Barclays Bank Plc v Unicredit Bank AG & Anor

This is a dispute about whether the Claimant (“Barclays”) exercised its discretion in a commercially reasonable manner in refusing to consent to the early termination of three synthetic securitisations of loan portfolios entered into in the wake of the financial crisis in autumn 2008.

The First Defendant (“HVB”) is a German bank. The Second Defendant (“Bank Austria”) is an Austrian bank. Since 2005 each has been a wholly owned direct subsidiary of UniCredit S.p.A., an Italian company based in Rome and Milan which is part of the UniCredit banking group. Save where it is necessary to distinguish between the individual companies, I shall refer to the Defendants, as the parties did, as “UniCredit”.

The securitisations were embodied in three deeds of guarantee as follows:

UniCredit sought consent for early termination of the Guarantees in June 2010 as a result of a change in their regulatory treatment, which adversely affected UniCredit’s ability to obtain the capital relief which the Guarantees were designed to achieve. Barclays determined that it would not consent unless it were paid the balance of five years’ fees, amounting to some €82 million (discounted to then present value). UniCredit was not prepared to pay this or any sum, and contends that Barclays’ insistence on five years’ fees was not a commercially reasonable ground for declining consent.

UniCredit S.p.A. is regulated in Italy by the Bank of Italy. HVB is regulated in Germany by the Bundesanstalt für Finanzdienstleistungsaufsicht (“BaFin”). Bank of Austria is regulated in Austria by the Österreichische Finanzmarktaufsicht (Financial Market Authority “FMA”).

HVB and BA were at all material times subject to the regulatory oversight of their local financial regulator because both entities acted as fully licensed banks in their respective countries. At the consolidated level for the UniCredit group, the Bank of Italy also had regulatory oversight in relation to the capital position of the group. At the time the Guarantees were entered into, Bank of Italy had a policy of following the local regulator.

A key function of banking regulators is to ensure that the banks under their supervision maintain adequate capital reserves. The Basel Committee on Banking Supervision, which is a committee of banking supervisory authorities established by the central bank governors of ten countries, has attempted to achieve a measure of international consistency in relation to minimum capital requirements for banks. The so-called "Basel Accords" were issued by the Basel Committee in 1988 ("Basel I") and 2004 ("Basel II"), and were implemented in Germany and Austria.

The nature of a bank's business is that its assets will consist mainly of debts of one sort or another which are owed to the bank by third parties. In order to assess the amount of capital reserves which must be held against each asset, these assets are assigned a percentage risk weight which reflects their credit risk. The composition of the Risk Weighted Assets ("RWA") in turn dictates the amount of capital (tier one and tier two) which a bank must hold in respect of them. Tier one capital, the more important of the two, consists largely of shareholders' equity and disclosed reserves. Tier two capital, comprises undisclosed reserves, revaluation reserves, general provisions, hybrid instruments and subordinated term debt.

Securitisation can be used to transfer or mitigate the credit risk on a pool of assets, and in turn reduce the regulatory capital requirements of the bank in relation to such pool. In a "synthetic securitisation", the credit risk of a pool of assets is transferred to a third party through the use of credit default swaps ("CDS") or guarantees, without actually removing the portfolio of assets from the balance sheet of the bank selling the credit risk. The CDS or guarantees are tranched: the credit risk of the pool is divided up into junior, mezzanine and senior tranches which are commonly rated by rating agencies. If effective risk transfer, as the regulators define it, has been achieved by transferring the risk on some or all of the tranche, the bank can hold less capital against the risk of it suffering losses on that pool of assets. How much risk has to be transferred to achieve effective risk transfer is precisely defined by the applicable regulatory rules for synthetic securitisation. The resulting regulatory capital relief is known as "RWA relief”. In Germany, an effective risk transfer leading to RWA relief needed to be approved by the regulator, either before entering into the deal, or afterwards on the basis of an information pack sent to the regulator. HVB's practice for the previous 10 years had been to inform BaFin and seek regulatory approval for all regulatory capital securitisations before concluding the transactions.

There are different methods used by regulators to assess the amount of tier one capital reserve that must be held against the securitised tranches which are retained by the originator of such a securitisation. One common approach is for the risk weight to be determined by external credit ratings. Another approach, which at the relevant time was little used and could only be adopted by a small number of approved banks, is known as the Supervisory Formula Approach ("SFA"). Basel II allows for two main methodologies for calculating RWA: the "standardised" approach and the Internal-Rating Based ("IRB") approach. In 2008, HVB in Germany and Bank Austria in Austria were already using the IRB method, whereas UniCredit Banca SpA and the other Italian banks which were part of the UniCredit group were still using the standardised method. The regulatory relief afforded in relation to a securitisation applying the SFA was only available to so-called Advanced IRB banks (which both HVB and BA were), and then only on portfolios which met certain criteria.

To determine a risk weight on each tranche, the SFA applies a formula which takes account of five inputs, which are: (1) the number of loans in the underlying pool taking into account their individual sizes, (2) the regulatory capital which the bank would have had to maintain against the pool without synthetic securitisation, (3) the pool's exposure weighted average loss given default number (i.e. the expected loss against a particular asset which would be suffered if there was to be a default), (4) where the securitised tranche sits in the structure (junior, mezzanine, senior) and (5) the size of the tranche. The output of the formula is an RWA percentage for the tranche for which it has been computed. The formula enables the bank flexibly to divide the pool into different tranches, computing the regulatory capital requirement each tranche must bear. To manage the capital charge against the whole pool the bank can transfer the risk in certain tranches to a counterparty, and thereby gain regulatory capital relief on the pool, if effective risk transfer is achieved according to the regulatory rules.

These rules for synthetic balance sheet deals were particularly well developed by BaFin for German banks due to the frequent usage of synthetic securitisation in Germany. Bank Austria was the largest Austrian bank, and as far as Dr Spaeth knew, it was the only bank in Austria conducting synthetic securitisations of the type contained in the Guarantees. In this respect, the FMA, and the auditors who approved the regulatory treatment, were content to follow BaFin's lead on the regulatory interpretation of Basel II, whose implementation in Austria was in substantially the same terms as in Germany.

All this was well known to Dr Spaeth, and her colleagues in the team at HVB who were involved in concluding the Guarantees. Those at Barclays involved in negotiating the Guarantees would have been aware of these general principles of regulatory capital relief, but were not familiar with the detail of the SFA formula, nor the particular attitudes and approach of the Austrian and German regulators to effective risk transfer. Nor were they aware of the UniCredit group’s overall regulatory capital position.

The immediate context for the HVB-1 Guarantee was that UniCredit as a consolidated group wished to improve its tier one capital ratio, and was under considerable pressure from the Bank of Italy to do so before the quarter date of 30 September 2008. The initiative to execute the HVB-1 Guarantee came from the Italian Head Office at UniCredit in Milan.

Negotiations for the HVB-1 Guarantee started on 16 September 2008 and were concluded within a fortnight. At the same time Dr Spaeth was conducting negotiations for synthetic securitisations of two other pools of assets, which were successfully concluded with other counterparties by 30 September 2008. These negotiations took place during considerable turmoil in the financial markets. A few of the headline events which were shaking the markets in September 2008 were the collapse of Fanny Mae and Freddy Mac; the rescue of Merrill Lynch by Bank of America; the bail out of AIG by the Federal Reserve providing a $85 billion credit facility; the rescue of HBOS by a government encouraged merger with Lloyds Bank; Washington Mutual filing for Chapter 11 bankruptcy; and the collapse of Wachovia and its proposed sale to Citigroup. On 29 September 2008, HVB’s real estate lending arm, Hypo Real Estate, needed a 35 billion euro bail out from the German Government and a consortium of banks. Many banks other than UniCredit were facing similar regulatory capital pressures. The cash securitisation market was effectively closed. Interbank lending was severely restricted.

The position in the markets was not significantly better in December 2008 when the BA and HVB-2 Guarantees were concluded. Their negotiation proceeded on the basis that the principles agreed for the HVB-1 Guarantee should continue to apply. Apart from the amounts involved, they were concluded on similar terms, with one material difference which I will identify below.

The Guarantees were amended and restated on 2 April 2009. Nothing turns on the amendments and I shall refer to the Guarantees in their restated form unless the context otherwise requires.

The essential structure of the Guarantees was that in exchange for quarterly payments of premium and a fixed fee, Barclays would make quarterly payments to UniCredit in respect of the first losses suffered by reason of credit defaults on a portfolio of obligations which had been designated by UniCredit for each of the transactions as its “Reference Portfolio”. The losses were referred to in the Guarantees as Credit Protection Payments. The notional values of the Reference Portfolios, when initially designated, were €9,965,235,219.18 for HVB-1; €6,663,757,405.85 for the BA Guarantee; and €3,982,760,904.00 for HVB-2. The first loss tranche covered by the Credit Protection Payments for the HVB-1 Guarantee was a little over 7%; Barclays’ liability for the first loss tranche was limited to €700 million. The first loss tranche was of a different width for the later transactions; the equivalent limits for the BA and HVB-2 Guarantees were €600 million and €420 million respectively.

The amount of the premiums was determined in such a way as to ensure that, over the lifetime of the Guarantees, the total premium would exceed the total Credit Protection Payments, so that Barclays was not exposed to the credit risk on the first loss tranche of the Reference Portfolio. The Guarantees potentially lasted for the period of the longest loan in the Reference Portfolio, which was 11, 19 and 19 years respectively. Over this period the premiums would be bound to exceed the Credit Protection Payments. Barclays did not, on the other hand, stand to gain from the performance of the first loss tranche of the Reference Portfolio, because any surplus premium was to be returned to HVB or Bank Austria. In the HVB-1 Guarantee the surplus was to be returned at the termination of the Guarantees; in the BA and HVB-2 Guarantees there was a mechanism for return of surplus annually during the currency of the Guarantees, with adjustment to ensure the same neutral effect.

In addition to premium, Barclays was entitled to a fee which was not subject to any adjustment. For the HVB-1 Guarantee, the fee was €3 million per quarter (€12 million per annum). For the BA and HVB-2 Guarantees it was €2 million and €1.2 million and respectively (€8 million and €4.8 million per annum).

The quarterly payment dates for premium, fees and Credit Protection Payments were 30 March, 30 June, 30 September and 30 December each year. The amounts due each way would be calculated, with the balance falling to be paid at the end of each quarter on a netted off basis. The calculation was to be based on an Accumulation Ledger, to be prepared by UniCredit, which recorded the premium and Credit Protection Payments. In essence the Accumulation Ledger represented the net balance of payments made and to be made by each party to the other, subject only to exclusion of Barclays’ quarterly fee which was payable in addition to whatever balance was shown on the Ledger. The Ledger was described as being positive if premiums exceeded Credit Protection Payments, so that UniCredit would pay Barclays the balance plus the fees. It was described as negative if the Credit Protection Payments exceeded the premiums, in which case Barclays would pay UniCredit the balance subject to deduction of its fees. Each Guarantee was accompanied by a Credit Support Agreement (“CSA”) under which, if the Accumulation Ledger was negative, UniCredit was required to post security for the deficit, such that Barclays was not taking a credit risk on UniCredit in relation to the Accumulation Ledger being negative at any stage of the running of the Guarantee.

One of the reasons for structuring the deals in a way which did not expose Barclays to any real credit risk on the first loss tranche of the portfolio, which was the risk notionally being protected by the Guarantees, was that Barclays would not have been prepared to take such risk, at any price, in the current turbulent market conditions; and because in any event the urgency of the deals and the short negotiating timescale gave it no opportunity to carry out due diligence on the Reference Portfolios so as to be able to assess the underlying credit risks of the borrowers. Nevertheless the structure of the transaction was apparently such as to persuade the regulators that there had been an “effective transfer of risk” on this tranche so as to qualify for the full RWA relief in respect of it.

It was not the case, however, that Barclays assumed no risk at all on the Guarantees, in return for its annual fees totalling almost €25 million. Barclays was at risk in two respects:

Although the life of the Guarantees was potentially 11, 19 and 19 years respectively, because of the longest single loan in each pool, both parties expected that UniCredit would wish to terminate each Guarantee when the portfolio had substantially reduced in size by the passage of time, and agreed that it should be allowed to do so. Accordingly, each Guarantee provided that UniCredit would be entitled to elect early termination after a period roughly equivalent to the weighted average life of the loans in the portfolio. This was referred to as “the WAL date” and the option to terminate as a “time call”. In the HVB-1 Guarantee the WAL date was 30 September 2013, reflecting a weighted average life of the portfolio rounded up to five years. In the BA and HVB-2 Guarantees it was again five years from the first quarter payment date. The Guarantees provided that upon such a time call, which could be made on the WAL date or any quarter date thereafter, UniCredit had the option to terminate, but only if the Accumulation Ledger were positive (by which I mean, to be more accurate, if the Accumulation Ledger, minus any Credit Protection Payments which would subsequently fall due as a result of already declared credit defaults, were positive). If the Ledger were negative, in the sense described, UniCredit would have to wait for it to turn positive before being able to exercise its time call. Accordingly, subject to the effect of other early termination options, Barclays would be guaranteed to earn at least five years’ fees under the Guarantee.

The time call was one of four circumstances designated in the HVB-1 and HVB-2 Guarantees (once amended and restated) as circumstances in which HVB could elect to terminate the Guarantees before their natural expiry. Each was designated an Optional Early Termination date and provided for in clause 12 of the Guarantee. The four circumstances were the occurrence of:

The BA Guarantee provided additionally for a fifth circumstance as an Optional Early Termination date if “the Rating Event Substitution Date occurs”.

I have explained the Weighted Average Life Termination Date. It was the time call available five years after the date of the Guarantee, and entitled UniCredit to terminate the Guarantee then, or on any subsequent quarter date, provided that the Accumulation Ledger were positive.

Clause 12.1(b) of the Guarantees related to the designation of an Optional Early Termination Date by UniCredit upon the occurrence of a Regulatory Change, and is the clause which is central to the current dispute. It provided:

“If:……..(b) a Regulatory Change occurs in respect of [UniCredit], provided that [UniCredit] has obtained the prior consent from [Barclays], such consent to be determined by [Barclays] in a commercially reasonable manner…….. [UniCredit] may, by not less than 5 Business Days notice to [Barclays], designate……… the next following Payment Date as an Optional Early Termination Date.”

Regulatory Change was defined in Clause 1.1 to occur when:

“in the determination of [UniCredit]…………[UniCredit] will be subject to less favourable regulatory capital treatment with respect to this Guarantee, [the portfolio obligations]………and/or the amount of regulatory capital freed up in respect of [the portfolio obligations]….”

The other three grounds of Optional Early Termination are also relevant:

Thus of the five grounds upon which UniCredit were able to designate an Optional Early Termination Date in any of the Guarantees, three expressly ensured that Barclays would receive at least five years’ fees under the Guarantees:

The provisions for termination following a Regulatory Change and a 10% Clean-up Call Event did not contain any express reference to a time period or the payment of fees, but were both subject to the need for Barclays to consent to termination of the Guarantees.

It was common ground that Barclays would have been entitled to refuse consent if and for so long as the Accumulation Ledger were negative. As explained more fully below, it was Barclays’ understanding, as a consequence of discussions during negotiations, that there would have been a regulatory problem with including an express provision for payment of the balance of five years’ fees in these two situations; but that the consent mechanism could be used by Barclays to achieve the same result, by refusing its consent unless the balance of five years’ fees were paid. By this mechanism, so Barclays understood, it would be able to ensure that whatever happened, it would receive a minimum of five years’ fees.

On that basis Barclays booked as profit, at the date of each Guarantee, the discounted present value of five years’ fees, on the understanding that this amount of fee income was guaranteed to be received come what may (subject to the credit risk of default by UniCredit).

There was one other relevant ground for early termination. Under clause 11.4(c) of the Guarantees, UniCredit was entitled to give notice terminating the Guarantees upon the occurrence of an Illegality, a Tax Event or a Tax Event Upon Merger in respect of UniCredit (subject to first attempting to restructure the transaction to avoid the event giving rise to the right to terminate). Each of these grounds of termination depended on subsequent events making performance illegal or subjecting the payment regimen to unexpected tax treatment. Where early termination occurred pursuant to clause 11.4(c) of the Guarantees, there was a requirement under clause 11.7(a)(i)(D) that UniCredit pay to Barclays the “Unwind Costs”. The Unwind Costs were defined as

“the amount determined by [Barclays] to be its costs in connection with the early termination of the Guarantee, as determined by [Barclays] in good faith and a commercially reasonable manner by reference to its books and records and including, without limitation, any costs associated with loss of profit or unwinding any hedges associated with this Guarantee”.

As with a termination on the ground of a Regulatory Change, the right of Barclays to require payment on termination of its “loss of profit” determined in a “commercially reasonable manner” was understood by Barclays as being a mechanism whereby, as with the other grounds for early termination, Barclays could ensure that it would receive five years’ fees for the Guarantees.

The Guarantees, as executed, contained an entire agreement and understanding clause (“the EAU Clause”) in the following terms:

“20.1

This Guarantee, together with the Credit Support Agreement, constitutes the entire agreement and understanding of the parties with respect to its subject matter and supersedes all oral communication and prior writings with respect thereto.”

On 14 June 2010 UniCredit wrote a letter in respect of each Guarantee seeking Barclays’ consent to an early termination on 30 June 2010 as a result of a Regulatory Change. This did not come wholly out of the blue. There had been discussion between the parties earlier in the year as to whether UniCredit would be able to terminate by reason of a Regulatory Change, and Barclays had suggested terms on which it would allow UniCredit to do so. I did not have the full extent of those discussions in the evidence before me.

By letters dated 23 June 2010 Barclays responded that it was not prepared to give its consent because

“Although early termination of the Agreement was contemplated in the Agreement, it is clear that the parties intended the Agreement to continue for a substantial period of time. Unicredit cannot reasonably expect Barclays to consent to termination so early in the term of the Agreement, in circumstances where this would deprive Barclays of a significant proportion of the overall revenue that it had bargained for and thus result in material economic detriment to Barclays.”

Although the letter did not say so, Barclays had internally determined that it would only consent if paid the balance of five years’ fees. An internal document, which was prepared to show the position at 30 June 2010, identified that the discounted present value of these fees was calculated as being €82,606,465. Mr Norfolk-Thompson gave evidence on behalf of Barclays explaining and verifying the document, and the calculation was not challenged.

The document also showed that it would cost Barclays some €5.65 million to unwind the interest rate swaps which it had entered into to hedge against the extension risk; and about €1.65 million to unwind credit hedges taken out in respect of the credit risk of UniCredit’s non payment of the fees (which were not protected by the CSAs). These were mark to market figures; Mr Norfolk-Thompson explained that it would have cost Barclays about €150,000 more to close out the hedge transactions, to reflect the bid/offer spread. The figures showed, therefore, that it would cost Barclays about €7.45 million to close out existing hedge transactions if the Guarantees were terminated on 30 June 2010. The costs which would be saved by Barclays by the early termination, as then estimated, comprised a maximum of some €6.5 million. That figure is made up of some €2.5 million, which was the present value of the future credit risk hedged by credit hedges (or to put it another way, the present value of the fees was some €2.5 million less than the €82,606,465 because there was a €2.5 million risk they would not be paid); and a figure of €4 million which had been reserved for the future costs of hedging, which would no longer be necessary. However Mr Norfolk-Thompson’s evidence, which I accept, was that this €4 million would almost certainly have been released from reserve if the Guarantees continued, because the credit and interest rate risks were already fully hedged in a way which made it unlikely that any further hedging would have been required. The extension risk on HVB-1 only arose in the event of heavy losses in the early years of the Guarantee and by June 2010 it was apparent that these had not occurred. The €4 million reserve did not, therefore, represent a saving which would result from early termination of the Guarantees. I find, therefore, that the savings to Barclays as a result of termination of the Guarantees at 30 June 2010 would have been a little over €2.5 million. In addition, of course, it would save the management time and expense in managing and monitoring the Guarantees over their lifetime.

The evidence was not well developed as to exactly which sum Barclays internally determined it would insist upon, in return for its consent. Mr Ali, Mr Mariani and Mr Wrobel each said in their witness statements that Barclays determined that it would insist on the balance of five years’ fees; but said nothing about discounting to present value, or taking account of hedge unwind costs, or savings. This evidence was not explored or challenged in cross examination. I find that the sum upon which Barclays internally determined it would insist, in return for providing consent, was €82,606,465 i.e. the balance of its fees for five years, discounted to then present value, without taking account of hedge unwind costs, which it would absorb, or savings from early termination, from which it would benefit. This was its approach in January 2010 when there were discussions between the parties about what Barclays would require for unwinding the Guarantees at that stage. Barclays’ demand at that time, reflected in an internal email of 20 January 2010 and communicated to UniCredit on 27 January 2010, was for the balance of its fees for five years, discounted to then present value, without taking account of hedge unwind costs, which it would absorb, or savings from early termination, from which it would benefit.

On the limited material available to me, I conclude that this is what Barclays’ letter was intended to convey, and how it was understood by UniCredit, in the light of Barclay’s previous stance in January 2010. Barclays was saying, and was understood by UniCredit to be saying, that it would not consent unless it were paid the balance of its fees for five years.

UniCredit responded to Barclays’ letters of 23 June 2010 by letters of the same date which said that it was not reasonable for Barclays to refuse consent in the light of the fact that the regulatory change resulted in UniCredit no longer receiving capital relief by reason of the Guarantees. UniCredit stated that Barclays’ unreasonable refusal of consent was being treated as a waiver of the consent requirement, and purported to designate 30 June as the Optional Early Termination date in accordance with the Guarantees. The letter went on to identify the financial consequences arising under various provisions of the Guarantees.

Barclays responded by letters dated 25 June 2010, repeating that

“Clause 12.1(b) clearly contemplates the possibility that consent may not be given to a proposed early termination. As outlined in our 23 June letter, the decision in the present case has been made by Barclays in a commercially reasonable manner, taking into account particularly the fact that an early termination at such an early stage in the scheduled life of the Agreement would deprive Barclays of a significant proportion of the overall revenue that it had bargained for and thus result in material economic detriment to Barclays.”

On 29 and 30 June 2010 Bank Austria wrote seeking payment from Barclays of what it alleged was the net payment due as a result of early termination. On 30 June 2010 HVB wrote to Barclays addressing Barclays’ reasons for refusing consent in its letter of 25 June 2010. UniCredit stated that what was commercially reasonable must be assessed in the light of the severity of the impact of the Regulatory Change on UniCredit. It pointed out that the Guarantee did not oblige UniCredit to make any additional payments to Barclays upon early termination on the grounds of a Regulatory Change; and went on “in the light of this, it is not commercially reasonable for Barclays to withhold its consent”. The letter concluded by offering to pay the amounts required to make the Accumulation Ledger neutral, without offering any payment as a price for Barclays’ consent to early termination. Mr Railton QC characterised this letter, in my view fairly, as a statement by HVB that it was not prepared to offer any sum as the price for Barclays’ consent.

Thereafter the dispute between the parties crystallised. There was an issue as to whether any Regulatory Change, as defined, had occurred, but by the time of the trial Barclays conceded that one had (although on a different basis from that put forward by UniCredit in its June 2010 letters and the earlier stages of the dispute). After June 2010 UniCredit did not calculate the quarterly payment due by the parties under the Guarantees, and there were no further payments made under the Guarantees.

Barclays contended as follows:

UniCredit contended as follows:

UniCredit had unsuccessfully applied to have the scope of the EAU Clause determined as a preliminary issue. I therefore heard evidence of the negotiations, and of the alleged understandings of the parties, without prejudice to UniCredit’s arguments as to their admissibility or relevance. The commercial negotiations for HVB-1 were conducted partly by email, but primarily by phone, for the most part between Mr Wrobel on behalf of Barclays and Dr Spaeth on behalf of UniCredit. In one relevant telephone call Mr Wrobel spoke to another member of Dr Spaeth’s team who was assisting her on the deal, Ms Streck. Mr Wrobel and Dr Spaeth were the main point of contact for the negotiations. More senior management was involved in the decision making on each side, and there were some communications during the negotiations through other channels, but none on which either side placed significant reliance. Both sides were being advised by lawyers, who were involved in negotiating and finalising the wording of the Guarantees in order to give effect to the commercial terms agreed.

The conversations between Mr Wrobel and Dr Spaeth (and Ms Streck) were recorded by Barclays in the normal course of business, unless, exceptionally, they were made by or to Mr Wrobel on his mobile phone when he was out of the office. They were conducted mainly in the German language. Translations of transcriptions of the recordings were before the Court. There were numerous conversations as the deal progressed, some very lengthy. They were analysed by the parties in considerable detail. Mr Wrobel and Dr Spaeth gave evidence of what their understanding was at each stage of the negotiations when something of potential significance was said.

UniCredit accepted that it was Mr Wrobel’s and Barclays’ understanding, as a result of the discussions during negotiations, that there would have been a regulatory problem with including an express provision for payment of the balance of five years’ fees in the event of a Regulatory Change Event, but that the consent wording was inserted to achieve the same result, so that Barclays could refuse its consent unless the balance of five years’ fees were paid. By this mechanism, so Barclays understood, it would be able to ensure that whatever happened, it would receive a minimum of five years’ fees. That Mr Wrobel and others at Barclays had such an understanding was put beyond doubt by Barclays’ internal documentation, including Mr Wrobel’s internal reporting of his conversations with Dr Spaeth, and the booking of five years’ fees as profit at the outset.

Mr Wrobel’s evidence was that he believed from what Dr Spaeth and Ms Streck said, and wrote, that they and UniCredit shared his understanding. Dr Spaeth’s evidence was that it was not an understanding she shared, nor had she conveyed to Mr Wrobel that she shared it. Both were honest witnesses and neither Mr Wolfson QC nor Mr Railton QC contended otherwise. UniCredit nevertheless accepted that Barclays genuinely believed that its understanding was a shared understanding i.e. that this was also what UniCredit understood when the Guarantees were entered into. Moreover it was accepted by UniCredit that this belief by Barclays, that UniCredit shared the understanding, was a reasonable one in the light of the terms of the communications between Mr Wrobel and Dr Spaeth and Ms Streck during the negotiations, albeit that UniCredit contended that this honest and reasonable belief was in fact mistaken, and that in any event the EAU Clause precluded Barclays from seeking to rely on its understanding of UniCredit’s understanding.

If there was a shared understanding as alleged by Barclays, and if reliance upon it were not precluded by the EAU Clause, UniCredit did not actively challenge the proposition that the determination to refuse consent was made by Barclays in a commercially reasonable manner. This was realistic. In those circumstances there would be no need to have resort to the doctrines of estoppel by convention or acquiescence. Even adopting an objective standard of reasonableness, it would have been commercially reasonable for Barclays to exercise its discretion in the way both parties understood from their negotiations that it would, and would be entitled to.

In these circumstances it is convenient to address the issues which arise in the following order:

In Socimer Bank Ltd v Standard Bank Ltd [2008] Bus LR 1304 Rix LJ reviewed the authorities governing the restrictions implicit in a discretion granted by one contracting party to the other:

“ 60 When a contract allocates only to one party a power to make decisions under the contract which may have an effect on both parties, at least two questions arise. One is, what if any are the limitations on the decision-maker's freedom of decision? ………….

61 The answer to the first question is illustrated by cases such as the following. In Abu Dhabi National Tanker Co v Product Star Shipping Ltd (The Product Star) (No 2) [1993] I Lloyd's Rep 397 the charterparty contained a clause which gave charterers the right to alter the destination of the cargo in circumstances where the contractual port of loading or discharge was blockaded owing to war and "the loading or discharging of cargo at any such port be considered by the Master or the owner in his or their discretion dangerous". The trial judge, upheld by the Court of Appeal, held that the owners' purported decision under this clause was wholly unwarranted, and that in fact they did not consider it dangerous to proceed to the contractual loading port. Leggatt LJ, with whom Balcombe and Mann LJJ agreed, said about the content of the owners' power, at p 404:

"For purposes of judicial review the court is concerned to judge whether a decision-making body has exceeded its powers, and in this context whether a particular decision is so perverse that no reasonable body, properly directing itself as to the applicable law, could have reached such a decision. But the exercise of judicial control of administrative action is an analogy which must be applied with caution to the assessment of whether a contractual discretion has been properly exercised. The essential question always is whether the relevant power has been abused. Where A and B contract with one another to confer a discretion on A, that does not render B subject to A's uninhibited whim. In my judgment, the authorities show that not only must the discretion be exercised honestly and in good faith, but, having regard to the provisions of the contract by which it must be conferred, it must not be exercised arbitrarily, capriciously, or unreasonably. That entails a proper consideration of the matter after making any necessary inquiries. To these principles, little is added by the concept of fairness: it does no more than describe the result achieved by their application."

62 Ludgate Insurance Co Ltd v Citibank NA [1998] Lloyd's Rep IR 221 concerned an agreement by which the London Market Letter of Credit Scheme was operated by Citibank. In certain circumstances the agreement gave to the bank the rights "to retain in the account(s) such additional margin as it considers appropriate in all the circumstances" and to "allocate the drawing(s) . . . in such manner as the bank considers appropriate in its sole discretion": p 221. Waller J and this court held that Citibank had exercised its decision-making rights in accordance with the purposes for which they were granted. Brooke LJ, with whom Mummery and Russell LJJ agreed, said, at paras 35-36:

By contrast, an objective standard of reasonableness is to be applied in cases where a lease provides that a tenant may sublet or assign with the consent of the landlord, such consent not to be unreasonably withheld. It is for the Court to determine whether the withholding of consent is reasonable in such cases: see Woodfall’s Law of Landlord and Tenant paragraph 11.127.

In International Drilling Fluids Ltd v Louisville Investments (Uxbridge) Ltd [1986] 1 Ch 513, at 519-521, Balcombe LJ summarised the following principles as established by the authorities relating to such clauses:

In Ashworth Frazer Ltd v Gloucester City Council [2001] UKHL 59, [2001] 1 WLR 2180 , Lord Bingham said :

It is apparent therefore that in the landlord and tenant context, where there is an objective requirement of reasonableness, the question is not whether the decision is justified, but whether the decision is one which might be reached by a reasonable man in the circumstances. The decision maker is entitled to take into account his own commercial interests. These will take precedence over the commercial interests of the other party. Indeed Balcombe LJ’s proposition (6) is to the effect that a landlord “need usually only consider his own relevant interests”, that is to say he can normally consider his own commercial interests to the exclusion of the interests of the tenant; the tenant’s interests only come into play where to ignore them would be so disproportionate as to be unreasonable.

These principles were applied in a commercial setting in Porton Capital Technology Funds & others v 3M Holdings Ltd [2011] EWHC Civ 2895 (Comm). In that case the Claimant vendors had sold their shareholding in Acolyte to 3M on terms which included an entitlement to an earn out payment based on Acolyte’s net sales for 2009. The Share Purchase Agreement included a provision that Acolyte should not cease to carry on business without the written consent of the vendors, such consent not to be unreasonably withheld. Consent was requested and refused, following which 3M terminated Acolyte’s business in December 2008. One issue was whether the consent was unreasonably withheld. Hamblen J referred to International Drilling Fluids and Ashworth Frazer and said:

“222.

In support of the applicability of such cases to commercial agreements, the Claimants relied upon the case of British Gas Trading Limited v Eastern Electricity, The Times, 29 November 1996, which concerned a long-term gas supply contract which required the customer's consent to any assignment of the supplier's rights and obligations under the contract, such consent not to be unreasonably withheld. The question for the Court was whether it was reasonable for the customer to withhold its consent, in circumstances where the supplier was undergoing a reorganisation (following a report by the Monopolies and Mergers Commission Report) and the resulting change in control would entitle the customer to terminate the contract in any event, unless the contract was first assigned. At first instance, Colman J made extensive reference to the landlord and tenant authorities and concluded that, in the circumstances of that case, consent to the assignment was being unreasonably withheld. That decision was upheld on appeal: [1996] EWCA Civ 1239.

223.

The Claimants submitted that of particular importance in this case are the following principles, to be derived from the above authorities:

i)

First, the burden is upon 3M to show that the Claimants' refusal to consent to the cessation of the Acolyte business was unreasonable.

ii)

Second, it is not for the Claimants to show that their refusal of consent was right or justified, simply that it was reasonable in the circumstances.

iii)

Third, in determining what is reasonable, the Claimants were entitled to have regard to their own interests in earning as large an Earn Out Payment as possible.

iv)

Fourth, the Claimants were not required to balance their own interests with those of 3M, or to have any regard to the costs that 3M might be incurring in connection with the ongoing business of Acolyte.

………………………………

228.

3M disputed the applicability of principles derived from landlord and tenant cases to a commercial agreement such as the SPA. However I accept, as Colman J did in the British Gas Trading Limited v Eastern Electricity case, that they provide some assistance and that the approach set out in paragraph 223 is appropriate in this case.”

Mr Railton QC argued that the contractual requirement in Clause 12.1(b) that the determination be made in a commercially reasonable manner did no more than to make express that which would have been implicit if the discretion had been unfettered and the clause had simply provided that the option could only be exercised with Barclays’ consent. The word “commercially” added nothing to the concept of reasonableness, and the use of the word “reasonable” was to import expressly the test of rationality which would have been implicit. He argued that to construe the clause as posing an objective test would be impractical because the Court would not be able to determine what objective criteria to apply.

Attractively as these arguments were presented, I am unable to accept them. The parties chose not to give Barclays an uncircumscribed discretion whether or not to consent, but chose to confine it to behaviour which was commercially reasonable. The presumption is that by using express words, the parties were seeking to achieve some greater restriction than would have applied if no words had been used. Although public lawyers are familiar with the concept of reasonableness in its Wednesbury sense of not irrational, that is not the sense in which the word would commonly be used or understood by businessmen in a commercial agreement. That conclusion is reinforced by the use of the adverb “commercially”, which is apposite for a concept of objective reasonableness but rather less so for a concept of rationality.

It is true that it is difficult to define objective criteria applicable in all cases in which the determination of consent under clause 12.1(b) falls to be considered, other than at a high level of generality. But the same is true of a provision that consent is not to be unreasonably withheld in landlord and tenant cases, as is apparent from Lord Bingham’s observations in Ashworth Frazer Ltd v Gloucester City Council and those of Lord Denning MR inBickel v Duke of Westminster which Lord Bingham endorsed. The difficulty is mitigated by the two aspects of the objective test which I have highlighted, namely that the question is not whether the decision is justified but whether the decision is one which might be reached by a reasonable man in the circumstances; and the decision maker is entitled to take into account his own commercial interests, in preference to those of the other party, and normally to their exclusion.

The primacy of Barclays’ commercial interests in the exercise contemplated by Clause 12.1(b), which arises by analogy with the landlord and tenant cases and the Porton decision, is reinforced by the following features of the Guarantees:

Barclays’ only commercial reason for entering into the Guarantees was to make profits from the transactions through its fees. It had no other commercial interest in them. Without the ability to protect and preserve profits by means of the exercise of its discretion, the discretion would have little purpose for Barclays. What is meant by Barclays’ commercial interests in the context of the discretion conferred by clause 12.1(b) is Barclays’ interest in keeping the Guarantees on foot so that it can earn the profits represented by its fees.

Accordingly in my judgment the correct approach to the application of the clause is as follows:

In my judgment it is not open to Barclays to seek to justify its refusal of consent on these grounds. The determination to refuse consent was in fact made on the basis that Barclays would insist upon being paid five years’ fees, either by the agreement continuing to run for that period or by being paid the equivalent as the price for its consent. Barclays did not suggest the latter in its letter refusing consent, or ask any price for its consent; the letter merely refused consent. But as I have found earlier in this judgment, the letter refusing consent was intended to convey, and would have conveyed to UniCredit, that the refusal of consent was based on an insistence on five years’ fees. If that was not commercially reasonable (which I address below), it was not incumbent upon UniCredit to make a proposal as to payment of some lesser sum as the price for consent.

Nevertheless, there is little doubt that Barclays would have been acting reasonably in refusing consent without receiving some payment or credit from UniCredit. UniCredit conceded that it would have been reasonable for Barclays to recover its out of pocket costs of unwinding the hedges already in place, at least if they exceeded costs saved by early termination, as I have found they did. It may be that if five years’ fees could not be justified as commercially reasonable, a smaller amount could. Dr Spaeth said in evidence that in the event of the seeking of consent before five years had expired, she would have expected there to be a commercial negotiation. I do not need to decide, at this stage, whether Barclays would have been acting reasonably in demanding a lesser amount than five years’ fees.

Applying the criteria which I have identified, I have little hesitation in concluding that Barclays was acting in a commercially reasonable manner in refusing its consent unless and until it recovered five years’ fees, for a number of reasons.

First, that was Barclays’ reasonable and legitimate expectation at the time the Guarantees were entered into. Whatever the state of mind of Dr Spaeth or others at UniCredit at that time, the state of mind of Mr Wrobel and Barclays’ management was that the commercial deal was that Barclays would be entitled to its fee income for a minimum of five years, come what may. It was a good deal for Barclays (as was readily conceded during the trial) but it was the basis on which Barclays was prepared to contract in the turbulent market conditions and given UniCredit’s pressing need for capital relief. The understanding which Barclays had, at the time of the Guarantees, of the way in which the consent mechanism would operate was that it could be used to protect their expectation that they would receive five years’ fees. Mr Wolfson QC conceded that that was a reasonable understanding. Moreover it was an understanding which had a practical effect in Barclays booking five years’ fees as profit in the 2008 year. If Barclays were unable to secure payment of these fees by refusal of consent, it would have to enter a loss in its accounts in 2010. Whilst in one sense this is an accounting loss rather than an economic loss, it nevertheless represents a real detriment to Barclays in circumstances where it would confound their reasonable expectations at the time of contracting.

This is so irrespective of any understanding which Barclays had as to UniCredit’s understanding, which Mr Wolfson QC contended was irrelevant or inadmissible by virtue of the EAU clause. For the reasons set out below I reject the argument that the EAU clause has that effect. It is relevant in this context that, as Mr Wolfson QC conceded, Barclays understanding was not merely an internal one, but one which they honestly and reasonably believed was shared by UniCredit. UniCredit’s case therefore involves the proposition that no reasonable person in Barclays’ position could have exercised the discretion to refuse consent in a way in which Barclays honestly and reasonably believed, at the time the Guarantees were entered into, that both parties expected it to be exercised. That would be a surprising conclusion.

Mr Wolfson QC argued that whether Barclays reasonably, but mistakenly, believed that there was a common understanding is irrelevant to commercial reasonableness. As he put it, a mistaken belief that you can charge an unreasonable sum does not make an unreasonable sum reasonable. I disagree. The primacy of Barclays’ interests, which is accorded by the discretion conferred on it by Clause 12.1(b), entitles Barclays to take account of its own reasonable expectations at the time of contracting. That is all the more so if it reasonably believes, at the time of contracting, that those expectations are shared, even if such belief is in fact mistaken. The question is not whether the decision was justified, but whether a reasonable person in Barclays’ position could have reached it.

Secondly, in the BA Guarantee, UniCredit expressly agreed to pay Barclays a minimum of five years’ fees if there were an optional early termination by UniCredit as a result of a Rating Event, i.e. if Barclays’ credit rating were downgraded. The relevance of such an event is that it might have the same effect as a Regulatory Change, in that the capital relief afforded to UniCredit was calculated in a way which included taking account of Barclays’ credit rating. A Rating Event would be outside the control of UniCredit and would or might deprive UniCredit of the entire benefit of the Guarantees. Yet in those circumstances, UniCredit agreed that Barclays should receive a minimum of five years’ fees. Ex hypothesi UniCredit regarded that as commercially reasonable. I do not see a material distinction between a Rating Event and a Regulatory Change for these purposes. Both are outside UniCredit’s control and both will trigger the exercise of the early termination option because they diminish or remove the benefit UniCredit receives under the Guarantees. Dr Spaeth was cross examined as what the relevant distinction was which made one commercially reasonable but not the other. I did not find her answer coherent or convincing. She explained the difference as being that a regulatory change was dictated by the regulator, whereas if Barclays were downgraded, the amount of capital relief thereby extinguished or impaired would depend upon HVB’s (sic) internal view of Barclays’ rating, not Barclays’ external rating. But that does not detract from the fact that the trigger would be something which was outside Bank Austria’s control, and if Bank Austria took the view that it removed or significantly impaired the capital relief, it would be entitled to terminate the Guarantees. Nevertheless in those circumstances Bank Austria agreed it would still have to pay Barclays five years’ fees.

Thirdly, Barclays’ expectation of a minimum of five years’ fees was protected under the terms of the other early termination options which expressly dealt with the point. Of the five grounds upon which UniCredit were able to designate an Optional Early Termination Date in any of the Guarantees, two were dealt with by the consent mechanism but the other three expressly ensured that Barclays would receive at least five years’ fees under the Guarantees. The Weighted Average Life Termination Date could occur only after five years; the Substitution Event Date occurred after four years but required payment of one year’s fees as a condition of termination; and the Rating Event Substitution Date in the BA Guarantee required a premium to be paid on termination equivalent to the balance of five years’ fees. The capital relief available to UniCredit from the Guarantees would diminish as loans expired before the WAL date, but that would not preclude Barclays earning five years’ fees. UniCredit might want to resecuritise the portfolio at any stage early in its history because it would result in more favourable capital relief, but it could not do so without paying Barclays five years’ fees. These events can be distinguished from a Regulatory Change in that they are not subsequent events outside UniCredit’s control. Nevertheless they reflect Barclays’ reasonable commercial interest in a minimum fee income of five years’ worth.

This attracts some, albeit limited, further support from the terms of Clause 12.3(c)(ii) of each of the Guarantees. This provides that the amounts recoverable on early termination under Clause 12 (i.e. the outstanding balance of five years’ fees where that is expressly provided for in the way I have identified) constitute a “reasonable pre-estimate of loss”. Whilst no doubt intended to preclude reliance by UniCredit on the penalty doctrine, this provision nevertheless treats the balance of five years’ fees as a reasonable estimate of Barclays’ loss from early termination where that arises from a Substitution Event or Rating Event. There is no reason to adopt a more restrictive view of what is reasonable when considering whether Barclays exercised the discretion given to it in connection with the Regulatory Call in a commercially reasonable manner.

In those cases in which there is express contractual provision for a pre-determined payment on early termination, the Guarantees themselves provide for a greater payment to Barclays than that which UniCredit contend would be reasonable in the case of an early termination on the ground of Regulatory Change. There is nothing in the contractual terms which indicates that Barclays should be exposed to the risk of losing some of its fees in the event of an early termination on the ground of Regulatory Change when it is not exposed to that risk in the event of an early termination on other grounds. Nor is the difference in the nature of the events which trigger the early termination option in each case sufficient for one to be able to say that no reasonable person in Barclays’ position would insist on being in the same financial position in each case.

Fourthly, I am reinforced in my conclusion by the difficulties which Mr Wolfson QC had in identifying what it was about the insistence upon five years’ fees which made it unreasonable. This was not due to any lack of forensic skill, but rather the impossibility of the task. His primary position was that any element of profit was unreasonable: Barclays could insist on any costs which would be incurred in unwinding hedges, but no more. This was not consistent with UniCredit’s pleaded case that “the purpose of the discretion was to enable Barclays to protect itself against losses which would crystallise if the Guarantee terminated early... or other detriment to Barclays”. Other detriment must envisage something other than losses. UniCredit’s argument was that once a regulatory change had removed from UniCredit any benefit under the Guarantees, Barclays could not reasonably keep them alive in order to earn any more fees by way of profit. Such a restrictive interpretation of what would be commercially reasonable would be to give UniCredit’s interests in terminating the Guarantees primacy over Barclays’ interests, and so subvert the clause. It would remove any consideration of Barclays’ interests and give primacy to those of UniCredit. If the Ledger were positive, Barclays’ only interest in the continuation of the Guarantees was the earning of its fees, as Mr Wolfson QC was at times concerned to emphasise; there would be no point in giving Barclays a discretion as to whether the Guarantees should continue (which was not expressed to be limited to circumstances in which the Ledger was negative) if Barclays could not exercise it for the purposes of earning future fees. Moreover, if Mr Wolfson QC were right in this argument, Barclays would be bound to consent to an early termination at any time, even if the regulatory change only had a marginal effect on diminishing UniCredit’s capital relief and the option were exercised simply because market conditions had changed in a way which made it cheaper for UniCredit to resecuritise the portfolio with another counterparty. Again that would be to give UniCredit’s interests primacy over those of Barclays.

It follows that it could not be said that no one in Barclays’ position would be acting reasonably in insisting on at least some element of profit. This is a conclusion shared by Dr Spaeth, who said that in the event of a regulatory change call she expected there to be a commercial negotiation over what future fees Barclays could charge.

Mr Wolfson QC’s alternative position was that on any view it would not be reasonable for Barclays to insist on the balance of five years’ revenue, rather than five years’ worth of profits. It was not that the five year period was too long, or the size of the sums disproportionate to the burden imposed upon UniCredit by the continuation of the Guarantees. Rather, he contended that by insisting on five years’ fees, Barclays was thereby seeking to secure more than it could have recovered by way of damages, because it was seeking revenue, not profit, for the period up to a point when UniCredit would have been entitled to exercise the time call at the WAL date. There are three answers to this point. The first is that Barclays in fact simply refused consent, rather than demanding a sum of money. The refusal of consent and continuation of the Guarantees would allow Barclays to earn its fee revenues but would not avoid the incurring of any costs. It would provide Barclays with its profit, not more. Secondly, as I have found on the evidence, Barclays’ decision was to insist on the then present value of its fees, which would not have exceeded its profit, given the hedge unwind costs and cost savings resulting from early termination. Thirdly, the sums to be paid in the event of a Substitution Event Date, or a Rating Event Substitution Date, are the balance of five years’ fees, undiscounted to present value and without taking account of saved costs: the Guarantees provide in these circumstances for the payment of revenue, not merely profit.

Mr Wolfson QC argued that if the regulatory change clause entitled Barclays to keep the Guarantees alive unless and until it received five years’ fees, the clause would confer no benefit on UniCredit and would be stripped of any commercial purpose: they could exercise the WAL call at five years anyway, but could not do any better if there were an earlier regulatory change. This is not so. The clause would be of value to UniCredit if Barclays did not insist upon payment of five years’ fees, which is only one response in a range of determinations which might have been made in a commercially reasonable manner. Moreover, an early termination on payment of five years’ fees, discounted to their present value, might be of benefit to UniCredit in reorganising and re-securitising their loan book. Such a benefit is contemplated by the provisions for early termination by reason of a Substitution Event Date and a Rating Event Substitution Date, each of which allowed termination before the expiry of five years, and envisaged that UniCredit might wish to opt for it, but required a payment to Barclays of the balance of five years’ fees.

For all these reasons I conclude that Barclays determined its refusal of consent in a commercially reasonable manner.

This issue does not arise, as a result of the conclusions I have already reached, and in light of my conclusions on Issue 4 below. I can express my views upon it briefly.

The clause provides: “This Guarantee, together with the Credit Support Agreement, constitutes the entire agreement and understanding of the parties with respect to its subject matter and supersedes all oral communication and prior writings with respect thereto.” It is concerned with identification of the contractual terms. It is intended to preclude investigation of, or reliance on, any communication in the course of negotiation of the Guarantee, as something which is to be treated as having contractual effect, unless it finds expression in the written document. It is for that purpose alone that any understandings are to be excluded or superseded. It would not preclude reliance on such communications to found an allegation of misrepresentation (cf Axa Sun Life Services Plc v Campbell Martin Ltd [2012] Bus LR 203). Nor does it, in my view, restrict the scope of the material which may be relied upon to inform the question whether a determination of refusal of consent is commercially reasonable. It is not concerned with limiting a discretion which is contractually conferred. The clause does not purport to introduce any limitation as to matters which could be considered in the exercise of a contractual discretion. Nor is this the place where one would expect to find such limitation or restriction: its obvious place would be in clause 12.1. If there is to be any restriction or limitation on matters which would otherwise be relevant to a contractual discretion, such a limitation would need to be clear and express and unequivocally directed to the exercise of the discretion. Clause 20.1 is not in such terms.

This conclusion accords with the purpose and function of entire agreement clauses as expressed by Lightman J in Inntrepreneur Pub Co v East Crown [2000] 2 Lloyds Rep 611 at [33], and approved by the Court of Appeal in Axa Sun Life at [63]:

“ The purpose of an entire agreement clause is to preclude a party to a written agreement from threshing through the undergrowth and finding in the course of negotiations, some (chance) remark or statement (often long-forgotten or difficult to recall or explain) upon which to found a claim, such as the present, to the existence of a collateral warranty. The entire agreement clause obviates the occasion for any such search, and the peril to the contracting parties posed by the need in its absence to conduct such a search. For such a clause constitutes a binding agreement between the parties that the full contractual terms are to be found in the document containing the clause and not elsewhere, and that, accordingly, any promises or assurances made in the course of the negotiations (which in the absence of such a clause, might have effect as a collateral warranty) shall have no contractual force, save in so far as they are reflected and given effect in the document.”

Whether an entire agreement clause precludes reliance on an estoppel by convention is a question on which there is little authority. Each case will depend upon the particular terms of the entire agreement clause and the nature of the estoppel contended for, which may be a common assumption as to an existing state of fact or law, which was the paradigm situation in which the doctrine was first developed, or a common assumption as to what is expected to happen in the future in relation to a transaction, to which the doctrine also now extends (ING Bank v Ros Roca SA [2012] 1 WLR 472 at [64(i)]).

In Sere Holdings Limited v Volkswagen Group United Kingdom Ltd [2004] EWHC 1551 Mr Christopher Nugee QC, sitting as a deputy High Court Judge, held that an entire agreement clause in wide terms precluded reliance on an estoppel by convention. That was a case in which the alleged common assumption was as to the future conduct of the defendant and therefore its effect was akin to a contractual promise. Mr Nugee QC said at [25] that “if the entire agreement clause is effective... to rob an express promise made in precontractual negotiations of any legal effect, it seems to me that it must equally be effective to prevent a promise from having any legal effect where that promise is said to arise out of an assumption shared by the parties when entering into the contract”. On the other hand in Dubai Islamic Bank v PSI Energy Holding Company [2011] EWHC 2718 (Comm) Hamblen J, in considering a summary judgment application, accepted at [83]-[84] that it was arguable that an estoppel by convention could take effect notwithstanding an entire agreement clause. The common assumption in that case did not involve the assertion of an additional contractual promise, but rather was relied on to prevent a party enforcing an existing contractual promise in a way contrary to the parties’ shared understanding. Hamblen J pointed out that an entire agreement clause did not preclude a claim for rectification, a principle which, like estoppel by convention, is based on considerations of unconscionability.

The particular estoppel by convention or acquiescence alleged by Barclays in this case, which involves an assumption as to future conduct which is not akin to an additional contractual promise, would not be precluded by the terms of Clause 20.1 of these Guarantees. The assumption relied upon is as to how one party will exercise a contractual discretion conferred upon it. Clause 20.1 is not addressed to the mode of exercise of that discretion.

Mr Railton QC submitted that upon a detailed analysis of the translations of the transcriptions of the phone conversations, and of the contemporaneous emails, it was clear that UniCredit, and in particular Dr Spaeth, shared Barclays’ understanding that in the event of a Regulatory Change, Barclays would be entitled to refuse consent unless or until it was paid five years’ fees. This analysis did not depend upon any single clear expression of that understanding in unequivocal terms in a document, but rather upon the language used in a large number of telephone calls, in the context of the place of each call in the evolution of the negotiation.

There are three reasons why I am unable to accept that submission, which can be expressed without recording at length the numerous passages in the communications which were relied upon in support of it.

The first is that the transcripts of the telephone conversations do not provide a full or reliable picture of the negotiations. They do not fall to be interpreted in the same way as written exchanges. In some places the conversation is referred to as inaudible; it is impossible to know how much is missing, or what the content or significance of the missing words might be. Moreover, there were additional conversations when Mr Wrobel was using his mobile phone, which were not recorded and are therefore not available for scrutiny. The material before the Court is therefore incomplete, and the witnesses could not reasonably have been expected to recall the detailed content of unrecorded conversations from unaided memory.

The conversations were conducted mainly in the German language. There were some disputes about translations, but that was perhaps the least of the relevant difficulties in interpreting them. When the transcription records someone as saying “yes”, it is not clear whether he or she is agreeing with a proposition put forward by the other, or merely indicating that he or she understands it, or understands that it is what the other party wants. Something which appears on the transcript as a statement may have been a question by reason of inflection of voice, and vice-versa. It may have been a summary of what the other person seemed to be saying to check it was understood. Transcription does not capture a speaker’s nuances of tone which may significantly affect how the words are to be interpreted. It does not capture surprise, or sarcasm or irony (Dr Spaeth said of one of the statements upon which Barclays relied that it was ironic). It does not capture whether an apparent statement was tentative or emphatic. The interlocutors were often talking over each other. Expressions upon which Barclays placed heavy emphasis, such as “guaranteed”, or “locked in” might have different meanings to different people, and were heavily dependent on their context for interpretation. Some were used to express different concepts at different times. A good example was the expression “make whole” which was used (in English) during some of the conversations. It might refer to Barclays being made whole by payment to make the Ledger neutral in circumstances where the Ledger were negative, without any connotation for future fees; or it might refer to Barclays being made whole by the payment of a minimum amount of future fees, especially if the context of the discussion posited a positive Ledger on termination. Unless it were clear that both interlocutors were talking of a situation in which the Ledger were positive, there was ample scope for ambiguity and misunderstanding.

All this meant that the transcripts which were subjected to detailed analysis during the hearing could give only an incomplete picture, and quite possibly a misleading picture, of the sense of the critical exchanges between Mr Wrobel and Dr Spaeth.

The negotiations for the HVB-1 Guarantee were extensive and fluid, with many conversations, sometimes taking place late at night. Topics were raised, discarded, parked and revisited, as often happens in such negotiations. It is dangerous to assume that something discussed without dissent at a particular stage of such negotiations was part of the overall final understanding of the deal. Commercial parties sometimes prefer not to try to tie down every detail which might give rise to a dispute, but to leave it for post contractual argument or negotiation if the dispute happens to eventuate. The approach may be “I hear what you say, but we’ll cross that bridge if we come to it”.

Moreover, the conversations took place against the background that there were multiple decision makers on each side. Although the discussions took place mainly between Mr Wrobel and Dr Spaeth, there were some additional channels of communication. The understanding of each party as to the deal was to be set out in a carefully drafted written document, to which people other than Mr Wrobel and Dr Spaeth would be expected to pay close attention. Mr Wrobel and Dr Spaeth were not concluding an oral agreement, and did not think that they were. What Mr Wrobel and Dr Spaeth knew, when having their discussions, was that the terms of the agreement would be the subject of careful drafting by lawyers; and would be considered by others in the management structure on both sides. Mr Wrobel and Dr Spaeth were concerned to agree the commercial terms of the deal, but could not be expected to have done so with the attention to precision of expression upon which their cross examination in Court focussed.

All these circumstances, a number of which Mr Wolfson QC characterised as “the fog of war”, render the extant translated transcriptions of conversations between Mr Wrobel and Dr Spaeth (and Ms Streck) an unreliable source from which to draw a conclusion as to UniCredit’s contemporaneous understanding.

Secondly, Barclays’ case as to the interpretation to be put upon the extensive passages relied upon was explored fully with Dr Spaeth in cross examination, with a view to establishing that they demonstrated her understanding to be the same as Barclays’. Her evidence, maintained under a sustained, vigorous and fair cross examination by Mr Railton QC, was that the understanding for which Barclays contends was not her understanding at the time. Mr Railton QC accepted that she was an honest witness and had not been seeking deliberately to mislead the Court. This realistic assessment illustrates the potential ambiguities, or nuances of interpretation, in the language upon which Barclays’ argument relied.

Mr Wolfson QC contended that the concession that Dr Spaeth was an honest witness was fatal to any case based on a shared understanding. Mr Railton QC contended that I could conclude that her evidence was honest but mistaken as to her contemporaneous understanding, for any of the reasons for which witnesses give honest but mistaken evidence. In particular he referred to witnesses being mistaken as a result of the passage of time, or because they have convinced themselves of a mistaken view after the event in the light of the dispute which has arisen. These were not possibilities which applied to Dr Spaeth’s careful evidence when taken through the transcripts of the telephone conversations and other documents passage by passage. Her honest evidence that she did not share Barclays’ understanding at the time at which the transactions were concluded cannot be explained by mistaken recollection. It was her honest evidence because it was true: she did not share Barclays’ understanding at the time the transactions were concluded.

Thirdly, there is one exchange in a telephone conversation late in the negotiations which is difficult to reconcile with Barclays’ allegation that Dr Spaeth understood Barclays were to get five years’ fees come what may. The conversation started at about midnight on 28/29 September 2008, the night before the HVB-1 deal closed. The subject matter of discussion in the relevant exchange was early termination in the event of illegality or a tax change. When Mr Wrobel indicated that in those circumstances Barclays would want five years’ fees, Dr Spaeth immediately responded with a sarcastic “oh that’s just great… bold actually”. She did not agree to it. Although this was in the context of an illegality or tax call, she could not have reacted in this way had it been her understanding up to that point that Barclays would be guaranteed to receive five years’ fees, come what may, including in the event of a regulatory change call.

Barclays contended that even if UniCredit did not share the understanding, so that there was no common assumption, nevertheless UniCredit are estopped from denying it by operation of an estoppel by convention or estoppel by acquiescence. Such an estoppel can arise where one party to a transaction acts on an assumed state of facts or law, which is communicated to the other party and acquiesced in by the other party, if it would be unjust to allow the acquiescing party to go back on the assumption: see Republic of India v India Steamship Co Ltd (The Indian Endurance and The Indian Grace) (No2) [1998] AC 878, per Lord Steyn at 913-4 and ING Bank v Ros Roca SA [2012] 1 WLR 472 per Carnworth LJ at [56]-[60]. Barclays argued that UniCredit knew the assumption Barclays was making; that it did not say anything to correct or contradict that assumption; and that it therefore acquiesced in Barclays’ assumption.

This argument fails on the facts. For the reasons given above, I am not able to conclude that Dr Spaeth, or anyone else at UniCredit, clearly understood that Barclays was making the assumption; nor that she or UniCredit said or did anything, which conveyed acquiescence in it to Mr Wrobel, or Barclays.

Barclays’ claim was for a declaration that its consent had validly been refused, and for orders for the taking of accounts to enforce periodic payments in accordance with a recreated Accumulation Ledger, which would amount in effect to an order for specific performance. As a result of discussion during the course of closing submissions, it was agreed that I should grant declaratory relief, but leave over any further consideration of specific performance or damages to enable the parties to consider their positions in the light of my determination, subject only to a point about whether the Guarantees had already automatically come to an end as a result of UniCredit’s unaccepted repudiatory breach. UniCredit relied on the decision of the House of Lords in White & Carter (Councils) Limited v McGregor [1962] AC 413 to support a submission that the Guarantees had already come to an end because Barclays was not entitled to insist on the ongoing performance of the Guarantees, even if Barclays had refused consent to early termination in a commercially reasonable manner.

This point emerged for the first time in UniCredit’s skeleton argument at the beginning of the hearing. The document did not identify with precision exactly what conduct constituted the repudiation relied on. In oral closing submissions Mr Wolfson QC identified it as either the letters in June 2010 in which UniCredit purported to terminate the Guarantees irrespective of Barclays’ refusal of consent; or alternatively UniCredit’s conduct since that date in failing to perform its obligations to calculate the payments due by each party under the Guarantees and seek or make payment in accordance with such calculations each quarter, i.e. on 30 September 2010 or at least by the time of the commencement of the trial. As a result of the way in which the point emerged, little argument was addressed to the question whether or when any repudiation took place, but Mr Railton QC made clear that Barclays did not accept that the UniCredit letters in June 2010 were repudiatory because they purported to be relying upon the contractual terms.

White & Carter confirmed the general rule that, if one party to a contract repudiates it by refusing to carry out its obligations, the innocent party has a right of election to accept the repudiation and sue for damages or to keep the contract in effect (see page 427). Lord Sumption recently articulated this “elective theory of repudiation”, as it is sometimes called, in Raphael Geys v Societe Generale [2012] UKSC 63:

There are two limitations to the principle. The first limitation is that in many cases the party in breach can compel the innocent party to restrict his claim to damages by refusing co-operation: see White & Carter at page 428 and Geys at [114]-[116].  This is because, if the contract is kept alive, it is kept alive for both parties and so the innocent party must also perform its contractual obligations if it is to earn the right to claim the price that is due to be paid by the party in breach.  If the innocent party cannot earn the right to claim the price due to it for its performance without the co-operation of the party in breach, it will not be able to pursue a debt claim and will be limited to a claim in damages. The limitation does not apply in this case because Barclays’ payment obligations are triggered only if positive steps are taken by UniCredit to claim payment upon the occurrence of a Credit Event; if UniCredit do nothing, their own payment obligations to Barclays will simply accrue on a quarterly basis but Barclays will not itself be obliged to make any payments. That is what has happened since 30 June 2010.

The second limitation arises from an obiter dictum by Lord Reid (at page 431) that “it might be said that, if a party has no interest to insist on a particular remedy, he ought not to be allowed to insist on it”.  This is the basis of the UniCredit’s submission advanced for the first time in UniCredit’s opening submissions in terms that Barclays “is not entitled to assert that it has any interest in insisting on performance”.

Later authorities make clear that this second limitation referred to by Lord Reid is a narrow one, which will be applied only in exceptional cases. These authorities were recently considered by Cooke J in The Aquafaith [2012] 2 Lloyd’s Rep 61. His conclusion at [44] was that “The effect of the authorities is that an innocent party will have no legitimate interest in maintaining the contract if damages are an adequate remedy and his insistence on maintaining the contract can be described as ‘wholly unreasonable’, ‘extremely unreasonable’ or, perhaps, in my words, ‘perverse’”.

This is not such a case. At the hearing, UniCredit advanced two reasons for the application of Lord Reid’s dictum. The first was that it has not been monitoring the Reference Portfolios, and hence does not know how they would have performed, or what Credit Events have occurred. In the opening submissions, where this point was raised for the first time, the effect was said to be that it would be “incredibly difficult, if not impossible” to recreate the “portfolios and credit events” since 30 June 2010. In the written final submissions this was elevated to being “impossible”. The unsatisfactory way in which this point emerged meant that there was no evidence given by UniCredit to support this allegation in either form, still less as to what the difficulty was or when the difficulty or impossibility arose. That is fatal to the submission.

If it were the case that recreation of the Accumulation Ledger would now be impossible, that might be a relevant consideration when the Court has to consider whether to exercise its discretion to grant equitable relief by way of specific performance or to confine Barclays to a remedy in damages. But that is a different issue from whether the Guarantees have already come to an end as a result of UniCredit’s repudiation, and not one with which I am concerned at this hearing.

The second ground advanced for the application of Lord Reid’s dictum was that there was no good reason for Barclays to insist upon performance, including calculation of the Ledger, when Barclays’ only interest was in its fees. Barclays accepted that its primary interest in the Guarantees is the fees it should have received, and will receive, under them; and was prepared to accept in principle that it might be compensated appropriately by way of damages for non-performance by UniCredit. But whether an award of damages would in fact give it full compensation for its losses would depend upon the basis on which damages were assessed, and in particular whether it would be by way of assumption that no further Credit Event Notices would be served, or whether some (and if so which, and when) would be served; whether it would recognise Barclays receiving the present value (as at the date of judgment) of its quarterly fees under the Guarantees until the Guarantees reached their WAL date; whether it would recognise that Barclays acted reasonably in the hedges it entered into both before and after June 2010; and whether Barclays would recover its costs in unwinding hedges which it has entered into.

These are matters which have not yet been explored between the parties and were not the subject matter of evidence or argument before me. The parties had agreed, before this point emerged, that should there be any need for a determination of any amount to be paid (by way of damages, or otherwise), following my determination of the issues of principle, then it was to be dealt with by way of an assessment to take place at a subsequent hearing. If there is any basis on which it might be argued that Barclays’ position by way of damages is less favourable to it than continued performance of the Guarantees, Barclays has a very clear interest in continuing to insist on performance of the Guarantees. In those circumstances, Barclays has not adopted an unreasonable approach in seeking to hold UniCredit to the performance of its contractual obligations.

Accordingly I reject UniCredit’s submission that the Guarantees have come to an end as a result of any unaccepted repudiatory breach by UniCredit.

Barclays determined its refusal of consent to early termination of the Guarantees in a commercially reasonable manner, and the Guarantees have not been validly terminated or come to an end.