Vossloh Aktiengesellschaft v Alpha Trains (UK) Ltd.

These proceedings are concerned with the true construction of an agreement in writing executed as a deed and dated 15 September 2009. It is between the claimant, Vossloh Aktiengesellschaft (“VAG”), and Angel Trains International Limited. The latter has since undergone a change of name to Alpha Trains (UK) Limited (“Alpha”). The agreement describes itself as a “Guarantee”. I shall refer to it as the “2009 Guarantee”. VAG is described in it as the “Guarantor”, and Alpha and its affiliates are collectively referred to as the “Angel Trains Group”. Members from time to time of that group “and their respective agents, assigns, directors, employees, officers, secondees and servants” are each described as a “Beneficiary” with “Beneficiaries” construed accordingly.

The issue which I am asked to decide is the basis on which, under the 2009 Guarantee, VAG can be required to make payment to a Beneficiary.

It is the contention of VAG, which has appeared by Geraldine Andrews QC and James Willan, that its liability under the 2009 Guarantee is triggered upon proof (whether by admission or by decision of a competent court of law) of a breach of contract by any one or more of those entities referred to in it as the “Guaranteed Party”. That expression is defined as meaning two named companies in the Vossloh Group “and each other member from time to time” of that group.

It is the contention of Alpha, which is sued in its own right and as representative defendant on behalf of all the Beneficiaries under the 2009 Guarantee and has appeared by Stephanie Barwise QC and Jennifer Jones, that VAG’s liability under it is triggered by a demand alone. By that is meant (as counsel described it in their skeleton argument) that the obligation undertaken by VAG is in the nature of an “on demand performance guarantee” in that it constitutes an unconditional independent promise to pay on demand all amounts demanded.

VAG was prompted to bring these proceedings with a view to determining the basis of its liability under the 2009 Guarantee following the receipt by it of a letter of demand to VAG claiming in excess of €17 million. That letter was replaced by a revised demand dated 22 February 2010 in the sum of €17,267,354. VAG contends that none, or at the most only a small part, of that sum is payable under 2009 Guarantee.

VAG is the parent company of the Vossloh Group of companies which are leaders in the rail infrastructure and technology market. One of its subsidiaries is Vossloh Locomotives GmbH (“VL”) which manufactures and supplies locomotives. It previously had another name.

Alpha, under its former name Angel Trains International Ltd, was a member of the Angel Trains Group which was formerly owned by the Royal Bank of Scotland Group plc (“RBS”). On completion of the sale of the Angel Trains Group by RBS in August 2008, its UK and international operations were separated and now function under separate ownership and separate management. From December 2009 the international businesses began using the Alpha Trains name.

The commercial dealings between the Alpha Group (as it is now called) and the Vossloh Group relevant to these proceedings date back to 2000. On 13 September 2000 a company in the (then) Angel Trains Group entered into a co-operation agreement with VAG to facilitate a mechanism for the sale of locomotives by Vossloh Group companies to Angel Trains Group companies and to create an operations and maintenance services facility for the locomotives. This led to the establishment that same year of two companies owned, albeit in unequal shares, by the two Groups. One of those two companies was called Locomotive Capital Ltd (“LCL”), 10% of whose shares were held by the VAG Group and the remaining 90% by the Angel Trains Group. LCL in its turn had three wholly-owned subsidiaries. One of them, referred to for short as LCUKL, was the purchasing party under a Master Purchase Agreement (the “MPA”) (also known as the “German MPA”) dated 23 April 2001 and made with VL. Another of the subsidiaries was the purchasing party under other purchase agreements including, in particular, two agreements referred to in evidence as the “German Agreements” and another which was referred to as the “2001 Agreement”.

The MPA acted as a framework agreement pursuant to which LCUKL could purchase locomotives from VL from time to time by issuing “Call Off Notices” (“CONs”). The German Agreements and the 2001 Agreement operated in a similar manner.

The other jointly owned company was set up to provide the maintenance and services operation. This company was owned 10% by the Angel Trains Group and the remaining 90% by the Vossloh Group. A representative of the minority shareholder was appointed to the board of each of the two jointly owned companies.

From at least 2002 onwards, locomotives were purchased from VL by various members of the Angel Trains Group, issuing CONs under the MPA. The purchasers included group affiliates in Belgium, Germany, Switzerland and Luxembourg. The same happened under the German Agreements and the 2001 Agreement.

Prior to the sale of the Angel Trains Group in the summer of 2008, RBS, the ultimate parent of the Group, used its own money to finance the purchase of locomotives pursuant to the MPA. The Group purchased locomotives on a speculative basis before any third party customers had committed to acquiring them. Payment was in stages prior to the delivery of the locomotives and before therefore any income could be generated from them. By the end of 2002, the Group was committed to the spending of €95 million on purchasing locomotives from VL.

In around late 2002, RBS, in the light of the level of its financial outlay and exposure, made it a condition of the grant of its approval to any further locomotive purchases from VL that VAG should provide a parent company guarantee of VL’s obligations under the MPA. The first such guarantee was entered into on 5 March 2003. It was given in favour of LCUKL alone. This guarantee was replaced by a guarantee dated 15 September 2006 which extended the guarantee to the obligations of a VL affiliate in respect of an order for the purchase of locomotives in Spain. The 2006 guarantee also enabled other companies within the Angel Trains Group to benefit from the guarantee and to make claims under it.

In consequence of the sale and restructuring of the Alpha/Angel Trains Group in the summer of 2008, the 2006 guarantee was replaced on 15 September 2009 by the further guarantee, namely the 2009 Guarantee which is the subject of these proceedings.

Alpha has complained that sixty-three G1206 cargo locomotives supplied to them by VL suffer from various defects in their engines and gearboxes. Fifty-two of the locomotives were supplied under CONs issued pursuant to the MPA. The remaining eleven were supplied under the German Agreements and the 2001 Agreement.

A formal notice of dispute under clause 21 of the MPA was given by Alpha on 12 September 2008. This was followed by a pre-action protocol letter of claim dated 29 January 2010, which also related solely to the locomotives supplied pursuant to the MPA. It claimed losses in excess of €14 million. On the same date, Alpha sent a letter of demand to VAG under the 2009 Guarantee claiming losses in excess of €17 million in respect of all sixty-three locomotives. This, as I have mentioned, was replaced by a revised demand dated 22 February 2010 in the sum of €17,267,354, which again related to all sixty-three locomotives. The bulk of that sum is made up of what Alpha estimates to be costs of repair.

Alpha accepts that it has not yet expended any money on repair, or incurred a liability to pay anything approaching the amount demanded. It also admits that if it is paid the €17 million and expends it, it cannot afford to pay the money back to VAG. Alpha says that it has purchased six new engines for its fleet at a cost of €1.8 million because the bulk of the fleet is undergoing maintenance or repair, and that it has compensated customers in the total sum of €687,554 for poor performance of the locomotives or for their withdrawal from service for longer than anticipated. Those two figures total €2,487,544.

VAG does not accept that Alpha would be successful in recovering either of these heads of damage from VL or VAG even if it were to succeed in proving liability. Proceedings in respect of the fifty-two locomotives supplied pursuant to the MPA have been issued against VL by LCUKL and other members of the Alpha Group in the Technology and Construction Court. Both liability and quantum are in dispute. There will be detailed technical issues (involving expert evidence) on both sides. There has been an order for service of the particulars of claim in those proceedings by 6 August 2010.

Before coming to the 2009 Guarantee I set out my understanding of the relevant law, starting with some general observations. In the summary that follows I have drawn from the section headed “In general” in chapter 44 (on suretyship) in volume 2 of Chitty on Contracts, 30^th edition, and from chapter 1 of Andrews and Millett on the Law of Guarantees, 5^th edition. The propositions which I set out accord with my experience of the law in this area and, in any event, are extensively supported by authority in the footnotes to both books. I am mindful of the fact that Ms Andrews is the co-author of the second of those textbooks.

Contracts of suretyship, of which the 2009 Guarantee is an example, are an area of law bedevilled by imprecise terminology and where therefore it is important not to confuse the label given by the parties to the surety’s obligation (although the label may be indicative of what the parties intend) with the substance of that obligation. Because the parties are free to make any agreement they like, each case must depend upon the true construction of the actual words in which the surety’s obligation is expressed. This involves “construing the instrument in its factual and contractual context having regard to its commercial purpose”, a task which the court approaches “by looking at it as a whole without any preconception as to what it is.” See Tuckey LJ in Gold Coast Ltd v Caja de Ahorros Del Mediterraneo [2002] EWCA Civ 1806, [2002] 1 Lloyd’s Rep 231 at [10] and [15]. Further, as Ms Andrews observed, the court must endeavour to avoid a construction which renders a clause otiose or duplicative.

A contract of suretyship is in essence a contract by which one person, the surety, agrees to answer for some existing or future liability of another, the principal (or principal debtor), to a third party, the creditor, and by which the surety’s liability is in addition to, and not in substitution for, the liability of the principal. Even the use of the expressions “creditor” and “debtor” (as in “principal debtor”) can be misleading: the liability which is “guaranteed” may consist of the performance of some obligation other than the payment of a debt, and it does not have to be a contractual liability.

Contracts of suretyship fall into two main categories: contracts of guarantee and contracts of indemnity. Because they have many similar characteristics, and similar rights and duties arise between the parties, it is not unusual to find the term “guarantee” used loosely to describe what is in reality an indemnity.

A contract of guarantee, in the true sense, is a contract whereby the surety (the guarantor) promises the creditor to be responsible for the due performance by the principal of his existing or future obligations to the creditor if the principal fails to perform them or any of them. Depending on its true construction, the obligation undertaken by the surety may be no more than to discharge a liability, for example a particular debt, if the principal does not discharge it so that if for any reason the principal ceases to be liable to pay that debt (it may have been discharged and replaced by some other debt or liability) the surety will not come under any liability to the creditor. The surety’s liability in such a case is conditional upon the principal’s failure to pay the particular debt so that if the condition is fulfilled the surety’s liability will sound in debt. In contrast to that is the more usual case (sometimes referred to as a “see to it” guarantee) where, on the true construction of the contract, the surety undertakes that the principal will carry out his contract and will answer for his default. In such a case, if for any reason the principal fails to act as required by his contract he not only breaks his own contract, but he also puts the surety in breach of his contract with the creditor, thereby entitling the creditor to sue the surety, not for the unpaid debt, but for damages. The damages are for the loss suffered by the creditor due to the principal having failed to do what the surety undertook that he would do. See Moschi v Lep Air Services Ltd. [1973] AC 331 at 344 to 345 (Lord Reid).

An essential distinguishing feature of a true contract of guarantee – but not its only one - is that the liability of the surety (i.e. the guarantor) is always ancillary, or secondary, to that of the principal, who remains primarily liable to the creditor. There is no liability on the guarantor unless and until the principal has failed to perform his obligation. The guarantor is generally only liable to the same extent that the principal is liable to the creditor. This has the consequence that there is usually no liability on the part of the guarantor if the underlying obligation is void or unenforceable, or if the obligation ceases to exist (to which principle – the so-called principle of co-extensiveness - there are, however, a number of exceptions). It will depend upon the terms of the contract of suretyship whether a demand must be made on the principal or on the guarantor (or on both) in order to trigger the guarantor’s obligation to pay. Many modern guarantees expressly negative the need for the creditor to make a demand on the principal or on the guarantor or to take any other given step before enforcing the guarantee.

In contrast to the contract of guarantee is the contract of indemnity. In one sense all contacts of guarantee (strictly so called) are contracts of indemnity (as indeed are many contracts of insurance) since, in its widest sense, an indemnity is an obligation imposed by operation of law or by agreement of the parties. In the narrower sense in which, in the current context, the expression occurs, a contract of indemnity denotes a contract where the person who gives the indemnity undertakes his indemnity obligation by way of security for the performance of an obligation by another. Its essential distinguishing feature is that, unlike a contract of guarantee (strictly so called), a primary liability falls upon the giver of the indemnity. Unless (as is quite possible) he has undertaken his liability jointly with the principal, his liability is wholly independent of any liability which may arise as between the principal and the creditor. It will usually be implicit in such an arrangement that as between the principal and the giver of the indemnity, the principal is to be primarily liable, so that if the indemnifier has to pay first he has a right of recourse against the principal. (It will not be so if, for example, the indemnifier has not undertaken his indemnity obligation at the request of the principal.) It is this feature which leads to the person giving the indemnity to be described as a “surety” although, strictly, the contract of indemnity cannot itself be a contract of suretyship.

The fact that the obligation to indemnify is primary and independent has the effect that the principle of co-extensiveness does not apply to a contract of indemnity. The indemnity not only shifts the burden of the principal’s insolvency on to the indemnifier but it also safeguards the creditor against the possibility that his underlying transaction with the principal is void or unenforceable. It also prevents the discharge of the principal or any variation or compromise of the creditor’s claims against the principal from necessarily affecting the liability of the indemnifier under his contract with the creditor. Otherwise, the rights and duties of the parties to a contract of indemnity are generally the same as those of the parties to a contract of guarantee.

So much for some of the essential differences. Whether a particular contract of suretyship is of the one kind or the other or, indeed, a combination of the two turns on its true construction. A contract which contains a provision preserving liability in circumstances where a guarantor would otherwise be discharged (for example, the granting of time by the creditor to the principal or a material variation of the underlying contract between the principal and the creditor, without (in either case) the guarantor’s consent) will usually indicate that the contract is one of guarantee because such a provision would be unnecessary if the contract were one of indemnity. On the other hand, a provision stating that the surety is to be liable in circumstances where the principal has ceased to be liable (for example, on the principal’s release by the creditor) may be indicative either of a guarantee (because the provision would be unnecessary in the case of a contract of indemnity) or of an indemnity (because it makes clear that the liability of the surety was intended to continue regardless of the liability of the principal). See, for example, Clement v Clement, (unreported, Court of Appeal, 20 October 1995). Context is important in deciding what the nature is of the obligation under consideration as even minor variations in language, plus a different context, can produce different results. See IIG Capital LLC v Van Der Merwe [2008] EWCA Civ 542, [2008] 2 Lloyd’s Rep 187 (“IIG”) at [20] (Waller LJ). But if, in a contract of guarantee (strictly so called), the parties are minded to exclude any one or more of the normal incidents of suretyship “clear and unambiguous language must be used to displace the normal legal consequence of the contract…” See IIG at [19] (Waller LJ).

This brings me to the so-called “performance bond”, sometimes known as a “performance guarantee”, often as a “demand bond” or “demand guarantee” or even as a “first demand guarantee”. In the context of the present dispute I prefer the expression “demand bond”. In essence it is a particularly stringent contract of indemnity. It is a contractual undertaking by a person, usually a bank, to pay a specified amount of money to a third party on the occurrence of a stated event, usually the non-fulfilment of a contractual obligation by the principal to that third party. Sometimes the wording of the contract has the result that the liability of the person who has given the bond arises on mere demand by the creditor, notwithstanding that it may be evident that the principal is not in any way in default or even that the creditor himself is in default under his contract with the principal. It all depends on the wording of the instrument. It is often a difficult question to determine whether, on its true construction, a particular contract which provides for payment on demand is a performance or demand bond (where the obligation to pay is triggered by a demand alone or by a demand accompanied by the provision of specified documents) or whether it is a guarantee (strictly so called) where the obligation to pay is of the “see to it” kind, i.e. conditional on proof by the creditor of default by the principal.

A good example of the need to distinguish between the two is provided by the recent decision of the Court of Appeal in Marubeni Hong Kong v Government of Mongolia [2005] EWCA Civ 395, [2005] 2 Lloyd’s Rep 255. In that case the claimant sold goods to a Mongolian company for a price payable by instalments. The agreement provided for a document described as a “guarantee” to be issued in a specified form by the Mongolian Central Bank on behalf of the defendant government. In the event a letter in a modified form (“the MMOF Letter”) signed on behalf of the Mongolian Ministry of Finance was provided. It stated that in consideration of the sale agreement “the undersigned Ministry of Finance unconditionally pledges to pay to you upon your simple demand all amounts payable under the Agreement if not paid when the same becomes due…and further pledges the full and timely performance by the Buyer of all the terms and conditions of the Agreement.” An accompanying legal opinion from the Deputy Minister of Justice of Mongolia referred to the “guarantee” and expressed the opinion that the “guarantor” had full power to enter into the “guarantee”. The contract in the letter was stated to be governed by English law and that disputes were subject to the jurisdiction of the English court. Goods were supplied under the sale agreement. Various payment rescheduling agreements were later entered into. Following default by the purchaser of the goods to make payments due, one of the questions which arose was whether the obligation of the Mongolian Ministry of Finance had been discharged by the variations to the sale agreement since, it was contended, this had occurred without its consent. The contention assumed that the MMOF Letter constituted a guarantee (strictly so called) and not, as was contended by the claimant seller, a performance bond. Cresswell J held that it was a guarantee and that the defendant government was discharged by the variation to the sale agreement. The government appealed. The appeal failed.

In the Court of Appeal Carnwath LJ gave the leading judgment with which the other two members of the court agreed. He felt able (at [23]) to distinguish previous authorities, where the court had found there to be a performance or demand bond, on the basis that performance or demand bonds, however described, “are a specialised form of irrevocable instrument developed by the banking world for its commercial customers” which the courts had accepted as equivalent to irrevocable letters of credit. “It cannot be assumed” he continued “that cases relating to such banking instruments provide any useful guidance when construing guarantees given outside the banking context.” He stated ([at 28]) that such cases “provide no useful analogy for interpreting a document which was not issued by a bank and which contains no overt indication of an indication to create a performance bond or anything analogous to it.” He pointed out that the MMOF Letter was not a banking instrument and that it was not described, either on its face or in the supporting legal opinion, in terms appropriate to a demand bond or something having similar legal effect. Accepting that the terminology used in the transaction was not conclusive he stated nevertheless (at [30]) that, in a transaction outside the banking context, the absence of such language, created “a very strong presumption” against the existence of such an obligation. He went on (at [31]) to consider whether there were sufficient indications in the wording of the MMOF Letter, to displace the presumption. Notwithstanding the words “unconditionally pledges” and “simple demand” in the words of obligation in the Letter, he observed that the obligation was only to arise if “the amounts payable under the agreement (are) not paid when the same becomes due”. He said that there was no reason not to give those words their ordinary meaning and (at [32]) that the pledge of “…the full performance and observance by the buyer of all the terms and conditions of the agreement” reinforced that sense. He noted that the Letter contained a primary obligation in the form of an indemnity against cost or damage resulting from the buyer’s default but considered that it did not qualify the ordinary meaning of the earlier part of the Letter.

I should refer to one other authority to which my attention was drawn. This was IIG (referred to briefly above). It is an example of a transaction, outside the banking context, where the presumption against the existence of a demand guarantee (or the like) was successfully rebutted. Indeed, it was the only one of which Ms Andrews said that she was aware.

In that case the claimant, a New York entity, gave financial assistance to a company of which Mr and Mrs van Der Merwe were directors. Each of them entered into a document described as a “guarantee” which stated that the guarantor “as principal obligor and not merely as surety” agreed that “it will immediately upon demand unconditionally pay to the Lender the Guaranteed moneys which have not been…paid…” The document provided that “A certificate in writing signed by a duly authorised officer…stating the amount at any particular time due and payable by the Guarantor…shall, save for manifest error, be conclusive and binding on the Guarantor for the purposes hereof.” Following a demand on the company which went unpaid the claimant sent letters to the Van Der Merwes reciting the company’s failure to pay and certifying the amount due and payable by each of them under the guarantee. It demanded payment within two days. The Van Der Merwes did not pay. The question was whether, properly construed, the obligation which each of the Van Der Merwes had undertaken was a demand bond or whether it gave rise to guarantees strictly so called. The relevance of this was that under New York law, which governed the loan agreement, there was on the evidence a defence which would provide the borrower (the company of which the Van Der Merwes were directors) with a complete answer to the claimant’s demand on it and would therefore (on the principle of co-extensiveness assuming the obligation was no more than a guarantee strictly so called) provide the Van Der Merwes with a defence as well. The Master gave summary judgment to the claimant. The Van Der Merwes’ appeal to Lewison J was dismissed. A further appeal to the Court of Appeal was likewise unsuccessful.

In the Court of Appeal, Waller LJ (with whom the other two members of the court agreed) considered, in agreement with Lewison J, that the obligation was in the nature of a demand bond or guarantee under which the Van Der Merwes were primary debtors. Acknowledging that the issue turned on the true construction of the agreement and that, following what had been said in Marubeni, there was, outside the banking context, a strong presumption against instruments of this kind being demand bonds unless there was clear wording to the contrary, Waller LJ was of the view that the clear language of the operative clauses was effective to rebut the presumption. He referred (at [20]) to the context in which the transaction arose, namely “a company run by two shareholders borrowing money for a business over which those shareholders have complete control.” He pointed out (at [31]) that the guarantor (Mr or Mrs Van Der Merwe) had agreed “as principal obligor and not merely as surety” that he (or she) would “immediately upon demand and unconditionally pay…the Guaranteed moneys” and that guaranteed moneys were defined as “all moneys and liabilities…which are now or may at any time be due, owing or payable, or expressed to be due owing or payable, to the lender from or by the borrower…” (emphasis added). He also pointed to the clause stating that a certificate in due form stating the amount due was, except for manifest error, conclusive and binding, noted that there was no manifest error in the certificate served on the Van Der Merwes, and, in agreement with Lewison J, considered that the presence of that clause put the matter beyond doubt.

The result of the foregoing brief survey is that, with the parties free to agree whatever terms they choose, there is in this field of law a spectrum of contractual possibilities ranging from the classic contract of guarantee, properly so called, at the one end, where the liability of the guarantor is exclusively secondary and will be discharged if, for example, there is any material variation to the underlying contract between principal and creditor, to the performance or demand bond (or demand guarantee) at the other end, where liability in the giver of the bond may be triggered by mere demand and without proof of default by the principal (and indeed where it may be apparent that the principal is not in default). There may be little to distinguish (and it may not matter) whether the obligation undertaken is in the nature of a guarantee (strictly so called) or an indemnity. Where it does matter, the question is whether the liability to be enforced is secondary (or ancillary) to that of the principal (however qualified that liability may be), in which case the obligation is in the nature of a guarantee, or primary, in which case it will be in the nature of an indemnity and, if the latter, may be enforceable merely on demand (as with a performance or demand bond) or conditional on proof of default by the principal or on satisfaction of some other event or requirement. Where on the spectrum a particular case falls may call for a nice judgment on the part of the court faced with the task of construing the instrument in question. The instant case calls for just such a judgment.

The following are the material terms of the 2009 Guarantee:

As VAG is not a bank or in any way equivalent to a bank and the 2009 Guarantee was not given in a banking or like context, the jurisprudence discussed above raises a strong presumption that the payment obligations undertaken by VAG do not constitute a demand bond. What grounds then are there in the context and wording of the document for rebutting that presumption?

Dealing first with context, Ms Barwise pointed to the following factors: (1) express requirements in a memorandum of understanding entered into between VAG, LCUKL and Angel Train Contracts Ltd on 11 February 2000 and also in a co-operation agreement between Angel Trains Ltd (formerly Angel Train Contracts Ltd) and VAG on 13 September 2000 to act in good faith towards each other in the furtherance of their mutual business interests, (2) the fact that the two groups jointly owned LCL which in turn owned LCUKL and LCD (the locomotive purchasers under the MPA and the other agreements – see paragraph 8 above), thereby demonstrating the extent to which the groups were bound up in each other’s activities and shared a community of interest, (3) the nature of the MPA as a framework agreement with no limit either on the number of locomotives to be purchased under it or on its duration, a circumstance matched in the 2009 Guarantee which states (by clause 4) that it is to “extend to the ultimate balance of the Secured Obligations” and (by clause 5.1) that VAG may not terminate it in any way, and (4) Alpha’s requirement which it says it communicated to VAG (although VAG disputes this) for a form of security which could be realised promptly to protect it from exposure arising from non-delivery of or defective locomotives. Ms Barwise did not suggest that these factors in themselves showed an intention that the 2009 Guarantee should operate as a performance bond but submitted that they made it more likely that, if the wording of the 2009 Guarantee justified such a construction, I should be the more willing to conclude that that is indeed what it is.

Coming next to the wording of the 2009 Guarantee Ms Barwise submitted that, properly understood, clauses 2.1, 3.1, 6.3, 6.4, 7 and 11.2 show that the document constitutes a demand bond. The opening words of clause 2.1 stating that VAG’s contractual promises were entered into “as a principal debtor and not merely as a surety, as a separate, continuing and primary obligation…” show that all that follows in that clause are primary and not secondary obligations. Clause 3.1 states in terms that all that is payable under the instrument is to be “paid on demand”. Clause 6.3, whereby VAG waives any right to require a Beneficiary to proceed against any other person before claiming against VAG, emphasises the immediacy and primacy of VAG’s obligations. Likewise clause 6.4, stating in effect “pay now, argue later”. This would be unnecessary in a true guarantee since, under a guarantee, the guarantor is always entitled to raise defences available to the principal. Clause 7, when read with clause 6.4, shows that the Beneficiary is entitled to payment by VAG irrespective of whether it has made a demand or proceeded against a third party. This is reinforced by clause 11.2 stating that the Beneficiary’s certificate of the amount due is to be “conclusive evidence of such amount” save for manifest error.

Taking these indications in the wording with the contractual context, the 2009 Guarantee, she submitted, is correctly to be interpreted as a demand bond.

Ms Andrews submitted that the key clause defining VAG’s obligations is clause 2.1. Subparagraphs (a) and (b) of that clause are secondary obligations (and are expressed in the classic language of a guarantee); subparagraph (c) is probably also secondary in nature; subparagraphs (d) (almost certainly) and (e) (certainly) are primary obligations. But, whether or not the obligations are primary or secondary in nature, the clause assumes that there has been default by the principal (i.e. VL) in performing the underlying MPA or in making a payment that is contractually due under it. Clause 2 cannot therefore be construed as giving rise to any liability to pay against a mere assertion of breach or a failure to pay money.

Ms Andrews submitted that Ms Barwise could derive no assistance from the other clauses on which she relied. Clause 3.1 is no more than a standard form demand clause found in guarantees. It does not define the circumstances in which VAG’s liability arises. Instead, its reference to “all sums payable hereunder” points to the liability which arises, if at all, under clause 2.1. In a performance or demand bond, by contrast, the demand is critical because it is that which triggers liability. Clause 3 does not seek to do this. Clause 6 contains provisions typically found in guarantees. The express reference in clause 6.4 to the Guarantor having first to comply with clause 2.1 begs the question of what that compliance entails. So far from precluding VAG from raising any defence available to the Guaranteed Party, it merely defers the right to raise the defence until after compliance with clause 2; in so doing it preserves the co-extensiveness of VAG’s liability with that of the principal. That is the antithesis of a performance or demand bond. Clause 11.2 is a standard form “conclusive evidence” clause which entitles the Beneficiary referred to in the 2009 Guarantee to certify the amount of any Secured Obligation which has not been discharged by the Guaranteed Party. The certificate goes to quantum not to liability.

The first point to note is that under the 2009 Guarantee the liability of the “Guarantor” (as VAG is described) extends to what are termed “Secured Obligations”. As appears from the definition of that expression set out above these are not confined to monetary obligations but include “all present and future monies, liabilities and obligations (whether for payment of money or otherwise…)”. They may be owed by “any member of the Guarantor Group to a Beneficiary”. “Beneficiary” is defined as “each member of the Angel Trains Group from time to time and their respective agents, assigns, directors, employees, officers, secondees and servants”. The obligations must arise “under or in connection with any Relevant Document”. “Relevant Document” is defined to mean each of (1) the MPA (2) each so-called “German Call-Off Notice” (i.e. each contract for the acquisition of a locomotive or other item as contemplated by the MPA) (3) the “Spanish MPA” (i.e. a master purchase agreement dated 11 September 2006 for the acquisition of locomotives) (4) each so-called “Spanish Call-Off Notice” (i.e. each contract for the acquisition of a locomotive or other item as contemplated by the Spanish MPA) and (5) “any other agreement for the sale and purchase of locomotives and/or related items entered into between any member of the Angel Trains Group and any member of the Guarantor Group, from time to time”. The reach therefore of VAG’s obligations under the 2009 Guarantee is exceedingly extensive both in terms of those entitled to its benefit and the scope of the obligations caught by it.

The next point to note is that by clause 1.2 (j) references to obligations “guaranteed” under the 2009 Guarantee are to “include a reference to indemnified obligations”. The contrast is between obligations that are “guaranteed” by the instrument and those that are “indemnified” by it and suggests therefore that “guaranteed” as used in the instrument includes secondary liability so that it would be wrong to construe the 2009 Guarantee as confined to indemnified obligations.

Clause 2 which sets out VAG’s obligations (it is headed “Guarantee and Indemnity” and is plainly intended to state what those obligations are) is of critical importance. Sub-clauses (a) and (b) of Clause 2.1 are, as Ms Andrews submitted, framed in the classic language of guarantee (i.e. of a secondary obligation triggered upon proof of a breach of obligation by the principal). Sub-clauses (d) and (e) are worded as primary obligations. Sub-clause (c), although not prefaced by the expression “guarantee” (as sub-clauses (a) and (b) are) has the appearance of a secondary obligation: it is premised upon a default in payment by “a Guaranteed Party”. The obligation is to make good on demand any “Secured Obligations which have not been paid at the time such demand [i.e. demand of VAG] is made”.

I do not consider that the opening words of clause 2 are effective to convert into purely primary obligations obligations which are otherwise secondary in nature. The words used are certainly capable of having that purpose and effect. But, if that had been their purpose, it is difficult to see why the clause then goes to the trouble of setting out sub-clauses which, as worded, have the appearance of secondary obligations. The opening words are more consistent with an intention to set out in what follows a mixture of primary and secondary obligations. This is achieved, as a matter of interpretation, by reading the opening words as if a comma appears after the words “as a continuing obligation”. The sub-clauses that follow show all the signs of an intention to subject VAG to an obligation to answer for every kind of default that could arise under or in connection with what is referred to in the definition of Secured Obligations as “a Relevant Document”. The possibility of overlap and duplication does not matter provided that all possibilities are covered.

But, as Ms Andrews observed, the critical question is whether, even if all that follows is by way of primary obligation, the liability thereby assumed by VAG is triggered merely by demand (or, as she put it, by assertion) on the part of someone who qualifies as a Beneficiary or whether in addition it is necessary to demonstrate the existence of a breach or failure of obligation under or in respect of a Relevant Document. The wording of sub-clauses (a) and (b) are difficult to read except as obligations premised upon a failure of observance or performance of an underlying obligation (in the case of sub-clause (a)) or of punctual payment under an underlying obligation (in the case of sub-clause (b)) by a Guaranteed Party under a Relevant Document. It is not realistically possible, in my view, to read sub-clause (c) except as conditional on a failure of payment of an underlying debt by a Guaranteed Party and as intended therefore to subject VAG to an obligation in debt to make good that failure. Likewise sub-clause (e). It assumes that the Beneficiary has suffered loss “in connection with or as a direct or indirect result of the failure of a Guaranteed Party to duly and punctually perform…” a term etc contained in a Relevant Document “...or a failure “to duly and punctually pay the Secured Obligations…” VAG’s liability is premised upon a failure of performance of some underlying obligation. The intention is to make good the loss thereby suffered. It is premised upon the establishment of that loss by the Beneficiary.

Clause 3.1 does not define the circumstances in which VAG’s liability arises. Instead it assumes that there is a liability “hereunder” (i.e. under the 2009 Guarantee) to make a payment. The liability, if it exists, must arise from some other term of the instrument. I agree therefore with Ms Andrews that this clause does not lend support to the case that Ms Barwise seeks to make. But then neither does it detract from it.

That brings me to clause 6. Headed “Waiver of Defences” sub-clauses 6.1 to 6.3 contain provisions I would expect to find in a guarantee strictly so called. In any event, Ms Barwise did not suggest that they give support to the existence of a performance bond. To my mind, the “pay now, argue later” terms of sub-clause 6.4 point to the existence of a secondary rather than to a primary liability. The sub-clause assumes that VAG may raise defences which the Guaranteed Party could have raised if the demand had been addressed to it. Its purpose is simply to postpone the exercise of that right by VAG until after the demand has been fully met. If VAG’s liability were primary, let alone one triggered by a demand alone, the fact that the Guaranteed Party has or may have a defence to the demand would be quite immaterial. I do not therefore consider that Ms Barwise can find support in this clause for the case that she advances. Nor do I accept her submission that its presence throws light on a proper understanding of the other clauses on which she relies.

The same is true of clause 7 which is headed “Demands”. It does no more than state that more than one demand may be made and that VAG’s liability may be enforced whether or not the Beneficiary has pursued to any extent its rights against the Guaranteed Party or in respect of any security it holds. In particular, I do not accept Ms Barwise’s submission that, when taken with clause 6.4, clause 7 shows that, in order to establish liability in VAG, it is unnecessary for the Beneficiary to establish any underlying liability in the Guaranteed Party.

Clause 11.2 is a conclusive evidence provision of a kind commonly seen in guarantees. As worded it is well capable of extending to the damages payable for breach of a Secured Obligation and is not confined merely to liquidated sums. It also assumes that there is a breach of the Obligation. The question is whether its reference to the certificate constituting “conclusive evidence... against the Guarantor in the absence of any manifest error” goes to the existence of the breach (i.e. liability) and not just to its monetary consequences (i.e. quantum). Ms Barwise did not suggest that on its own the clause would have the effect of converting the document into a demand bond but submitted that, when taken with clauses 2.1, 3, 6.3, 6.4 and 7, it should be taken as having that result and, in effect, as extending to liability as well as to quantum. In my view it is confined to the latter. Its reference is to a certificate “setting forth the amount ” of the Secured Obligation in question. I agree with Ms Andrews that it does not have the effect (whether taken on its own or in combination with the other clauses to which Ms Barwise referred) of transforming this contract into a demand bond.

In Bache & Co (London) Ltd v Banque Vernes et Commerciale de Paris [1973] 2 Lloyd’s Rep 437 (mentioned in Van Der Merwe ) the conclusive evidence clause provided that the notice of default should be “as between [the claimants and the defendants] conclusive evidence that [the defendants’] liability hereunder has accrued in respect of the amount claimed.” The Court of Appeal in that case held that the clause was valid. In Van Der Merwe it was pointed out, in reference to a passage in The Modern Contract of Guarantee English Edition (2003) by O’Donovan and Phillips, that

Adopting that strict approach it seems reasonably clear that clause 11.2 is not to be construed as having the effect for which Ms Barwise contended. I do not need to rely on any ambiguity in its expression to reach this conclusion.

Viewing the instrument as a whole I do not consider that the presumption against construing it as a demand bond has been rebutted. Clause 2, which sets out what the obligations are which VAG has undertaken, is premised upon the establishment of a failure of performance (including a failure of payment) under a Relevant Document. None of the other provisions, viewing them individually or in combination, alter that conclusion.

That brings me back to context. I am not persuaded that any of the points urged by Ms Barwise (and summarised at paragraph 37 above) should lead to a different conclusion. Turning in particular to the second of Ms Barwise’s points, an acceptance that, at any rate until late 2004 (which is long prior to the 2009 Guarantee), there was a measure of community of interest does not make it more likely that VAG should be willing in 2009 to enter into a demand bond than if the two groups had at all times been wholly at arm’s length in their dealings. In any event, as Ms Andrews pointed out, there are several factors which point against VAG’s obligations under the 2009 Guarantee being in the nature of a demand bond. They include the fact that they are unlimited in amount and duration and can be the subject of multiple demands (VAG’s only escape from liability being to bring about a termination of the underlying purchase agreements), the fact that there is no express right of recourse (which might have been expected if this were indeed a demand bond) to enable VAG, having satisfied a demand, to effect any recovery from the person primarily liable, the fact that, as the very extensive definition of “Beneficiary” shows, a demand can come from such a wide range of persons and the fact that the benefit of the instrument is expressed (by clause 11.1) to be assignable to any person.

For completeness I should mention two further points made by Ms Barwise. The first, drawing on the width of definition of “Beneficiary” and the fact that the underlying purchase agreements (the so-called Relevant Documents) involve overseas entities, is that the instrument was intended, as she put it, “to be used internationally by Beneficiaries who might not understand the subtleties of English law in relation to guarantees/indemnities” and that those underlying purchase agreements might well be subject to civil law jurisdictions and thus to good faith obligations, with references to “demand” and the like being read literally. I am unable to give any weight to this argument. The instrument is expressed by clause 12.1 to be governed by and construed in accordance with English law. (So also, as it happens, is the MPA which is the principal underlying purchase agreement). The parties must therefore take English law as they find it. It cannot be right to give the instrument a meaning different from what English law fairly requires merely because one of the parties affected by it has been accustomed to a different system of law where the approach to contractual interpretation may be different.

The other point raised by Ms Barwise concerned the fact that the instrument was preceded by two earlier instruments, namely the guarantee dated 15 September 2006, which itself superseded the earlier guarantee dated 5 March 2003 (see paragraph 13 above). Drawing on the evidence before the court that there was no substantial renegotiation of terms after the first of the guarantees was executed, she addressed submissions on the nature and effect of that guarantee, pointing out that, by clause 2.1(b), VAG undertook to make payment to LCUKL “upon first written demand by Locomotion Capital [i.e. LCUKL] without further proof or condition” and that by clause 13 it was agreed that “A certificate of Locomotion Capital setting forth the amount of any Guaranteed Amounts not then paid by the Contractor shall be conclusive evidence of such amount against the Guarantor in the absence of any manifest error.” Accepting that those words do not appear in that precise form in the 2009 Guarantee but pointing out that clause 2.1 of that instrument contains provisions similar to what were clauses 2.1 and 3.1 of the 2003 guarantee she submitted that the 2003 guarantee gave rise to a demand bond and that, as there had been no renegotiation of its terms before execution of the later instruments, the 2003 guarantee shows what the parties were intending to achieve by the 2009 Guarantee. As I understood it this was with a view to approaching the 2009 Guarantee with a greater willingness to construe it as a demand bond.

But, as Ms Andrews submitted in reply, there were changes in wording in the later instruments, not least in the scope between the 2003 guarantee and the later instruments. The plain fact is, she said, that they are not the same.

In my judgment, it is not a permissible approach to the construction of an agreement to draw comfort from, let alone be guided by, an earlier and different agreement which the later agreement has replaced, any more than it is permissible to do so by reference to earlier drafts of the later agreement. The terms of the earlier instrument might be material to rectification of the later instrument, but I am not having to deal with any such question.

I shall make the declaration sought.

Alpha contended that, if I should reach the conclusion that I have concerning the nature of the 2009 Guarantee, there was evidence that VL was in breach of its obligations under the MPA in that, as Mr Althen, the legal director of Alpha Trains (UK) Limited, put it in paragraph 39 of his first witness statement, “the beneficiaries to the 2009 Guarantee have to date incurred the sum of €2,487,544 (broken down as follows: €1.8 million for the purchase of new engines and €687,554 arising from losses incurred by locomotive operators) which is recoverable under clause 6.1 of the MPA…” with the consequence that VAG is obliged to indemnify the beneficiaries in that amount under clause 2.1(c) of the 2009 Guarantee in any event. (See paragraph 17 above.) Judgment in that sum was sought or, at the least, a declaration that VAG is liable under the 2009 Guarantee to pay that sum. I understood from this that the Beneficiaries (as they are described in the 2009 Guarantee) had incurred those amounts as a result of defects in the locomotives supplied under the MPA, in part by replacing defective locomotives and in part as a consequence of the defects in those locomotives.

I feel a difficulty about acceding to this alternative claim. As mentioned at paragraph 18 above LCUKL and other members of the Alpha Trains Group have issued proceedings against VL (and possibly others – I am unaware of the precise details) in the Technology and Construction Court for relief, including damages, under the MPA. The proceedings arise out of the same matters that form the basis of the alternative claim before me. Liability and quantum are in dispute in those proceedings. Subject to any further argument that counsel may wish to address on the point (the matter was only lightly touched on in argument before me) the better course would be to see what happens in those proceedings. If liability in VL (or in some other member of the Guarantor Group as defined by the 2009 Guarantee) is established (whether by decision of the court or otherwise) then, as it seems to me, VAG’s liability under the 2009 Guarantee to effect an indemnity against that liability will arise (subject only to demand). As I understand matters, however, that point has not yet been reached.