submissions. The meaning of the paragraph is slightly obscure. A possible construction (as Vroon submits) is that the paragraph does no more than make clear that interest is not to be treated as a penalty. On the other hand, the drafting links interest to the shipbuilding contract, and I think that Kwangju Bank is right to say that it lends support to its construction, though not in my view a great deal of support when seen against the other provisions.

Kwangju Bank’s argument on the sixth paragraph of the guarantee (point (4)) is to the effect that the reference to an arbitration award is an indication that its liability is secondary. The paragraph provides that where, within 30 days from the date of the buyer’s claim on the builder, the bank receives notification that the claim for cancellation/refund has been disputed and referred to arbitration, payment will be of the sum adjudged due in the award, and made upon demand together with a copy of the award. The provision envisages that in the event that the builder refers the issue to arbitration, the guarantee will respond to the amount adjudged to be due under the award. It is true that there may be an issue on the construction of the guarantee where there has been a prior demand under the fourth paragraph, but that does not arise in this case, where there was no reference to arbitration. This provision does not in my view imply in any way that the guarantor’s liability is secondary. The instrument is conditioned either upon the certification in the demand, or (if the dispute as to cancellation/refund goes to arbitration) the amount of the award. Essentially the same argument on very similar provisions was rejected in Meritz at [76].

As to point (5), the same clause was found in the Meritz guarantees (set out at [24]), though it was not argued in that case that it had any bearing on the nature of the instruments. It is fair to say that the “null and void” provisions are predicated on facts ultimately referable to the underlying contract, but read with the instrument as a whole, I accept Vroon’s submission that this provision is consistent with primary liability.

As to point (6), as stated in Gold Coast at [21], nothing turns on whether the instrument is called a guarantee. This is no indication whatever that the instrument imposes a secondary obligation (whether the term “guarantee” is repeated in the body of the instrument or not). Its nature depends on the terms of the instrument in question.

The paragraph of the instrument dealing with liquidated damages notwithstanding, in agreement with Vroon, my view is that on their true construction, these guarantees were clearly demand guarantees, not “see to it” guarantees under which Kwangju’s liability was conditional on GEO being liable under the shipbuilding contracts.

Kwangju Bank however contends that a number of further factors point in favour of its construction. It is said that “usual language of primary obligations - such as “unconditionally”, “primary obligor”, “not as surety”, “forthwith” - is not present here”. Kwangju Bank points out that the words “and unconditionally” were present in paragraph 2 of the Meritz instruments, but are not present in the instruments in this case. These are marginal points in my view. There are no special words necessary to constitute a demand guarantee. In Kwangju Bank’s closing submissions it is contended that its “liability under the refund guarantee is co-extensive with GEO’s liability. Certainly it is not greater than GEO’s liability”—but that is to state, rather than make good, the conclusion.

Further, Kwangju Bank contends that the fourth paragraph of the guarantee is expressly made subject to the second paragraph. The relevant words are, “The payment by the undersigned under this guarantee (subject to the second and third paragraph hereof)…”. (The words in parentheses were not present in the Meritz guarantees.) Kwangju Bank argues that the fourth paragraph is made subject to the conditionality of the second paragraph, which it says is the operative provision (it is the provision which spells out Kwangju Bank’s promise to guarantee repayment). The fourth paragraph, it argues, deals with the mechanics of payment (describing how the buyer is to make a claim, assuming that the conditions in the second paragraph have been met). For the reasons given by Vroon, I reject this contention. The words “subject to …” in parentheses are concerned with the amount of the demand. Otherwise it would not be necessary to refer to the third paragraph. The purpose of the provision that “…payment… shall be made upon simple receipt by us of written demand from you including a signed statement certifying … (etc.)” is (as Vroon submits) to provide documents which can be checked by a bank employee without any knowledge of the underlying facts. This is consistent with a demand guarantee conditioned upon documents. Contrary to Kwangju Bank’s submissions, it makes no difference that there is no “conclusive evidence” wording such as is sometimes found in these instruments.

I conclude that the refund guarantees in the present case are demand guarantees. Despite the number of points raised, I do not think that this is a doubtful issue. SGIC which issued materially identical instruments does not support the bank’s contention. In its opening, it states that, “It is accepted that the RGs are similar to those in [Meritz] and that the RGs provided by SGIC appear to fall within the category of instruments that have often been regarded as “demand” guarantees and thus respond upon presentation of documents/certificates. Such liabilities are classically primary, or in any event not eroded by any underlying dispute …”. In my view, this concession was rightly made.

There is a further point worth mentioning. In the Court of Appeal in Meritz ([2011] EWCA Civ 827 at [19]), Longmore LJ said in dismissing the appeal:

“One of Meritz 's main arguments below was that the APGs were not like performance bonds in respect of which money was automatically due on the beneficiary's say-so but were traditional "see to it" guarantees pursuant to which the beneficiary had to prove that the principal debtor was truly liable to his counter party under the original contract. In the light of the incorporation of the Uniform Rules No. 458, which expressly state that the terms of the underlying contract are of no concern to the beneficiary and the guarantor, this argument is extremely difficult and was, in my view, rightly rejected by the judge.”

As indicated in this passage, the incorporation of the ICC Uniform Rules for Demand Guarantees (UDRG) is likely to be conclusive as to the nature of the instrument. But its omission is in itself (in my view) a neutral factor. I have described above what happened as regards the UDRG in the present case. In summary, the guarantee as first issued by Kwangju Bank in respect of GMS-102 incorporated the UDRG. The rules do not appear in the form annexed to the shipbuilding contract, and it is a fair inference that this reference must have been added by the bank’s international department when issuing the guarantee and sending it through the SWIFT system to Vroon’s bank in Amsterdam. This might be taken to indicate what the bank thought it was issuing. As it happened, upon receipt, a member of Vroon’s treasury (not legal) department noticed the reference, and since he was unfamiliar with the rules, asked for it to be deleted. The reference was duly removed by Kwangju Bank by amendment. When the guarantee in respect of GMS-103 (upon which Vroon sues) was issued a few weeks later, it was also amended to omit the reference. However both parties accept that this cannot be relevant to the construction issue, and I have left it out of account.

The construction defence as to the limit of Kwangju Bank’s guarantee

This is the other issue that arises between Vroon and Kwangju Bank on the construction of the Kwangju Bank refund guarantee. The bank contends that on a true construction of the terms, its liability is limited to US$1,571,000 plus interest. If correct, this is a complete answer to the claim against it. Again, SGIC does not take this point on its guarantees, although the instruments are in the same form.

The issue turns on the interaction between the second and third paragraphs of the guarantee (which is set out above). The second paragraph provides that if the buyer “ … shall become entitled to a refund of the advance payments made to the builder prior to the delivery of the vessel, we hereby irrevocably guarantee the repayment of the same to the buyer within thirty (30) days after demand not exceeding USD1,571,000.00 … together with interest thereon …”. The third paragraph provides that “the amount of this guarantee will be automatically increased upon the builder’s receipt the second, third, fourth and fifth instalments, plus interest thereon as provided in the contract, but in any eventuality the amount of this guarantee shall not exceed the total sum of USD12,568.000 …”. The point essentially that Kwangju Bank argues is that the guarantee is limited to the amounts originally specified in the second paragraph, whereas Vroon argues that the effect of the third paragraph is that the amount payable on demand automatically increased as each subsequent instalment of the relevant contract price was received by the builder.

An initial point taken by Kwangju Bank is as to the correct approach to construction of the refund guarantee. It submits that guarantees are to be strictly construed so that no liability is imposed on the guarantor which is not clearly and distinctly covered by their terms and that any ambiguity in the construction of an instrument should be construed in favour of the guarantor: in support Chitty on Contracts (2008) 30^th edn) para 44-056 is cited. It submits that indemnities are subject to the strictest degree of construction recognised, citing McMeel, The Construction of Contracts (2011) 2^nd edn, para 20.65.

The passage in Chitty which is relied on is dealing with suretyship in the sense of secondary liability, whereas I have held that Kwangju Bank’s guarantee was a demand guarantee. On that basis, this point does not arise. It is worth however noting that a reason for a strict approach to construction given in Chitty is that, in general, a surety receives no benefit from the contract which is, so far as he is concerned, gratuitous. This is far removed from a situation such as the present in which a financial institution issues a demand guarantee for a fee. As to indemnities, as Vroon says, the passage cited from Professor McMeel’s book is concerned with indemnity clauses, rather than indemnities. It is clearly not dealing with instruments like the Kwangju Bank guarantee. In any case, as I have suggested above, the analogy between indemnities and demand guarantees is not necessarily apposite one. No other authority has been cited in favour of Kwangju Bank’s approach to construction, and I would reject it. Kwangju Bank’s demand guarantee was a commercial instrument which should be construed without reference to principles such as strict construction or contra proferentem. (In fact, I do not think that this conclusion makes any difference to the result, since, as I shall explain, the meaning of the bank’s guarantee appears to me to be clear.)

Both parties have made points on the factual matrix. The most compelling one is that made by Vroon—the shipbuilding contracts provided for a single guarantee covering all instalments, and that is what both parties originally envisaged. In this context, the “automatic increase” wording in the form annexed to the contracts (which was the form used in the demand guarantees as issued) would apply so as to increase the amount of the guarantee as instalments were paid. It is correct, as Kwangju Bank points out, that when ABN AMRO notified Vroon of the issue of the refund guarantee on 19 April 2007, it stated the amount as USD1,571,000. However, that was the amount of the refund guarantee at that time, and this shows nothing as regards the effect of the “automatic increase” wording.

Among the other points taken by Kwangju Bank is the fact that when a different financial institution (Hanwha Non-Life Insurance) issued advance payment guarantees in respect of vessels GMS-106 and GMS-107 on 4 March 2008, the “not exceeding” figure in paragraph 2, was the same as the figure in paragraph 3, namely USD12,568,000. Perhaps Vroon is correct to say that there must have been an error in the way the Hanwha guarantees were filled in, because to have the same figure in both paragraphs renders otiose the “automatic increase” wording. However, this was long after the Kwangju Bank guarantee had been given, and cannot affect its construction. Some of the other points raised appear to go to Kwangju Bank’s contention that the shipbuilding contract between Vroon and GEO was varied as regards refund guarantees, a contention that I have rejected for reasons set out above.

The matter comes down to a question on the construction of the wording of the guarantee. Kwangju Bank points to the use of the term “advance payments” in the plural in the second paragraph—the guarantee responds if the buyer shall become entitled to a refund of the “advance payments” made prior to the delivery of the vessel”. It submits that the use of the plural shows that the paragraph sets out the extent of Kwangju’s liability not just in respect of the first instalment, but in respect of all of the advance payments which might fall to be repaid. But as Vroon points out, the wording is equally consistent with its own construction. As a matter of construction, I do not think that the use of the plural in the second paragraph has much effect on the meaning to be given to the whole instrument in the light of the third paragraph.

I have explained above, when the first Kwangju Bank guarantee was issued in relation to the first vessel (GMS-102), the figure in paragraphs two and three was the same, namely USD 1,571,000. Someone in Vroon’s treasury department realised that this did not allow for the automatic increase provisions in the third paragraph, and took objection. Accordingly, Kwangju Bank amended the guarantee so that the figure of USD 12,568,000 (that is to say the total of all the pre-delivery instalments) appeared in the third paragraph. It may be, as was submitted on behalf of Kwangju Bank, that in construing the instrument, this is inadmissible as part of the negotiations. But it is nevertheless the explanation of why the figure appears. Kwangju Bank has to explain why the guarantee should be read in a way that effectively construes it in its un-amended form.

It seeks to do so by invoking a distinction that can be relevant, for example, in a bankruptcy of the principal debtor, between a guarantee of the whole of a debt with a limit on the extent of the guarantor’s liability, and a guarantee of a part of the debt. This can make a difference to the guarantor’s right to prove in the debtor’s bankruptcy in competition with the creditor. Kwangju Bank cites the bankruptcy case of InRe Sass [1896] 2 QB 12 in that regard. In its closing submissions, it suggested that the third paragraph of the refund guarantee was aimed at, or at least dealt with, this point, because a guarantee of the whole of the instalments (but with a limit of liability) was to the commercial advantage of Vroon.

There is however no apparent reason why the rules as to proof in an English bankruptcy should have been of concern to any of the parties in this case (none of the parties have any connection with England). Kwangju Bank’s contention is intended to provide some justification for a construction that effectively renders the third paragraph of the guarantee otiose, by containing its liability within the second paragraph. I do not accept that this is a proper construction of the contractual provisions of the guarantee as a whole. On their true construction, and in agreement with Vroon, I find that the guarantee provides that the amount payable on demand automatically increases as each subsequent instalment of the contract price is received by the builder. As I have noted, although its refund guarantees are for these purposes in identical form, SGIC did not adopt Kwangju Bank’s construction argument. In my view, it was right not to do so.

Kwangju Bank’s defence that Vroon terminated the shipbuilding contract prematurely

Kwangju Bank argues that it is not liable on its guarantee because the buyers’ termination of the shipbuilding contract in relation to GMS 103 (upon which the return of the advance payments was premised) was premature. This point is not taken by SGIC. Kwangju Bank accepts that it only arises if the refund guarantee imposes secondary, not primary, liability. I have held that the guarantee was a demand guarantee imposing primary liability, not a “see to it” guarantee imposing secondary liability. It follows that this defence does not arise, and I can deal with it shortly.

Kwangju Bank’s argument is that Vroon purported to cancel the shipbuilding contract on 27 January 2010 on the basis that the delivery of the vessel was delayed for a period of more than 210 days beyond the date when delivery was due, namely 30 June 2009 (the date set out in Addendum No 1). However, because Vroon paid the second instalment on 29 October 2007 instead of 30 June 2007, it was deemed to be in default. The contractual “Delivery Date” (it is argued) was thereby automatically extended under Article XI.2 of the shipbuilding contract, with the result that the 210 days ran from 29 October 2009, and did not expire until 28 May 2010. Vroon’s cancellation was wrongful, amounting to a repudiatory breach of contract, which GEO accepted by ceasing to perform the shipbuilding contract, thereby also discharging Kwangju Bank.

In response, Vroon relies on Article X.3 of the contract by which the builder is to give the buyer at least fourteen days notice prior to the date the payment falls, and that no such notice was given until GEO sent an invoice on 12 October 2007. (Vroon also had a point which it pleaded in its reply to the effect that the buyer and GEO agreed to extend the time for payment under the shipbuilding contract to provide for a later payment date for the second instalment, but as I shall explain it was not its case at trial that there was a variation.) Kwangju Bank responds that the Article X.3 argument could only succeed if compliance with the article by GEO was a condition precedent to payment falling due, which it was not.

There is plainly an issue in this regard, which could in itself be one of considerable complexity on the construction of the shipbuilding contract, and on the facts. A financial institution issuing a demand guarantee is not required to engage with such a question. As has been said in the case law, the “on demand” nature of the instrument is intended to avoid it having to do so. Notwithstanding, my view on balance is that Vroon’s case is correct, and that it was not in default in paying the second instalment when it did. As Vroon points out, it is clear that GEO was telling Vroon that it would not have to pay the second instalment until further refund guarantees were procured (I have set out the history above), and it would not sit well with the facts to treat Vroon as being in default in that regard.

This issue is, of course, one that arises between the buyer and builder under the shipbuilding contract. The natural party to have raised the issue of premature termination (had it been a defence to raise) was GEO through arbitration provisions in the shipbuilding contracts. Had GEO wished to challenge the cancellation notices, then under the provisions of the guarantee that I have set out above, notice of arbitration would have suspended the claim under the guarantee unless and until Vroon’s claim had been upheld in the arbitration. It was submitted in closing on behalf of Kwangju Bank that GEO ought to have taken the point, and its failure to do so is its own look out. In fact, I infer that Kwangju Bank itself required confirmation from GEO that it was not going to refer the matter to arbitration before it made payment to Vroon (which it did on 5 March 2010). By an agreement of 23 February 2010, GEO specifically confirmed that it was not going to contest Vroon’s termination, and had no objection to Kwangju Bank making payment. This is hardly surprising. As Mr Bong-Jung Lee explained when asked about this agreement in cross-examination, by then the workers had gone, and the project had stopped. He said that he had been advised by Kwangju Bank not to pursue arbitration. In its opening submissions, SGIC stated that it no longer took issue with the cancellation of the shipbuilding contracts. In my view, it was right to abandon this defence.

The variation defence

Both defendants argue that they are discharged from liability on the basis that the buyers and GEO varied the shipbuilding contract in relation to GMS-103 with regard to the timing of the second instalment due under that contract. (This is a distinct point from the more general case advanced by the defendants that the shipbuilding contracts were varied so as to provide for “discrete” rather than “blanket” refund guarantees). The defence is advanced even if the guarantees in question were demand guarantees creating primary liability (which I have held they are and, as I have explained, SGIC accepts they are).

Although both defendants took this point, it was Kwangju Bank that presented submissions in support. It arises as follows. Kwangju Bank’s case is that there was no variation as regards the timing of the second instalment in relation to GMS-103. As I have said, it relies on default in the form of late payment and consequent extension of the delivery date to mount its premature termination defence. However, in its reply as against Kwangju Bank, Vroon pleaded that it and GEO “agreed to extend the time for payment of the second instalment …”. Kwangju Bank submitted that Vroon thereby asserted a variation of the shipbuilding contract to extend the time for payment of the second instalment. It submitted that, if (contrary to its primary submission) such was the case, then it was discharged as guarantor under the principle that a surety is discharged by a material variation to the contract to which he has not consented.

However, Vroon had made its position clear before the trial started. In its written opening submissions it stated that in practice what would happen is that GEO would raise an invoice and Vroon would pay that invoice. Payment of the second instalment was not made until an invoice was raised on 12 October 2007 without complaint from GEO. Specifically, it stated: “This delay did not involve any variation of the GMS-103 Contract. As a matter of analysis, either the Builder was precluded from complaining about late payment when it resulted from its own breach in failing to give proper notice pursuant to Art. X.3, or it was granting a short forbearance by not raising an invoice. But there was no agreement between the Builder and C in this regard.” (The same passage appears in Vroon’s written closing submissions.) In the circumstances, when the position became contentious during Kwangju Bank’s oral closing, I gave Vroon permission to amend the reply to reflect its opening submissions. Both defendants submit that the same defence arises in the case of forbearance as in the case of variation.

I have set out in detail above the factual position between the parties at the material time, and need not repeat it here. Payment of the second instalment under the contract in respect of GMS-103 was changed from 15 April 2007 to 30 June 2007 by Addendum No 1 made on 9 February 2007. However, an invoice was not provided by GEO until 12 October 2007, and the second instalment was paid on 29 October 2007. Plainly, the date in Addendum 1 was not adhered to, but over the summer of 2007, Vroon and GEO both thought that further refund guarantees would have to be obtained to cover the second instalments under the various contracts. It was only when Mr Marks looked into the position that he saw that the value of the existing guarantees increased automatically and told GEO in his email of 17 October 2007. At that point, he authorised payment. I accept Vroon’s factual case as set out in its submissions, and am satisfied (as indeed is Kwangju Bank’s own primary case) that there was no variation of the contract. At most, there was a short forbearance in the sense that GEO did not raise an invoice until 12 October 2007. I am also not convinced that even if there was such a variation (or forbearance) it was a material one in the sense of prejudicing the defendants. Although Kwangju Bank is right to say that the delay in payment would adversely affect GEO’s cashflow, this has to be seen in its overall commercial context. The price was payable over time and in instalments, of which this was only the second. The principal effect so far as the defendants were concerned was, as Vroon points out, to put back the date on which their exposure pursuant to the refund guarantees automatically increased. If that is so, the defendants’ guarantees would not be discharged even if (contrary to my conclusion) they created secondary liability only.

There is however a more fundamental reason why a defence based on the variation of, or forbearance in respect of, the shipbuilding contract does not provide the defendants with a defence on their guarantees. In equity, a guarantor is relieved of liability in the case of a material variation of the contract which could prejudice him and to which he has not consented. This is because the liability of debtor and surety is, in principle, co-extensive. Holme v Brunskill (1878) 3 QBD 495 contains a classic statement of the rule. At p.505, Cotton LJ cites Lord Loughborough in Rees v Berrington 2 Ves. J. 540 to the effect that: “It is the clearest and most evident equity not to carry on any transaction without the knowledge of him [the surety], who must necessarily have a concern in every transaction with the principal debtor. You cannot keep him bound and transact his affairs (for they are as much his as your own) without consulting him”. (For a recent example see Hackney Empire Ltd v. Aviva Insurance UK Ltd [2011] EWHC 2378 (TCC)). As with variation, so with forbearance in the sense of a binding agreement by the creditor to give time to the debtor (see Chitty on Contracts, ibid, at 44-094). But in the case of a demand guarantee, the guarantor has no such concern, because his liability under the instrument is autonomous of the underlying contract.

The point goes slightly further, because as Chitty puts it at 44-099, the effect of these rules is so technical and inconvenient that in practice any well-drawn contract of suretyship will expressly permit variation of obligations or the giving of time without discharging the surety. As explained above, the absence of such clauses has been held to be one of the indicia pointing towards a demand guarantee (there were none in the guarantees issued by Kwangju Bank or SGIC). It would be an odd result if the material variation/forbearance principles potentially operated to invalidate a demand guarantee, where exclusions are omitted precisely because they are not required.

I do not accept Kwangju Bank’s submission that Marubeni Hong Kong and South China Ltd v Government of Mongolia [2004] 2 Lloyd’s Rep 198, is authority in support of its submission. In that case, the instrument in question was not a demand guarantee. At para [142], Cresswell J said that it did not follow that the principles underlying Holme v Brunskill had no application if the defendant did undertake a primary liability—but he was considering the scope of the primary liability rather than questions of variation. On appeal, it is clear that the Court of Appeal proceeded on the basis that Holme v Brunskill only applied if the contract was one of secondary suretyship (see [2005] 1 WLR 2497 at [9], Carnwath LJ). In my opinion, the rules as to discharge of a surety on the basis of material variation or forbearance have no application to demand guarantees (see for example Andrews & Millett, The Law of Guarantees, 4^th edn, at paragraph 16-001 on p.526). Any other finding would be contrary to principle and create much commercial uncertainty. Accordingly, I reject the defendants’ case in this regard.

The non-disclosure defence

In North Shore Ventures Ltd v. Anstead Holdings Inc [2011] 2 Lloyd’s Rep. 45 at [14], the Chancellor, Sir Andrew Morritt, summarised the case law as regards disclosure in the case of guarantees as follows:

“(1)

the creditor is obliged to disclose to the surety any contract or other dealing between creditor and debtor so as to change the position of the debtor from what the surety might naturally expect, but (2) the creditor is not obliged to disclose to the surety other matters relating to the debtor which might be material for the surety to know. This is consistent with the fact that a contract of guarantee is not ordinarily a contract uberrimae fidei, such as insurance, whereunder the insured is required to disclose all facts material to the risk …”.

The discussion is in the context of the line of authority in which the guarantor’s liability was secondary in nature.

Both defendants rely on these principles, arguing that they are discharged from liability on the basis that Vroon failed to disclose various loans made to GEO by the buyers over the course of the contractual relationship between them. Both argue that this is so whether their guarantees were primary or secondary (though SGIC took a different position in its opening submissions, accepting that “if the liabilities under the RGs, properly construed, are primary liabilities then non-disclosure and material variation(s) are not relevant”). I have held that the guarantees issued by Kwangju Bank and SGIC were demand guarantees imposing primary liability.

There are three loans at issue. The first was made by agreement between Vroon and GEO on 5 March 2007 relating to GMS-103 in the sum of USD 785,500 (see paragraph 22 above). The second was a loan of USD 2,000,000 by Vroon to GEO on 16 August 2007 (see paragraph 73 above). The third were loan agreements entered into on 24 July 2008 by which Vroon loaned GEO the sum of USD 2,125,000.00 in respect of GMS-102 and GMS-103 (see paragraph 99 above). The last two are relevant as regards SGIC only. Whilst Vroon is right (in my view) to say that the insistence of both defendants that they did not know about these loans was not entirely convincing, their evidence in this respect was effectively unchallenged, and I accept it.

As is pointed out in The Law of Guarantees, ibid, 4^th edn, at paragraph 5-015, the jurisprudential basis of the creditor’s duty of disclosure to the surety is somewhat unclear. In some cases, the duty has been said to arise from the presumed basis of the guarantee, in others, on the basis that asking the surety to enter into the contract involves a representation that there is no unusual feature about the transaction affecting the risks that he is being asked to undertake. The authors conclude that the principle is an independent one, though its precise boundaries have yet to be defined. For Kwangju Bank, Mr Henry Byam-Cook invokes a similar implied representation, arguing that it applies equally as regards demand guarantees. However, as I have explained, a demand guarantee issued by a financial institution for a fee is very different from the traditional contract of suretyship. There is nothing inherent in the relationship that places a duty of disclosure on the beneficiary, and there seems no reason to put an onus on the beneficiary in this regard. The defendants accept that there is no authority in support of their contention. In my view, the disclosure obligation owed by a creditor to a surety has no application to the beneficiary of a demand guarantee as regards the issuer of the guarantee.

In any case, I am not convinced that these loans would fall within the disclosure obligations which apply as regards a surety. The defendants were well aware that GEO was a start-up operation, and must have had at least a general understanding of the close relationship which of necessity existed with Vroon, as GEO’s first customer, in relation to the vessels that Vroon had ordered. In these circumstances, I am not convinced that the making of the loans changed the position of GEO from what the financial institutions “might naturally expect”. I am also sceptical as to SGIC’s assertion that if any of the loans made or procured by Vroon to GEO had been disclosed to it, it would not have issued any refund guarantees, or that any issued would have been on materially different terms. However, its evidence in this respect was not challenged (Vroon was not in a position to do so).

Kwangju Bank’s Statute of Frauds defence

Section 4 of the Statute of Frauds 1677 (An Act for prevention of Frauds and Perjuries) provides that, “No action shall be brought … whereby to charge the Defendant upon any special promise to answer for the debt default or miscarriages of another person unless the Agreement upon which such Action shall be brought or some Memorandum or Note thereof shall be in Writing and signed by the party to be charged therewith or some other person there unto by him lawfully authorised”. Kwangju Bank submits that it has a defence on this ground. This only arises if the guarantee is of a secondary nature, since the statute does not apply to indemnities or demand guarantees. Since I have found that Kwangju Bank’s guarantee was a demand guarantee, it follows that this defence has no application. Should, contrary to that finding, the guarantee be secondary in nature, the result in my view is as follows.

I have set out above how Kwangju Bank’s guarantee was issued. It was sent by Kwangju Bank’s international department in Seoul by SWIFT message to the buyers’ bank, which was ABN AMRO Bank, Amsterdam. SWIFT is the provider of secure financial messaging services to banks and other institutions internationally. I do not think that it is in dispute that the method employed by Kwangju Bank was entirely conventional, and enabled ABN AMRO Bank to be sure that the guarantee in fact came from the bank that purported to issue it.

It is accepted on behalf of Kwangju Bank (to quote its closing submissions) that the word “signed” in the Statute of Frauds “does not necessarily involve signature by an individual using an ink pen and that it suffices that the guarantor’s name is written or printed in the document”. As it happens, in the body of the guarantee (which I have set out in paragraph 54 above) the words “Kwangju Bank” do not appear. The bank is referred to as “we”. The defence is to the effect that it is thereby unsigned, and the bank is not bound.

Such a conclusion would, in my view, be a surprising one. The bank accepts that its guarantee was properly issued, fully authorised and intended to be relied upon by the beneficiary. It was sent by conventional means by way of the secure messaging system used between banks. As a matter of common sense, authentication by sending was equivalent (in modern terms) to authentication by signing, and so within the spirit, if not the letter of section 4 of the Statute of Frauds.

In fact however, the conclusion is in my view erroneous, because the words “Kwangju Bank Ltd” are contained in the header to the SWIFT message. It is said on behalf of Kwangju Bank that this is not text which it typed in, but an output message header, that is, text generated by the SWIFT messaging system. That may be correct, but the name appears, and in my opinion it is a sufficient signature for the purposes of the Statute of Frauds. The words “Kwangju Bank Ltd” appear in the header, because the bank caused them to be there by sending the message. They were “voluntarily affixed” in the words of the old cases (c.f. J Pereira Fernandes SA v. Mehta [2006] 1 WLR 1543 dealing with email addresses). Whether or not automatically generated by the system, and whether or not stated in whole, or abbreviated (in fact the name of the bank appeared here in complete form), this is in my judgment a sufficient signature for the purposes of the Statute of Frauds. The position is analogous to that considered by Christopher Clarke J in Golden Ocean Group Ltd v. Salgaocar Mining Industries Pvt Ltd [2011] EWHC 56 (Comm) who at [103] observed that “an e-mail, the text of which begins “Paul/Peter”, may be regarded as signed by Peter because by that form of wording Peter signifies that he is addressing Paul and authenticates the content of the whole of what follows”. Therefore, I reject Kwangju Bank’s submissions in this regard.

The rectification and estoppel defences

The rectification and estoppel defences are advanced by both defendants in similar terms, and raise similar factual issues. Broadly, the defendants say that, if their refund guarantees increase automatically on the payment of instalments (which SGIC, but not Kwangju Bank, accepts they do, and as I have held they do) the instruments do not properly reflect the parties’ agreement, intentions and assumptions—by parties, I include GEO as builder, though the guarantees are contracts entered into between the financial institutions and Vroon as buyer. The defendants submit that the guarantees fall to be rectified so that the defendants’ liability is limited to the amount already paid under the guarantees. Further or alternatively, they say that Vroon is estopped by convention (SGIC says by acquiescence also) and/or by representation from contending that their liability is more than the amount already paid under the guarantees. This defence is linked to a contention that the shipbuilding contracts were varied to the effect that the guarantees provided by GEO to the buyers would be limited to single instalments of the advance payments, rather than covering them all as the contracts had envisaged. So that the history can be seen as a whole, I have set out my detailed findings in the factual section of this judgment, and shall avoid repetition here. In considering the facts, I have had the rectification and estoppel defences particularly in mind. In setting out my conclusions in relation to the defences, I shall take Kwangju Bank’s case first because the bank’s guarantees came first in time.

There was no dispute at trial as to the test to be applied. A party seeking rectification must produce convincing proof that the document to be rectified was not in accordance with the parties’ true intentions at the time of its execution, and that the document in its proposed form does accord with their intentions (see Chitty on Contracts, ibid, paragraph 5-122). Such common intention must be objectively manifested, rather than only being what the parties subjectively thought, and continue up to the time of execution of the document of which rectification is sought (see e.g. The Aktor [2008] 2 Lloyd’s Rep. 246 at [38] - [40], Christopher Clarke J). In Chartbrook Ltd v. Persimmon Homes Ltd [2009] 1 AC 1101 at [48], Lord Hoffmann approved the requirements for rectification as summarised by Peter Gibson LJ in Swainland Builders Ltd v Freehold Properties Ltd [2002] 2 EGLR 71, paragraph 33: “The party seeking rectification must show that: (1) the parties had a common continuing intention, whether or not amounting to an agreement, in respect of a particular matter in the instrument to be rectified; (2) there was an outward expression of accord; (3) the intention continued at the time of the execution of the instrument sought to be rectified; (4) by mistake, the instrument did not reflect that common intention.”

Kwangu Bank’s primary case is that on its true construction, its refund guarantee was limited to US$1,571,000 (plus interest), a case that I have rejected. Alternatively it submits that the refund guarantee it issued should be rectified so that it does provide that Kwangju Bank’s liability is limited to US$1,571,000 (plus interest). This can be achieved, it is said, by adding the words “(but not our liability under it)” after the words “The amount of this guarantee” at the beginning of the third paragraph, or by deleting the third paragraph entirely. It is Kwangju Bank’s case that as at late March and up to 18 April 2007 Kwangju Bank and Vroon had a continuing common intention that Kwangju Bank’s refund guarantee was to cover only the 1^st instalment of the price of GMS-103 and that subsequent instalments would be covered by separate refund guarantees. It submits that the relevant “cut off date” for these purposes is 18 April 2007 because that was the day on which Vroon first saw the refund guarantee (when GEO sent it to them).

In his oral closing, Mr Henry Byam-Cook, counsel for Kwangju Bank, identified four headings in support of the bank’s case on common intention. The points are equally relevant so far as its case in estoppel is concerned. The first was Mr Marks’ oral evidence, combined with the documents (particularly emails). The second was that the second instalment in respect of GMS-102 was contractually due on 15 April 2007, but not in fact paid until 29 October 2007. The third related to the description of the refund guarantee which Vroon itself used in an email. The fourth is the variation of the ship building contract which the defendants allege that Vroon and GEO agreed between them to the effect that “discrete” rather than “blanket” refund guarantees would be provided.

I shall begin with the fourth heading, namely the alleged agreement between GEO and Vroon to amend the ship building contracts. I have dealt with the facts and my findings in this regard in paragraphs 24 to 47 of this judgment. I have rejected the contention that the shipbuilding contract was varied in this way, and need not repeat the reasons.

As regards the second instalment under GMS-102, Kwangju Bank’s case is that the instalment was not paid any earlier because GEO was to provide further refund guarantees. Mr Marks’ evidence was that had an invoice been raised at that time, then the terms of the guarantee would have been checked, and the invoice processed and paid. Whilst there is force in the defendants’ point that their explanation is a more simple one, I do not accept that Mr Marks’ explanation was “simply ludicrous”. That fails to take account of the fact that Mr Marks did, in due course, authorise payment of the second instalment prior to receiving further refund guarantees. In his email of 17 October 2007 (which I have analysed in detail in the factual section of this judgment), he stated his reason, namely that having looked into the matter, he realised that the amount of the guarantees increased automatically. So I cannot give the weight to this point that the defendants urge.

The third heading relates to an email which shows that the Kwangju Bank’s refund guarantee for GMS-103 was saved in a PDF document named “RG 1st Instalment origineel GMS 103”. I do not think that the title of the PDF document takes matters very far. As Vroon pointed out, a description of the guarantee as being for the first (or indeed any particular) instalment is not in itself inconsistent with a term of the guarantee which provides for subsequent “automatic increase”.

As regards the first heading, I have rejected the suggestion that it was Mr Marks’ intention from about 19 March 2007 that Kwangju Bank’s refund guarantee was to cover only the first instalment, and that subsequent instalments would be covered by separate guarantees. I need not repeat the reasons. It is not in dispute that that was the understanding of Kwangju Bank, but such understanding was based on what it had been told by GEO, which was not (as I have found) justified by what Vroon had told GEO.

There is a further obstacle to the rectification defence. When Kwangju Bank issued its first guarantee in respect of GMS-102 on 30 March 2007, it included the figure of USD 1,571,000 (being the amount of the first instalment) both in the second and in the third paragraph of the guarantee—the latter containing the “automatic increase” provision. Had this figure remained in the third paragraph of the guarantee, then this would have been a strong indication that despite the reference to “automatic increase”, the intention was that the issuer’s liability was capped. But as I have explained, Mr Waterman in Vroon’s treasury department spotted this, and rejected the guarantee. He pointed out that once the second and subsequent instalments were paid, the amount guaranteed would need to increase. Kwangju Bank did not actually see the email exchanges, but it was asked by GEO to correct the guarantee. It duly issued an amendment on 2 April 2007 substituting the figure of USD 12,568.000 in the third paragraph which was the amount of all five instalments prior to the delivery instalment for the vessel. This was then replicated a few weeks later when Kwangju Bank issued its guarantee in respect of GMS-103 on 17 April 2007.

Kwangju Bank seeks to meet this objection by pointing out that Mr Marks was the individual at Vroon with conduct of the transaction. That is correct, but does not affect the fact that in this respect, Mr Waterman intervened on Vroon’s behalf. In its response to him, GEO apologised for sending “incorrect document [which] I have sent it without checking because of late night”. GEO said it would contact the bank and resubmit the guarantee, which it duly did. Kwangju Bank objects that it did not understand that the effect of the amendment was that its liability under the refund guarantee was being increased. I accept that that was its understanding, but it was an erroneous understanding. The rectification defence amounts to a withdrawal of the amendment, and a substitution of the guarantee as first issued. I conclude that the refund guarantee issued on 17 April 2007 was not drawn up in a manner contrary to the parties’ continuing common intention, nor were they mistaken in concluding it in the terms which they did. I accordingly reject Kwangju Bank’s rectification defence.

Much of the above applies also in the case of SGIC, and again I will not repeat it. SGIC’s basic contention is that a mistake has been made in the drafting of the refund guarantees. Once that is accepted, it submits, all falls into place.

As a preliminary point, SGIC places particular emphasis on the submission that the shipbuilding contracts had been varied post-execution to provide for more than one refund guarantee per vessel, and more than one guarantor, a submission which I have rejected as set out above in the factual section of the judgment. It does however further submit that in order to obtain rectification of the guarantees, it is not necessary to rectify each shipbuilding contract as well because (to quote from its closing submissions): “Firstly the common understanding that preceded each RG was in respect of the RG to be provided. Each are autonomous instruments. … each is … a separate contract between Vroon and SGIC. Providing those two parties had the same common understanding up until issue and acceptance of each RG, then rectification is available. The fact that this common understanding is also shared by another party is not strictly relevant, but may, as here, provide the reason for A and B, between whom rectification is sought, sharing the same understanding”. As a matter of principle, I do not think that this approach is in dispute.

In opening, it was indicated on behalf of SGIC that its guarantees should be rectified by deleting the words in the third paragraph, “shall automatically extend”, and substituting USD 1,571,000 for USD 12,568,000, the figure in the guarantees as actually issued. The thrust of SGIC’s argument, which as Mr Cogley QC said was a simple one, is as follows. SGIC issued ten guarantees in all in relation to the four vessels. Its clear understanding was that each guarantee related to a single instalment as regards the vessel in question. In its case, emphasis is placed upon the events over the summer of 2007. As it was put it in SGIC’s closing submissions, these objectively show that by October 2007 (if not earlier), there had been a variation to the ship building contracts. Its case is that in consequence, Mr Marks’ email of 17 October 2007 came too late. The variation was not susceptible of a unilateral revocation. In any case, there was no communication of the 17 October email to SGIC. All this explains why SGIC continued to issue guarantees. Accordingly, the guarantees (or at the least those subsequent to the 25 May 2007 guarantees) fall to be rectified. This is in essence also its case as regards estoppel, though it is submitted that this without question catches the first guarantees, since even if the estoppel arose later, it related to all the guarantees, including the earliest ones. Counsel recognised at trial that the rectification and estoppel defences are different ways of framing the same fundamental submission.

As the factual account set out above shows, in issuing the guarantees SGIC adopted the template that had already been established by Kwangju Bank. Its understanding derived from what it had been told by GEO. It is not in dispute that, so far as material, Vroon had no contact with SGIC.

So far as events up to 18 April 2007 (the cut off date in respect of Kwangju Bank) are concerned, these are covered by my findings above. I do not think that it is suggested that anything further was said or done by Vroon during the period up to 25 May 2007, when SGIC issued its first two guarantees, which materially affects the argument as to rectification (or estoppel). In any case, I am satisfied that it did not. There is, I am satisfied therefore, no basis for finding that any common intention as asserted by SGIC had been manifested by 25 May 2007. This rules out a case in rectification as regards SGIC’s first two guarantees (which were in respect of GMS-104 and GMS-105).

Further, I have found that there was no variation of the ship building contracts by October 2007, or at all, and I reject SGIC’s case so far as it is based on such a variation. I have however found that over the period June to October 2007, both GEO and Vroon were working on the assumption that further refund guarantees would need to be provided before the second instalments on the contracts fell due. SGIC points out that there was no communication to it of Mr Marks’ email of 17 October 2007. Thus, it submits, it is entitled to rely on the common intention as continuing, whether or not the shipbuilding contracts were varied, and whatever the position between Vroon and GEO.

This point was forcefully put, and requires careful consideration. However, though it is correct that there was no communication to SGIC of Mr Marks’ email of 17 October 2007, there was no communication at all between SGIC and Vroon at any material time. Effectively, SGIC’s case is that the common continuing intention as between GEO and Vroon is sufficient to found the common continuing intention between itself and Vroon which is necessary so as to found a claim for rectification of the refund guarantees. On that basis, my view is that SGIC must abide by the consequences of the fact that such common continuing intention as there was between GEO and Vroon came to an end on 17 October 2007. On that date, Mr Marks made it clear that the “automatic increase” language of the guarantees meant that further instalments were already covered. He went ahead to authorise payment of the second instalments. These were paid prior to (albeit just prior to) the issue of the next four SGIC guarantees. There was clearly no common continuing intention at that time.

It is correct, as SGIC says, that it went on to issue four more guarantees (this includes those issued in relation to GMS-102 which were returned when Vroon took delivery of that vessel in an unfinished state). It invites the court to find that the intention to treat the guarantees as discrete guarantees continued until at least after issue of the last one, which was on 17 October 2008 in the sum of US$3,142,000 in respect of GMS-104. I have set out the facts in this regard above. So far as those relating to the vessels the subject of this claim are concerned, a further guarantee for GMS-103 was issued on 30 May 2008, but Vroon had already paid the 3^rd instalment on 20 May 2008. A further guarantee for GMS-104 was issued on 17 October 2008, but Vroon had already paid the 3^rd instalment on 21 August 2008. The correlation between guarantees and instalments is therefore not made out. Further, instalments continued to be paid by Vroon thereafter without any further issuance. Mr Mark’s evidence was that Vroon felt secure that the payments made were covered under the existing refund guarantees. In my view, the evidence from GEO fell considerably short of establishing any such understanding or agreement to the contrary (see paragraphs 101 and 102 above).

A final point that again applies in the case of SGIC (but not Kwangju Bank) relates to the way the third paragraph of the guarantees was filled in. In several cases, SGIC altered the wording that refers to “the second, third, fourth and fifth instalments”. This was not correctly or consistently done, and shows, so SGIC submitted, that something had clearly gone wrong with the language of the instruments. In my view, this does not advance matters. It goes to show that SGIC thought that it was issuing its guarantees in relation to individual instalments, but that is not in dispute. Accordingly, I reject SGIC’s case in rectification.

As an alternative to rectification, Kwangju Bank asked the court to find that Vroon is estopped (by convention or by representation) from contending that its liability under the refund guarantee is greater than US$1,571,000 plus interest. Kwangju Bank relied in support of this defence on the same facts as have been extensively analysed above in the context of rectification. Understandably, it dealt with this defence briefly in its closing.

It submits that from late March up to 18 April 2007, it shared an assumption with Vroon that its refund guarantee was to cover only the 1^st instalment of the price of GMS-103. This assumption passed between the parties via GEO, which only asked Kwangju Bank to issue the refund guarantee after its exchanges with Vroon established that both GEO and Vroon shared an assumption that there would be multiple refund guarantees. It would now be unconscionable, it is submitted, to allow Vroon to go back on that assumption because Kwangju Bank acted on it by only taking security from GEO on the basis that the refund guarantee covered the 1^st instalment and nothing more.

Given my findings, there was no such shared assumption, and an estoppel by convention does not arise. The amendment of the guarantee issued on 30 March 2007 to increase the figure in the third paragraph is equally an obstacle as in the claim for rectification. Further, I reject the contention that the communications and conduct from 19 March up to 18 April 2007 and Vroon’s acceptance of the refund guarantee on 18 April 2007 amounted to a representation that the refund guarantee would only cover the 1^st instalment of the price of GMS-103. In my view, Vroon (which was the beneficiary) made no representation when accepting the refund guarantee. In particular, it did not by accepting the guarantee make any representation that it would not rely on its terms.

The estoppel defence does not require the instruments to be rectified, instead preventing reliance on their terms. SGIC submits that Vroon is estopped by convention (SGIC says by acquiescence also) and/or by representation from contending that its liability is more than the amount already paid under the guarantees. As already pointed out, counsel accepted that the rectification and estoppel defences are in essence different ways of framing the same fundamental submission. I have set out SGIC’s case in this regard under rectification, and the same factual issues arise.

As to estoppel by convention/acquiescence, there was no dispute as to the legal test. In Republic of India v India Steamship Co [1998] AC 878 , 913–4 (The “Indian Endurance”) Lord Steyn said:

“It is settled that an estoppel by convention may arise where parties to a transaction act on an assumed state of facts or law, the assumption being either shared by them both or made by one and acquiesced in by the other. The effect of an estoppel by convention is to preclude a party from denying the assumed facts or law if it would be unjust to allow him to go back on the assumption: The August Leonhardt [1985] 2 Lloyd's Rep. 28; The Vistafjord [1988] 2 Lloyd's Rep. 343; Treitel, Law of Contracts, 9th ed., at 112–113. It is not enough that each of the two parties acts on an assumption not communicated to the other. But it was rightly accepted by counsel for both parties that a concluded agreement is not a requirement for an estoppel by convention.

…

That brings me to estoppel by acquiescence. The parties were agreed that the test for the existence of this kind of estoppel is to be found in the dissenting speech of Lord Wilberforce in Moorgate Mercantile Co. Ltd. v. Twitchings [1977] A.C. 890. Lord Wilberforce said that the question is “… whether, having regard to the situation in which the relevant transaction occurred, as known to both parties, a reasonable man, in the position of the ‘acquirer’ of the property, would expect the ‘owner’ acting honestly and responsibly, if he claimed any title to the property, to take steps to make that claim known… ” at 903. Making due allowance for the proprietary context in which Lord Wilberforce spoke, the observation is helpful as indicating the general principle underlying estoppel by acquiescence.”

In ING Bank NV v. Ros Roca SA [2011] EWCA Civ 353 at [59], Carnwath LJ referred to Spencer Bower, Estoppel by Representation (4^th edn 2004) para VII.2.3 to the effect that the only distinction between the two forms of estoppel lies in the manner in which the party to be estopped assumes responsibility for the proposition from which he is to be estopped from departing, namely by mutual assent rather than unilateral assertion. He said at [66] that under Lord Steyn's formulation the issue can be reduced to two questions:

“i)

Was there a relevant assumption of fact or law, either shared by the two parties, or made by Ros Roca and acquiesced in by ING?

ii)

If so, would it be unjust (or “unconscionable”) to allow ING to go back on the assumption?”

SGIC placed considerable reliance on the Ros Roca case. The claim in that case was in respect of a fee for financing services provided by ING Bank as adviser to Rosa Roca as its customer. This turned on the construction of a particular clause, it being held that the bank’s construction was correct, and a fee which was agreed to be based on “EV/EBITDA 06” could not be construed as referring to “EV/EBITDA 07”. However the parties’ shared assumption was that the bank’s fee would not be charged on the basis of “EV/EBITDA 06”. The bank had a contractual obligation to Ros Roca to share information and to develop a genuine and reasonable estimate of transaction costs (see Rix LJ at [88]). In fact, it deliberately chose to keep quiet ([95]). In the circumstances, it was held that it would be unjust or unconscionable to allow the bank to go back on the shared assumption and the bank was estopped from doing so ([71])

The facts of the present case are far removed from those in Ros Roca. Vroon plainly did not have advisory duties as regards SGIC, the financial institution that issued the guarantees. Indeed, it had no relationship with SGIC at all, other than to receive the guarantees, and claim on them should occasion arise. It was simply the beneficiary of guarantees that SGIC issued. SGIC rightly accepts that these were demand instruments the value of which (by their terms) increased automatically as the instalments were paid. I can see the force of the point that, in the language of estoppel, SGIC suffered “detriment” in that it secured itself on the basis that the guarantees’ value was in fact limited by reference to a particular instalment. On the other hand, SGIC—Seoul Guarantee Insurance Company—is in the business of issuing such instruments. As Mr Stephen Hofmeyr QC for Vroon pointed out, the fact that the company appears to have taken no steps to ascertain the meaning of the instruments, and consequently misunderstood them, is not, on the face of it, promising ground to mount an estoppel argument against the beneficiary seeking to enforce the guarantees according to their terms.

The question however is whether SGIC can make good its case on the facts. On one aspect of the approach to be taken, I think that it is correct. Vroon appeared to argue that because it had no relationship with SGIC then, even if there was a shared assumption as between Vroon and GEO, such assumption did not “cross the line” between them (The “August Leonhardt” [1985] 2 Lloyds’ Rep 28 at 34), with the effect that the assumption was not “communicated” to Vroon as required by the case law. In principle, it is correct to say that it is not enough for the purposes of an estoppel by convention that each of two parties acts on an assumption not communicated to the other (see The “Indian Endurance” at p.913F). In the present case however, the relationship with Vroon as buyers, and with the financial institutions as guarantors, was exclusively with GEO as builder. If buyer, builder and guarantor proceeded on the shared assumption that (despite the terms) each guarantee was limited to a single instalment, and it is unconscionable for the buyer now to claim otherwise (which is the defendants’ case), I do not see it as a bar that the buyer and guarantor did not communicate directly with each other. On the other hand, as I have indicated already, they cannot take an inconsistent position on this point, and treat some aspects of GEO’s understanding as “crossing the line”, but not others.

I accept that SGIC (and Kwangju Bank) believed that the guarantees they were issuing were limited to individual instalments. This was one of the misunderstandings that have given rise to this litigation. As I say above (paragraph 89), GEO assumed that all parties were proceeding on the basis that there would be separate guarantees for each instalment, but it did not in fact get Vroon’s assent to this. The bank and SGIC took it on trust that their liability on each was limited to a single instalment, without seemingly focusing on the language of the instruments that they were issuing, because that is what GEO told them. For the reasons set out in detail earlier in this judgment, I am satisfied that SGIC’s defence does not succeed, whether framed as an estoppel by convention or an estoppel by acquiescence. There was no common assumption as to the limited effect of the guarantees at the time SGIC issued its first guarantees on 25 May 2007. The email exchanges over the period June to October 2007 show both GEO and Vroon working on the assumption that further refund guarantees would need to be provided before the second instalments on the contracts fell due. However in October 2007, Mr Marks reviewed the position, realised that the amount of the guarantees increased automatically, and authorised payment of the second instalments. He made this clear to GEO, and thereby put an end to any common assumption there may have been. After that, it was not unjust or unconscionable for Vroon to rely on the terms of the guarantees as issued.

On the face of it, SGIC’s contention appears to be supported by the fact that it continued to issue further guarantees. But on careful examination, that point loses much of its force, because the sequence was not in fact one of guarantee followed by payment. As explained above, even where a further guarantee was provided, the instalment payment had already been made by Vroon before it was received (sometimes just before, sometimes a considerable period before). Furthermore, after October 2008 there were no further guarantees, but Vroon continued to pay instalments. This is not consistent with a common assumption that each guarantee was limited to a particular instalment.

SGIC nevertheless maintains that Mr Marks has lied to the court in his evidence. For reasons set out above, I reject that submission, and have generally accepted his evidence as truthful. In particular, I accept his explanation of the circumstances in which the email of 17 October 2007 came to be written. I accept that he felt secure that the later instalments were covered under the existing refund guarantees: the evidence does not establish any understanding or agreement to the contrary (see paragraphs 101 and 102 above). It is in respect of these later instalments that Vroon brings its claim, and in my view, it is not estopped from doing so. So far as it has been advanced separately, that conclusion includes the argument based on estoppel by representation. In my view, Vroon made no representation when accepting the refund guarantees. In particular, it did not by accepting the guarantees make any representation that it would not rely on their terms.

The counterclaim

The defendants have a counterclaim on the basis that if any of the “guarantee” defences succeed, namely premature termination (Kwangju Bank only), variation, nondisclosure and/or the Statute of Frauds (Kwangju Bank only), it follows that the payments which the defendants have made under their guarantees following the claimants’ demands are recoverable, as having been paid under a mistake of fact and law, or on the basis of a total failure of consideration. Since I have rejected each of these defences, the counterclaim does not arise.

I should add that, even on the premise that any of these defences was sound, I am sceptical as to the defendants’ case that the payments made were made by mistake in the legal sense. After demand was made by Vroon, it is to be inferred that the defendants carefully considered their position, and determined to make payments on the basis of their understanding as to the limited effect of the guarantees, namely that they extended to individual instalments only. The payments must have been made regardless of any uncertainties there may have been as to legal liability. The financial institutions simply took a view as to their liability. This is particularly clear in the case of SGIC, which made a payment on 7 September 2010 in respect of the GMS-105 guarantee. This was the day after its defence in the GMS-103 action was signed on its behalf. The issues at this point in time had generally been flagged up in the pleadings. Be that as it may, the counterclaim does not arise.

For the reasons set out above, the claimants succeed on their claims, and the counterclaims are dismissed. I am grateful to the parties for their assistance, and will hear them on any consequential matters arising.